University of Southampton Research Repository ePrints Soton
Copyright © and Moral Rights for this thesis are retained by the author and/or other copyright owners. A copy can be downloaded for personal non-commercial research or study, without prior permission or charge. This thesis cannot be reproduced or quoted extensively from without first obtaining permission in writing from the copyright holder/s. The content must not be changed in any way or sold commercially in any format or medium without the formal permission of the copyright holders.
When referring to this work, full bibliographic details including the author, title, awarding institution and date of the thesis must be given e.g.
AUTHOR (year of submission) "Full thesis title", University of Southampton, name of the University School or Department, PhD Thesis, pagination
http://eprints.soton.ac.uk
UNIVERSITY OF SOUTHAMPTON FACULTY OF ENGINEERING AND THE ENVIRONMENT BIOENGINEERING SCIENCES AND ENGINEERING MATERIALS RESEARCH GROUPS
OBSERVATION AND PREDICTION OF BIOCIDE RELEASE WITH FLUORESCENCE TECHNIQUES AND MATHEMATICAL MODELLING
BY Liam Rhys Goodes
A thesis submitted in partial fulfillment of the requirements for the degree of
Doctor of Philosophy
JUNE 2013
UNIVERSITY OF SOUTHAMPTON
ABSTRACT
FACULTY OF ENGINEERING AND THE ENVIRONMENT
NATIONAL CENTRE FOR ADVANCED TRIBOLOGY AT SOUTHAMPTON RESEARCH GROUP
Doctor of Philosophy
OBSERVATION AND PREDICTION OF BIOCIDE RELEASE WITH FLUORESCENCE TECHNIQUES AND MATHEMATICAL MODELLING
Liam Rhys Goodes
Antifouling coatings are crucial for protection of vessels’ hulls against marine biofouling. A range of technologies is available, although biocidal coatings - containing toxic or deterrent compounds – still represent a majority of the market. A long-term goal is the development of less environmentally harmful and persistent compounds; one of many potential avenues is that of synthetic analogues of natural products from marine organisms. The development of coatings using natural products has been hampered by poor performance in the field without sufficient work on their leach rates and behaviour. Furthermore, little work has been carried out on the leach rate of traditional organic biocides as used in modern coatings. Prediction of biocide diffusion is crucial to estimation of antifouling efficacy. However, diffusion in glassy polymers is a complex and oft-neglected topic; the chemically and physically changeable environment of the ocean and swelling of the polymer in such a ternary system also increase the complexity of models.
A test matrix of antifouling paint coatings was composed, including polymethylmethacrylate (pMMA), an erodible rosin-based commercial binder and a novel trityl methacrylate/butylacrylate copolymer (pTrMA/BA) as binders. Copper (I) oxide and usnic acid, a natural product biocide of interest, were incorporated into the binders and the coatings were subjected to 10 months of natural immersion and 6 months of accelerated rotor immersion tests (17 knots, 25 °C). A novel application of fluorescence microscopy was developed, allowing quantification of the usnic acid content within the test coatings from both immersion schemes. This fluorescence technique and optical microscopy techniques were applied to these coatings before and after immersion, allowing quantification of the organic biocide and pigment distribution. Existing literature models for diffusion in glassy systems were adapted with a
i
novel method for taking into account the presence of seawater as a diluent, to obtain effective diffusion coefficients for usnic acid. These have been integrated into mathematical models of diffusion to predict biocide lifetime. These data were compared with experimental data for biocide leaching from the long term immersions.
The biocide leached completely from the p(TrMA/BA) binder during rotor testing, compared to 35% from the pMMA binder. For pontoon immersions, 61% of the additive was lost from the pMMA coating, and 53% from the rosin-based binder. An accelerated loss of usnic acid occurred in the surface of the rosin-based binder, due to rosin depletion. In all samples, release of the biocide was inhibited beyond the cuprous oxide front, which was congruent with the leached layer in samples where cuprous oxide release occurred. The erodible binder was the only one which demonstrated synchronous depletion of both additives, and it demonstrated a good resistance to fouling in immersion trials. Results of the mathematical modelling of the biocide diffusion were in good agreement with the observed data in the case of pMMA, highlighting in particular the importance of water uptake with respect to biocide diffusion. However, there was poor agreement in the case of p(TrMA/BA), for which the model under- predicted the release rate by about three orders of magnitude.
ii
Contents 1. Introduction ...... 1 1.1. The need for durable marine coatings ...... 1 1.2. Special requirements for navy vessels ...... 2 1.3. Aims and objectives ...... 3 1.4. References ...... 6 2. Literature Review ...... 9 2.1. The marine biofouling process ...... 9 2.1.1. The fouling challenge ...... 9 2.1.2. The biofouling ‘sequence’ ...... 10 2.1.3. Natural defences against fouling ...... 16 2.2. Antifouling strategies – historical, recent and emerging technologies ...... 21 2.2.1. Historical antifouling solutions ...... 21 2.2.2. Modern antifouling systems...... 22 2.2.3. A summary of antifouling technologies ...... 38 2.3. Natural products as antifouling agents: challenges and considerations ...... 39 2.3.1 Integration of natural product ...... 39 2.3.2. Challenging aspects of natural product usage ...... 41 2.4. Consideration of biocide release rate ...... 44 2.5. References ...... 48 3. Materials and methodology ...... 75 3.1 Fluorescence microscopy techniques: an overview...... 75 3.1.1. Confocal and fluorescence microscopy theory ...... 75 3.1.2. Summary of fluorescence techniques and application to paint coatings ..... 80 3.2. Selection of binders and biocides for LSCM screening ...... 81 3.2.1. Usnic Acid ...... 83 3.2.2. Juglone ...... 83 3.2.3. Bulk testing of binder properties ...... 84 3.3. Preparation of test matrix with selected binders and biocides ...... 84 3.4. Formulation and application of coatings ...... 85 3.5. Dynamic and static testing of coatings...... 88 3.5.1. Dynamic testing protocol ...... 88 3.5.2. Static testing protocol ...... 90 3.6. LSCM stack image acquisition and binary processing ...... 91 3.7. Preparation of cross-sections for microscopy analysis ...... 94 3.8. References ...... 96
iii
4. Results and discussion ...... 99 4.1 Determination of excitation/emission spectra for binders and biocides...... 99 4.1.1. pMMA ...... 99 4.1.2. Usnic acid ...... 99 4.1.3. Copper (I) oxide ...... 103 4.1.4. Rosin ...... 104 4.1.5. Poly(triphenyl methacrylate/butylacrylate) ...... 106 4.1.6. Summary of emission data ...... 107 4.2. Roughness/thickness assessment of primer and topcoat ...... 107 4.1.1. Implications of primer and topcoat roughnesses for LSCM imaging ...... 111 4.2. Pre-immersion usnic acid distribution ...... 113 4.2.1. Assessment of vertical distribution of crystals with LSCM ...... 113 4.2.2. Assessment of vertical distribution of usnic acid with fluorescence microscopy ...... 118 4.2.3. Summary of pre-immersion usnic acid distribution ...... 121 4.3. Results of panel immersion tests ...... 121 4.3.1. Static immersion (NOCS pontoon) ...... 121 4.3.2. Dynamic immersion (TNO, Den Helder) ...... 135 4.4. Post-immersion natural product distribution ...... 146 4.4.1. pMMA coatings ...... 146 4.4.2. Metamare coatings ...... 150 4.4.3. p(TrMA/BA) coatings ...... 153 4.4.4. Summary of post-immersion usnic acid distribution ...... 155 4.5. Overall discussion ...... 156 4.5.1. Erosion, copper oxide distribution and leaching ...... 156 4.5.2. Natural product distribution and leaching ...... 158 4.6. References ...... 162 5. Mathematical modelling of NP diffusion ...... 165 5.1. Background ...... 165 5.1.1. The challenge of modelling marine paint systems ...... 165 5.1.2. Modelling biocide release ...... 166 5.1.2. Selection of a model, and its limitations ...... 168 5.1.3. Description and modification of the Gray-Weale model ...... 171 5.2. Determination of diffusion coefficients for usnic acid in polymers ...... 174 5.3. Determination of diffusion coefficients for usnic acid in seawater ...... 181 5.4. Integration of diffusion coefficients into a simple mathematical model ...... 181
iv
5.5. Comparison of modelling and experimental data ...... 183 5.5.1. Determination of ‘release coefficients’ for experimental data ...... 183 5.5.2. Discussion ...... 184 5.6. Summary of modelling work and results ...... 186 5.7. References ...... 187 6. Conclusions ...... 191 6.1 Further work ...... 193
v
List of Publications and Conferences
Fluorescence microscopy techniques for quantitative evaluation of organic biocide distribution in antifouling paint coatings: application to model antifouling coatings. Biofouling 2012;28(6):613-25. doi: 10.1080/08927014.2012.696103 L R Goodes, S P Dennington, H Schuppe, J A Wharton, M Bakker, J W Klijnstra, K R Stokes.
Designing biomimetic antifouling surfaces. Philos Trans A Math Phys Eng Sci. 2010; 368(1929):4729-54. doi: 10.1098/ rsta.2010.0195. M Salta, J A Wharton, P Stoodley, S P Dennington, L R Goodes, S Werwinski, U Mart, R J Wood, K R Stokes.
Investigation of Chondrus crispus as a potential source of new antifouling agents. International Biodeterioration & Biodegradation doi:10.1016/j.ibiod.2011. 07.002. L D Chambers, C Hellio, K R Stokes, S P Dennington, L R Goodes, R J K Wood, F C Walsh.
Estimation of the leaching rate of an organic biocide using a modified cavity jump diffusion model. Submitted to Biofouling. L R Goodes, J A Wharton, S P Dennington, K R Stokes.
Effect of saturated fatty acid chain length on fouling behaviour In preparation L R Goodes, J A Wharton, S P Dennington
Copper oxide depletion and polishing in pMMA binders In preparation S P Dennington, L R Goodes
15th International Conference on Marine Corrosion and Fouling, June 2010, Newcastle UK. Determination of distribution of paint additives and assessment of their leaching rates using Laser Scanning Confocal Microscopy.
vi
16th International Conference on Marine Corrosion and Fouling, June 2012, Seattle USA. Experimental observation and modelling of biocide release from model antifouling coatings.
Diffusion in Materials 2011, Dijon, France: Organic natural products as biocides for antifouling paints: assessing their integration, leaching and diffusion.
EUROCORR 2010, Nice, France, as delegate
vii
List of Tables
Table 3.1: Test matrix for biocide/binder combinations. ✓represents four coated panels for immersion at NOCS, and two coated panels for rotor erosion at TNO.
Table 4.1: Specimen area sizes for various magnifications across the different instruments.
Table 4.2: Fouling progression summary.
Table 4.3: Calculated polishing rates for pMMA-based coatings.
Table 4.4: Calculated thickness change rates for p(TrMA/BA)-based coatings.
Table 5.1: contribution of atoms to overall Lennard-Jones diameter.
Table 5.2: calculated diffusion coefficients for usnic acid relative to each side group type at relevant swelling values and temperature values.
Table 5.3: Overall calculated diffusion coefficients for each coating type at 10 °C (pontoon immersion) and 25 °C (rotor immersion).
Table 5.4: Estimated diffusion coefficients from modelling results, and experimentally derived release coefficients.
ix
List of Figures
Fig. 1.1: Flow chart demonstrating work areas carried out within this thesis. Green denotes novel areas of study.
Figure 2.1: The flow regime associated with various entities is correlated to their size and speed; at higher Re values, turbulent flow will dominate.
Figure 2.2. Erosion of a contact-leaching type paint system over time.
Figure 2.3: Abietic acid, the primary constituent of rosin.
Figure 2.4. Erosion of a CDP type paint system over time.
Figure 2.5: The saponification process of the TBT monomer in seawater that leads to the self- polishing effect. The polymer shown is a typical self-polishing methylmethacrylate/butylacrylate/TBT copolymer.
Figure 2.6. Erosion of an SPC type paint system over time.
Figure 2.7: Baier curve for relative bioadhesion vs. surface energy (See Baier, 1971, 1973).
Figure 2.8: A timeline for the development of antifouling systems towards modern systems.
Fig 2.9: Where biocide (green) release rate exceeds the erosion of the polymer (blue), the effective lifespan of the coating is reduced.
Figure 3.1: Conventional optical microscope image of natural product crystals in a poly(methyl methacrylate) film: a) surface; b) ~20 µm depth. Note that certain features that are in plane in only one image but are visible in both, despite their spatial separation.
Figure 3.2: Principle of confocal microscopy.
Figure 3.3: Mechanism leading to emission of a lower energy (higher wavelength) photon after absorption of a photon.
Figure 3.4: 1PE illumination of specimen area. Note that all fluorophores in the sample are activated; the pinhole aperture must be deployed to eliminate light from outside the focal plane.
Figure 3.5: Mechanism by which photoactivation is limited to the confocal point during two- photon emission, decreasing laser light attenuation and absorption out of the focal plane.
xi
Figure 3.6: Methyl methacrylate (A), butylacrylate (B) and triphenyl methacrylate (C) monomers.
Figure 3.7: Chemical structure of usnic acid.
Figure 3.8: Chemical structure of juglone.
Figure 3.9: Change in the weight of a Metamare-coated panel over time due to loss of solvent, and the concomitant increase in hardness.
Figure 3.10: TNO Rotor system with panels attached.
Figure 3.11: Coated panels for rotor testing at TNO.
Figure 3.12: Examples of pMMA and Metamare formulation coated panels prior to the first immersion at NOCS pontoon (May 2010). The coatings on individual panels are binder + Cu2O +
FD (leftmost panels), binder + Cu2O (top centre panels), binder + FD (bottom centre panels) and blank binder only (rightmost panels). Each scheme was repeated on the PVC board’s front and back and at two depths.
Figure 3.13: Binary processing of a z-stack to quantify crystal distribution.
Figure 3.14: Top-down view of the 3D reconstruction of a pMMA/furan derivative film (left) on a microscope slide, with reconstructed cross-section (right).
Figure 3.15: LSCM measurements are carried out at three locations on each panel.
Figure 3.16: Representation of paint chip in various orientations, and the possible resulting distortion of measured features.
Figure 3.17: Percentage increase in apparent cross-section thickness resulting from deviation of the paint chip from vertical (i.e. 90˚ to cutting direction).
Figure 4.1: Emission spectra for usnic acid for corresponding excitation wavelengths: 458 nm (violet circles); 476 nm (dark blue squares); 488 nm (light blue point-up triangle) and 514 nm (green point-down triangle). Arrows demonstrate excitation wavelengths.
Figure 4.2: Usnic acid crystals on a microscope slide observed through LSCM. Excitation wavelength is 476 nm.
Figure 4.3: Fluorescence images of a single usnic acid crystal, embedded in pMMA, across a range of excitation wavelengths. Detectors for all excitation wavelengths were set between 450 and 550 nm.
Figure 4.4: 2PE emission spectrum for usnic acid at 920 nm excitation.
xii
Figure 4.5: Emission spectra for juglone at 476 nm (dark blue circles), 488 (light blue squares) and 498 (green-blue triangles). Arrows demonstrate excitation wavelengths.
Figure 4.6: Emission spectrum for copper (I) oxide at 840 nm excitation.
Figure 4.7: Copper (I) oxide agglomerates on a microscope slide.
Figure 4.8: Emission spectrum for rosin at 730 nm excitation.
Figure 4.9: The edge of a Metamare film on a glass slide. Excitation wavelength is 730 nm. Figure 4.10: Rosin crystals viewed by fluorescence microscopy.
Figure 4.11: Emission spectrum for pTrMA at 730 nm excitation.
Figure 4.12: Summary of excitation/emission spectrum data determined for each component. Vertical lines correspond to excitation wavelengths; horizontal lines correspond to the recommended detection ranges.
Fig. 4.13: Decoupled primer roughness (a, b) and waviness (c, d) provided by optical profilometry. Scale bars 1 mm.
Figure 4.14: Optical profilometer microscope image of a small section of primer.
Figure 4.15: Alicona 3D image of a 741 mm x 514 mm primer surface area overlaid with false colour contouring. Total peak-to-trough roughness = 17 µm.
Figure 4.16. Specimen area size against measured peak-to-trough roughness. Error bars are given as 3σ. Linear regression is represented by the black dashed line and 99% confidence levels represented by the red dashed line. The grey lines represent predicted range of the peak-to- trough roughness over the working area for LSCM.
Figure 4.17: Exaggerated representation of film structure demonstrating the effect of primer and film roughness heterogeneities on z-stack scanning.
Figure 4.18: Usnic acid crystal distribution in three locations within pMMA film. The coloured lines delineate the start of the primer interface with the topcoat.
Figure 4.19: Top down composite view of crystals in the pMMA film (1.5 mm x 1.5 mm area). Figure 4.20: Proposed settling mechanism for usnic acid crystals in the binder.
Figure 4.21: Usnic acid crystal distribution in three locations within Metamare film. The coloured lines delineate the start of the primer interface with the topcoat.
Figure 4.22: Top down composite view of crystals in the Metamare film (1.5 mm x 1.5 mm
xiii
area).
Figure 4.23: Usnic acid crystal distribution in three locations within p(TrMA/BA) film. The coloured lines delineate the start of the primer interface with the topcoat.
Figure 4.24: Top down composite view of crystals in the p(TrMA/BA) film (1.5 mm x 1.5 mm area).
Fig. 4.25: Average fluorescence intensity measurements for 10 month aged, unimmersed pMMA (blue, short dashed), CDP (black, solid) and p(TrMA/BA) (red, long dashed) coatings containing usnic acid (left) and for blank binders (right). Boundaries represent 90% confidence intervals. FITC filter block was employed for all measurements.
Fig. 4.26: Fluorescence micrograph of p(TrMA/BA)/FD coating on PVC substrate (FITC filter). FI profile associated with the solid line (drawn perpendicular to coating) is demonstrated below. Figure 4.27a: Fouling progression of pMMA panels (front) over 10 months from May to March.
Figure 4.27b: Fouling progression of pMMA panels (back) over 10 months from May to March.
Figure 4.27c: Fouling progression of Metamare panels (front) over 10 months from May to March.
Figure 4.27d: Fouling progression of Metamare panels (back) over 10 months from May to March
Figure 4.28: The back of a board, demonstrating proliferation of sea squirts around panels. Note that tubeworms dominate the area between panel schemes.
Figure 4.29: Cross-sections of CDP/Cu2O/FD panels from rear (left) and front (right) of the exposed board. Scale bars = 50 μm.
Figure 4.30: Thickness change of pMMA and reference epoxy coatings with time: a) pMMA
Blank; b) pMMA/FD; c) pMMA/Cu2O; d) pMMA/Cu2O/FD; e) reference non-polishing epoxy coating.
Figure 4.31: Cross-sections from pMMA/Cu2O/FD (top) and pMMA/FD (bottom) coatings after 6 months of rotor immersion.
Figure 4.32: Surface of pMMA/Cu2O coating, with pale copper-depleted areas clearly visible (solid circle) and surface debris from rotor immersion (dotted circle). Scale bar 50 μm.
Figure 4.33: Thickness change of Metamare coatings with time: CDP/Cu2O (left) and
CDP/Cu2O/FD (right).
Figure 4.34: Cross-section of wax-embedded CDP/Cu2O/FD from the rotor exposure trials. Note the attached primer layer that remains adhered to the coating in the second image. Scale bar
xiv
100 μm.
Figure 4.35: Thickness change of p(TrMA/BA) copolymer coatings with time: a) p(TrMA/BA)
Blank; b) p(TrMA/BA/FD; c) p(TrMA/BA)/Cu2O and d) p(TrMA/BA)/Cu2O/FD.
Fig. 4.36: Cross-sections from p(TrMA/BA)/Cu2O (top) and p(TrMA/BA)/FD coatings after 6 months of rotor immersion. Figure 4.37: FTIR reflectance spectrum for eroded p(TrMA/BA) coating, before cleaning.
Figure 4.38: FTIR reflectance spectrum for eroded p(TrMA/BA) coating, after cleaning with tissue paper (A) and scourer (B). Receding peaks compared to the un-cleaned panel marked by black arrows.
Figure 4.39: a): Optical cross-section image of pMMA/FD film after 6 months of rotor immersion; b): fluorescence microscopy image of unimmersed pMMA/FD film cross-section; c): fluorescence microscopy image of pMMA/FD film cross-section after 6 months of rotor immersion. The fluorescent signal from the mean of three cross-sectioned samples for unimmersed (solid line, black) 6 month eroded (dotted line, grey) pMMA/FD coatings is shown right. Shaded areas represent 90% prediction intervals for the mean lines.
Figure 4.40: a) Fluorescence cross-section of pMMA/FD coating before-immersion, b): after 10 months of pontoon immersion from the shaded back side of the board, c): after 10 months of pontoon immersion from the lit front side of the board. Scale bars 200 μm. The average signal from line profiles associated with three cross-sectioned samples of unimmersed control (solid line, black), rear-mounted pMMA/FD panels (long dashed line, red), front-mounted panels (short dashed line, grey) and blank pMMA panel (solid line, blue) is shown (right) including 90% prediction intervals.
Figure 4.41: Fluorescence cross-sections of pMMA/Cu2O (a) and pMMA/Cu2O/FD (b) films after
10 months of static immersion. Optical microscope image of the pMMA/Cu2O/FD cross-section is also shown (c), with an attached thin algal film on the surface (~10 μm in thickness). Scale bars 200 μm. The average signal from line profiles associated with three cross-sectioned samples of pMMA/Cu2O/FD (long dashed line, grey), pMMA/Cu2O (short dashed line, red) and blank pMMA panel (solid line, blue) is demonstrated (right) including 90% confidence intervals.
Figure 4.42: CDP/FD cross-sections from unimmersed control (top left) and rear-mounted (middle left) and front-mounted (bottom right) after 10 months of static immersion. Mean fluorescence intensity profiles for two front (short dashed, grey) and two back panels (medium dashed line, red) are demonstrated relative to the non-FD bearing blank binder (long dashed, blue) with 90% confidence intervals. The fluorescence intensity relating to the unimmersed control sample is denoted by the solid black line. Scale bars 200 μm.
Figure 4.43: Fluorescence cross-sections of various CDP films after 10 months of pontoon immersion: a) Blank CDP (rhodamine), b) CDP/Cu2O (rhodamine), c) CDP/Cu2O/FD (rhodamine), d) CDP/Cu2O/FD (GFP) demonstrating leached layer not visible with rhodamine filter block. All
xv
scale bars 200 μm. The average signal from line profiles associated with three samples of each binder type are shown (right) including 90% prediction intervals: Blank CDP binder (short dashed line, blue), CDP/Cu2O (long dashed line, red), 10 month immersed CDP/Cu2O/FD (solid line) and unimmersed CDP/Cu2O/FD (dotted line, yellow).
Figure 4.44: p(TrMA/BA)/FD coating – un-immersed (a), 3 month erosion scheme (b) and 6 month erosion scheme (c). The mean signal derived from multiple line profiles across three cross-sections from each panel are shown right – un-immersed (solid line, grey) 3 months (long dashed line, red) and 6 months (dotted line, cyan).The fluorescence intensity from blank p(TrMA/BA) is denoted by the short dashed line (black). Scale bars are 100 μm.
Figure 4.45: Fluorescence intensity in rotor-aged p(TrMA/BA)/Cu2O/FD coatings; after 3 months of immersion (solid red line, 90% confidence intervals shown by dashed red lines) and 6 months of immersion (solid black line, 90% confidence intervals shown by dashed black lines).
Figure 4.46: Calculated Hansen solubility parameters for pMMA, TrMA and usnic acid.
Figure 5.1: Paint binder pores created by rosin and pigment dissolution (adapted from Yebra et al.1).
Figure 5.2: In heavily congested networks where the diluent comprises a low volume of the overall system, the properties of the network dominate diffusivity. Here, the yellow ball represents the migrant, blue the diluent, and black the polymer or protein network. A: free diffusion of a migrant in a liquid; B: diffusion in a network containing a low volume content of fibres or barriers; C: diffusion in a network containing a high volume content of fibres or barriers.
Fig 5.3: Reactant-transition-reactant cavity jump demonstrating the number of interacting units (4-3-4) within each stage. Note that owing to limitations in the model, only one side group type can be considered at a time.
Figure 5.4: Chemicalize simulations of molecular geometry for usnic acid (left) and juglone (right).
Figure 5.5: Lennard-Jones and deformation potentials for CH3 and CO2CH3 side groups in the case of pMMA (0% swelling).
Figure 5.6: Effect of the geometric swelling parameter (expressed as a percentage volume increase) on the calculated diffusion coefficient for pMMA.
Figure 5.7: Modelled FD depletion from a static pMMA film 100 µm in thickness, at 10 ˚C.
Figure 5.8: Modelled FD depletion from a static pMMA film 100 µm in thickness, at 25 ˚C.
xvi
Declaration of Authorship
I, Liam Rhys Goodes declare that the thesis entitled ‘OBSERVATION AND PREDICTION OF BIOCIDE RELEASE WITH FLUORESCENCE TECHNIQUES AND MATHEMATICAL MODELLING’ and the work presented in the thesis are both my own, and have been generated by me as the result of my own original research. I confirm that:
this work was done wholly or mainly while in candidature for a research degree at this University;
where any part of this thesis has previously been submitted for a degree or any other qualification at this University or any other institution, this has been clearly stated;
where I have consulted the published work of others, this is always clearly attributed;
where I have quoted from the work of others, the source is always given. With the exception of such quotations, this thesis is entirely my own work;
I have acknowledged all main sources of help;
where the thesis is based on work done by myself jointly with others, I have made clear exactly what was done by others and what I have contributed myself;
parts of this work have been published as described in the List of Publications.
Signed: ……Liam Goodes………………………………………………………..
Date:………30/06/2013…………………………………………………………….
xvii
Nomenclature
1PE = one-photon excitation ACWS = antifouling coatings for warships AF = antifouling BA = butylacrylate CDP = controlled depletion polymer (i.e. erodible, rosin-based binder system) EDX = energy dispersive x-ray spectroscopy FD = furan derivative (i.e. usnic acid) FI = fluorescence intensity FM = fluorescence microscopy FRC = foul release coating FTIR = Fourier transform infrared spectroscopy HPLC = high performance liquid chromatography HSP = Hansen solubility parameter LSCM = laser scanning fluorescence microscopy MMA = methylmethacrylate NP = natural product pMMA = poly(methylmethacrylate) p(TrMA/BA) = poly(tritylmethacrylate/butylacrylate) SEM = scanning electron microscopy SPC = self-polishing copolymer TBT = tributyltin TrMA = tritylmethacrylate ( = triphenylmethacrylate) h = Planck constant, 6.6261 × 10-34 m2 kg s-1
E0 = critical energy, J
Tb: boiling point, K ε: well depth, K σ: Lennard-Jones diameter, Å
-23 -1 kB: Boltzmann constant, 1.381 × 10 J K -1 Vdef: deformation potential, kJ mol -1 VLJ: Lennard-Jones potential, kJ mol ν: Poisson’s ratio
xix
µ: Lamé constant E: Young’s modulus, Pa ɸ: percentage volume increase of film from swelling
RS ri : initial reactant state cavity radius, Å TS ri : initial transition state cavity radius, Å
Tg: glass transition temperature, ˚C RS 3 V0 : initial reactant state cavity volume, Å TS 3 V0 : initial transition state cavity volume, Å h0: initial height of transition state cavity, Å RS rɸ : reactant state cavity radius with respect to swelling, Å TS rɸ : transition state cavity radius with respect to swelling, Å
xx
Chapter 1 Introduction
1. Introduction
1.1. The need for durable marine coatings
The need for robust and long-lived durable coatings to protect ships and boats from marine fouling has been a primary concern of seafarers since the earliest shipping endeavours. The marine environment is aggressive in terms of its dynamicity and the properties of the seawater itself, so the development and optimisation of coatings that can protect a vessel from corrosion and biofouling over extended periods is a permanent goal of the marine paint industry. Biofouling is the adhesion of marine organisms to the submerged part of a surface, resulting in severe speed and fuel penalties and significantly increasing operational costs and harmful emissions in the case of fouling on ship hulls, as well as possibly transferring invasive species to a new area.
Effective paint coatings for the protection of artificial submerged surfaces have been previously developed. In the simplest terms, an antifouling (AF) coating comprises one or more antifouling agents (biocides) held with a polymer matrix (binder) which acts as a delivery system. The biocides are selected to provide an effective and broad AF spectrum, and the binder is tailored to enable a slow release of the biocides over a prolonged period; coatings are expected to last in excess of three to five years in service, depending on type. Although fouling has been a concern for generations of seafarers (see chapter 2), AF coatings originally featured metallic copper as the primary active biocide, and latterly organic tin. However, as use of these highly toxic tin-based paint systems became more widespread, the lasting adverse effects of tin contamination on marine ecosystems became evident. As a consequence, most countries have been forced to revert to new iterations of original copper-based coating technologies - at a cost to the economic efficiency of the naval and commercial shipping industries. Furthermore, long- standing concerns about the long-term effects of copper pollution in environmentally important estuaries and offshore areas are providing an incentive to study new fouling solutions. Coatings based on low surface energy - which result in the pessimal adhesion of marine organisms, facilitating their removal - have grown in popularity over the last decade. However, the development of alternative biocidal systems as complementary AF solutions is overdue. This lag in technology results from complex and stringent legislation, as well as the numerous complexities surrounding elucidation and testing of new biocides. One of the main focal points over the previous fifteen years has been natural products (NPs) derived from the
1
Chapter 1 Introduction evolutionary adaptations of marine biota, some of whom possess effective chemical defences. The pursuit of coating development with these compounds has been hindered largely by a lack of understanding of the mechanisms by which they are removed from the binder, a process known as leaching. In fact, even the leaching of key commercial biocides is a subject rarely encountered in the literature; coating development has taken the route of very high throughput methods to locate the ideal composition and lifetime of coatings1-3.
1.2. Special requirements for navy vessels
The majority of the work within this thesis was carried out under the auspices of the European ACWS (Antifouling Coatings for Warships) project and funded by the DSTL (Defence Science and Technology Laboratory). This endeavour is contributing to the investigation and development of an NP-based antifouling coating system for use on European Naval vessels - in particular the French Navy - with the British Royal Navy keeping a watch on emerging industrial technologies.
As well as additional costs incurred by biofouling, the manoeuvrability penalty and increased sound signature resulting from accumulated biomass are particularly problematic for navy vessels4. Even slight fouling of sensitive areas containing sensor apparatus at the aft of a vessel can result in suboptimal performance of the equipment5 The operational profile for these ships is uniquely demanding, since they must have the potential to traverse nearly any latitude, withstanding the associated dramatic changes in temperature and sea state, whilst resisting any fouling species that could be encountered while in service. Naval ships are thus potentially exposed to a wider range of fouling and sea conditions than any other category of seagoing vessel.
In addition, protection must be offered whilst the vessel is standing in-dock as prolonged stand-by periods are characteristic of navy warships, which spend over 50% of their life alongside4. The vessel must also be capable of rapid deployment, which is hindered by the in- dock accumulation of foulers. A survey of 41 US naval vessels brought alongside for cleaning was carried out in 2009/20106. They exhibited a range of fouling states from unfouled to extensive calcareous fouling communities - an ‘artificial reef’. As well as the performance of the vessel, its dock-side appearance is also important, and the growth of unsightly algal slime or hard foulers is undesirable4. For these reasons, current coating technology which relies on a reduction of adhesive strength of foulers, facilitating their removal through motion of the vessel (see Section 2.2.2.4) is not always well-tailored for purposes on navy warships, as time
2
Chapter 1 Introduction and energy investments are required to remove accumulated fouling from each ship’s hull during and after docking. Despite these concerns, foul release coatings (FRCs) are becoming more widely used by naval authorities, owing to constant improvements in the technology. Nonetheless, biocidal mechanisms of fouling removal are likely to remain in widespread use for certain operational profiles for the foreseeable future.
All seagoing vessels require routine maintenance and repair. Within the French and UK navies, the interval between in-dock refitting and maintenance sessions is scheduled to double from 2- 3 to 5-6 years. The development of an ‘environmentally acceptable’ AF coating capable of lasting at least 5 years before reapplication is therefore desirable.
1.3. Aims and objectives
NPs are a promising branch of potential antifoulants, but their viability as active AF agents is hampered by the difficulties in integrating them into an effective and long-lived coating. Coatings containing NPs are often observed to perform badly in the field, fouling within a couple of months, despite promising laboratory assays7, 8. As a result of these setbacks, research interest in NPs has declined slightly over the last few years. The shortcoming of these coating systems lies not in the efficacy of the NP, but in the many unknown quantities regarding its distribution, leaching rate and affinity with the polymer binder. Information on distribution and leaching rate would allow an explanation of the poor performances observed in field trials, subsequent mitigation of the formulations’ weaknesses, and finally optimization to prolong in-situ coating lifetime.
The poor field performance of NPs has not been sufficiently studied in the literature; although an appreciation of the difficulties in NP integration has been gained, virtually nothing is known about the distribution or leach rates of the NPs once incorporated into binders. The distribution of traditional pigments and heavy metal biocides is readily assessed through use of scanning electron microscopy and energy dispersive x-ray spectroscopy (SEM-EDX) applied to coating cross-sections9. The presence of a ‘leached layer’ also permits distinction of the leaching boundary for copper oxide via optical microscopy of the cross-section; the leached layer will bleach due to loss of pigment. These are largely time-consuming and laborious procedures necessitating the fracturing of the test specimen, embedding and mounting in resin, polishing to a fine finish, and finally microscope observation, but the worth of these procedures in demonstrating the development of copper oxide leach fronts has been shown.
3
Chapter 1 Introduction
However, these methods are unsuitable for the detection of lighter, non-metallic atoms or molecules, such as natural biocides consisting generally of hydrocarbons, N, O, and in rare cases, halogens or thiol groups.
The principal aim of this study was to develop a method of assessing the distribution of NPs within a polymer binder, and to use this knowledge to calculate the leach rate of the NP from immersed samples. The first involved the attempted staining of fatty acids with barium sulphate in combination with elemental mapping by EDX, in freeze-fractured cross sections of polymer films. However the method proved unsuccessful, and the fatty acids were also found to provoke an increase in fouling compared to blank and control panels. More success has been achieved with the use of fluorescence microscopy techniques as a mapping technique for characterisation of additive distribution within the binder. The development and employment of this technique within the context of a long term immersion trial was one of the two main work elements within this thesis.
Within the present study, several binders and biocides identified within the ACWS project have been used to formulate paint mixtures for simple AF coatings. The lichen metabolite usnic acid10-15 and walnut extract juglone16-20 were selected as two potential AF agents for incorporation into simple binders based on successful bioassay trials carried out by other ACWS partners. The usnic acid was the first isolated NP extract considered for large-scale testing and further exploration; therefore, the majority of the work carried out as part of this thesis was performed using that biocide. These coatings were studied on raft immersions (static testing, carried out at the National Oceanography Centre, Southampton) and on a rotor system (dynamic testing, carried out at the TNO Maritime Materials Performance Centre, Den Helder, NL). Novel applications of fluorescence microscopy techniques have been employed to characterise the distributions of these additives within the binder. The work described in this thesis demonstrates that the distribution of biocides exhibiting fluorescent behaviour may be characterised by studying cross-sections of coatings21. Furthermore, a first attempt at mathematical modelling of NP release has been attempted, through the adaptation of existing literature models for diffusion of small molecules in glassy polymeric systems22, 23. The results from these models have been compared with natural and accelerated immersion regimes to gain an understanding of the mechanisms controlling biocide release. A flow chart following the major blocks of work carried out in this thesis is presented in Fig. 1.1.
4
Fig. 1.1: Flow chart demonstrating work areas carried out within this thesis. Green denotes novel areas of study.
5
Chapter 1 Introduction
1.4. References
1. Majumdar, P., Crowley, E., Htet, M., Stafslien, S.J., Daniels, J., Van der Wal, L., and Chisholm, B.J., Combinatorial Materials Research Applied to the Development of New Surface Coatings XV: An Investigation of Polysiloxane Anti-Fouling/Fouling-Release Coatings Containing Tethered Quaternary Ammonium Salt Groups. ACS Combinatorial Science, 2011. 13(3): p. 298-309. 2. Majumdar, P., Lee, E., Patel, N., Stafslien, S., Daniels, J., and Chisholm, B., Development of environmentally friendly, antifouling coatings based on tethered quaternary ammonium salts in a crosslinked polydimethylsiloxane matrix. Journal of Coatings Technology and Research, 2008. 5(4): p. 405-417. 3. Majumdar, P., Lee, E., Patel, N., Ward, K., Stafslien, S.J., Daniels, J., Chisholm, B.J., Boudjouk, P., Callow, M.E., Callow, J.A., and Thompson, S.E.M., Combinatorial materials research applied to the development of new surface coatings IX: An investigation of novel antifouling/fouling-release coatings containing quaternary ammonium salt groups. Biofouling, 2008. 24(3): p. 185 - 200. 4. ACWS, Anti-fouling Coatings for Warships - First Selection of Products - D001, 2008, DSTL. p. 92. 5. Ramotowski, T., Tucker, W., Rice, M., Haslbeck, E., and Maranda, L. New, biofouling- resistant elastomers for acoustic applications. in 15th International Congress on Marine Corrosion and Fouling. 2010. Newcastle, England. 6. Haslbeck, E., Griggs, D., Gardner, J., Holm, E., Stamper, D., Lynn, D., Peyton, C., Curran, J., Vestal, J., and Lawrence, S. US Navy evaluation of a foul release coating: biofouling control, physical performance, and impact on fuel economy. in 15th International Congress on Marine Corrosion and Fouling. 2010. Newcastle, England. 7. Chambers, L., The development of a marine antifouling systems using environmentally acceptable and naturally occurring products, in Surface Engineering and Tribology Research Group2008, University of Southampton. p. 210. 8. Burgess, J.G., Boyd, K.G., Armstrong, E., Jiang, Z., Yan, L., Berggren, M., May, U., Pisacane, A., Granmo, Å., and Adams, D.R., The development of a marine natural product-based antifouling paint. Biofouling, 2003. 19: p. 197-205. 9. Faӱ, F., Linossier, I., Langlois, V., Haras, D., and Vallée-Réhel, K., SEM and EDX analysis: Two powerful techniques for the study of antifouling paints. Progress in Organic Coatings, 2005. 54: p. 216-223. 10. Burkholder, P.R., Evans, A.W., McVeigh, I., and Thornton, H.K., Antibiotic activity of lichens. Proceedings of the National Academy of Sciences of the United States of America, 1944. 30(9): p. 250-255. 11. Cocchieto, M., Skert, N., Nimis, P.L., and Sava, G., A review on usnic acid, an interesting natural compound. Naturwissenschaften, 1982. 89(4): p. 137-146. 12. Francolini, I., Norris, P., Piozzi, A., Donelli, G., and Stoodley, P., Usnic acid, a natural antimicrobial agent able to inhibit bacterial biofilm formation on polymer surfaces. Antimicrobial agents and chemotherapy, 2004. 48(11): p. 4360-4365. 13. Ingólfsdóttir, K., Usnic acid. Phytochemistry, 2002. 61(7): p. 729-736. 14. Kristmundsdóttir, T., Jónsdóttir, E., Ögmundsdóttir, H.M., and Ingólfsdóttir, K.,
6
Chapter 1 Introduction
Solubilization of poorly soluble lichen metabolites for biological testing on cell lines. European Journal of Pharmaceutical Sciences, 2005. 24(5): p. 539-543. 15. Hauck, M. and Jürgens, S.-R., Usnic acid controls the acidity tolerance of lichens. Environmental Pollution, 2008. 156(1): p. 115-122. 16. Cutler, H.G., Belson, N.A., Dawson, R., and Wright, D.A., Method for treating aquatic pests, 2000, Pharmacogenetics, Inc. 17. Dana, M.N. and Lerner, B.R., Black walnut toxicity, P.U.D.o. Horticulture, Editor 2001, Purdue University Cooperative Extension Service: West Lafayette, IN, USA. p. 1-2. 18. Khodzhibaeva, S., Filatova, O., and Tyshchenko, A., New aspects of the preparation and control of juglone. Chemistry of Natural Compounds, 2000. 36(3): p. 281-283. 19. Randall, V.D. and Bragg, J.D., Effects of juglone (5'-hydroxy-l,4-naphthoquinone) on the algae Anabaenaflos aquae, Nostoc commune and Scenedesmus acuminatus. Proceedings: Arkansas Academy of Science, 1986: p. 52-55. 20. Uchimiya, M. and Stone, A.T., Reversible redox chemistry of quinones: Impact on biogeochemical cycles. Chemosphere, 2009. 77(4): p. 451-458. 21. Goodes, L.R., Dennington, S.P., Schuppe, H., Wharton, J.A., Bakker, M., Klijnstra, J.W., and Stokes, K.R., Fluorescence microscopy techniques for quantitative evaluation of organic biocide distribution in antifouling paint coatings: application to model antifouling coatings. Biofouling, 2012. 28(6): p. 613-625. 22. Gray-Weale, A.A., Henchman, R.H., Gilbert, R.G., Greenfield, M.L., and Theodorou, D.N., Transition-state theory model for the diffusion coefficients of small penetrants in glassy polymers. Macromolecules, 1997. 30(30): p. 7296-7306. 23. Tonge, M.P. and Gilbert, R.G., Testing models for penetrant diffusion in glassy polymers. Polymer, 2001. 42: p. 501-513.
7
Chapter 2 Literature Review
2. Literature Review
2.1. The marine biofouling process
2.1.1. The fouling challenge
Marine biofouling – the undesired attachment of organisms to structures immersed wholly or partially in the sea1, 2 – is a well-documented phenomenon dating back to the earliest shipping endeavours3. No known artificial surface yet created is immune to fouling4. Marine organisms have evolved in a number of diverse ways to withstand the aggressive marine environment and are capable of attaching to exposed manmade structures such as struts and buoys, and even ships moving at high speeds. Severe biofouling can result in a large increase in weight; studies suggest that 505 - 1506 kg m-2 of biofouling may accumulate over a 6 month period in particularly fouling-prone areas such as the tropics.
Adhesion of smaller pioneering organisms (microfoulers) can facilitate the recruitment of larger organisms (macrofoulers), decreasing the hydrodynamicity of the vessel and causing heightened drag7-10. Macrofoulers such as barnacles are particularly likely to penetrate and grow into the paint and primer layers if left unchecked, causing a plaque of paint to be lifted away11. The stress induced in the paint film around the circumference of barnacle ‘feet’ can induce cracking, and the adhesion of the coating to the primer is perturbed. Damage to the primer layer may also result, exposing the hull surface to the effects of corrosion. The fouling process and mechanisms of adhesion are discussed in detail in the next section.
As additional fuel is required to compensate drag resulting from fouling, the overall voyage costs increase. The additional costs implied by a fouling community, in the hypothetical case of an unprotected ship, were modelled by Gitlitz (1981): a very large cargo carrier (VLCC) spends around 300 days at sea per year, using 170 tonnes of heavy marine fuel (fuel oil) per day at an average speed of 15 knots (roughly 7.5 m/s). The cost of a tonne of marine fuel in 2008 averages around $25012, compared to $80 in 198113. The cost of fuel alone is therefore in the region of $12-13 million dollars per year. Given that moderate to severe fouling on the vessel imposes an arbitrary 30-50% penalty on fuel consumption to maintain an average of 15 knots14, the additional fuel costs incurred can be estimated at $4-6.5 million/y, based
9
Chapter 2 Literature Review on the model of Gitlitz as described above. In contrast, the cost of coating a similar vessel with a modern AF paint coating is less than $140,000/y15, representing a theoretical saving of at least $3.86 million/y for each vessel in full service. This excludes the costs that could be incurred by the possible corrosive or structural damage caused to the hull by heavy fouling. The US Navy alone is estimated to expend more than $1 billion/y on combating fouling1, with estimates of over $10 billion/y expenditure by marine industries worldwide16. Combined with costly dry-docking periods for maintenance and manual removal of organisms, the costs incurred by biofouling will reduce profit margins significantly17, 18.
Biofouling is also undesirable in that it may result in the dispersal of allochthonous (i.e. alien or non-native) organisms or their larvae and juveniles between ports or countries19-21. These alien species may colonise and dominate foreign ecosystems, particularly in the case of aggressive, opportunistic species such as Crepidula fornicata22. The problem of the transport of alien species via fouled hulls has only now begun to be confronted, with traditional legislative drivers targeting translocation of larval species and spores in ballast water23. Despite this, numerous recent studies have indicated that in many locations, the impact of hull fouling on species transport may be greater than that of ballast water 24 and references therein, 25; an untreated hull demonstrates a broad species richness and diversity of fouling species, of which more than 40% may be of an invasive nature26. On a properly protected vessel, fouling can still occur in regions that escape initial paint coverage (e.g. seachests25), or upon depleted or damaged AF coatings27. Even scratches as small as 5 mm2 on the surface of a coating are sufficient to harbour fouling taxa which can then spread24. Damage to the outer surface of vessels’ hulls during service is inevitable (e.g. from rope guard/anchor chain abrasion), and can easily provide a niche for foulers’ colonisation27. This threat to coastal ecosystems, which are already at risk from anthropogenic inputs related to boating and industrial activities28-30, is another driving factor for the development of efficient and environmentally friendly AF coatings.
2.1.2. The biofouling ‘sequence’
Any substrate immersed in seawater will rapidly attract adsorbed solutes and small particles, preceding colonisation by microscopic marine bacteria such as Marinobacter hydrocarbonoclasticus and Cobetia marina31. Pioneer organisms initially coming into contact with the substrate by means of turbulence and Brownian motion32, 33 develop adsorptive Van der Waals forces within the viscous boundary layer, permitting attachment. There is
10
Chapter 2 Literature Review evidence to suggest that certain bacteria can sense their proximity to a surface and actively colonise in a species-specific manner34-36. No surface, artificial or natural, is entirely immune to this kind of attachment by macromolecules and bacteria. The process of macromolecule and bacterial absorption is often termed ‘conditioning’, as it facilitates the settlement of higher flora and fauna7.
Diatoms - a major class of phytoplankton - colonise immersed surfaces and secrete extracellular polymeric substances (EPS1) within a short time scale, generally a few hours to a day32, 37-45. EPS include polysaccharides, lipids, nucleic acids, proteins and general elastic polymers which are secreted by the collective of diatoms and bacteria, and comprise 95- 98% of a mature biofilm’s volume; only the remaining 2-5% is made up of the organisms themselves46. EPS are the primary enablers of bacterial and diatomic adhesion to ‘inert’ and smooth surfaces, as they conform extremely closely to the topography of the colonised surface and form a sticky, gel-like matrix that holds the organisms in place34, 47. EPS are excreted via slits in the siliceous valves, and larger, elongated slits called ‘raphes’ in the frustules of these organisms48, 49. Some diatom species, such as Amphora, are able to strongly adhere even to Foul Release Coatings (see Section 2.2.2.6) at relatively high speeds by producing great quantities of EPS, and are often seen to dominate biofilm populations on these surfaces. Other behavioural adhesion methods include ‘rafting’, the process of accumulation of a large number of individuals of a single species such as Achnanthes that produce large mats of overlapping adhesive matter that facilitates adhesion and resistance to shear stress. Amphora, Navicula and Achnanthes spp. are considered the major fouling diatom species 48.
The EPS matrix provides other benefits to the microorganism; it may also isolate it from the toxic biocide50, 51, and allow mechanical trapping of nutrients in the water column. The EPS matrix renders the bacterial and diatomic component of the fouling community extremely resistant to many toxic surfaces52, 53. The accumulated organisms and EPS form a visible, slippery layer known as a biofilm or ‘slime’, the architecture of which is a complex system of towers and channels, and which depends on the order of settlement, differing motility of the colonising bacterial species, as well as competition between sub-colonies and species31. The biofilm and slime may generally be removed easily by physically wiping the surface of the hull clean, but is highly resistant to shear stress resulting from movement of the
1 EPS is taken as a contraction of exopolysaccharides in some literature. 11
Chapter 2 Literature Review vessel54.
The biofilm surface provides an excellent substrate for 'macrofouling' - the colonisation by higher fauna or larvae, ranging from seaweeds such as Ulva sp. to hard foulers such as Mylitus edulis (common mussels) and Balanus sp. (barnacles). Owing to their size, shape and roughness, macrofouling organisms impart a far greater drag on the vessel than slime alone, requiring a large excess of fuel expenditure to maintain the desired speed (Fig. 2.1). The larvae of these organisms are highly motile and are between 5-10 µm in size. They ‘explore’ surfaces until they find a suitable surface for colonisation, preferably with a niche similar to their body size55, 56. Once a suitable area is found, the larvae spin and secrete an elastic material on a patch known as a ‘footprint’. They will generally settle on this patch, secreting further material along with a glycoprotein to enable permanent adhesion. Footprints are believed to attract further individuals once secreted57, 58. It should be emphasized that many macrofoulers do not require the presence of a biofilm in order to colonise a surface; barnacle cyprids, ascidians and bryozoans can settle on ‘clean’ surfaces almost instantaneously59. Macrofoulers have been observed to colonise surfaces before or during the development of biofilm60, and bacterial colonisation is not necessarily a prerequisite of diatom settlement61. However, it is true that certain species like tubeworms are rarely observed without a biofilm precursor (J. Callow, pers. comm.). It is often incorrectly assumed that biofouling follows the ‘classical’ sequence strictly, in all cases. For example, the Baltic Sea is extremely prone to rapid barnacle fouling, but algal fouling is very uncommon62. Such changes in the composition of regional biofouling communities require a different approach to fouling; in the case of the Baltic Sea, the proliferation of barnacles has led to the development of a highly effective anti-barnacle biocide in Sweden63-67.
12
Chapter 2 Literature Review
Figure 2.1: The flow regime associated with various entities is correlated to their size and speed; at higher Re values, turbulent flow will dominate.
As a fouling community matures and macrofoulers settle and grow on the film, the weight of the vessel increases and the rough surface generates additional frictional resistance whilst in motion. Ludgate68 estimated that given 33% barnacle fouling cover, 50% additional fuel expenditure would be incurred to maintain speed when compared to the unfouled hull. Estimates for the amount of additional fuel consumption required vary widely, however; values of 30-40% additional fuel expenditure are commonly cited in recent literature14, whilst Schultz9 demonstrated that an extra 86% shaft power is required in the case of a vessel at cruising speed (15 knots) with heavy calcareous fouling, assuming homogeneous distribution of fouling. The influence of fouling distribution on local and overall drag increase is currently under investigation by the same author. Townsin54 claims that 5% barnacle coverage is sufficient to cause a 66% increase in drag, whereas 75% barnacle coverage gives an 85% increase. The increased drag imparted by calcareous fouling communities with barnacles and mussels is relatively well established, though the contribution of the intermediate biofilm, comprising bacteria and diatomaceous slime, is highly variable8-10, 69. Traditional estimates of mature biofilm contribution to drag range from 1-10%, although the most recent evaluations place the additional resistance incurred at 11- 13
Chapter 2 Literature Review
20%9. The main areas for future study include the relative contribution of vessel hull areas to overall drag, and the modelling of hydrodynamics relating to both hard and soft fouling on the scale of individual organisms9, 70, permitting a more realistic model of a non- heterogeneously fouled vessel. Although modelling flow around a static hard fouler should be relatively simple, modelling of flow in the case of spongy and amorphous biofilm and weed fouling promises to be extremely challenging.
Over 4000 species are now estimated to contribute to persistent biofouling71. Not all of these species are present in any given water mass, although certain species, such as various barnacles and Ulva sp., are prevalent in many coastal regions72. The threat posed by a given fouler can be considered to be relative to – firstly - its abundance in the water mass frequented in a vessel’s operational profile and, secondly, the severity of the increase in drag or other problems that result from its attachment to a vessel. The ‘high impact’ foulers are therefore those that not only occur readily in a large majority of water bodies, but also have a significant effect on the performance of the vessel.
2.1.2.2. Controls on biofouling occurrence and intensity
2.1.2.2.1. Temporal variability of oceanic production
The range and abundance of species encountered in a location is determined by the physical and oceanographic properties of their host water body; the foremost of these are temperature, salinity, light availability (for phototrophic organisms) and nutrient availability (for heterotrophic organisms). The abundance of most species follows an annual cycle, as they are dependent upon the production of phytoplankton - at the bottom of the food- chain - which in turn follow cycles of growth and depletion according to seasonal light variation and nutrient availability, the latter generally being greater in coastal regions where run-off from precipitation is fed through estuaries to the sea, as well as mineral erosion products from the fluvial system. Nutrients such as nitrate, phosphate and silicic acid as well as numerous important micronutrients (e.g. iron) are crucial to the metabolic pathways involved in photosynthesis73 The main phytoplankton bloom occurs in spring in the northern hemisphere, as these key nutrients accumulate during the winter months, when light intensity and duration is lowest, and uptake from plankton communities is lowest (the growth of phytoplankton is light-limited). As light availability increases towards the end of spring (April-June), planktonic communities are no longer limited by either nutrient supply
14
Chapter 2 Literature Review or light, and can begin to thrive. Carbon dioxide from the atmosphere is partially drawn down into the sea as it reaches pressure equilibrium, where it partitions among carbon species such as carbonic acid, bicarbonate, and dissolved ‘free’ carbon dioxide.
Phytoplankton use this dissolved CO2 and energy from sunlight to produce energy, with oxygen ejected as a toxic by-product318. Nutrients in the water are then depleted throughout the summer until very low (in the order of picomolar) concentrations remain; phytoplankton die back in response as they run out of nutrients to fuel their metabolic processes, becoming nutrient-limited. Generally, natural nutrient supplies are depleted by the end of summer, resulting in falling plankton biomass concentrations despite good weather conditions. Plankton communities may therefore be nutrient-limited as light exposure decreases throughout the autumn, and switch to light-limitation whilst nutrient supply increases with higher precipitation during this period. Nutrients continue to accumulate throughout the dark winter months, completing the yearly cycle318.
The abundance of the phytoplankton’s main predator species – zooplankton - increases in keeping with primary (i.e. autotrophic) production, often displaying a short lag as seed zooplankton communities take time to adapt to an increase in their food source. The abundance of most fouling organisms in ocean environments (and therefore higher species in the food chain) is at its peak at the end of spring, and at its lowest in the winter months.
2.1.2.2.2. Spatial variability of oceanic production
For photosynthesis to occur, nutrients must be present in sufficient concentration, as well as enough light. The main nutrients are generally limiting to open-oceanic plankton communities, whereas in coastal regions new production and natural nutrient recycling are higher, as well as anthropogenic inputs. For this reason, both micro- and macro-fouling are significantly greater in coastal waters. Wind-driven upwelling of cold, nutrient-rich deep water also increases primary productivity in certain coastal regions, these numbering among the most productive in the world. Tropical latitudes are also generally richer in terms of production than temperate latitudes as the light intensity at the ocean surface, the main driver of phytoplanktonic production, is greater and more consistent318. Higher fouling occurrence is therefore observed to occur closer to the equator and near coasts, but is equally dependant on supply of nutrients via upwelling or aeolian (terrestrial) input from erosion products carried in rivers or from dust carried on the wind.
15
Chapter 2 Literature Review
In many estuaries and some littoral zones, anthropogenic inputs of nutrients (i.e. effluent from agricultural pesticide, sewage) may provide a constant surplus of nutrients all year round, essentially permanently light-limiting the plankton community. In the case where too many nutrients are present and favourable light conditions are sustained, phytoplankton growth can rapidly bloom and self-shading occurs, whereby thick algal mats prevent exchange of atmospheric gases with the water body. This results in eutrophication and hypoxia or even anoxia of the water mass, resulting in the asphyxia or emigration of larger marine organisms and favouring the development of toxic dinoflagellate blooms. A prolonged peak of primary production is therefore very common in coastal regions and ports, causing fouling to become both more rapid and sustained throughout the year. Thus, the problems encountered by harboured vessels are two-fold, in that static vessels are easy targets for the attachment and adherence of motile algal larvae, and that the abundance of fouling organisms is often at its highest when at port.
2.1.3. Natural defences against fouling
Marine organisms themselves suffer pressures from the settlement of other species (i.e. epibiosis). In high production areas of the ocean such as reefs or upwelling zones, coverage of the primary substrate – the sea bed – may be close to 100%74. Fouling by secondary organisms has deleterious effects upon the settled creature, restricting movement and access to light and food, and increasing the probability of tissue damage and disease75. Through this intense competition for space, sedentary and slow-moving marine species have evolved methods to deter epidermal fouling76-78. Fouling is also detrimental to larger, more mobile organisms, such as fish, dolphins and whales, reducing hydrodynamic efficiency and increasing the chance of disease79; these creatures have also adapted methods of deterrence.
Natural antifouling mechanisms are extremely diverse and can be categorised into several types:
a) Production of secondary metabolites; b) Metabolic responses; c) Mechanical mechanisms; d) Adaptation of dermis to repel or prevent foulers; e) Structural defences;
16
Chapter 2 Literature Review
f) Behavioural or reproductive adaptation.
2.1.3.1. Production of secondary metabolites
Complex and diverse chemical compounds are produced by a vast number of marine species in response to fouling pressures80, 81. It has been suggested that every known class of chemical compound is represented in Rhodophyta (red algae) alone - and that Laurencia sp. in particular is possibly the most chemically complex genus in the world82. Owing to the fact that these compounds serve no apparent metabolic purpose other than their antifouling activity and are often costly to produce83, they are termed secondary metabolites. Many compounds produced as secondary metabolites are also unpalatable to prospective predators, and therefore serve multiple ecological roles82. Due to the variation in fouling pressures between locales, intraspecial variation (i.e. variation between different members of the same species) in the quantity of active chemical produced is common where specimens from geographically remote sites are obtained84. Secondary metabolites are often characterised by narrow fouling spectra, and there is evidence to suggest that they are produced to target individual fouling species85, 86. The scale of secondary metabolite production increases directly in response to fouling pressure, culminating in a peak at the end of the fouling season in June87.
Over the last decade, great efforts have made to isolate and identify the secondary metabolites produced by marine biota. Hundreds of these chemicals have been compiled in reviews by Blunt et al.88-90. Classes of compounds with AF potential include (among others) fatty acids, lipopeptides, amides, alkaloids, terpenoids, lactones, pyrroles, polyphenols, steroids and halogenated furanones74, 80, 81, 87-89, 91-104. Many of these active molecules are large and highly complex. These metabolites may be produced at the surface of the organism where fouling occurs105, 106, or may be dispersed throughout the organism in the case of sponges83, 107. Ulva reticulata - itself a major fouling species - was considered undefended until recently, but is now known to produce polar macromolecules at surface boundary layer which are concentrated at the exterior106. This results in a steep concentration gradient of the deterrent chemical. Extracts of red algae such as Delisea pulchra, Bonnemaisonia hamifera, Ceramium botryocarpum and Chondrus crispus have shown promise, deterring fouling in bioassays and field experiments15, 74, 105, 108-117. D. pulchra’s halogenated furanone is perhaps the most studied and best documented secondary metabolite, with potent antibacterial activity, but has yet to be incorporated into
17
Chapter 2 Literature Review a practical long-life coating96, 118-122. Bioactive compounds from bacteria, such as the complex branched fatty acids isolated from Streptomyces sp.123, have also being investigated as potential AF agents. Primary metabolites (i.e. those compounds which serve a distinct metabolic function) have also been the subject of limited study124 and have been found to have some anti-settlement potential against barnacle larvae.
2.1.3.2. Metabolic responses
Certain species react to fouling by producing inorganic acids instead of (or in addition to) secondary metabolites, as most predators and epibionts are deterred by pH of three or lower125. However, these acids demonstrate low efficacy compared to most secondary metabolites and are considered a remnant of ancestral AF mechanisms. Another mechanism common to several species of algae is the ‘oxidative burst’ defence triggered by fragments of cell wall, indicating colonisation or reception of signal molecules from the invading organisms74. This involves a rapid production of superoxide, hydroxyl ions or hydrogen peroxide with cytotoxic effects on the invading pathogens126.
2.1.3.3. Mechanical mechanisms
Certain organisms have developed the ability to regularly shed or slough outer layers, removing foulers. One such example is the red alga Dilsea carnosa, which uses a combination of secondary metabolites and outer-layer sloughing127. Various echinoderms have also been found to slough in defence of fouling, alongside other mechanisms118, and references therein. Many echinoderms and arthropods groom their surfaces to manually remove macrofoulers. However, these methods of defence are limited to areas that the organism can physically reach, and are generally complimented by at least one other defence mechanism.
2.1.3.4. Surface topography
The microtopography of the skin or shell of marine biota is often well adapted to deter settling biota42, 58, 128-136. Such surfaces are difficult to achieve artificially. Settling organisms often prefer a niche for settling roughly the size of their bodies, or slightly larger; however, the wide variety of size scale of fouling biota137 would require a fractal surface encompassing many integrated size scales. Studies involving microtextured
18
Chapter 2 Literature Review poly(methylmethacrylate) meshes have demonstrated that a surface that impairs settlement for one species can encourage another138. However, in the same study, 92% of settlement was deterred by a surface with 30-45 µm surface roughness. To deter a full spectrum of fouling organisms, one can postulate that a fractal surface would need to include scale ranges less than the critical spore size for each individual organism. To further complicate matters, the results of studies can be inconsistent – for example, barnacles have been observed to prefer smooth surfaces in at least one study138.
Accurate resin cast replication of microtextured shells such as Mytilus edulis and Ophiura texturata resulted initially in an efficient repulsion of fouling larvae58, 139. However, the efficacy of the resin surfaces decreased after three weeks. The authors suggested that a build-up of biofilm over the first weeks might mask the topography and present a more uniform surface; alternately, the increasing age of larvae as the fouling season continues may increase their disposition to unfavourable surfaces. The protein used to affix to surfaces is at its highest concentration during the larval stage, and it also acts as an ‘encourager’ for subsequent exploring larvae57. It is likely that once settlement commences, the other larvae that explore the surface will preferentially colonise the ‘footprinted’ area.
Disparities between studies are commonplace. In one study, two bivalve shells of different microtopography were compared. The shell with a microrelief of 1-5 µm demonstrated good resistance to fouling with >90% inhibition140. However in other studies, 5 µm was found to encourage barnacle settlement and reliefs of 50 µm, 100 µm and 2-4 mm demonstrated more thorough inhibition of settlers141. A further study133 compared the performance of a high-resolution synthetic replica to the living bivalve shell. The replica shell and living shell were more repellent to fouling than the smooth blank control shell, but the replica began to become fouled after 6-8 weeks’ immersion, whereas the living shell did not. The authors concluded that other factors besides the topography of the shell must contribute to the observed fouling resistance. Interestingly, a roughly sanded blank ‘shell’ (with random topography) performed better than the smooth blank control.
Surface wettability is another factor that is important in controlling settlement42, 142. Certain fouling species such as Ulva are more likely to colonise artificial hydrophobic surfaces in tightly packed groups as opposed to individually, increasing adhesion strength and susceptibility to further fouling 55, 143. Arrays of artificial surfaces of varying wettability – from superhydrophobic to superhydrophilic – provide a discontinuous gradient of exposure
19
Chapter 2 Literature Review in bioassays enabling characterisation of species-specific responses 143, 144. It is possible to integrate microtexture and repellent substances in a single coating in biomimicry of killer whale or dolphin skin, which use oils and microtopography to deter foulants57, as well as enzymes in defence of outer cells and peptides as an anti-fungal agent145. The pore size of some marine mammals is less than 0.2 µm2, smaller than the larval size of most fouling species and also resistant to contaminant ingress79. One study found that integrated non- network forming poly(dimethylsiloxane) oils in a 5 µm microtextured surface proved effective, although it was concluded that the effect of microtopography type and spacing was greater than that of additive oils137. Conversely, 5 µm niches have shown in other studies to encourage settlement of species such as Ulva intestinalis, whose spore size is in the order of 5 µm128.
2.1.3.5. Structural adaptations
Some species of sponges subject to high fouling pressure seem to be neither fouled nor predated, despite in certain cases the apparent lack of secondary metabolite production (with notable exceptions146-149) and their high nutritional value. Sponges develop siliceous, needle-like structures called spicules which, when they reach a sufficient size or volume, are indigestible to predators150. Large-sized spicules render the sponge physically unpalatable or painful to eat, whereas a large volume of smaller spicules reduces the nutritional value of the normally protein-rich and delicious sponge tissue. Although they do not contribute to the prevention of fouling on sponges, in some species where chemical and structural defences are both present83, 149-151, the spicules provide a synergistic anti-predation effect with the same secondary metabolites that counteract fouling86. Ascidians (sea squirts) have highly acidic inclusions in their outer layers that are believed to deter fouling and predation152, 153. There is a considerable body of literature on sponge metabolites and other defences outside of the scope of this work83, 99, 146-151, 154-170.
2.1.3.6. Other adaptations
Some organisms possess no evident physical or chemical defence mechanisms, yet do not seem to be fouled to the extent that one would expect. Crabs have developed traits such as intermittent burial or exposure to air that disrupts the development of biofilm and prevents larval settlement on their carapace171. They also benefit from mechanical adaptations - their eyestalks are retractable and are cleaned by small cilia to remove microfoulers and debris -
20
Chapter 2 Literature Review and a shell surface topography detrimental to diatom settlement172. Other behavioural adaptations include maintaining nocturnal activity when photosynthetic activity is nil and fouling larvae are sedentary. Bivalves also bury themselves in sand or mud152. Other ‘undefended’ species include some varieties of gorgonian corals and sponges which rely on accelerated rates of growth, healing and reproduction to counteract the effects of fouling173. Some sponge species seem to be unadapted and, being immobile, must rely purely on hiding to avoid being eaten150.
Yet another method of fouling and predation deterrence in marine biota is the sequestration and concentration of naturally occurring heavy metals like vanadium in body tissue152. Saxidomus sp. (butter clam) appears to actively filter and concentrate toxins from toxic dinoflagellate blooms that are known to cause paralysis and death in many shellfish. These species have developed tolerance to high concentrations and prolonged exposure to such toxins152. The presence of high PSP (paralytic shellfish poisoning) toxin concentrations is a deterrent to most predators and settling organisms.
Finally, symbiotic relationships are common between species such as Perna viridis, which filter feeds to remove macrofouler larvae, whilst co-occurring gastropods graze microfoulers that develop on its shell174. Some sedentary species exhibit a symbiotic relationship with smaller organisms which produce secondary metabolites to protect the host. These symbionts have been observed on organisms from hydrothermal vent clusters as well as reef systems175, 176. Unfortunately, cultivation of these organisms for further testing is difficult due to their dependence upon on the host organism176.
2.2. Antifouling strategies – historical, recent and emerging technologies
2.2.1. Historical antifouling solutions
Due to the rapid and irreversible damage caused by fouling to untreated surfaces177, efforts have been made to prevent biofouling for over 2500 years3, 71. Early attempts to curtail biofouling ranged from rough applications of arsenic or lime, to lead sheets and copper sheathing3. These efforts were geared towards preventing burrowing and boring organisms such as tubeworms from attacking or ‘gribling’ (boring into) the exposed wooden hulls of early ships. Copper sheathing was found to be the most effective at reducing gribling. The earliest examples of AF coatings (in the modern sense) began to be developed with the
21
Chapter 2 Literature Review advent of iron hulls, which restricted the use of copper sheathing due to enhanced galvanic corrosion of the iron which resulted from contact between the two materials. The accelerated corrosion of the iron had not been predicted at all, and 3500 lives were lost in 1782 when several iron-hulled, copper-sheathed British and French vessels sank off the coast of Newfoundland3.
Wooden panels were used to replace copper sheathing. When the wooden panels became too heavily fouled, they could be discarded and replaced. In order to prolong the service life of the wooden panels, hundreds of different combinations of compounds were tested, including powdered iron, cement, copper compounds, beeswax, tar, oil, lime, zinc salts, arsenic, resin, suet, chalk, fish oil, asphalt, pitch, brimstone, mercury, guano and salt, amongst others, many of which were useless as coatings, and some of which actually nourished fouling communities rather than hindering them3. More effective copper sulphate and copper oxide coatings first appeared in late 19th century, often in a matrix of gum shellac or resin. Coal tar-resin paints and shellac-based coatings with cuprous oxide and mercuric oxide toxins might be considered the primitive ancestors of modern day CDP- type (controlled depletion polymer; see 2.2.2.2) coatings, and were used until the mid 20th century. Although they featured potent biocides, these paints required heating and awkward, elaborate apparatus to apply, and shellac supplies soon became limiting. Nonetheless, the industrial production and quality control of these AF coatings was an important precursor for the development of modern AF solutions.
2.2.2. Modern antifouling systems
Typical modern AF systems comprise a series of coatings which, as a whole, resist diffusion of ions and ingress of water to the hull surface and are capable of contraction/expansion with the substrate to endure movement of the vessel and temperature variation. Most modern-day paint systems are applied in the form of polymeric paint coatings, generally comprising a primer, whose role is to facilitate adhesion to the substrate, an anti-corrosive layer to prevent corrosion of the hull, and an AF ‘topcoat’, which contains the biocide. A tie- coat between dissimilar coating layers is often necessary to prevent spalling and separation. The topcoat is the surface in contact with the aggressive marine environment; it must be able to withstand fouling, constant wind/wave action and possible abrasion, and must also meet the aesthetic demands of the consumer. For this reason, pigment molecules are embedded in the matrix of the paint (e.g. various iron or copper oxides for red or yellow,
22
Chapter 2 Literature Review titanium dioxide and fillers for white). These pigment molecules are often relatively large and may be soluble, and can themselves have biocidal properties integral to the operation of the coating (see Section 2.2.2.2), most notably in the case of copper (I) oxide, which exhibits an excellent toxicity spectrum. The loss of pigment particles as a result of reaction to soluble (or insoluble) derivatives as the ingress of seawater into the paint matrix occurs is inevitable. In paints with insoluble matrices (such as contact leaching coatings), this process is an essential mechanism of biocide delivery.
The different types of modern AF coatings discussed in the following sections include contact leaching coatings, soluble matrix type controlled depletion polymers (CDP), and self-polishing copolymer (SPC) containing TBT (tributyltin) or other alternative biocides.
2.2.2.1. ‘Contact-leaching’ coatings
Contact-leaching coatings were developed in the 1950s178 and consist of a hard, vinyl acrylic binder. In these paints, the binder is generally a largely insoluble resin. Contact-leaching coatings contain a very high concentration of cuprous oxide (Cu2O) that functions as both an effective biocide and a brownish-red pigment. Above a certain threshold of concentration (the critical pigment volume concentration – cPVC), the biocide/pigment molecules are considered to be present in a sufficiently high concentration that it is considered to be a continuous phase within the binder (Fig. 2.2). As pigment concentration is increased, the physical properties of the coating change abruptly when the cPVC is reached179. The reaction and dissolution of the pigment with the seawater results in the formation of channels – the formation of a contiguous network of channels throughout the paint coating is ensured by the high pigment content, allowing in principal for a constant release of the additive. The copper (II) chloride derivatives of the pigment/biocide Cu2O are slightly soluble and dissolve slowly into seawater180. The generation of soluble derivatives is governed by reaction of H+ and Cl- ions with the solid copper (I) oxide particle to form chloro-copper (I) complexes (Eq. 2.1 and 2.2181, 182) whose oxidation to Cu2+ species such as
CuCl2 is carried out during transport of copper (I) species out of the film. The accumulation of biocidal copper (I) and copper (II) species at the interface of the paint film with the seawater results in the inhibitory effect of the pigment.
+ − − Cu2O + 2H + 4Cl ⇄ 2CuCl2 + H2O (Eq. 2.1) − − 2− CuCl2 + Cl ⇄ CuCl3 (Eq. 2.2) 23
Chapter 2 Literature Review
Both reactions are reversible, although the first one is heavily influenced by kinetics 50, 182. As the pigment derivatives leach out from the binder, a biocide-depleted leached layer is formed. This leached layer presents the dual problems of increased roughness and a hollow structure (Fig. 2.2). This hollow leached layer firstly lengthens the pathway for biocide delivery to attached organisms, and secondly provides a trap in which insoluble particles and copper derivatives can accumulate, reducing efficacy178, 183. For example, impurities from seawater and insoluble copper salts can accumulate within this layer, which restricts the leaching rate of biocide184.
Figure 2.2. Erosion of a contact-leaching type paint system over time.
These types of AF systems generally have service lives in the order of 2 years, and are easily 24
Chapter 2 Literature Review replenished at the end of their lives; the hollow leached layer can simply be scoured away and fresh coating applied directly on top. They are rarely used in modern shipping.
2.2.2.2. Soluble matrix/CDP coatings
Erodible/soluble matrix or CDP paints feature rosin-based, partially erodible binders (up to 30% rosin) that have been in use since the 1950s. Rosins are a family of natural-origin brittle resins obtained from pine and fir trees which have been used for over 100 years in AF coatings3, 185. Rosins are mainly comprised of abietic acid (Fig. 2.3), although most contain a large fraction (as much as 20%) of hydrocarbon and ester impurities. Even good quality rosins are liable to contain at least 10% by weight of impurities.
Figure 2.3: Abietic acid, the primary constituent of rosin.
The pigment (normally Cu, Zn or Fe oxides and previously As and Hg) and other biocidal additives are ‘free-associated’ in the organic matrix, and are slowly leached out from the surface of the paint into the seawater body (Fig. 2.4) by similar mechanisms to the contact leaching type coatings. Rosin is sparingly soluble at the pH of seawater (7.8-8.2), allowing for a gradual dissolution of the binder component in sea water. This creates channels for ingress of water and leaching of biocidal pigment particles from the matrix. The formation of a leached layer is mitigated by the slow dissolution of the rosin component of the binder. The biocide release profile is non-linear, exhibiting a much higher initial release rate which rapidly declines (Fig. 2.4). This is due to the high surface concentration of biocide in the coating; as the biocide is depleted over time, the leached layer grows in thickness and biocide release becomes inhibited. As the biocide delivery drops below a critical threshold, some fouling will start to occur on the surface of the coating, even though the paint still
25
Chapter 2 Literature Review holds biocide below the leached layer (Fig. 2.4). The thickness of the leached layer is variable depending on ship speed and hydrological conditions, and the relative leaching rate of the biocide.
Figure 2.4. Erosion of a CDP type paint system over time.
The lifetime of early paints of this kind was only in the order of a year, as the poor mechanical properties of rosin meant that only a thin film could be applied178. However, modern versions, known as controlled depletion polymers (CDP), have more complex formulations with additives tailored to the operational profile of a specific vessel type, and hence obtain lifetimes of up to around three years. Commercial erodible matrix or controlled depletion polymer (CDP) coatings such as International’s Interspeed 340 continue to hold an important niche in the AF market. These are generally based on poly(methyl methacrylate)/poly(butylacrylate) polymer blends with rosin added to provide the soluble element. Higher quantities of rosin will increase the performance of the coating by increasing the rate of physical dissolution, but will diminish its cohesive and adhesive
26
Chapter 2 Literature Review qualities due to the poor mechanical properties of rosin itself185.
2.2.2.3. TBT self-polishing copolymers
During the 1960s, the marine paint industry was able to produce more efficient and cost- effective AF coatings featuring organotin compounds toxic to a wide range of marine organisms, including molluscs and algae186. Organotins such as TBT (tributyltin) had been previously used as fungicide, bactericide and wood preservative187, and could be optimised for activity against the target range of species by altering the length of the carbon chain pendant to the tin atom188. The early organotin-based antifouling mixtures, using ‘free- association’ TBT embedded in an insoluble or partially-soluble matrix, proved effective but unreliable in their delivery6.
The combination of a TBT monomer in an integrated polymer delivery system was conceived in the late 1970s, and is generally considered to be the most effective AF system ever developed. A large, hydrophobic TBT monomer is linked to a polymer chain by means of an ester group, the other components of the copolymer being commonly methylmethacrylate and butylacrylate. The strongly polar Sn-O bond is vulnerable to attack by readily available electrophilic species in seawater, the most common of which are sodium cations (Na+). The TBT monomer is thus substituted by saponification of this bond (Figs. 2.5 and 2.6). Its replacement by Na+ on the polymer chain causes an increase in hydrophilicity, encouraging a local ingress of water and permitting ‘erosion’ of the polymer chain from the binder matrix. A chloride ion (Cl–) is assumed to accept the spare electron at the tin atom, forming TBTCl185. A new ‘active’ surface, replete with biocide, is therefore always available at the coating/seawater interface. This ‘polishing’ mechanism confines the reaction to within a few nm of the surface189. The roughness of the surface is also superior to that of other erodible paints, as the nature of the paint ensures that the roughness resulting from application is soon eroded away and a smooth surface results54, 190. The active lifetime of SPCs is effectively only limited by the coating thickness; lifetimes of >10 years can be in achieved in theory, although the weight and thickness of the paint coating may cause cohesion problems and is therefore impractical for most vessels.
27
Chapter 2 Literature Review
Figure 2.5: The saponification process of the TBT monomer in seawater that leads to the self-polishing effect. The polymer shown is a typical self-polishing methylmethacrylate/butylacrylate/TBT copolymer.
The leached TBT monomer is rapidly absorbed by the cells within the fouling organism, and interferes with energy transfer in photosynthetic and respiratory processes1. The erosion of the binder is continual, and consequently there is no leached layer, or only a very thin one (Fig. 2.6). The effective lifetime of this coating is therefore far superior to the soluble matrix or CDP-type systems, as the biocide-replete layer is always in proximity to fouling species, and the surface roughness decreases with erosion, giving an improved hydrodynamic profile. SPC systems are commonly observed to last up to five or six years71, roughly twice the length of a modern CDP. The life of the coating system as a whole may be limited by certain areas on the vessel, such as the stern, which are prone to increased turbulent flow and a high polishing rate. Rapid erosion of coatings in these areas is a phenomenon known as ‘polish-through’, and results in fouling on the depleted regions27.
28
Chapter 2 Literature Review
Figure 2.6. Erosion of an SPC type paint system over time.
Such was the outstanding performance of TBT-SPC systems that they are estimated to have been employed on 70-80% of the world shipping fleet at their zenith17, from leisure boats to commercial carrier vessels, saving an estimated $5.7 billion dollars per annum191, and reducing world fleet consumption by 7 million tons per year, corresponding to a reduction
188 in emissions of 22 million tons of CO2 and 0.6 million tons of SO2 . In addition, the longer effective life of TBT coatings reduced the costs incurred by dry-docking and increased time spent in service. However, the environmental impacts of the eroded TBT side group on non- target species were soon recognised during the late 1970s and 1980s; in particular, imposex in dog-whelks192, 193, shell deformities and population decline in commercial oysters, totalling around $150 million in loss194, 195, long-term sterilisation and population decline in gastropods in areas of previous high TBT usage196, 197, and bio-accumulation in seals and fish198. TBT has been observed to inhibit growth in zoo- and phytoplankton and inhibit photosynthesis in phytoplankton at concentrations of 0.4 ngL-123; that is, 1 g TBT per 25
29
Chapter 2 Literature Review billion litres of water – equivalent to one gram of TBT diluted in 1,000 Olympic swimming pools – and a concentration of >2 ngL-1 adversely affects shell formation in Crassostrea gigas23. Such is its potency that TBT has been described as “perhaps the most toxic substance ever deliberately introduced to the marine environment by mankind” (Goldberg, 1996, in Champ, 200117).
Preliminary legislation prohibiting TBT usage was introduced in 1986 and 1987, and UK and USA legislation restricting the use of TBT to vessels longer than 25 metres was passed in the early 1990s18, on the back of similar restrictions a decade before in France6, 199 which was hard-hit by the effects of TBT on ostreiculture194. Following observation of harmful TBT concentrations in sediments in the UK, France, Mediterranean, Tyrrhenian Sea, Bay of Naples, Bahrain, USA, NZ and Australia, a total ban on TBT was agreed in 1999, to be gradually implemented during the 2000s200. TBT has also recently been found to be present in high concentrations in sediments of regions hitherto thought uncontaminated, such as the Antarctic, along with some invasive species26. This is thought to be due to the dislodgement of species and paint from the hulls of boats during ice-breaking. From 2003, the application of new coats of TBT to any vessel was forbidden, and a total ban on TBT, scheduled originally for the beginning of 2008, is now in action since September 2008 in most countries186. This forbids an exposed coat of TBT on any vessel; either removal of old TBT coatings is required, or reapplication of a different paint coating over the old layer, preventing TBT leaching. Issues presented by TBT toxicity are compounded by the persistence of TBT and its derivatives, mono and dibutyltin, in water bodies and estuarine sediment201, which could be as long as 30 years according to some estimates, as decomposition of organotins in sediments is slow6, 188. Remobilisation of sediment during storms, dregding or by bottom-feeding organisms acts as a secondary source of TBT emission to the water column. In particular, dredging and disposal of sediments in heavily TBT-contaminated zones is an area for great ecological concern. Furthermore, the relocation or destruction of such contaminated sediments is extremely costly under current legislation, which disallows the near-shore dumping of TBT-contaminated material18. The legacy of TBT will therefore by far outlive the active use of the paint; furthermore, TBT is probably employed illegally in some areas in even in IMO-compliant countries6.
Despite the environmental concerns, there was opposition to the TBT ban from some scientists and industry representatives, who were unconvinced of the specificity of the imposex reaction, arguing that it could result from environmental stress or exposure to
30
Chapter 2 Literature Review copper and zinc, nonylphenols or certain parasites188, 192, 200, 202, 203 and that the vast economic benefits of TBT deployment outweighed the environmental concerns. They highlighted in particular the good life-span of TBT-SPCs, the broad toxicity of the biocide, and the fact that on reverting to other biocides as solutions, the problem is merely being shifted, with build-up of other harmful biocides an inevitable consequence187, 204, 205. Nonetheless, the eventual banning of TBT was welcomed by most.
2.2.2.4. TBT-free self-polishing copolymers
In the wake of the abolition of TBT186, biocides have become heavily regulated206, 207. The biocidal products directive (BPD) requires the lengthy and expensive registration of biocides, including full toxicology and ecotoxicology assays and environmental risk assessments, with a total cost of $7 – $10 million208. As of 2008, only ten biocides are currently registered for use, with no more having been added to date (August 2010). These permitted biocides are copper, copper (I) oxide, copper thiocyanate, copper pyrithione, zinc pyrithione, tolylfluanid, dichlorfluanid, cybutryne (N-cyclopropyl-N′-(1,1-dimethylethyl)-6- (methylthio)-1,3,5-triazine-2,4-diamine; trade-name Irgarol 1051), Zineb (zinc ethane-1,2- diylbis(dithiocarbamate); trade-names Dithane Z-78, Parzate and Polyram Z) and DCOIT (4,5-Dichloro-2-n-octyl-3(2H)-isothiazolone; trade-name Sea-NineTM)208. Other biocides, such as Diuron, were not accepted into the BPD209-213. Many of these biocides are employed in combinations, in order to broaden the AF spectrum and potency against resistant organisms. Over 50% of AF formulations employ two biocides, with roughly 20% employing three and about 5% with four or more214, 215. Self-polishing copolymer technology, highly effective as a delivery mechanism for TBT, is currently adapted for use of copper, which replaces tin as the biocide in the form of polymeric metal-acrylate complexes. These do not directly mimic the hydrolysis mechanism of the TBT-based SPCs, as they do not feature a biocide that is entirely covalently linked to the polymer; polishing and biocide depletion are interrelated, but not directly related. Copper has well-documented AF properties and toxicological profiles, having been used for centuries209. It was also present in TBT-SPC coatings as a pigment, and also widened the AF spectrum and reduced the amount of organotin required216. However, long-standing concerns are beginning to re-emerge: sedimentary accumulation28, 217, bio-accumulation218, and the tolerance of several particular fouling species to high and sustained Cu exposure, such as Ulva sp., Ectocarpus sp., and Achnanthes sp.209. These issues are stimulating the need for development of alternative biocides and non-biocidal coatings.
31
Chapter 2 Literature Review
Other commercial SPC technologies besides the Cu/Zn acrylates exist, including copolymers with complex di/tri-block architectures featuring hydrolysable trialkylsilyl side groups219, 220. Other types of delivery system are under research but are not yet commercialised. A recent example is the development of block copolymers based on polyesters, enabling an incremental hydrolysis of the binder backbone itself, rather than of side groups. Polylactides reacted in block with methacrylic acid demonstrate controlled biocide release and polishing rate, but only over short time periods221. Other biodegradable copolymers have been developed from poly(ε-caprolactone), poly(sebacic acid), isophthalic acid, ricinoleic acid and terephthalic acid222. Although the films developed from these polymers remain fouling-free, the erosion is too rapid and the coatings suffer from excessively fast polish-through. Work is ongoing to retard the hydrolysis of the binder in these cases.
Another type of hydrolysable mechanism to emerge in the last decade is the use of organotitanium (tetraalkoxytitanium) compounds that react with carboxylic acid side groups on multiple polymer chains, forming several possible structures223, 224. However, no further research or commercial developments of this technology have evolved since then, despite the original promise shown.
Recently emerged commercial technologies in the SPC field include Transocean’s biocide- free SPC Futureline 95.60, whose functionality is dependent upon the binder hydrolysis mechanism and a combination of hydrophilic and hydrophobic domains at the polymer surface (http://www.transocean-coatings.com/site/home.html, accessed 13.10.09).
Another example is Hempel’s Globic NCT, which achieves a self-polishing effect by the means of micelle-like acrylate nanocapsules which feature a hydrophobic outer layer and a reactive, hydrophilic inner core (www.hempel.com, accessed 14.10.09). The carboxylic acid groups at the core are saponified by seawater, forming alkali metal salts. This disrupts the micelle-like structure, triggering the polishing process. This new technology has demonstrated an excellent correlation of biocide leaching and polishing. The biocide Sea- NineTM has also been encapsulated and commercialised225. The biocide itself has a plasticising effect on the binder which is negated by encapsulating it, resulting in higher potential loadings in the binder.
Not all microcapsules need to be chemically complex; Nydén et al.226, 227 demonstrated that
32
Chapter 2 Literature Review incorporation of active agents in simple pMMA microcapsules helped to reduce their diffusion, reducing the initial release of biocide from film during the first few days of immersion from ~50 to ~10% of the total amount, compared to the free-associated biocide. Microcapsules also provide the secondary benefits of improved handling safety, higher potential loading in the film, reduction of photolytic degradation of the biocide and better shelf stability. They are likely to be a key element of emerging paint technologies in the near future.
2.2.2.5. Hybrid SPC coatings
Hybrid SPC coatings are intended as a midpoint in terms of cost vs. performance, between traditional CDP erodible coatings and more effective but expensive TBT-free SPCs. They contain quantities of both rosin and self-polishing copolymer components, and present a leached layer that develops more slowly than a traditional erodible coating.
2.2.2.6. Foul release coatings
Many operators of high-activity vessels such as ferries, which travel at high speed and have rapid turnovers, have turned to recently emerged foul release coatings (FRCs). Although FRCs were developed almost simultaneously with TBT-based SPC systems (Fig. 2.8)228, FRCs would not come into commercial use until the late 1990s; not only were they more expensive to apply, but also the harmful effects of TBT were not fully understood until the early 90s, and legislation on their use would not come into effect for around two decades. Modern FRC coatings are applied by conventional airless spray as two or three-pack applications229. Although FRCs cost 2-7 times more than SPCs on average, they offer considerable benefits to many types of operational profile178.
Foul release coatings operate by producing an ultra-smooth polydimethylsiloxane (PDMS) or fluoropolymer surface. Such a coating may be enhanced by incorporation of silicone oils. A successful FRC must be free from polar side groups, and chemically and physically stable in seawater, as well as being ultra-smooth and having an optimal surface energy for repelling adhesion14. Surface energy is the excess energy of molecules at the surface/electrolyte interface compared with the interior, measured by analysis of contact angle of various liquids229. The surface energy can be altered by changing the type of groups pendant to the
230 polymer backbone; surface energy decreases in the order CH2 > CH3 > CF2 > CF3 . The
33
Chapter 2 Literature Review pessimal surface energy for biological adhesion is known to be around 23 mNm-2 231, 232 although this differs for certain species233-235. This value is about equivalent to the critical surface tension of water. This reduces the ability of any fouling organism larger than a bacterium to adhere to the vessel, and shear stress at the surface imparted by motion of the vessel can easily dislodge any weakly bonded foulers. In addition, the ultra-smooth surface provides a relative decrease in frictional resistance, thereby decreasing fuel consumption and increasing maximum speed, when compared to CDP or SPC coatings228. Some writers have highlighted the effect of coating modulus on adhesion, and have suggested that surface energy should be lowered as far as possible, rather than to the optimal part of the Baier curve (Fig. 2.7)14.
Baier Curve for bioadhesion
Relative Adhesion Relative
0 10 20 30 40 50 60 70 80
-2 Surface free energy / mN m
Figure 2.7: Baier curve for relative bioadhesion vs. surface energy (See Baier, 1971, 1973).
As the dislodgment of organisms is dependent upon the motion of the vessel, accumulation of biofouling may readily occur whilst the vessel is standing in dock. The vehicle must reach a 'critical' speed in order to remove biofouling and avoid transfer of alien species, generally advertised as being in the order of 7 knots for barnacles and 30 knots to remove large diatoms, which are highly resistant to shear forces when forming mats or tight-knit communities49, 178, 236-238. Field trials demonstrate a great intraspecific variation in detachment, however, with large barnacles of a certain species detaching at speeds as low as 3 knots, whilst smaller examples of the same species held fast until 15 knots were reached239. The reinforcing effect of surrounding slime and algae, presumably protecting the point of barnacles’ adhesion with the substrate from the shear effect of water, was also noted in the latter study.
34
Chapter 2 Literature Review
FRCs have been used to great effect by boats with a fast turnover (e.g. ferries), but may be unsuited for slow vessels or vessels experiencing extended in-dock periods, such as naval vessels. FRCs are also easily damaged due to the poor mechanical properties of the PDMS or fluoropolymer coatings, and an abrasion or scratch provides a perfect shelter for settling organisms118. The use of microfibers to internally reinforce the film has been implemented by Hempel in their Hempasil X3 line. The mechanical properties and damage resistance of most FRC lines are now far superior to their original iterations.
Shipyard owners are beginning to realise the practical applications of using of another AF system in conjunction with FRCs, as certain areas on a vessel’s hull are designed to receive minimal flow and shear stress, and as such do not reach the critical flow rate required to oust foulers from the surface. These areas include sea chests, which house sensory equipment or structures such as valves which regulate, for example, the flow of water to and from ballast tanks183. Although heavy fouling in sea chests, intakes and grillages is unlikely to contribute to increased drag, access to these areas is awkward and translocation of species between ports is likely20, 21, 24, 25, 240. FRCs are clearly unpractical for these low-flow areas, so a compatible AF paint system for use on these parts would also be of commercial and ecological interest. However, contrary to observation of accumulation on stationary FRC vessels, some workers48 have not only demonstrated a 100-fold decrease in bacterial adhesion compared to SPC in dynamic experiments, but also observed a lower adhesion in static immersion experiments. This could be explained by immersion in an area with high ambient flow rate, such as an estuary or river, rendering adhesion to the FRC surface difficult and adhesion to the traditional coating preferable.
Despite the traditional problems associated with FRCs, some navies are now switching to or trialling this technology. Although paint companies claim that fouling can currently be prohibited by about a month in dock before cleaning or movement is required (C. Birkert, International Paint, pers. comm.), many navies have experienced poor results with FRCs alongside, particularly in extremely productive areas such as the USA west coast and the Australian eastern and northern coasts241. FRCs represent the most rapidly growing area of AF technology in terms of commercial employment, with coatings such as International’s fluoropolymer Intersleek 970 and Hempel’s PDMS Hempasil X3 being adopted as flagship coatings by their respective companies.
35
Chapter 2 Literature Review
2.2.2.7. Novel emerging fouling solutions
Lindgren et al.242 discuss a novel binder system incorporating up to 43% industrial proteins. Through natural microbial degradation that occurs in-situ, a hypoxic layer evolves in the surface/water interfacial layer of the paint and around 200 µm into the surrounding water mass. This creates unfavourable conditions for the attachment of barnacles and bryozoans due to the inhibition of respiration, as well as decreasing pH at the coating-water interface and encouraging reduction of sulphate species to toxic hydrogen sulphide and particulate organic matter to organic acids, which may damage organisms and adversely affect test (shell) formation in macrofoulers. A commercial product (Neptune Formula, Ekomarine AB, Sweden) has been launched in the low-fouling, environmentally sensitive Baltic Sea242.
Much focus has recently been placed on nano-functionalised polymer surfaces featuring customized side groups which form self-assembling layers243, 244. The latter review articles discuss in depth these types of coatings, which operate via adjacent domains of hydrophobicity and hydrophilicity (known as amphiphilic) that ‘confuse’ fouling organisms, deterring settlement, although the mechanisms by which they do so are not fully understood245, 246. Examples of side groups generating amphiphilic zones include hydrophobic fluorinated alkyl groups and PEGylated fluoroalkyl side chains. Due to the hydrophobic extremities of these side groups and the hydrophilic nature of the PEGylated area adjacent to the backbone, these side groups self-align parallel to the substrate. Adjacent, heterogeneous regions of hydrophilicity and hydrophobicity are thus generated on the nanoscale (1-10 nm). The surface therefore encompasses the low surface energy effect of the hydrophobic fluoropolymer segments, whilst benefitting from the resistance to protein adsorption characteristic of the hydrophilic PEG section247.
Other types of surface functionality currently under investigation include polymers featuring Zwitterionic side groups, hyperbranched PEGylated/fluoropolymers, xerogels (cross-linked sol-gel coatings) featuring tuneable wettability-tailored side groups, siloxane- polyurethane copolymers, phosphazene polymers and polyoxazolines244. Grozea and Walker243 again discussed many of the same systems as well as microtopographic patterned systems such as the SharkletTM PDMS antifouling coating. SharkletTM, a bioinspired surface topography based on shark skin, features a gradient of micro-scale change in three dimensions that renders firm attachment difficult.
36
Chapter 2 Literature Review
A novel type of coating consisting of a biocide-free, extra smooth, extra hard surface has recently been developed248. The surface permits adhesion up to an extent, but is designed to operate in tandem with very aggressive cleaning regimes that the coating is designed to withstand. A typical cleaning regime entails the deployment of an underwater rotary brush- type device, piloted by divers, to remove fouling from the hull. Traditional AF coatings would not withstand this ‘grooming’ protocol. Other cleaning methods recently investigated involve the localised heating of water to temperatures in excess of 70 °C to kill fouling communities, and the deployment of ‘spargers’ underneath the hull, which release a steady stream of air bubbles, deterring settlers241. These latter cleaning methods can be applied to most hull and coating types.
Finally, Murosaki et al.249 describe the performance of tough synthetic, polymeric hydrogels in long term immersion experiments. These gels have a high water content (80-90 vol.%). Fouling is repelled for about 2 months by most gels, and about 11 months in the case of poly(acrylamide)/poly(2-acrylamide-2methyl-1-propanesulfonate) double-networked gel (i.e. two interpenetrating films). Furthermore, the elasticity of the double networked PAAm/PAMPS gels allowed for relatively easy removal of barnacles at the end of the immersion test. It is proposed that the smoothness of the gels could reduce friction on moving vessels as well249.
2.2.2.8. Alternative and environmentally acceptable future fouling solutions
The ideal goal of coating development is to locate a biocide that not only matches the TBT- based paints for efficacy and life-span, but also that has little effect on non-target species, and degrades rapidly once leaching has occurred. However, this is likely to be an impossible goal. Recent research is focusing on developing new biocide-free coatings250, though ‘environmentally acceptable’ biocides, often either derived from natural marine products or synthesised as analogues, are increasingly the target of research throughout the last decade1, 251. Many different interpretations of what constitutes an environmentally acceptable coating or deterrent have been posited. The term ‘environmentally acceptable’ is defined here as an AF surface or paint coating that meets or exceeds demands on drag reduction and efficacy against target species, and which could be used for the foreseeable future without risk of environmental harm, bio-accumulation or sedimentary accumulation.
Avenues of current exploration include extracts from marine bacteria123, 252, enzymes
37
Chapter 2 Literature Review designed to prevent adhesion by hydrolysing adhesive proteins and polymers or catalyse in situ production of repellents46, 253-259 and biomimetic polymers designed to resist protein adsorption at an early stage of biofouling42, 260-266. The use of micro-structured synthetic analogues of natural shell or dermis surfaces is also under investigation as earlier discussed, and shows promise (See Section 2.1.3). It has been suggested that for true environmental acceptability, a ‘biocide’ should act via chemical signalling rather than direct toxicity267.
Other possible non-chemical, non-biological AF deterrents have been investigated. Radioactive isotopes such as thallium 204 and technetium 95 or 99 in AF coatings have been found to be highly effective. However, the use of these isotopes would expose workers and boat crews to hazardous radiation, and would be difficult to implicate practically71.
2.2.3. A summary of antifouling technologies
A summary of the historical progression of the development of antifouling systems is outlined in Fig. 2.8.
Figure 2.8: A timeline for the development of antifouling systems towards modern systems.
1. Copper is found to have excellent AF properties, and preliminary trials are highly successful. British navy vessels are outfitted with copper sheathing, which is costly but saves on repairs of wooden boards. 2. Iron hulls are introduced; Cu is found to accelerate corrosion of iron hulls and is
38
Chapter 2 Literature Review
rapidly abandoned. First patented AF coating introduced in 1625 (William Beale3). Hundreds of different coatings are tested. Copper sulphate and copper oxide paint coatings appear in the late 19th century, often in a matrix of gum shellac or resin3. Coal tar-resin paints and shellac based coatings with cuprous oxide and mercuric oxide toxins are used until the mid 20th century. Soluble matrix and then insoluble ‘contact-leaching’ type coatings become more popular. Free-associated TBT paints soluble matrix systems are found to be effective, although AF activity is unreliable. 3. A patent for TBT-based SPC is approved. In the same year, a patent is obtained for silicone-based FRCs (International Paint), although the technology is not applied until later. 4. Trials of FRCs commence, although they did not see widespread use until the 21th century. First legislation restricting TBT use is introduced soon after. 5. International Maritime Organisation186 responds to mounting concern over TBT toxicity and persistence. New application of TBT banned on any vessel shorter than 25 m. 6. Further application of TBT on any vessel is banned. SPC technology applied to less hazardous biocides such as copper acrylates. Research commences on biocide-free and alternative AF methods. 7. TBT is banned in most countries. Copper/zinc-based SPCs and FRCs now in prevalent use and research into environmentally acceptable alternatives is increasing.
2.3. Natural products as antifouling agents: challenges and considerations
2.3.1 Integration of natural product
Secondary metabolites are incredibly complex and diverse compounds88-90, and the elucidation of an NP and its stereochemistry is a lengthy and difficult process. There are many different approaches to employing the AF abilities of an organism in order to maximise the efficacy of the coating whilst maintaining cost-efficiency.
2.3.1.1. Crude extract approach
The first approach is the incorporation of a ‘whole’ or ‘crude’ extract of processed algae or other biomatter, which can entail any one of a vast array of methodologies. The method involves crushing or pulping a large quantity of biomass, which may have been oven-dried,
39
Chapter 2 Literature Review freeze-dried, frozen in liquid nitrogen, simply left to dry in the sun, or any combination thereof. The ground product is subjected to a solvent - commonly xylene, hexane, methanol, ethanol or diethyl ether - for a predetermined period of time, generally less than 24 hours. The extraction may occur whilst standing, stirring or under sonication or maceration, at room temperature or sub-zero268, or even by supercritical fluid extraction269, 270. The solvent is removed, a process which is generally accelerated by rotary evaporator or ultrasound bath, leaving a small quantity of solid extract from the plant. This process can be repeated with new batches of fresh biomatter until the desired amount of extract is procured. Total extract yield data is not frequently given in literature, but is around 0.25% for observed examples carried out within the ACWS project for ethanolic C. crispus extract. Examples where this procedure have been carried out include Amsler et al.146 Chambers et al. 15, 271, da Gama et al.272, Harrison and Chan273, Hellio et al.274, Henrikson and Pawlik 157 Lam et al.85, Lima-Filho et al.275, Maréchal et al.87, Nylund and Pavia108, Tuney et al.276 and Wiltshire et al.268.
This crude extract can be integrated into paint coatings. However, difficulties can arise as the extract contains many compounds, often upwards of 100277, 278. Therefore, the crude extract will contain any of these compounds that were soluble in the solvent of choice - which may not include the secondary metabolites responsible for AF activity – and many of these compounds may exhibit incompatibility when integrated into a polymer film. Separation of natural chlorophyll-based pigment or other components is seen to occur vary rapidly on mixing with binders and additives. In addition, due to the fact that a low proportion of the crude extract will have any AF activity, an impractical amount of extract compared to the weight of polymer solids may be required to observe the desired level of AF performance15. Despite these issues, crude extract addition is not without merit as an approach; the promising early results obtained in many of the previously quoted studies demonstrate the presence of potential AF sources.
Unfortunately, many crude extracts are not integrated into coatings and trialled in in-situ immersion in the sea, and although they may prove effective in single-species bioassays, it is difficult to evaluate performance based on these data alone. Bioassays have an important role, however, in the elucidation of compounds with high AF potential, allowing for high throughput down-selection of biocides before coating integration124, 274. Of the crude extracts that have been integrated into coatings and tested at sea, many have proven effective for up to 3 months, but have become rapidly fouled afterwards 15. More often, the
40
Chapter 2 Literature Review effective life of coatings is restricted to only a couple of weeks. This could be due, in part, to the water solubility of certain active components of the crude extract.
2.3.1.2. Isolated extract approach
An alternative method of employment involves identifying the compound (or compounds) responsible for the AF properties of the organism. This involves fractionation of the crude extract, determination of the structure and stereochemistry of the compound, and extraction of the single useful compound. However due to the specific nature of the extraction and the number of steps required, total yield may be extremely low, in the order of 0.00047%; i.e. 2.13 kg of crude product is required to produce one gram of isolated secondary metabolite277. Bioassays play a crucial role in the determination of where the active compounds lie, during the fractionation procedure. This procedure of fractionation is time-consuming and costly, and requires a greater quantity of crude product as a result of the low yield. However, it provides several advantages over the crude extract method; the lack of superfluous non-active compounds and specificity of the biocidal component allow integration in the coating at a much lower concentration level. For example, concentrations of sponge isolates as low as 0.01% – 0.1% in a coating reduced Balanus amphitrite and Crepidula fornicata recruitment by ~50 and 89%, respectively164.
Another benefit of a single, isolated biocide is that binder and solvent selection can be tailored to the chemical properties of the biocide molecule. For example, biocidal medetomine was found to form strong associations with the carboxylic acid group on the alkyd polymer backbone67. It is suggested that this type of biocide-binder interaction could help to control the biocide release rate via a mechanism analogous to self-polishing technologies.
2.3.2. Challenging aspects of natural product usage
Even given the successful elucidation of a secondary metabolite or crude extract with AF activity, there are still several considerations, both general and specific to the EDA project, to be taken into account in terms of coating design. Some of these considerations are outlined in this section.
41
Chapter 2 Literature Review
2.3.2.1. Can the necessary fouling spectrum be addressed?
The AF spectrum of most secondary metabolites is quite limited, and there is rarely only one present in an organism86, 95. Furthermore, most fouling organisms are mainly adapted against foulers that cause pressure in a certain region. The ability of organisms to shift to a certain brand of AF defence is proven by intraspecial variation – the ability of two geographically remote individuals of a same species to produce different concentrations of certain product, dependant on the local fouling community84. It seems unlikely that a single AF compound from a single species could totally prevent fouling in a single given locale, let alone anywhere in the world. Therefore, it is probable that several active compounds will be required for naval vessels that could be called for duty anywhere. For a complete AF spectrum, these active compounds may need to be sourced from organisms from more than one location worldwide.
2.3.2.2 Can biocide distribution be controlled?
The physical distribution of the biocide within the coating is unstudied before present, but is likely to be crucial to the functioning, longevity and application of the coating. A biocide that is poorly compatible with its binder or solvent may rise to the top or sink to the bottom of the drying film, or form discrete crystals or other structures that may form weak points in the coating that act as focal points for ingress of water or contaminant trapping. In the case of a crude extract, it is conceivable that all of these scenarios could occur where different fractions of the extract have very different behaviour.
2.3.2.3 How is the rate of depletion to be controlled?
Assuming that a homogeneous and compatible distribution has been achieved, it is important to recognise that the depletion rate of the NP in relation to the binder may differ. If the rate of depletion of the biocide is greater than the rate of erosion of the polymer, then a leached layer will form, impeding the release of further biocide, collecting impurities from the water, and lowering the lifespan of the coating (Fig. 2.9). In the study of Burgess et al.279, for example, five natural AF extracts were successfully integrated into a coating. However, the biocides leached out too rapidly, resulting in a very short term efficacy.
42
Chapter 2 Literature Review
Fig 2.9: Where biocide (green) release rate exceeds the erosion of the polymer (blue), the effective lifespan of the coating is reduced.
The success of TBT paints was founded on the depletion of biocide and binder being controlled and synchronous. A similar controlled self-polishing behaviour is likely to be necessary to obtain the long-life (5+ years) coating that is required. The erosion rate and durability of the polymer can be altered by modifying Tg and molecular weight of the repeating unit280, but it is likely that the AF compounds will be present as separate entities in the paint coating, unless a method of cross-linking or association with the binder is determined. It is recognised that a non-covalent association of biocide and binder may be achieved67, although it is uncertain if this process alters the erosion rate of the polymer, or could potentially provide a self-polishing effect by providing a hydrolysable linkage to the polymer. However, enabling such association of the biocide molecule to the binder becomes exponentially more difficult as more biocides are integrated, as a high specificity of binder and biocide molecule would be required to allow a strong interaction.
Differing water solubilities of the various film components will likely alter the individual leaching rate of each component and overall leaching rate. The characterisation of compounds in terms of solubility parameters is discussed in Section 1.4.
2.3.2.4 Is cultivation of the target organism feasible?
The amount of paint required to coat a warship can be estimated. Given a Royal Navy Type 45r class destroyer with length 133 m and beam 16.1 m, from the commonly used formula for wetted hull area L x B x 0.85 it can be estimated that the area to be bottom coated is approximately 1820 m2. Assuming an average coverage of 4 m2/litre of paint (International Marine Paint product data sheets for Intersmooth 465 SPC), roughly 455 litres of paint would be required to completely coat the ship. Intersmooth 465 marine paint weighs
43
Chapter 2 Literature Review approximately 34 kg per 20 L including pigmentation, equating to roughly 774 kg per ship. Considering the yields given for crude extract and isolated compounds (~0.25% and ~0.00047%277 respectively), we can estimate the amount of algae that would need to be processed to make 774 kg of paint:
For crude extract: assuming that the crude extract is integrated at 10% on the polymer solids15, and that the polymer is at 40% wt., 31 kg of crude extract would be required. At a yield of 0.25%, 12400 kg of raw NP would need to be processed. In the case where individual secondary metabolites are used, assuming a concentration of 0.1% on the polymer164 and that the polymer is at 40% wt., 310 g of isolated secondary metabolite would be needed. At a yield of 0.00047%277, 660 kg of raw NP would be processed for each biocide to be extracted.
In terms of quantity of material to cultivate and process, it can be seen that the second route is the more viable, due largely to the much smaller amount of material required to have the desired level of efficacy. As trawling coastal areas for the selected species of algae would be environmentally devastating and economically impractical, large-scale cultivation in a dedicated site may be the only viable option. To quote Armstrong et al.281,
“Because of both environmental concerns and economic limitations, it is unlikely that natural populations can provide sufficient tissue for extensive research or exploitation of interesting natural products. While it may be possible to synthesise some of the natural products or artificially produce analogues, many compounds are likely to be too costly to synthesise. This leaves culture as the only feasible method of producing sufficient quantities of these compounds.”
The cultivation of a very large, high biomass algal field for this purpose would require large- scale feasibility studies, planning and research in its own right. It is suggested by several authors that isolation of bacterial extract would be economically more feasible than that of algal extract, owing to relatively higher yields of bioactive compound from bacteria60, 123.
2.4. Consideration of biocide release rate
As described above, the delivery of biocides to the surface of the paint to the encroaching organisms is key to the efficacy of a coating system. A considerable body of work has been
44
Chapter 2 Literature Review published on the mechanics of copper (I) oxide release, and to a lesser extent, zinc (II) oxide181, 182, 185, 282, 283, owing to their widespread use as biocidal pigments, as well as TBT from self-polishing paints283-287. Work by Yebra et al.181, 185 focused on estimation of copper and rosin reaction rates in precisely calibrated model binders, including modelling of seawater and copper speciation within the coating leached layer. The release of copper oxide and its derivatives, incorporated as relatively large particles in the order of 5 µm, is largely governed by chemical speciation and reaction at the seawater front; however, the release rate of organic biocides dispersed finely in the polymer matrix are more likely driven by diffusivity, their solubility in water, and affinity with the polymer binder. An understanding of the interaction of AF biocides within the binder matrix is crucial in predicting their capacity for leaching. The mechanics of diffusion in polymers are significantly more complex than in true solids, liquids or gases, because of the polymer medium’s inherent physicochemical heterogeneity and variable temperature dependant conformation. Nonetheless, a good understanding of the key mechanics governing diffusion in polymers has been achieved in areas such as pharmaceutical applications for drug delivery, and in food packaging288-294. Direct monitoring and measurement of diffusion rates is a lengthy and arduous procedure, particularly when considering materials with an intentionally low release rate. Unfortunately, literature efforts in modelling of diffusion have focused primarily on amorphous polymers above glass transition temperature295-300.
301, 302 Consideration of below Tg acrylic polymers is limited and models designed to estimate penetrant diffusion in glassy polymers have been found to err significantly for penetrants whose size scale differs from that of the repeating monomer unit303. Despite the limitations of the available models, they have been applied in a modified form to a model coating and biocide in this current work as later discussed.
When considering diffusion of solute from packaging to food, the total solute amount in both systems is constant throughout. In this case, diffusion will drive the migrant to an equilibrium with a constant concentration ratio in both food and packaging, the balance of which depends on its solubility in both systems. In the case of AF systems, another well- studied diffusion case is a much more pertinent analogy; in the area of drug delivery, the aim of delivering an active substance is the same, and once the active substance has diffused out of its binder, it no longer plays a part in the system. However, AF coatings present a more complex, multicomponent system. Water ingress occurs into these coatings in the marine environment, resulting in a degree of plasticisation of polymer chains. The inclusion of large pigment particles, whose size scales are within an order of magnitude of
45
Chapter 2 Literature Review the total binder thickness, renders the medium effectively anisotropic and further complicates any one-dimensional representation of the system. There is modelling and experimental evidence demonstrating the limitation of diffusion below the pigment front in AF paints181, 282, 283, 304. As a result, some assumptions must be made in order to treat the system as a traditional one-dimensional diffusion problem.
As earlier discussed, efforts in experimenting with new biocides or crude extracts often rely on addition of a compound or mixture to a pre-existing acrylic and/or rosin-based paint formulation 15, 252, 271, 279, 305-308. Although bioassays might demonstrate favourable results, these coatings rarely produce favourable long-term results in static immersion. In the case of crude algal or naturally-derived extracts, a common method for preparation is maceration of biomatter in various solvents, which are down-selected to identify which of these demonstrate some form of AF capacity. Assuming excellent activity for a lone compound within a crude extract, the amount of crude extract required to achieve a critical level of this compound may be impractically high to achieve the required level of both AF efficacy and coating cohesion or adhesion. Clearly, identification and isolation of the compound and a full understanding of its interactive capacity with other coating components is required for optimisation. This process is highly complex and time- consuming104, 269, 270, 307-311. The most logical method of creating an entire coating system seems to be to begin with a combination or ‘cocktail’ of biocides with established efficacy, and then working backwards to formulate a (co)polymeric binder with desirable properties based on the choice of biocide, in order to optimise release rate. Indeed, consideration of the biocide in relation to its binder has increasingly been observed312. As such, positive results for coatings bearing active ingredient concentrations as low as 0.1 wt.% - 1 wt.% dry film thickness have been noted, such as in the study of Pinori et al.312, wherein the low water solubility of a large macrocyclic lactone biocide was considered in relation to the rosin-based binder in terms of its octanol-water partition coefficient, concluding a negative affinity for water and the potential for intermolecular interaction with the rosin. In the latter study, a good anti-barnacle efficacy was observed despite the low level of active substance. There remains more room for expansion of this concept, however; one possible avenue for consideration is via estimation of solubility parameters, which allow characterisation and quantitative prediction of interaction between molecules. Hansen313-315 expanded on the concept of the lone Hildebrand solubility parameter, δ, noting that compounds with similar solubility parameters would not necessarily interact completely in the case of highly polar compounds or those that hydrogen bond, resulting in anomalies for many organic
46
Chapter 2 Literature Review molecules. The result of this work was the development of three Hansen solubility parameters (HSPs). The capacity of a molecule to form hydrogen bonds (δh) as well as its polarity (δp) and predisposition to dispersion forces (i.e. instantaneous dipole-induced dipole, δd) may all be assessed separately. It is worth noting that for the δp term, uniquely, a value of zero can be achieved for non-polar, symmetrical molecules such as benzene or straight chain alkanes. These three parameters may be considered in terms of absolute values, though the relative contributions of each term to a total solubility parameter are often of more use. Considerable effort has been made to provide a database for commonplace solvents and polymers314, which is of considerable use in the formation of solvent blends, ably predicted by the HSP, for acrylic resins and copolymers in AF paints. For larger and more complex NPs or organic biocides, determination of HSP is much less straightforward. Experimental evaluation is possible by determination of solubility in a number of chemically different solvents in order to narrow down the target’s solubility. Accurate determination of the concentration in the solvent at potentially low levels (i.e. via HPLC or LC-MS) is important. Stefanis and Panayiotou’s group contribution model316 allows estimation via analysis of the chemical structure and its possible conjugates, which provides a good match for experimentally derived HSP for even complex compounds. Compounds with very large molar volumes and complex stereochemistry are likely to propagate error, as the increased structural complexity can result in shielding at the interior of the molecule.
For example, this could lead to overestimation of the contribution to δp of a carboxylic acid moiety at the centre of a molecule, and hence has reduced potential for intermolecular interaction. However, the group contribution model could be expected to give accurate results for most AF biocides. In the case of multiple biocides being employed within a single binder system, corroboration of the biocides’ HSPs to the binder’s HSP would ensure a maximal affinity between the components, and would be an elegant way to control release rate. Work on diffusion of specific compounds from packaging into food has demonstrated corroboration of these solubility parameters with diffusion rates depending on the fat or water content of the solvent food289.
Evaluation of biocide distribution has only recently been considered in the literature. Certain biocides, in the case of transition metal pigments, with AF capacity, have been analysable within coating cross-sections by simple optical microscopy. The 2005 paper of Faӱ et al.317 on SEM-EDX analysis of eroded coating cross-sections provided a method of corroboration for investigating pigment fronts and movement of pigment derivatives in coating leached layers, proving an excellent technique for analysis of worn cross-sections in
47
Chapter 2 Literature Review particular or those where multiple pigments are employed.
2.5. References
1. Callow, M.E. and Callow, J.A., Marine biofouling: a sticky problem. Biologist 2002. 49 (1): p. 1-5. 2. Konstantinou, I.K. and Albanis, T.A., Worldwide occurrence and effects of antifouling paint booster biocides in the aquatic environment: a review. Environment International, 2004. 30: p. 235-248. 3. WHOI, Woods Hole Oceanographic Institute (Massachusetts) : Marine Fouling and its Prevention1952: US Naval Institute (Annapolis, Maryland). 4. Krug, P.J., Defence of benthic invertebrates against surface colonization by larvae; a chemical arms race. Progress in Molecular and Subcellular Biology, 2006. 42: p. 1-53. 5. Lambourne, R., The painting of ships, in Paint and Surface Coatings - Theory and Practice, R. Lambourne and T.A. Strivens, Editors. 2004, Woodhead Publishing Limited/William Andrew Publishing. 6. Bray, S., Tributyltin pollution on a global scale. An overview of relevant and recent research: impacts and issues., 2006, WWF UK. p. 54. 7. Wieczorek, S.K. and Todd, C.D., Inhibition and facilitation of settlement of epifanual marine invertebrate larvae by microbial biofilm cues. Biofouling, 1998. 12(1-3): p. 81- 118. 8. Schultz, M.P., Frictional resistance of antifouling coating systems. Journal of Fluids Engineering, 2004. 126: p. 1039-1047. 9. Schultz, M.P., Effects of coating roughness and biofouling on ship resistance and powering. Biofouling, 2007. 23(5): p. 331-341. 10. Schultz, M.P. and Swain, G.W., The influence of biofilms on skin friction drag. Biofouling, 2000. 15(1-3): p. 129-139. 11. Moss, B. and Woodhead, P., The breakdown of paint surfaces by Enteromorpha sp. New Phytologist, 1970. 69(4): p. 1025-1027. 12. Meech, R. Impact of marine emission legislation on the bunker industry: pricing. Pre- Conference Workshop (MARPOL annex VI: regulations on marine fuels/air emissions). in EnviroArabia. 2007. 13. Gitlitz, M.H., Recent developments in marine antifouling coatings. Journal of Coatings Technology, 1981. 53: p. 46-52. 14. Brady, R.F., A fracture mechanical analysis of fouling release from nontoxic antifouling
48
Chapter 2 Literature Review
coatings. Progress in Organic Coatings, 2001. 43: p. 188-192. 15. Chambers, L., The development of a marine antifouling systems using environmentally acceptable and naturally occurring products, in Surface Engineering and Tribology Research Group2008, University of Southampton. p. 210. 16. Zhao, Q., Effect of surface free energy of graded NI-P-PTFE coatings on bacterial adhesion. Surface & Coatings Technology, 2004. 185: p. 199-204. 17. Champ, M.A., New IMO convention to ban harmful antifouling systems on ships. Sea Technology, 2001. 42(11): p. 48-51. 18. Champ, M.A., Economic and environmental impacts on ports and harbors from the convention to ban harmful marine anti-fouling systems. Marine Pollution Bulletin, 2003. 46: p. 935-940. 19. Mineur, F., Johnson, M.P., and Maggs, C.A., Hull fouling on commercial ships as a vector of macroalgal introduction. Marine Biology, 2007. 151: p. 1299-1307. 20. Minchin, D., Gollasch, S., and Wallentinus, I., Vector pathways and the spread of exotic species in the sea, in ICES Cooperative Research Report. 2005. p. 25. 21. Minchin, D. and Gollasch, S., Fouling and ships' hulls: how changing circumstances and spawning events may result in the spread of exotic species. Biofouling, 2003. 19(1S1): p. 111-122. 22. De Montaudouin, X., Labarraque, D., Giraud, K., and Bachelet, G. La crépidule Crepidula fornicata (Linné, 1758) dans le bassin d'Arcachon: caractérisation du stock. in Colloque international d'océanographie du Golfe de Gascogne No7. 2001. Biarritz. 23. Howell, D.J. and Evans, S.M., Antifouling Materials, in Encyclopedia of Ocean Sciences, J.H. Steele, K.K. Turekian, and S.A. Thorpe, Editors. 2009, Elsevier. p. 3463-3470. 24. Piola, R.F. and Johnston, E.L., The potential for translocation of marine species via snall-scale disruptions to antifouling surfaces. Biofouling, 2008. 24(3): p. 145-155. 25. Hewitt, C.L. Biofouling as a modern vector of invasions: risky behaviours and management opportunities. in 15th International Congress on Marine Corrosion and Fouling. 2010. Newcastle, England. 26. Lewis, P.N., Riddle, M.J., and Hewitt, C.L., Management of exogenous threats to Antarctica and sub-Antarctic islands: balancing risks from TBT and non-indigenous organisms. Marine Pollution Bulletin, 2004. 49: p. 999-1005. 27. Otani, M., Oumi, T., Uwai, S., Hanyuda, T., Prabowo, R.E., Yamaguchi, T., and Kawai, H., Occurrence and diversity of barnacles on international ships visiting Osaka Bay, Japan, and the risk of their introduction. Biofouling, 2007. 23(3/4): p. 277-286. 28. Obhodas, J., Kutle, A., and Valkovic, V., Concentrations of some elements in coastal
49
Chapter 2 Literature Review
sea sediments: bays with marinas. Journal of Radioanalytical and Nuclear Chemistry, 2006. 270(1): p. 75-85. 29. Voulvoulis, N., Scrimshaw, M.D., and Lester, J.N., Comparative environmental assessment of biocides used in antifouling paints. Chemosphere, 2002. 47: p. 789- 795. 30. Voulvoulis, N., Scrimshaw, M.D., and Lester, J.N., Partitioning of selected antifouling biocides in the aquatic environment. Marine Environmental Research, 2002. 53: p. 1- 16. 31. Geesey, G.G., Bacterial behaviour at surfaces. Current Opinion in Microbiology, 2001. 4: p. 296-300. 32. Wahl, M., Marine epibiosis. I. Fouling and antifouling: some basic aspects. Marine Ecology Progress Series, 1989. 58: p. 175-189. 33. Baier, R.E., Initial Events in Microbial Film Formation, in Marine Biodeterioration: An Interdisciplinary Study, J.D. Costlow, and Tipper, R.C., Editor 1984, Naval Institute Press: Annapolis, MD. 34. Costerton, J.W., Introduction to biofilm. International Journal of Antimicrobial Agents, 1999. 11: p. 217-221. 35. Parsek, M.R. and Fuqua, C., Biofilms 2003: Emerging themes and challenges in studies of surface-associated microbial life. Journal of Bacteriology, 2004. 186(14): p. 4427- 4440. 36. Lynch, M.J., Swift, S., Kirke, D.F., Keevil, C.W., Dodd, C.E.R., and Williams, P., The regulation of biofilm development by quorum sensing in Aeromonas hydrophila. Environmental Microbiology, 2002. 4(1): p. 18-28. 37. Cooksey, K.E. and Wigglesworth-Cooksey, B., Diatoms in Biofilms, in Encyclopedia of Environmental Microbiology, 2003, John Wiley & Sons, Inc. 38. Edyvean, R.G.J., Rands, G.A., and Moss, B.L., A comparison of diatom colonization on natural and artifical substrata in seawater. Estuarine Coastal and Shelf Science, 1985. 20: p. 233-238. 39. Gebeshuber, I.C., Stachelberger, H., Drack, M., D. Dowson, M.P.G.D., and Lubrecht, A.A., Diatom biotribology, in Tribology and Interface Engineering Series, 2005, Elsevier. p. 365-370. 40. Patil, J.S. and Anil, A.C., Quantification of diatoms in biofilms: Standardisation of methods. Biofouling, 2005. 21(3): p. 181-188. 41. Patil, J.S. and Anil, A.C., Biofilm diatom community structure: Influence of temporal and substratum variability. Biofouling, 2005. 21(3): p. 189-206.
50
Chapter 2 Literature Review
42. Salta, M., Wharton, J.A., Stoodley, P., Dennington, S.P., Goodes, L., Werwinski, S., Mart, U., Wood, R.J.K., and Stokes, K.R., Designing Biomimetic Antifouling Surfaces. Philosophical Transactions of the Royal Society A, 2010. 368: p. 4729-4754. 43. Smith, A.M., Callow, J.A., Chiovitti, A., Dugdale, T.M., and Wetherbee, R., Diatom adhesives: molecular and mechanical properties, in Biological Adhesives, 2006, Springer Berlin Heidelberg. p. 79-103. 44. Wigglesworth-Cooksey, B. and Cooksey, K.E., Can diatoms sense surfaces?: State of our knowledge. Biofouling, 1992. 5(3): p. 227-238. 45. Cooksey, K.E., Cooksey, B., Evans, L.V., and Hoagland, K.D., Adhesion of fouling diatoms to surfaces: some biochemistry, in Studies in Environmental Science, 1986, Elsevier. p. 41-53. 46. Kristensen, J.B., Meyer, R.L., Laursen, B.S., Shipovskov, S., Besenbacher, F., and Poulsen, C.H., Antifouling enzymes and the biochemistry of marine settlement. Biotechnology Advances, 2008. 26: p. 471-481. 47. Roeselers, G., van Loosdrecht, M.C.M., and Muyzer, G., Heterotrophic pioneers facilitate phototrophic biofilm development. Microbial Ecology, 2007. 54: p. 578-585. 48. Cassé, F. and Swain, G.W., The development of microfouling on four commercial antifouling coatings under static and dynamic immersion. International Biodeterioration and Biodegration, 2006. 57: p. 179-185. 49. Molino, P.J. and Wetherbee, R., The biology of biofouling diatoms and their role in the development of miocrobial slimes. Biofouling, 2008. 24(5): p. 365-379. 50. Yebra, D.M., Kiil, S., Weinell, C., and Dam-Johansen, K., Effects of marine microbial biofilms in the biocide release rate from antifouling paints – A model-based analysis. Progress in Organic Coatings, 2006. 57: p. 56-66. 51. Chapman, J.S., Biocide resistance mechanisms. International Biodeterioration and Biodegration, 2003. 51: p. 133-138. 52. Silver, S. and Phung Le, Bacterial heavy metal resistance: new surprises. Annual Review of Microbiology, 1996. 50: p. 753-789. 53. Morton, L.H.G., Greenway, D.L.A., Gaylarde, C.C., and Surman, S.B., Consideration of some implications of the resistance of biofilms to biocides. International Biodeterioration and Biodegration, 1998. 41: p. 247-259. 54. Townsin, R.L., The ship hull fouling penalty. Biofouling, 2003. 19S: p. 9-15. 55. Callow, M.E., Callow, J.A., Ista, L.K., Coleman, S.E., Nolasco, A.C., and Lopez, G.P., Use of self-assembled monolayers of different wettabilities to study surface selection and primary adhesion processes of green algal (Enteromorpha) zoospores. Applied and
51
Chapter 2 Literature Review
Environmental Microbiology, 2000. 66(8): p. 3249-3254. 56. Scardino, A.J., Biomimetic fouling control, in School of Marine Biology and Aquaculture. 2006, James Cook University. p. 188. 57. Matsumura, K., Nagano, M., Kato-Yoshinaga, Y., Yamazaki, M., Clare, A.S., and Fusetani, N., Immunological studies on the settlement-inducing protein complex (SIPC) of the Barnacle Balanus amphitrite and its possible involvement in larva-larva interactions. Proceedings: Biological Sciences, 1998. 265(1408): p. 1825-1830. 58. Bers, A.V. and Wahl, M., The influence of natural surface microtopographies on fouling. Biofouling, 2004. 20(1): p. 43-51. 59. Rittschof, D., A perspective from 2 decades of academic lab and field-testing of experimental and commercial coatings, in 15th International Congress on Marine Corrosion and Fouling2010: Newcastle, England. 60. Dobretsov, S., Dahms, H.-U., and Qian, P.-Y., Inhibition of biofouling by marine microorganisms and their metabolites. Biofouling, 2006. 22(1): p. 43-54. 61. Mitbavkar, S. and Anil, A.C., Seasonal variations in the fouling diatom community structure from a monsoon influenced tropical estuary. Biofouling, 2008. 24(6): p. 415- 426. 62. Pinori, E., Berglin, M., Sjövali, P., Dahlström, M., and Elwing, H. Post-settlement inhibition (PSI) inhibition of barnacle growth, Balanus improvisus. A novel approach in marine antifouling control. in 15th International Congress on Marine Corrosion and Fouling. 2010. Newcastle, England. 63. Bellas, J., Hilvarsson, A., Birgersson, G., and Granmo, A., Effects of medetomidine, a novel antifouling agent, on the burrowing bivalve Abra nitida (Müller). Chemosphere, 2006. 65(4): p. 575-582. 64. Bellas, J., Hilvarsson, A., and Granmo, Å., Sublethal effects of a new antifouling candidate on lumpfish (Cyclopterus lumpus L.) and Atlantic cod (Gadus morhua L.) larvae. Biofouling, 2005. 21(3-4): p. 207-216. 65. Dahlstrom, M., Jonsson, P.R., Lausmaa, J., Arnebrant, T., Sjogren, M., Holmberg, K., Martensson, L., and Elwing, H., Impact of polymer surface affinity of novel antifouling agents. Biotechnology and Bioengineering, 2004. 86(1): p. 1-8. 66. Dahlström, M., Mårtensson, L.G.E., Jonsson, P.R., Arnebrant, T., and Elwing, H., Surface active adrenoceptor compounds prevent the settlement of cyprid larvae of Balanus improvisus. Biofouling, 2000. 16(2-4): p. 191-203. 67. Shtykova, L.S., Ostrovskii, D., Handa, P., Holmberg, K., and Nydén, M., NMR diffusometry and FTIR in the study of the interaction between antifouling agent and
52
Chapter 2 Literature Review
binder in marine paints. Progress in Organic Coatings, 2004. 51: p. 125-133. 68. Ludgate, J.W. The economic and technical impact of TBT legislation on the USA marine industry. in Oceans 87 International Organotin Symposium. 1987. Institute of Electronic and Electrical Engineering, New York. 69. Schultz, M.P. and Swain, G.W., The influence of biofilms on skin friction drag. Biofouling, 2000. 15(1): p. 129 - 139. 70. Schultz, M.P., Bendick, J.A., Holm, E.R., and Hertel, W.M. Economic impact of biofouling on a naval ship. in 15th International Congress on Marine Corrosion and Fouling. 2010. Newcastle, England. 71. Yebra, D.M., Kiil, S., and Dam-Johansen, K., Antifouling technology – past, present and future steps towards efficient and environmentally friendly antifouling coatings. Progress in Organic Coatings, 2004. 50: p. 75-104. 72. Callow, M.E., Fouling algae from 'in-service' ships. . Botanical Materials 1986. 24: p. 351-357. 73. Libes, S.M., An Introduction to Marine Biogeochemistry1992: John Wiley & Sons. 74. Steinberg, P.D. and De Nys, R., Chemical mediation of colonization of seaweed surfaces. Journal of Phycology, 2002. 38: p. 621-629. 75. Bryan, P.J., Rittschof, D., and McClintock, J.B., Bioactivity of echinoderm ethanolic body-wall extracts: an assessment of marine bacterial attachment and macroinvertebrate larval settlement. Journal of Experimental Marine Biology and Ecology, 1996. 196: p. 79-96. 76. Fusetani, N., Biofouling and antifouling. Natural Product Reports, 2004. 21: p. 94-104. 77. Guenther, J. and De Nys, R., Differential community development of fouling species on the pearl oysters Pinctada fucata, Pteria penguin and Pteria chinensis (Bivalvia, Pteriidae). Biofouling, 2006. 22(3): p. 151 - 159. 78. Guenther, J., Walker-Smith, G., Warén, A., and De Nys, R., Fouling-resistant surfaces of tropical sea stars. Biofouling, 2007. 23(6): p. 413-418. 79. Baum, C., Meyer, W., Stelzer, R., Fleischer, L.-G., and Siebers, D., Average nanorough skin surface of the pilot whale (Globicephala melas, Delphinidae): considerations on the self-cleaning abilities based on nanoroughness. Marine Biology, 2002. 140: p. 653- 657. 80. Bakus, G.J. and Kawaguchi, M., Toxins from marine organisms: studies on antifouling, in Toxins, Drugs, and Pollutants in Marine Animals L. Bolis, J. Zadunaisky, and R. Gilles, Editors. 1984, Springer-Verlag: Berlin. p. 43-46. 81. Bakus, G.J., Wright, M., Khan, A.K., Ormsby, B., Gulko, D.A., Licuanan, W., Carriazo, E.,
53
Chapter 2 Literature Review
Ortiz, A., Chan, D.B., Lorenzana, D., and Huxely, M.P., Experiments seeking marine natural antifouling compounds, in Recent Developments in Biofouling Control, M.-F. Thompson, et al., Editors. 1994. 82. Pereira, R.C., Da Gama, B.A.P., Teixeira, V.L., and Yoneshigue-Valentin, Y., Ecological roles of natural products of the Brazilian red seaweed Laurencia obtusa. Brazilian Journal of Biology, 2003. 63(4): p. 665-672. 83. Burns, E., Ifrach, I., Carmeli, S., Pawlik, J.R., and Ilan, M., Comparison of anti-predatory defenses of Red Sea and Caribbean sponges. I. Chemical defense. Marine Ecology Progress Series, 2003. 252: p. 105-114. 84. Da Gama, B.A.P., Pereira, R.C., Carvalho, A.G.V., and Coutinho, R., The effects of seaweed secondary metabolites on biofouling. Biofouling, 2002. 18(1): p. 13-20. 85. Lam, C., Grage, A., Schulz, D., Schulte, A., and Harder, T., Extracts of North Sea macroalgae reveal specific activity patterns against attachment and proliferation of benthic diatoms: a laboratory study. Biofouling, 2008. 24(1): p. 59-66. 86. Puglisi, M.P., Paul, V.J., Biggs, J., and Slattery, M., Co-occurrence of chemical and structural defenses in the gorgonian corals of Guam. Marine Ecology Progress Series, 2002. 239: p. 105-114. 87. Maréchal, J.-P., Culioli, G., Hellio, C., Thomas-Guyon, H., Callow, M.E., Clare, A.S., and Ortalo-Magné, A., Seasonal variation in anti-fouling activity of crude extracts of the brown alga Bifurcata bifurcata (Cystoseiraceae) against cyprids of Balanus amphitrite and the marine bacteria Cobetia marina and Psuedomonas haloplanktis. Journal of Experimental Marine Biology and Ecology, 2004. 313: p. 47-62. 88. Blunt, J.W., Copp, B.R., Munro, M.H.G., Northcote, P.T., and Prinsep, M., Marine natural products. Natural Product Reports, 2004. 21: p. 1-49. 89. Blunt, J.W., Copp, B.R., Munro, M.H.G., Northcote, P.T., and Prinsep, M., Marine natural products. Natural Product Reports, 2006. 23: p. 26-78. 90. Blunt, J.W., Copp, B.R., Hu, W.-P., Munro, M.H.G., Northcote, P.T., and Prinsep, M.R., Marine natural products. Natural Product Reports, 2008. 25(1): p. 35-94. 91. Bhadury, P. and Wright, P.C., Exploitation of marine algae: biogenic compounds for potential antifouling applications. Planta, 2004. 219: p. 561-578. 92. Dahms, H.-U., Ying, X., and Pfeiffer, C., Antifouling potential of cyanobacteria: a mini- review. Biofouling, 2006. 22(5): p. 317-327. 93. Dembitsky, V.M. and Rozensvet, O.A., Lipids of macrophyte marine algae. I. Fatty-acid and phospholipid composition of Rhophyceae. Khimiya Prirodnykh Soedinenii, 1990. 1: p. 113-115.
54
Chapter 2 Literature Review
94. Devi, K.P., Suganthy, N., Kesika, P., and Pandian, S.K., Bioprotective properties of seaweeds: In vitro evaluation of antioxidant activity and antimicrobial activity against food borne bacteria in relation to polyphenolic content. Complementary and Alternative Medicine, 2008. 8(38): p. 1-11. 95. Cho, J.-Y., Kwon, E.-H., Choi, J.-S., Hong, S.-Y., Shin, H.-W., and Hong, Y.-K., Antifouling activity of seaweed extracts on the green alga Enteromorpha prolifera and the mussel Mytilus edulis. Journal of Applied Phycology, 2001. 13: p. 117-125. 96. Baveja, J.K., Willcox, M.D.P., Hume, E.B.H., Kumar, N., Odell, R., and Poole-Warren, L.A., Furanones as potential anti-bacterial coatings on biomaterials. Biomaterials, 2004. 25: p. 5003-5012. 97. Gershenzon, J. and Dudareva, N., The function of terpene natural products in the natural world. Nature Chemical Biology, 2007. 3(7): p. 408-414. 98. El-Naggar, M.E.E., A comparative study of the fatty acids in some marine algae from Saudi Arabia. Microbios, 1989. 96(385): p. 177-187. 99. Goto, R., Kado, R., Muramoto, K., Kamiya, H., Fatty acids as antifoulants in a marine sponge. Biofouling, 1992. 59: p. 1953. 100. Vaskovsky, V.E., Khotimchenko, S.V., Bangmei, X., and Li, H., Polar lipids and fatty acids of some marine macrophytes from the yellow sea. Phytochemistry, 1996. 42(5): p. 1347-1356. 101. Fusetani, N., Clare, A.S., Nys, R., Givskov, M., Kumar, N., Kjelleberg, S., and Steinberg, P.D., Furanones, in Antifouling Compounds, W.E.G. Müller, et al., Editors. 2006, Springer Berlin Heidelberg. p. 55-86. 102. Hume, E.B.H., Baveja, J., Muir, B., Schubert, T.L., Kumar, N., Kjelleberg, S., Griesser, H.J., Thissen, H., Read, R., Poole-Warren, L.A., Schindhelm, K., and Willcox, M.D.P., The control of Staphylococcus epidermidis biofilm formation and in vivo infection rates by covalently bound furanones. Biomaterials, 2004. 25: p. 5023-5030. 103. Kjelleberg, S. and Steinberg, P., Surface warfare in the sea, in Microbiology Today2001. p. 134-135. 104. Xu, Y., He, H., Schulz, S., Liu, X., Fusetani, N., Xiong, H., Xiao, X., and Qian, P.-Y., Potent antifouling compounds produced by marine Streptomyces. Bioresource Technology, 2010. 101(4): p. 1331-1336. 105. Dworjanyn, S.A., De Nys, R., and Steinberg, P., Chemically mediated antifouling in the red alga Delisea pulchra. Marine Ecology Progress Series, 2006. 318: p. 153-163. 106. Harder, T., Dobretsov, S., and Qian, P.-Y., Waterborne polar macromolecules act as algal antifoulants in the seaweed Ulva reticulata. Marine Ecology Progress Series,
55
Chapter 2 Literature Review
2004. 274: p. 133-141. 107. Faimali, M., Sepcic, K., Turk, T., and Geraci, S., Non-toxic antifouling activity of polymeric 3-alkylpyridnium salts from the Mediterranean sponge Reneira sarai. Biofouling, 2003. 19(1): p. 47-56. 108. Nylund, G.M. and Pavia, H., Inhibitory effects of red algal extracts on larval settlement of the barnacle Balanus improvisus. Marine Biology, 2003. 143: p. 875-882. 109. Nylund, G.M., Cervin, G., Hermansson, M., and Pavia, H., Chemical inhibition of bacterial colonisation by the red alga Bonnemaisonia hamifera. Marine Ecology Progress Series, 2005. 302: p. 27-36. 110. Nylund, G.M., Chemical mediation of fouling in seaweeds, 2005. 111. de Nys, R., Dworjanyn, S.A., and Steinberg, P.D., A new method for determining surface concentrations of marine natural products on seaweeds. Marine Ecology Progress Series, 1998. 162: p. 79-87. 112. Bazes, A., Silkina, A., Defer, D., Bernede-Baudin, C., Quemener, E., Braud, J.-P., and Bourgougnon, N., Active substances from Ceramium botryocarpum used as antifouling products in aquaculture. Aquaculture, 2006(258): p. 664-674. 113. Silkina, A., Bazes, A., Vouvé, F., Le Tilly, V., Douzenel, P., Mouget, J.-L., and Bourgougnon, N., Antifouling activity of macroalgal extracts on Fragilaria pinnata (Bacillariophyceae): A comparison with Diuron. Aquatic Toxicology, 2009. 94(4): p. 245-254. 114. Cassano, V., De-Paula, J.C., Fujii, M.T., Da Gama, B.A., and Teixeira, V.L., Sesquiterpenes from the introduced red seaweed Laurencia caduciramulosa (Rhodomelaceae, Ceramiales). Biochemical systematics and ecology, 2008(36): p. 223-226. 115. Cho, J.-Y., Choi, J.-S., Kang, S.E., Kim, J.-K., Shin, H.-W., and Hong, Y.-K., Isolation of antifouling active pyroglutamic acid, triethyl citrate and di-n-octylphthalate from the brown seaweed Ishige okamurae. Journal of Applied Phycology, 2005(17): p. 431-435. 116. Kubanek, J., Lester, S.E., Fenical, W., and Hay, M.E., Ambiguous role of phlorotannins as chemical defenses in the brown alga Fucus vesiculosus. Marine Ecology Progress Series, 2004. 277: p. 79-93. 117. Sudatti, D., Rodrigues, S., Coutinho, R., da Gama, B., Salgado, L., Amado Filho, G., and Pereira, R., Transport and defensive role of elatol at the surface of the red seaweed Laurencia obtusa (Ceramiales, Rhodophyta). Journal of Phycology, 2008. 44: p. 548- 591. 118. Ralston, E. and Swain, G.W., Bioinspiration - the source for biofouling control?
56
Chapter 2 Literature Review
Bioinspiration and Biomimetics, 2009. 4: p. 1-9. 119. de Nys, R., Steinberg, P.D., Willemsen, P., Dworjanyn, S.A., Gabelish, C.L., and King, R.J., Broad spectrum effects of secondary metabolites from the red alga Delisea pulchra in antifouling assays. Biofouling, 1995. 8: p. 259-271. 120. de Nys, R., Wright, A.D., Konig, G.M., and Sticher, O., New halogenated furanones from the marine alga Delisea pulchra (cf. fimbriata). Tetrahedron, 1993. 49(48): p. 11213-11220. 121. Martinelli, D., Grossmann, G., Séquin, U., Brandl, H., and Bachofen, R., Effects of natural and chemically synthesised furanones on quorum sensing in Chromobacterium violaceum. BMC Microbiology, 2004. 4: p. 25. 122. Steinberg, P.D., de Nys, R., and Kjelleberg, S., Chemical inhibition of epibiota by Australian seaweeds. Biofouling: The Journal of Bioadhesion and Biofilm Research, 1998. 12(1): p. 227 - 244. 123. Xu, Y., Li, H., Li, X., Xiao, X., and Qian, P.-Y., Inhibitory effects of a branched-chain fatty acid on larval settlement of the polychaete Hydroides elegans. Marine Biotechnology, 2009. 11: p. 495-504. 124. Hellio, C., Simon-Colin, C., Clare, A.S., and Deslandes, E., Isethonic acid and floridoside isolated from the red alga, Grateloupia turuturu, inhibit settlement of Balanus amphitrite cyprid larvae. Biofouling, 2004. 20(3): p. 139-145. 125. Pisut, D.P. and Pawlik, J.R., Anti-predatory chemical defenses of ascidians: secondary metabolites or inorganic acids? Journal of Experimental Marine Biology and Ecology, 2002. 270: p. 203-214. 126. Kupper, F.C., Kloareg, B., Guern, J., and Potin, P., Oligoguluronates elicit an oxidative burst in the brown algal kelp Laminaria digitata. Plant Physiology, 2001. 125: p. 278- 291. 127. Nylund, G.M. and Pavia, H., Chemical versus mechanical inhibition of fouling in the red alga Dilsea carnosa. Marine Ecology Progress Series, 2005. 299: p. 111-121. 128. Scardino, A., Biomimetic fouling control, in School of Marine Biology and Aquaculture2006, James Cook University. p. 188. 129. Scardino, A.J., Guenther, J., and de Nys, R., Attachment point theory revisited: the fouling response to a microtextured matrix. Biofouling, 2008. 24(1): p. 45-53. 130. Scardino, A.J., Harvey, E., and De Nys, R., Testing attachment point theory: diatom attachment on microtextured polyimide biomimics. Biofouling, 2006. 22(1): p. 55-60. 131. Scardino, A.J., Hudleston, D., Peng, Z., Paul, N.A., and de Nys, R., Biomimetic characterisation of key surface parameters for the development of fouling resistant
57
Chapter 2 Literature Review
materials. Biofouling 2009. 25(1): p. 83-93. 132. Scardino, A.J., Zhang, H., Cookson, D.J., Lamb, R.N., and de Nys, R., The role of nano- roughness in antifouling. Biofouling, 2009. 25(8): p. 757-767. 133. Scardino, A. and De Nys, R., Fouling deterrance on the bivalve shell Mytilus galloprovincialis: a physical phenomenon? Biofouling, 2004. 20(4/5): p. 249-257. 134. Hoipkemeier-Wilson, L., Schumacher, J.F., Carman, M.L., Gibson, A.L., Feinberg, A.W., Callow, M.E., Finlay, J.A., Callow, J.A., and Brennan, A.B., Antifouling potential of lubricious, micro-engineered, PDMS elastomers against zoospores of the green fouling alga Ulva (Enteromorpha). Biofouling, 2004. 20(1): p. 53-63. 135. Schumacher, J.F., Aldred, N., Callow, M.E., Finlay, J.A., Callow, J.A., Clare, A.S., and Brennan, A.B., Species-specific engineered antifouling topographies: correlations between the settlement of algal zoospores and barnacle cyprids. Biofouling, 2007. 23(5): p. 307-317. 136. Schumacher, J.F., Carman, M.L., Estes, T.G., Feinberg, A.W., Wilson, L.H., Callow, M.E., Callow, J.A., Finlay, J.A., and Brennan, A.B., Engineered antifouling microtopographies - effect of feature size, geometry, and roughness on settlement of zoospores of the green alga Ulva. Biofouling, 2007. 23(1): p. 55-62. 137. Hoipkemeier-Wilson, L., Schumacher, J.F., Carman, M.L., Gibson, A.L., Feinberg, A.W., Callow, M.E., Finlay, J.A., Callow, J.A., and Brennan, A.B., Antifouling potential of lubricious, micro-engineered, PDMS elastomers against zoospores of the green fouling alga Ulva (Enteromorpha). Biofouling, 2004. 20(1): p. 53-63. 138. Berntsson, K.M., Jonsson, P.R., Lejhall, M., and Gatenholm, P., Analysis of behavioural rejection of micro-textured surfaces and implications from recruitment by the barnacle Balanus improvisus. Journal of Experimental Marine Biology and Ecology, 2000. 251: p. 59-83. 139. Berntsson, K.M., Andreasson, H., Jonsson, P.R., Larsson, L., Ring, K., Petronis, S., and Gatenholm, P., Reduction of barnacle recruitment on micro-textured surfaces: Analysis of effective topographic characteristics and evaluation of skin friction. Biofouling, 2000. 16(2-4): p. 245-261. 140. Scardino, A., De Nys, R., Ison, O., O'Connor, W., and Steinberg, P., Microtopography and antifouling properties of the shell surface of the bivalve molluscs Mytilus galloprovincialis and Pinctada imbricata. Biofouling, 2003. 19S: p. 221-230. 141. Callow, M.E., Jennings, A.R., Brennan, A.B., Seegert, C.E., Gibson, A., Wilson, L., Feinberg, A., Baney, R., and Callow, J.A., Microtopographic cues for settlement of zoospores of the green fouling alga Enteromorpha. Biofouling, 2002. 18: p. 237-245.
58
Chapter 2 Literature Review
142. Genzer, J. and Efimenko, K., Recent developments in superhydrophobic surfaces and their relevance to marine fouling: a review. Biofouling, 2006. 22(5): p. 339-360. 143. Finlay, J.A., Callow, M.E., Ista, L.K., Lopez, G.P., and Callow, J.A., The influence of surface wettability on the adhesion strength of settled spores of the green alga Enteromorpha and the diatom Amphora. Integrative and Comparative Biology, 2002. 42: p. 1116-1122. 144. Ista, L.K., Callow, M.E., Finlay, J.A., Coleman, S.E., Nolasco, A.C., Simons, R.H., Callow, J.A., and Lopez, G.P., Effect of substratum surface chemistry and surface energy on attachment of marine bacteria and algal spores. Applied and Enviornmental Microbiology, 2004. 70(7): p. 4151-4157. 145. Meyer, W. and Seegers, U., A preliminary approach to epidermal antimicrobial defense in the Delphinidae. Marine Biology, 2004. 144: p. 841-844. 146. Amsler, C.D., Moeller, C.B., McClintock, J.B., Iken, K.B., and Baker, B.J., Chemical defenses against diatom fouling in Antarctic marine sponges. Biofouling, 2000. 16(1): p. 29 - 45. 147. Becerro, M.A., Lopez, N.I., Touron, X., and Uriz, M.J., Antimicrobial activity and surface bacterial film in marine sponges. Journal of Experimental Marine Biology and Ecology, 1994. 179: p. 195-205. 148. Berlinck, R.G.S., Braekman, J.C., Daloze, D., Hallenga, K., Ottinger, R., Bruno, I., and Riccio, R., Two new guanidine alkaloids from the Mediterranean sponge Crambe crambe. Tetrahedon Letters, 1990. 31(45): p. 6531-6534. 149. Braekman, J.C. and Daloze, D., Chemical and biological aspects of sponge secondary metabolites. Phytochemistry Reviews, 2004. 3(3): p. 275-283. 150. Burns, E. and Ilan, M., Comparison of anti-predatory defenses of Red Sea and Caribbean sponges. II. Physical defense. Marine Ecology Progress Series, 2003. 252: p. 115-123. 151. Bobzin, S.C. and Faulkner, D.J., Chemistry and chemical ecology of the Bahamian sponge Aplysilla glacias. Journal of Chemical Ecology, 1992. 18: p. 309-332. 152. Pawlik, J.R., Marine invertebrate chemical defenses. Chemical Reviews, 1993. 93(5): p. 1911-1922. 153. McClintock, J.B., Amsler, M.O., Amsler, C.D., Southworth, K.J., Petrie, C., and J, B.B., Biochemical composition, energy content and chemical antifeedant and antifoulant defenses of the colonial Antarctic ascidian Distaplia cylindrica. Marine Biology, 2004. 145: p. 885-894. 154. Armstrong, E., McKenzie, J.D., and Goldsworthy, G.T., Aquaculture of sponges on
59
Chapter 2 Literature Review
scallops for natural products research and antifouling. Journal of Biotechnology, 1999. 70: p. 163-174. 155. Clark, R.J., Field, K.L., Charan, R.D., Garson, M.J., Brereton, I.M., and Willis, A.C., The haliclonacyclamines, cytotoxic tertiary alkaloids from the tropical marine sponge Haliclona sp. Tetrahedon, 1998. 54: p. 8811-8826. 156. Faimali, M., Sepčić, K., Turk, T., and Geraci, S., Non-toxic antifouling activity of polymeric 3-alkylpyridinium salts from the mediterranean sponge Reniera sarai (Pulitzer-Finali). Biofouling, 2003. 19(1): p. 47-56. 157. Henrikson, A.A. and Pawlik, J.R., A new antifouling assay method: results from field experiments using extracts of four marine organisms. Journal of Experimental Marine Biology and Ecology, 1995. 194: p. 157-165. 158. McCaffery, E.J., and Endean, R., Antimicrobial activity of tropical and subtropical sponges. Marine Biology, 1985. 89: p. 1-8. 159. Paul, V., Puglisi, M., and Ritson-Williams, R., Marine chemical ecology. Natural Product Reports, 2006. 23: p. 153-180. 160. Puyana, M., Fenical, W., and Pawlik, J.R., Are there activated chemical defenses in sponges of the genus Aplysina from the Caribbean? Marine Ecology Progress Series, 2003. 246: p. 127-135. 161. Sarma, N.S., Rambabu, M., Anjaneyulu, A.S.R., and Rao, C.B.S., Strucutre and stereo- chemistry of herbacin, new furanosesquiterpene from the marine sponge, Dysidea herbacea. Industrial Journal of Chemistry, 1986. 25(B): p. 1001-1003. 162. Schoenfeld, R.C. and Ganem, B., Synthesis of ceratimamine and moloka'iamine: antifouling agents from the marine sponge Pseudoceratina purpurea. Tetrahedron Letters, 1998. 39: p. 4147-4150. 163. Sears, M.A., Gerhart, D.J., and Rittschof, D., Antifouling agents from marine sponge Lissodendoryx isodictyalis Carter. Journal of Chemical Ecology, 1990. 16: p. 791-799. 164. Sjögren, M., Dahlström, M., Göransson, U., Jonsson, P.R., and Bohlin, L., Recruitment in the field of Balanus improvisus and Mytilus edulis in response to the antifouling cyclopeptides barettin and 8,9-dihydrobarettin from the marine sponge Geodia barretti. Biofouling, 2004. 20(6): p. 291-297. 165. Thompson, J.E., Walker, R.P., and Faulkner, D.J., Screening and bioassays for biologically-active substances from forty marine sponge species from San Diego, California, USA. Marine Biology, 1985. 88: p. 11-21. 166. Tomono, Y., Hirota, H., and Fusetani, N., Antifouling Compounds against (Balanus amphitrite) larvae from the marine sponge Acanthella cavernosa., in Sponge Sciences
60
Chapter 2 Literature Review
- Multidisciplinary Perspectives, Y. Watanabe, & Fusetani, N., Editor 1996, Springer- Verlag Tokyo: Lake Biwa Research Institute. p. 413-424. 167. Tsukamoto, S., Kato, H., Hirota, H., and Fusetani, N., Antifouling and metamorphosis- promoting compounds from the marine sponges Pseudoceratina purpurea and Agelas mauritina, in Sponge Sciences: Multidisciplinary Perspectives Y.F. Watanabe, N., Editor 1996, Springer-Verlag Tokyo: Lake Biwa Research Institute. p. 399-412. 168. Tsukamoto, S., Kato, H., Hirota, H., and Fusetani, N., Ceratinamine: an unprecedented antifouling cyanoformamide from the marine sponge Pseudoceratina purpurea. Journal of Organic Chemistry, 1996. 61: p. 2936-2937. 169. Turk, T., Frangež, R., and Sepčić, K., Mechanisms of toxicity of 3-alkylpyridinium polymers from marine sponge Reniera sarai. Marine Drugs, 2007. 5: p. 157-167. 170. Willemsen, P.R., and Ferrari, G.M., The use of anti-fouling compounds from sponges in anti-fouling paints. Surface Coatings International, 1993(10): p. 423-427. 171. Becker, K. and Wahl, M., Behaviour patterns as natural antifouling mechanisms of tropical marine crabs. Journal of Experimental Marine Biology and Ecology, 1996. 203: p. 245-258. 172. Greco, G., Torrassa, S., Svaldo Lanero, T., Young, R., Vassalli, M., Rolandi, R., Pelucchi, E., Faimali, M., and Davenport, J. How do crabs keep their eyes clean? The synergistic antifouling approach of Carcinus maenus. in 15th International Congress on Marine Corrosion and Fouling. 2010. Newcastle, England. 173. O'Neal, W. and Pawlik, J.R., A reappraisal of the chemical and physical defenses of Caribbean gorgonian corals against predatory fishes. Marine Ecology Progress Series, 2002. 240: p. 117-126. 174. Becker, K., A study of the natural antifouling mechanisms of the green mussel Perna viridis L. (Mytlilidae). Biofouling, 1995. 8: p. 233-242. 175. Kicklighter, C.E., Fisher, C.R., and Hay, M.E., Chemical defense of hydrothermal vent and hydrocarbon seep organisms: a preliminary assessment using shallow-water consumers. Marine Ecology Progress Series, 2004. 275: p. 11-19. 176. Hildebrand, M., Waggoner, L.E., Lim, G.E., Sharp, K.H., Ridley, C.P., and Haygood, M.G., Approaches to identify, clone, and express symbiont bioactive metabolite genes. Natural Product Reports, 2004. 21: p. 122-142. 177. Jelic-Mrcelic, G., Sliskovic, M., and Antolic, B., Biofouling communities on test panels coated with TBT and TBT-free copper based antifouling paints. Biofouling, 2006. 22(5): p. 293-302. 178. Almeida, E., Diamantino, T.C., and de Sousa, O., Marine paints: The particular case of
61
Chapter 2 Literature Review
antifouling paints. Progress in Organic Coatings, 2007. 59: p. 2-20. 179. Lobnig, R.E., Villalba, W., Goll, K., Vogelsang, J., Winkels, I., Schmidt, R., Zanger, P., and Soetemann, J., Development of a new experimental method to determiner critical pigment-volume concentrations using impedance spectroscopy. Progress in Organic Coatings, 2006. 55: p. 363-374. 180. Palmer, D.A. and Bénézeth, P. Solubility of copper oxides in water and steam. in 14th International Conference on the Properties of Water and Steam. 2004. Kyoto. 181. Yebra, D.M., Kiil, S., and Dam-Johansen, K., Mathematical modeling of tin-free chemically active antifouling paint behavior. American Institute of Chemical Engineers' Journal, 2006. 52(5): p. 1926-1940. 182. Yebra, D.M., Kiil, S., Weinell, C., and Dam-Johansen, K., Supplementary material to 'Analysis of chemically-active antifouling paints by mathematical modelling'. Progress in Organic Coatings, 2006: p. 20. 183. ACWS, Anti-fouling Coatings for Warships - First Selection of Products - D001, 2008, DSTL. p. 92. 184. Anderson, C.D. and Dalley, R., Use of organotins in antifouling paints, 1986, International Paint, PLC.: Felling, Tyne and Wear. p. 6. 185. Yebra, D.M., Kiil, S., Dam-Johansen, K., and Weinell, C.E., Reaction rate estimation of controlled-release antifouling paint binders: Rosin-based systems. Progress in Organic Coatings, 2005. 53: p. 256-275. 186. IMO, International Marine Organisation - International Convention on the Control of Harmful Anti-Fouling Substances on Ships, 2001. 187. Evans, S.M., Leksono, T., and McKinnell, P.D., Tributyltin pollution: a diminshing problem following legislation limiting the use of TBT-based anti-fouling paints. Marine Pollution Bulletin, 1995. 30(1): p. 14-21. 188. Omae, I., Organotin antifouling paints and their alternatives. Applied Organometallic Chemistry, 2003. 17: p. 81-105. 189. Anderson, C.D. Self-polishing antifoulings: a scientific perspective. in Ship Repair Conv. '93. 1993. 190. Howell, D.J. and Behrends, B., A methodology for evaluating biocide release rate, surface roughness and leach layer formation in a TBT-free, self polishing antifouling coating. Biofouling, 2006. 22(5): p. 303-315. 191. Rouhi, A.M., The squeeze in tributyltins. Chemical Engineering News, 1998. 27: p. 41- 42. 192. Mattheissen, P. and Gibbs, P.E., Critical appraisal for the evidence for TBT-mediated
62
Chapter 2 Literature Review
endocrine disruption in mollusks. Environmental Toxicology and Chemistry, 1998. 17: p. 37-43. 193. Oehlmann, J., Fiorini, P., Stroben, E., and Markert, B., Tributyltin (TBT) effects on Ocinebrina aciculata (Gastropoda, Muricidae) - imposex development, sterilization, sex-change and population decline. Science of the Total Environment, 1996. 188: p. 205-223. 194. Alzieu, C.L., Sanjuan, J., Deltriel, J.P., and Borel, M., Tin contamination in Arcachon Bay: Effects on oyster shell anomalies. Marine Pollution Bulletin, 1986. 17: p. 494-498. 195. Langston, W.J., Tributyl tin in the marine environment: a review of past and present risks. Pesticide Outlook, 1995. 6: p. 18-24. 196. Bryan, G.W., Gibbs, P.E., Hummerstone, L.G., and Burt, G.R., The decline of the gastropod Nucella lapillus around south-west England: evidence of the effect of tributyltin from antifouling paints. Journal of the Marine Biological Association of the United Kingdom, 1986. 66: p. 611-640. 197. Gibbs, P.E., Long-term tributyltin (TBT)-induced sterilization of neogastropods: persistence of effects in Ocenebra erinacea over 20 years in the vicinity of Falmouth (Cornwall, UK). 2008. 198. Strand, J. and Jacobsen, J.A., Accumulation and trophic transfer of organotins in a marine food web from the Danish coastal waters. Science of the Total Environment, 2005. 350: p. 72-85. 199. Terlizzi, A., Fraschetti, S., Gianguzza, P., Faimali, M., and Boero, F., Environmental impact of antifouling technologies: state of the art and perspectives. Aquatic Conservation: Marine and Freshwater Ecosystems, 2001. 11: p. 311-317. 200. Evans, S.M., Leksono, T., and McKinnell, P.D., Tributyltin Pollution: A diminishing problem following legislation limiting the use of TBT-based anti-fouling paints. Marine Pollution Bulletin, 1995. 30(1): p. 14-21. 201. Bryan, G.W. and Langston, W.J., Bioavailability, accumulation and effects of heavy metals in sediments with special reference to United Kingdom estuaries: a review. Environmental Pollution, 1992. 46: p. 89-131. 202. Champ, M.A., A review of organotin regulatory strategies, pending actions, related costs and benefits. The Science of the Total Environment, 2000. 258: p. 21-71. 203. Evans, S.M. and Nicholson, G.J., Assessing the impact of antifouling compounds in the marine environment. Lessons to be learned from the use and misuse of biological indicators of TBT contamination. Oceans 1999, 1999. 3: p. 1115-1118. 204. Evans, S.M., & Smith, R., The effects of regulating the use of TBT-based antifouling
63
Chapter 2 Literature Review
paints on TBT contamination. Oceans 1999, 1999. 3: p. 1119-1122. 205. Evans, S.M., Birchenough, A.C., and Brancato, M.S., The TBT ban: out of the frying pan and into the fire? Marine Pollution Bulletin, 2000. 40(3): p. 204-211. 206. Adrias, A.Q. and Matias, J.R., Validating eco-friendly claims. Polymers Paint Colour Journal, 2010. 200(4550): p. 18-20. 207. Matias, J.R., Perspectives on biocides. PPCJ, 2011: p. 27-29. 208. Poppleton, J. How the BPD and REACH will impact the anti-fouling coatings industry. in International Marine & Offshore Coatings Conference. 2007. 209. Omae, I., General aspects of tin-free antifouling paints. Chemical Research, 2003. 103: p. 3431-3448. 210. Thomas, K.V., Blake, S.J., and Walsh, F.C., Antifouling booster biocide contamination in UK sediments. . Marine Pollution Bulletin, 2000. 40(9): p. 739-745. 211. Thomas, K.V., McHugh, M., and Waldock, M.J., Antifouling paint booster biocides in UK waters: inputs, occurrence and environmental fate. Science of the Total Environment, 2002. 293(1-3): p. 117-127. 212. Voulvoulis, N., Scrimshaw, M.D., and Lester, J.N., Alternative antifouling biocides. Applied Organometallic Chemistry, 1999. 13: p. 135-143. 213. Chesworth, J.C. and Donkin, M.E., The interactive effects of the antifouling herbicides Irgarol 1051 and Diuron and the seagrass Zostera marina (L.). Aquatic Toxicology, 2004. 66: p. 293-305. 214. Backhaus, T., Arrhenius, A., Hilversson, A., Holm, K., Wendt, I., and Blanck, H. Employing classical mixture toxicity concepts for the optimization of biocide combinations for antifouling paints. in 15th International Congress on Marine Corrosion and Fouling. 2010. Newcastle, England. 215. Blanck, H., Backhaus, T., Förlin, L., Nydén, M., and Dahlbäck, B. A proposed strategy for risk assessment of antifouling combinations in paint products. in 15th International Congress on Marine Corrosion and Fouling. 2010. Newcastle, England. 216. Evans, L.V. and Clarkson, N., Antifouling strategies in the marine environment. Journal of Applied Bacteriology, 1993. Symposium Supplement 74: p. 119S-124S. 217. Turner, A., Fitzer, S., and Glegg, G.A., Impacts of boat paint chips on the distribution and availability of copper an English ria. Environmental Pollution, 2008. 151: p. 176- 181. 218. Baudrimont, M., Schäfer, J., Marie, V., Maury-Brachet, R., Bossy, C., Boudou, A., and Blanc, G., Geochemical survey and metal bioaccumulation of three bivalve species (Crassostrea gigas, Ceratoderma edule and Ruditapes philipinarum) in the Nord-
64
Chapter 2 Literature Review
Médoc salt marshes (Gironde Estuary, France). Science of the Total Environment, 2005. 337: p. 265-280. 219. Nguyen, M.N., Bressy, C., and Margaillan, A., Controlled radical polymerisation of a trialkylsilylated methacrylate by reversible addition-fragmentation chain transfer polymerization. Journal of Polymer Science: Part A. Polymer Chemistry, 2005. 43(22): p. 5680-5689. 220. Bressy, C., Margaillan, A., Faӱ, F., Linossier, I., and Vallée-Réhel, K., Tin-free self polishing antifouling coatings, in Advances in marine antifouling coatings and technologies, 2009, Woodhead Publishing in Materials (CRC Press). p. 445-491. 221. Faÿ, F., Linossier, L., Langlois, V., and Vallée-Réhel, K., Degradation of controlled release behaviour of caprolactone copolymers in biodegradable anti-fouling coatings. Biomacromolecules, 2006. 7(3): p. 445-491. 222. Faÿ, F., Linossier, L., Langlois, V., Vallée-Réhel, K., Krasko, M.Y., and Domb, A.J., Protecting biodegradable coatings releasing anti-microbial agents. Journal of Applied Polymer Science, 2007. 106(6): p. 3768-3777. 223. Camail, M., Humbert, M., Margaillan, A., Riondel, A., and Vernet, J.L., New acrylic titanium polymers: 1. Synthesis and characterisation of new titanium trialkoxide methacrylate monomers prepared via the esterification of methacrylic acid by titanium tetraalkoxides. Polymer, 1998. 39(25): p. 6525-6531. 224. Camail, M., Humbert, M., Margaillan, A., and Vernet, J.L., New acrylic titanium polymers: 2. Synthesis and characterisation of organotitanium polymers. Polymer, 1998. 39(25): p. 6533-6539. 225. Ashmore, J. Sea-Nine (TM) CR: A new, microencapsulated marine antifouling product. in 15th International Congress on Marine Corrosion and Fouling. 2010. Newcastle, England. 226. Nordstierna, L., Movahedi, A., and Nydén, M. New approach to microcapsule synthesis - replacement of a hazardous chemical. in 15th International Congress on Marine Corrosion and Fouling. 2010. Newcastle, England. 227. Nydén, M., Mok, A., and Nordstierna, L. Release from painted surfaces: free and encapsulated biocides. in 15th International Congress on Marine Corrosion and Fouling. 2010. Newcastle, England. 228. Candries, M., Altar, M., and Anderson, C.D. Considering the use of alternative antifoulings: the advantages of foul-release systems. in ENSUS 2000. 2000. Departments of Marine Technology and Marine Sciences, University of Newcastle upon-Tyne, UK.
65
Chapter 2 Literature Review
229. Anderson, C.D., Atlar, M., Callow, M., Candries, M., Milne, A., and Townsin, R.L., The development of foul-release coatings for seagoing vessels. Journal of Marine Design and Operations, 2003. B4: p. 11-23. 230. Adkins, J.D., Mera, A.E., Roe-Short, M.A., Pawlikowski, G.T., and Brady, R.F., Novel non- toxic coatings designed to resist marine fouling. Progress in Organic Coatings, 1996. 29: p. 1-5. 231. Baier, R.E. Influence of the initial surface condition of materials on bioadhesion. in 3rd International Congress on Marine Corrosion and Fouling. 1973. Northwestern University Press. 232. Baier, R.E. and DePalma, V.A., The relation of the internal surface of grafts to thrombosis, in Management of Occlusive Arterial Disease, W.A. Dale, Editor 1971, Yearbook Medical: Chicago. p. 147. 233. Brady, R.F., In search of non-stick coatings. Chemistry and Industry, 1997. 6: p. 219- 222. 234. Brady, R.F., Properties which influence marine fouling resistance in polymers containing silicon and fluorine. Progress in Organic Coatings, 1999. 35: p. 31-35. 235. Liu, Y. and Zhao, Q., Influence of surface energy of modified surfaces on bacterial adhesion. Biophysical Chemistry, 2005. 117: p. 39-45. 236. Molino, P.J., Campbell, E., and Wetherbee, R., Development of the initial diatom microfouling layer on antifouling and fouling-release surfaces in temperate and tropical Australia. Biofouling, 2009. 25(8): p. 685-694. 237. Molino, P.J., Childs, S., Eason Hubbard, M.R., Carey, J.M., Burgman, M.A., and Wetherbee, R., Development of the primary bacterial microfouling layer on antifouling and fouling release coatings in temperate and tropical environments in Eastern Australia. Biofouling, 2009. 25(2): p. 149-162. 238. Molino, P.J. and Wetherbee, R., The biology of biofouling diatoms and their role in the development of microbial slimes. Biofouling, 2008. 24(5): p. 365-379. 239. Swain, G.W. The mechanics and hydrodynamics of fouling release coatings. in 15th International Congress on Marine Corrosion and Fouling. 2010. Newcastle, England. 240. Neil, K.M., Hilliard, R., Russell, B., and Clark, P., Introduced marine species: management arrangements of consideration for the Torres Strait. Continental Shelf Research, 2008. 28: p. 2317-2323. 241. Scardino, A. Novel technologies to reduce biofouling on vessels when in port. in 15th International Congress on Marine Corrosion and Fouling. 2010. Newcastle, England. 242. Lindgren, J.F., Haeffner, M., Ericsson, C.T., and Jonsson, P.R., Oxygen-depleted
66
Chapter 2 Literature Review
surfaces: a new anti-fouling technology. Biofouling, 2009. 25(5): p. 455-461. 243. Grozea, C.M. and Walker, G.C., Approaches in designing non-toxic polymer surfaces to deter marine biofouling. Soft Matter, 2009. 5: p. 4088-4100. 244. Krishnan, S., Weinman, C.J., and Ober, C.K., Advances in polymers for anti-biofouling surfaces. Journal of Materials Chemistry, 2008. 18: p. 3405-3413. 245. Morra, M., On the molecular basis of fouling resistance. Journal of Biomaterials Science, Polymer Edition, 2000. 11(6): p. 547-569. 246. Morra, M. and Cassinelli, C., Force measurements on cell repellant and cell adhesive alginic acid coated surfaces. Colloids and Surfaces B: Biointerfaces, 2000. 18(3-4): p. 249-259. 247. Banerjee, I., Pangule, C.P., and Kane, R.S., Antifouling coatings: recent developments in the design of surfaces that prevent fouling by proteins, bacteria, and marine organisms. Advanced Materials, 2011. 23(6): p. 690-718. 248. Candries, M. The environmental benefits of surface treated coatings (STCs). in 15th International Congress on Marine Corrosion and Fouling. 2010. Newcastle, England. 249. Murosaki, T., Noguchi, T., Hashimoto, K., Kakugo, A., Kurokawa, T., Saito, J., Chen, Y.M., Furukara, H., and Gong, J.P., Antifouling properties of tough gels against barnacles in a long-term marine environment experiment. Biofouling, 2009. 25(7): p. 657-666. 250. Watermann, B.T., Daehne, B., Sievers, S., Dannenberg, R., Overbeke, J.C., Klijnstra, J.W., and Heemken, O., Bioassays and selected chemical analysis of biocide-free antifouling coatings. Chemosphere, 2005. 60: p. 1530-1541. 251. Chambers, L., Stokes, K.R., Walsh, F.C., and Wood, R.J.K., Modern approaches to marine antifouling coatings. Surface & Coatings Technology, 2006. 201: p. 3642-3652. 252. Peppiatt, C.J., Armstrong, E., Pisacane, A., and Burgess, J.G., Antibacterial activity of resin based coatings containing marine microbial extracts. Biofouling: The Journal of Bioadhesion and Biofilm Research, 2000. 16(2): p. 225 - 234. 253. Olsen, S.M., Pedersen, L.T., Laursen, M.H., Kiil, S., and Dam-Johansen, K., Enzyme- based anti-fouling coatings: a review. Biofouling 2007. 23(5): p. 369-383. 254. Pettitt, M.E., Henry, S.L., Callow, M.E., Callow, J.A., and Clare, A.S., Activity of commercial enzymes on settlement and adhesion of Cypris larvae of the barnacle Balanus amphitrite, spores of the green alga Ulva liza, and the diatom Navicula perminuta. Biofouling, 2004. 20(6): p. 299-311. 255. Leroy, C., Delbarre, C., Ghillebaert, F., Compere, C., and Combes, D., Effects of commercial enzymes on the adhesion of a marine biofilm-forming bacterium. Biofouling, 2008. 24(1): p. 11-22.
67
Chapter 2 Literature Review
256. Pettitt, M.E., Henry, S.L., Callow, M.E., Callow, J.A., and Clare, A.S., Activity of commercial enzymes on settlement and adhesion of cypris larvae of the barnacle Balanus amphitrite, spores of the green alga Ulva linza, and the diatom Navicula perminuta. Biofouling, 2004. 20(6): p. 299-311. 257. Kristensen, J.B., Olsen, S.M., Laursen, B.S., Kragh, K.M., Poulsen, C.H., Besenbacher, F., and Meyer, R.L., Enzymatic generation of hydrogen peroxide shows promising antifouling effect. Biofouling, 2009. 26(2): p. 141-153. 258. Olsen, S.M., Kristensen, J.B., Laursen, B.S., Pedersen, L.T., Dam-Johansen, K., and Kiil, S., Antifouling effect of hydrogen peroxide release from enzymatic marine coatings: Exposure testing under equatorial and Mediterranean conditions. Progress in Organic Coatings, 2010. 68(3): p. 248-257. 259. Olsen, S.M., Pedersen, L.T., Laursen, M.H., Kiil, S., and Dam-Johansen, K., Enzyme- based antifouling coatings: a review. Biofouling, 2007. 23(5): p. 369-383. 260. Dalsin, J.L. and Messersmith, P.B., Bioinspired antifouling polymers. Materials Today, 2005. 8(9): p. 38-46. 261. Dalsin, J.L., Hu, B.-H., Lee, B.P., and Messersmith, P.B., Mussel adhesive protein mimetic polymers for the preparation of nonfouling surfaces. Journal of American Chemical Society, 2003. 125: p. 4253-4258. 262. Statz, A., Finlay, J., Dalsin, J., Callow, M., Callow, J.A., and Messersmith, P.B., Algal antifouling and fouling-release properties of metal surfaces coated with a polymer inspired by marine mussels. Biofouling, 2006. 22(6): p. 391-399. 263. Bhushan, B., Biomimetics: lessons from nature - an overview. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, 2009. 367(1893): p. 1445-1486. 264. Bhushan, B., Jung, Y.C., and Koch, K., Micro-, nano- and hierarchical structures for superhydrophobicity, self-cleaning and low adhesion. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, 2009. 367(1894): p. 1631-1672. 265. Koch, K., Bhushan, B., and Barthlott, W., Multifunctional surface structures of plants: An inspiration for biomimetics. Progress in Materials Science, 2009. 54(2): p. 137-178. 266. Nosonovsky, M. and Bhushan, B., Green tribology: principles, research areas and challenges. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, 2010. 368(1929): p. 4677-4694. 267. Hellio, C., Berge, J.-P., Beaupoil, C., Le Gal, Y., and Bourguognon, N., Screening of marine algal extracts for anti-settlement activities against microalgae and
68
Chapter 2 Literature Review
macroalgae. Biofouling, 2002. 18(3): p. 205-215. 268. Wiltshire, K.H., Boersma, M., Moller, A., and Buhtz, H., Extraction of pigments and fatty acids fron the green alga Scenedesmus obliquus (Chlorophyceae). Aquatic Ecology, 2000. 34: p. 119-126. 269. Lang, Q. and Wai, C.M., Supercritical fluid extraction in herbal and natural product studies - a practical review. Talanta, 2001. 53: p. 771-782. 270. Herrero, M., Cifuentes, A., and Ibanez, E., Sub- and supercritical fluid extraction of functional ingredients from different natural sources: Plants, food by-products, algae and microalgae. Food Chemistry, 2006. 98(1): p. 136-148. 271. Chambers, L., Hellio, C., Stokes, K.R., Dennington, S.P., Goodes, L.R., Wood, R.J.K., and Walsh, F.C., Investigation of Chondrus Crispus as a potential source of new antifouling agents. International Biodeterioration and Biodegradation, 2011. 65(7): p. 939-946. 272. da Gama, B.A.P., Pereira, R.C., Carvalho, A.G.V., Ricardo, C., and Yocie, Y.-V., The effects of seaweed secondary metabolites on biofouling. Biofouling, 2002. 18(1): p. 13-20. 273. Harrison, P.G. and Chan, A.T., Inhibition of the growth of micro-algae by extracts of eelgrass (Zostera marina) leaves. Marine Biology, 1980. 61: p. 21-26. 274. Hellio, C., De La Broise, D., Dufossé, L., Le Gal, Y., and Bourguognon, N., Inhibition of marine bacteria by extracts of microalgae: potential use for environmentally friendly antifouling paints. Marine Environmental Research, 2001. 52: p. 231-247. 275. Lima-Filho, J.V.M., Carvalho, A.F.F.U., Frietas, S.M., and Melo, V.M.M., Antibacterial activity of extracts of six macroalgae from the Northeastern Brazilian coast. Brazilian Journal of Microbiology, 2002. 33: p. 311-313. 276. Tuney, I., Cadarci, B.H., Unal, D., and Sukatar, A., Antimicrobial activities of the extracts of marine algae from the coast of Urla (Izmir, Turkey). Turkish Journal of Biology, 2006. 30: p. 171-175. 277. Cho, J.-Y., Choi, J.-S., Kang, S.-E., Kim, J.-K., Shin, H.-W., and Hong, Y.-K., Isolation of antifouling active pyroglutamic acid, triethyl citrate and di-n-octylphthalate from the brown seaweed Ishige okamurae. Journal of Applied Phycology, 2005. 17: p. 431-435. 278. Iyapparaj, P., Palavesam, A., Immanuel, G., Ramasubburayan, R., and Hellio, C. A comparative study on the antifouling activity of Indian and Caribbean seagrasses extracts. in 15th International Congress on Marine Corrosion and Fouling. 2010. Newcastle, England. 279. Burgess, J.G., Boyd, K.G., Armstrong, E., Jiang, Z., Yan, L., Berggren, M., May, U., Pisacane, A., Granmo, Å., and Adams, D.R., The development of a marine natural product-based antifouling paint. Biofouling, 2003. 19: p. 197-205.
69
Chapter 2 Literature Review
280. Zhang, M., Cabane, E., and Claverie, J., Transparent antifouling coatings via nanoencapsulation of a biocide. Journal of Applied Polymer Science, 2007. 105: p. 3824-3833. 281. Armstrong, E., McKenzie, J.D., and Goldworthy, G.T., Aquaculture of sponges on scallops for natural products research and antifouling. Journal of Biotechnology, 1999. 70: p. 163-174. 282. Yebra, D.M., Kiil, S., Weinell, C.E., and Dam-Johansen, K., Dissolution rate measurements of sea water soluble pigments for antifouling paints: ZnO. Progress in Organic Coatings, 2006. 56(4): p. 327-337. 283. Kiil, S., Weinell, C.E., Yebra, D.M., Dam-Johansen, K., and Ka M. Ng, R.G., Marine biofouling protection: design of controlled release antifouling paints, in Computer Aided Chemical Engineering, 2007, Elsevier. p. 181-238. 284. Kiil, S., Dam-Johansen, K., Weinell, C.E., Pedersen, M.S. & Codolar, S.A., Estimation of polishing and leaching behaviour of antifouling paints using mathematical modelling: a literature review. Biofouling, 2003. 19 (supplement). p. 37-43. 285. Kiil, S., Dam-Johansen, K., Weinell, C.E., and Pedersen, M.S., Seawater-soluble pigments and their potential use in self-polishing antifouling paints: simulation-based screening tool. Progress in Organic Coatings, 2002. 45(4): p. 423-434. 286. Kiil, S., Weinell, C.E., Pedersen, M.S., and Dam-Johansen, K., Mathematical modelling of a self-polishing antifouling paint exposed to seawater: a parameter study. Chemical Engineering Research and Design, 2002. 80(1): p. 45-52. 287. Kiil, S., Weinell, C.E., Pedersen, M.S., and Dam-Johansen, K., Analysis of self-polishing antifouling paints using rotary experiments and mathematical modeling. Industrial & Engineering Chemistry Research, 2001. 40(18): p. 3906-3920. 288. Chan, R.K.S., Anselmo, K.J., Reynolds, C.E., and Worman, C.H., Diffusion of vinyl chloride from PVC packaging material into food simulating solvents. Polymer Engineering & Science, 2004. 18(7): p. 601-606. 289. Sanches Silva, A., Cruz Freire, J.M., Sendon, R., Franz, R., and Pasiero Losada, P., Migration and diffusion of diphenylbutadiene from packages into foods. Journal of Agricultural and Food Chemistry, 2009. 57: p. 10225-10230. 290. Canellas, E., Aznar, M., Nerin, C., and Mercea, P., Partition and diffusion of volatile compounds from acrylic adhesives used for food packaging multilayers manufacturing. Journal of Materials Chemistry, 2010. 20: p. 5100-5109. 291. Tehrany, E.A. and Desobry, S., Partition coefficients in food/packaging systems: a review. Food Additives and Contaminants, 2004. 21(12): p. 1186-1202.
70
Chapter 2 Literature Review
292. Bajpai, A.K., Shukla, S.K., Bhanu, S., and Kankane, S., Responsive polymers in controlled drug delivery. Progress in Polymer Science, 2008. 33(11): p. 1088-1118. 293. Reis, R.A., Nobrega, R., Oliveira, J.V., and Tavares, F.W., Self- and mutual diffusion coefficient equation for pure fluids, liquid mixtures and polymeric solutions. Chemical Engineering Science, 2005. 60: p. 4581-4592. 294. Verros, G.D. and Malamataris, N.A., Multi-component diffusion in polymer solutions. Polymer, 2005. 46: p. 12626-12636. 295. Coughlin, C.S., Mauritz, K.A., and Storey, R.F., A general free volume based theory for the diffusion of large molecules in amorphous polymers above Tg. 3. Theoretical conformational analysis of molecular shape. Macromolecules, 1990. 23(12): p. 3187- 3192. 296. Coughlin, C.S., Mauritz, K.A., and Storey, R.F., A general free volume based theory for the diffusion of large molecules in amorphous polymers above Tg. 4. Polymer- penetrant interactions. Macromolecules, 1991. 24(7): p. 1526-1534. 297. Coughlin, C.S., Mauritz, K.A., and Storey, R.F., A general free volume based theory for the diffusion of large molecules in amorphous polymers above Tg. 5. Application to dialkyl adipate plasticizers in poly(vinyl chloride). Macromolecules, 1991. 24(8): p. 2113-2116. 298. Mauritz, K.A., Storey, R.F., and George, S.E., A general free volume based theory for the diffusion of large molecules in amorphous polymers above Tg. 1. Application to di- n-alkyl phthalates in PVC. Macromolecules, 1990. 23(2): p. 441-450. 299. Mauritz, K.A. and Storey, R.F., A general free volume based theory for the diffusion of large molecules in amorphous polymers above Tg. 2. Molecular shape dependance. Macromolecules, 1990. 23(7): p. 2033-2038. 300. Karlsson, O.J., Stubbs, J.M., Karlsson, L.E., and Sundberg, D.C., Estimating diffusion coefficients for small molecules in polymers and polymer solutions. Polymer, 2001. 42: p. 4915-4923. 301. Gray-Weale, A.A., Henchman, R.H., Gilbert, R.G., Greenfield, M.L., and Theodorou, D.N., Transition-state theory model for the diffusion coefficients of small penetrants in glassy polymers. Macromolecules, 1997. 30(30): p. 7296-7306. 302. Greenfield, M.L. and Theodorou, D.N., Geometric analysis of diffusion pathways in glassy and melt atactic polypropylene. Macromolecules, 1999. 26: p. 5461-5472. 303. Tonge, M.P. and Gilbert, R.G., Testing models for penetrant diffusion in glassy polymers. Polymer, 2001. 42: p. 501-513. 304. Yebra, D.M., Kiil, S., Dam-Johansen, K., and Weinell, C., Reaction rate estimation of
71
Chapter 2 Literature Review
controlled-release antifouling paint binders: rosin-based systems. Progress in Organic Coatings, 2005. 53: p. 256-275. 305. Chambers, L.D., Stokes, K.R., Walsh, F.C., and Wood, R.J.K., Modern approaches to marine antifouling coatings. Surface and Coatings Technology, 2006. 201(6): p. 3642- 3652. 306. Hellio, C., De La Broise, D., Dufossé, L., LeGal, Y., and Bourgougnon, N., Inhibition of marine bacteria by extracts of macroalgae: potential use for environmentally friendly antifouling paints. Marine Environmental Research, 2001. 52: p. 231-247. 307. Armstrong, E., Boyd, K.G., and Burgess, J.G., Prevention of marine biofouling using natural compounds from marine organisms, in Biotechnology Annual Review, 2000, Elsevier. p. 221-241. 308. Armstrong, E., Boyd, K.G., Pisacane, A., Peppiatt, C.J., and Burgess, J.G., Marine microbial natural products in antifouling coatings. Biofouling, 2000. 16: p. 215. 309. Hansen, O.C. and Stougaard, P., Hexose oxidase from the red alga Chondrus crispus. Purification, molecular cloning, and expression in Pichia pastoris. The Journal of Biological Chemistry, 1997. 272(17): p. 11581-11587. 310. Kamatani, A. and Matsudaira, C., Extraction and determination methods of organic acids in sea water and marine sediment. The Journal of the Oceanographical Society of Japan, 1966. 22(3). 311. Savary, B.J., Hicks, K.B., and O'Connor, J.V., Hexose oxidase from Chondrus crispus: improved purification using perfusion chromatography. Enzyme and Microbial Technology, 2001. 29(1): p. 42-51. 312. Pinori, E., Berglin, M., Brive, L.M., Hulander, M., Dahlström, M., and Elwing, H., Multi- seasonal barnacle (Balanus improvisus) protection achieved by trace amounts of a macrocyclic lactone (ivermectin) included in rosin-based coatings. Biofouling, 2011. 27(9): p. 941-953. 313. Hansen, C.M., 50 Years with solubility parameters - past and future. Progress in Organic Coatings, 2004. 51: p. 77-84. 314. Hansen, C.M., Hansen solubility parameters: a user's handbook. 2nd ed. 2007, Boca Raton, Florida: CRC Press. 224. 315. Hansen, C.M., Polymer additives and solubility parameters. Progress in Organic Coatings, 2004. 51: p. 109-112. 316. Stefanis, E. and Panayiotou, C., Prediction of Hansen solubility parameters with a new group-contribution method. International Journal of Thermophysics, 2008. 29: p. 568- 585.
72
Chapter 2 Literature Review
317. Faӱ, F., Linossier, I., Langlois, V., Haras, D., and Vallée-Réhel, K., SEM and EDX analysis: Two powerful techniques for the study of antifouling paints. Progress in Organic Coatings, 2005. 54: p. 216-223. 318. Libes, S. M., Introduction to Marine Biogeochemistry. 2nd ed. 2009. Elsevier. 928pp.
73
Chapter 3 Materials and Methodology
3. Materials and Methodology
3.1 Fluorescence microscopy techniques: an overview
Laser scanning confocal fluorescence microscopy (LSCM) has gained appreciation over the previous decade in the field of life sciences as a means of imaging sub-microscopic features, even within complex cellular structures, with a high degree of specificity. As a major part of the work carried out for this thesis, LSCM has been investigated as a method of mapping the distribution of a broad range of additive types in polymeric AF binders.
3.1.1. Confocal and fluorescence microscopy theory
3.1.1.1. Confocal microscopy
Confocal microscopes offer the benefit of isolating individual features or structures within a medium. It is often employed to visualise complex cell structures. Traditional optical microscopy allows observation of structures within transparent media, but the resulting images are unsuitable for 3D reconstruction, modelling or quantitative analysis of the structures due to lack of a well-defined optical plane. For example, in Fig. 3.1, the crystals outside the focal plane are clearly visible. This results in severe feature distortion in the z- axis, and difficulty in assessing the features’ locations relative to the optically conjugate plane.
75
Chapter 3 Materials and Methodology
a b
Figure 3.1: Conventional optical microscope image of natural product crystals in a poly(methyl methacrylate) film: a) surface; b) ~20 µm depth. Note that certain features that are in plane in only one image but are visible in both, despite their spatial separation.
Confocal microscopes allow the isolation of a narrow plane by use of a variable width pinhole aperture in front of the detectors (see Fig. 3.2)1. Laser illumination light is expanded to fill the rear aperture, and then condensed at the objective lens to a focal point in the optical plane. Light from above or below the focal plane is excluded by the pinhole (Fig. 3.2). Change of the focal plane is achieved by adjusting the stage height relative to the objective lens.
Figure 3.2: Principle of confocal microscopy.
76
Chapter 3 Materials and Methodology
3.1.1.2. Fluorescence microscopy and one-photon excitation
Some compounds have an intrinsic fluorescence known as autofluorescence. Rather than flooding the sample with ‘white’ light, by selecting a single wavelength of light, the active parts of fluorescent molecules – fluorophores - can be stimulated to emit light of a lower energy, which is often in the range of the visible spectrum. This process is known as one photon excitation (1PE).
Within autofluorescent molecules, absorption of light may cause an electron to jump to a higher energetic state, as the electron absorbs the energy of the exciting photon (Fig. 3.3). A full range of transitions between possible excited states and vibrational levels is possible, some being more common than others; the absorption spectrum of the molecule is effectively composed of this total range of possibilities across different wavelengths, as all vibrational levels among the higher energy states are populated. A second stage of vibrational relaxation to a lower excitation state occurs, followed by a further loss of energy by photon emission, returning to the ground state (Fig. 3.3). The photon emitted is of a lower energy than the exciting photon, owing to the loss of energy from vibrational relaxation in the interim. It is this photon that provides the light that characterises the fluorescent molecule. Discrepancies in the energies emitted by ‘falling’ photons on their return to their ground state (caused by different vibrational states) result in the total emission spectrum of the material. This fluorescence microscopy can be combined with the pinhole aperture of the confocal microscope – the excitation allows isolation of the fluorescent compounds, whilst the pinhole allows exclusion of light from excited molecules in all but the optical plane, providing an optical sectioning effect (Fig. 3.4). As a result, the distribution of fluorescent compounds in a given xy plane can be assessed without interference from features above or below the point of interest.
77
Chapter 3 Materials and Methodology
Figure 3.3: Mechanism leading to emission of a lower energy (higher wavelength) photon after absorption of a photon.
Figure 3.4: 1PE illumination of specimen area. Note that all fluorophores in the sample are activated; the pinhole aperture must be deployed to eliminate light from outside the focal plane.
Scanning of the desired sample area by this confocal point is achieved by a raster motion of two high-speed oscillating mirrors in the scan head.
One photon fluorescence spectroscopy has some disadvantages. The main disadvantage is the attenuation of excitation light by fluorophores above the focal plane; even though the light emitted by them is excluded, the potential penetration depth into the sample is reduced. Secondly, some materials are sensitive and likely to be damaged in the process of
78
Chapter 3 Materials and Methodology fluorescing over time. If so, photodamage or bleaching takes place throughout the thickness of the sample, regardless of the optical plane. However, the development of two photon excitation (2PE) as a tool for sample analysis has done much to alleviate these problems.
3.1.1.3. Multi-photon excitation
Multi-photon fluorescence spectroscopy has gained popularity for use in sensitive materials as described above. Materials that respond to 1PE can be stimulated by absorbing two or more photons or equivalent energy within a short time frame in the order of 10-18 s. For example, a fluorophore that responds to UV light (380 nm) can be excited by two near- infrared photons (760 nm). Although the principal of two photon absorption was described during the 1930s in the thesis of Göppert-Mayer2, the requirement for photons to be absorbed within such a narrow temporal window restricted the experimental confirmation of 2PE until the 1960s, when lasers became sufficiently developed to provide enough photon density to produce a visible effect3. Excitation is achieved by short, powerful laser pulses between 100 femtoseconds (10-13 s) to 1 picosecond (10-12 s) apart; although the irradiation is still intense, the average irradiance to which the specimen is subjected is only slightly higher than that used for one-photon excitation3. The two excitation photons need not necessarily be of the same energies, but in practice, the emission of equivalent photons simply requires fewer laser sources and is much more commonly employed3.
Owing to the reliance on simultaneous absorption, the excitation effect provided by this method is non-linear; it is proportional to the square of the photon density. As the cone of excitation light also expands upwards and downwards outside the focal plane, so the probability of excitation decreases in a manner proportional to the fourth power of the cone diameter, resulting in a very thin plane of fluorophore stimulation (Fig. 3.5)3. The pinhole aperture can therefore be widened inconsequentially, as no excitation occurs outside the plane of interest. However, there may be some interest in employing the pinhole to remove ambient room light if noise is an issue3.
79
Chapter 3 Materials and Methodology
Figure 3.5: Mechanism by which photoactivation is limited to the confocal point during two- photon emission, decreasing laser light attenuation and absorption out of the focal plane.
Owing to the restricted field of photoactivation during multiphoton excitation, photobleaching is restricted to the focal plane only, making it the preferred technique for materials prone to this phenomenon. The fact that no absorption of excitation light by fluorophores outside the focal plane occurs allows for much greater sample depth penetration than 1PE, where excitation light is absorbed and attenuated above the focal plane and must be physically excluded from the image by the pinhole (Fig. 3.4).
3.1.2. Summary of fluorescence techniques and application to paint coatings
The three-dimensional analysis offered by LSCM is advantageous as it allows for ‘averaging’ of heterogeneous or inconsistent leached layer depth. LSCM is a suitable tool for the characterisation of paint coatings, being compatible with a range of objective lenses (OLs) offering different magnifications. The horizontal resolution achieved by LSCM is in the region of 100 – 200 nm, whilst the vertical resolution is generally between 1 – 3 µm, although this is dependent on the OL used (higher magnification OLs will yield better vertical resolution). LSCM has been applied to NPs, which usually feature aromatic structures that can absorb photons and reemit at a higher wavelength. The isolated NP biocides of interest within the ACWS project, usnic acid (a terrestrially derived dibenzofuran)4, 5 and juglone (a quinone from walnut extract)6-8, have been characterised in
80
Chapter 3 Materials and Methodology terms of their excitation/emission spectra. The determination of excitation and emission spectra and the employment of LSCM (Leica TCSSP2) to examine pure compounds and the distribution of usnic acid crystals in AF coatings is discussed in the next section. Furthermore, traditional fluorescence microscopy (Zeiss Axiophot) was been employed for further analysis of the distribution of dispersed usnic acid in the cross-sections of eroded coatings.
3.2. Selection of binders and biocides for LSCM screening
Three types of polymer coating were selected to provide a wide range of behaviours. The first of these, poly(methylmethacrylate) (pMMA), provides a hard, smooth surface that is non-hydrolysable (Fig. 3.6). Methylmethacrylate (MMA) is also a common co-monomer in AF formulations. Secondly, a CDP-type partially erodible binder, Metamare®, was chosen (Fig. 3.6). The rosin provides a soluble component, allowing the partial degradation of the film surface in the water column. The binder consists of an acrylic resin based on MMA and butylacrylate (BA) (60:40 wt.% blend) as is dissolved in xylene for application. Finally, a poly (triphenyl methacrylate/butylacrylate) p(TrMA/BA) (50:50 wt.% blend) copolymer developed within the current ACWS project was selected (Fig. 3.7). TrMA was developed as a co-monomer in the 1980s as an analogue to the TBT monomer (S. Dennington, Pers. Comm.), with an ester link attaching a large triphenyl side group to the polymerisible acrylic part of the monomer. The ester link may be saponified in seawater, allowing the hydrophobic triphenyl group to be dispersed and hydrolysis to occur, resulting in coating erosion. p(TrMA/BA) was hence expected to present a self-polishing behaviour analogous to the TBT-SPCs.
81
Chapter 3 Materials and Methodology
B C A
Figure 3.6: Methyl methacrylate (A), butylacrylate (B) and triphenyl methacrylate (C) monomers.
Several novel potential biocides are being studied within the ACWS project, including two isolated NP extracts (usnic acid and juglone). In addition, copper (I) oxide (99% purity, 95% <5 μm) is being used as a control as both biocide and model coating pigment. Importantly, copper oxide is the a subject of the few studies that have been carried out on the mechanics and kinetics of agent leaching as discussed in the second chapter9-15. All of these biocides were applied to microscope slides, under cover slips, to be examined by LSCM. For selection of biocides for inclusion in coatings, decisions would be taken based on bioassay results from other parallel studies, and from preliminary LSCM data. The data are discussed later within this chapter.
As discussed in section 2.3, crude extracts suffer from difficulties in their integration and complexity in their AF spectrum as a result of the disparate non-active compounds present alongside the biocide. As the interest of the present study is to develop a method of distribution analysis, the isolated compounds provide a more logical starting point and also are more representative of products that are likely to be employed in AF coatings.
82
Chapter 3 Materials and Methodology
3.2.1. Usnic Acid
5 6 5a 4a 4 7 3 9a 9b
2 8 9 1
Figure 3.7: Chemical structure of usnic acid.
Usnic acid (2,6-diacetyl-7,9-dihydroxy-8,9b-dimethyl-1,3(2H,9bH)-dibenzo-furandione; Fig. 3.7) is a dibenzofuran commonly found in terrestrial lichens (Usnea sp.). It is a fine, bright yellow powder which forms needle-like crystals which are visible to the naked eye when sufficiently separated. First isolated in 1844, it has long been considered a useful antibacterial, antipathogen and antiviral agent18. It has been considered for the ACWS project as an anti-microbial and anti-diatom agent, and has shown promise in bioassays4, 19. For reasons of confidentiality, usnic acid has been referred to as ‘furan derivative’ in outside communications throughout the ACWS project. As a consequence, it is referred to the acronym FD in certain figures throughout the remainder of this document. The characterisation of the excitation/emission spectrum of usnic acid is fully discussed in Section 4.
3.2.2. Juglone
Figure 3.8: Chemical structure of juglone.
83
Chapter 3 Materials and Methodology
Juglone (5-hydroxy-1,4-naphthalenedione), is a walnut extract that has also been studied within the ACWS project (Fig. 3.8). It is a fine reddish-brown powder under standard temperature and pressure conditions.
3.2.3. Bulk testing of binder properties
Bulk measurements of wettability and Persoz hardness were performed on the dry coatings (units for Persoz hardness are seconds passed for oscillations to fall from 12° to 4°). Blank pMMA yields a very hard coating (183 s) whereas the rosin-based CDP coatings are soft (70 s). Addition of FD to pMMA coatings resulted in an increase in the hydrophobicity and hardness, which was confirmed by comparison of the water droplet contact angle on blank pMMA coatings (contact angle = 68°) and with additive (contact angle = 89°). A concomitant increase in Persoz hardness was observed by pendulum testing; 10 wt.% FD content increased the relative hardness of pMMA film by 22% (to 223 from 183). No hardness increase was observed for the very soft CDP coating, although a similar increase in hydrophobicity occurred in CDP/FD. Addition of copper oxide rendered all the films 22% softer on average. Blank p(TrMA/BA) copolymer had a contact angle of 85.9°, which increased with addition of copper oxide (99.7°) and FD (92.5°). Water uptake for the blank coatings was assessed gravimetrically and weight increase was used to determine volume change based on the density of the polymers. Volume increase was 1% for pMMA and CDP coatings and 10% for the p(TrMA/BA) copolymer. The finding of such a high water uptake was surprising given the hydrophobicity of this binder. However, the conformation of the polymer chains is probably highly unusual due to the size of the triphenyl side group relative to the polymer chain. There is literature evidence demonstrating the formation of spiral chains for TrMA homopolymer; a similar situation may be the case here16, 17.
3.3. Preparation of test matrix with selected binders and biocides
Usnic acid was selected as the most promising NP with regard to bioassay results4, 19 and having also been characterised by fluorescence microscopy in the LSCM. Emission profiles had been obtained for both 1PE and 2PE. Copper oxide was also selected, for the following reasons: 1. Copper (I) oxide is most common biocide employed in AF formulations;
84
Chapter 3 Materials and Methodology
2. The presence of a pigment in experimental films is more representative of ‘real’ paint films; 3. A biocidal control will be required to isolate and compare any benefits of NPs in the coatings. 4. There is a small amount of published literature on copper (I) oxide leaching from AF coatings9-15. These two biocides were integrated into the three binders pMMA, Metamare® and p(TrMA/BA) in a combinatorial approach for immersion at the NOCS pontoon, and in the accelerated erosion rotor at TNO, Den Helder, in the Netherlands. The full test matrix is described below (Table 3.1).
Binder
Biocide Combination pMMA Metamare p(TrMA/BA)
Usnic acid ✓ ✓ ✓
Copper (I) oxide ✓ ✓ ✓
Usnic acid + copper (I) oxide ✓ ✓ ✓
Blank ✓ ✓ ✓
Table 3.1: Test matrix for biocide/binder combinations. ✓represents four coated panels for immersion at NOCS, and two coated panels for rotor erosion at TNO.
3.4. Formulation and application of coatings
pMMA was dissolved in chloroform to a solids content of 30% to provide a solution with similar viscosity to Metamare®. It should be noted that chloroform is quite a poor solvent for use in paints, due to its toxicity and high volatility (boiling point at 61oC). However, chloroform was one of the few solvents able to dissolve the polymer and biocide together. The p(TrMA/BA) was dissolved in toluene to a solids content of approximately 28%.
For addition of pigments and NPs, the copper was added to 30% by volume of the dry film. Usnic acid was added to 10% weight of the polymer solids in the dry film in order to
85
Chapter 3 Materials and Methodology replicate similar quantities as employed in crude extract trials (see section 2.3). It was also reasoned that if no anti-adherent effect could be observed for such a high NP loading, then further investigation could easily be ruled out. Copper oxide and usnic acid were incorporated by ball milling with 3/8” solvent-resistant stainless steel (type 316) balls after pre-cleaning the steel balls, first with acetone, and secondly with whichever solvent was used in the binder to be milled (i.e. chloroform or xylene). The balls were added to the flask containing the paint mixture, which was rotated on a roller bed for around 24 hours. There should be just enough steel balls to allow them to freely ‘turn over’ each other, resulting in pigment agglomerates being ground up in the mixture. Ball milling has the advantage of being performed in a closed container, so that solvent is not lost and solids content is maintained.
It was noted that when adding FD to the binder formulations, it could not be entirely dissolved - in the pMMA/chloroform solution, for example, about half of the FD was dissolved into the chloroform, resulting in a viscous clear yellow-brown varnish, whilst the other half settled visibly at the bottom of the bottle, forming a bright yellow particulate layer. A similar phenomenon was observed with Metamare-based solutions. The colour of the varnish changed slightly from golden-brown to golden-yellow. For this reason, all formulations were aggressively mixed by ball mill and/or stirring before application to panels, to ensure the dispersion of the NP solids. In order to determine the FD content dissolved in the polymer solution and the amount that precipitated, FD content in the polymer solution was determined by UV-visible spectroscopy. Subsamples of each paint mixture were vacuum filtered and an aliquot of the clear resin was extracted for comparison with a calibration range of usnic acid solutions in chloroform with known concentrations and polymer/solvent blanks. The resin aliquots were allowed to dry for four weeks. In the pMMA binder, of the 10 wt.% (on the dry film weight) added to the solution, 41% dissolved, and 59% precipitated out as needle-like crystals. In Metamare, only 16% of the added FD dissolved. The discrepancy between binders results from the greater solubility of FD in chloroform than in xylene. Although absolute values could not be determined for the p(TrMA/BA) binder owing to short supply, by visual inspection solubility of FD in this binder was very similar to that in the Metamare binder. This might be expected because of the similar solvents employed in these two coatings (toluene and xylene). These values are useful in the estimation of coating lifetime as the solid is expected to quickly dissolve, ensuring short term saturation of the binder until the majority of crystals are depleted20.
86
Chapter 3 Materials and Methodology
The three binder types would be employed for coating formulations to be used for dockside immersions and dynamic erosion tests. Each binder type was also applied to microscope slides by cube applicator to acetone-degreased microscope slides at a wet film thickness of 300 (Metamare®) and 400 µm (pMMA and p(TrMA/BA) formulations). The different wet film thicknesses were used to compensate the lower solids contents of the pMMA and p(TrMA/BA) formulations compared to the Metamare®. This method was found to produce consistent film thicknesses and allowed for rapid coating of multiple samples. Slides were left to dry inside a fume cupboard for at least a week.
Coatings were applied to 100 mm x 150 mm PVC panels (for immersion tests at NOCS) and mild steel panels (for rotor tests at TNO) to a final thickness of between 100-150 µm by using a doctor blade, and allowed to dry within a fume cupboard at room temperature. An excess of thickness was applied in the case of pigmented coatings (150-250 µm) to allow for greater than expected polishing rate as a result of copper (I) oxide depletion. The drying time of the Metamare-based formulations was found to last in excess of 6 weeks (Fig. 3.9), although the high volatility of chloroform resulted in the drying of pMMA-based coatings being achieved within hours. p(TrMA/BA)-based coatings dried within a few days to a week. The topcoat surface roughness for all specimens was investigated with a 3D optical profilometer (Alicona InfiniteFocus) microscope. Optimal values for cutoffs, objective lens magnifcation and Ra values were obtained from the Alicona user guide.
87
Chapter 3 Materials and Methodology
127.3 80
127.2
127.1 60
127.0
126.9 40 126.8
Weight / g / Weight
126.7 Hardness Persoz
20 126.6
126.5 Panel Weight Hardness 126.4 0 100 101 102 103 104 105 Time elapsed since coating / min Figure 3.9: Change in the weight of a Metamare-coated panel over time due to loss of solvent, and the concomitant increase in hardness.
3.5. Dynamic and static testing of coatings
3.5.1. Dynamic testing protocol
ACWS project partners TNO Maritime Material Performance Centre, Den Helder (Netherlands) houses a rotor setup designed for the accelerated dynamic testing of AF paint coatings (Fig. 3.10). Coatings for testing are applied to specialised mild steel panels and affixed to a drum, immersed in 1000 L of natural seawater, which is constantly replenished from the sea and maintained at 25 ± 2˚C. This is the main factor contributing to the artificial acceleration of polishing rates, along with the constant high speed of rotation. The seawater passes through a succession of increasingly fine filters, down to a pore size of 5 µm. This ensures that microorganisms and sand are removed from the water in the tank, although a fine layer of silt and some fine organic matter has been observed to coat the parts of the rotor that have been immersed for long periods. The drum is rotated by an electric motor that can be controlled to give various rotation speeds of in between 5 and 30 knots; an intermediate speed of 17 knots (8.7 ms-1) was selected for the present study. A generic anticorrosive primer was applied by TNO in-house to an average thickness of ~100 µm, before coating with the test formulations was carried out at UoS. The primer layer was investigated with a 3D optical profilometer, to determine the roughness characteristics.
88
Chapter 3 Materials and Methodology
Optimal values for cutoffs, objective lens magnifcation and Ra values were obtained from the Alicona user guide.
Figure 3.10: TNO Rotor system with panels attached.
Coated panels were attached to the rotor with clasps. The thickness measurements were carried out after panels have been removed from the water for an hour, allowing water to run off and for drying to occur. Ten locations on the panel surface are tested, and each measurement is the mean of three readings at each location. The elcometer (magnetic thickness gauge) is positioned by a mechanical arm and the rotor is rotated by the user to match grooves in the holding clasps with a light beam emitted from a source on the elcometer arm. This ensured that the same point on the panel is processed every time. The main source of error is by the operator’s judgement in aligning the light beam with the groove on the clasp, resulting in an error of about 0.5 mm left and right of the intended specimen site. However, testing either side of the intended area resulted in little or no change to the result. Panels attached for the first time were immersed for 24 hours before rotation began (at 17 knots). Thickness mesurements were performed before this
89
Chapter 3 Materials and Methodology conditioning period. Rotation is then commenced and measurements were performed at 7 days and every 14 days thereafter.
The coated panels for dynamic erosion are shown in Fig. 3.11.
H1
3.11
3.5.2. Static testing protocol
Coatings were also applied to PVC panels for immersion at the NOCS pontoon, using the same formulations. Four panels of each formulation type were attached to PVC backboards; two at the water’s surface (one on the front and one on the back), and two at 1 m depth (one front and one back; Fig. 3.12). These were immersed in May 2010. A visit was carried
90
Chapter 3 Materials and Methodology out each month to check on and photograph the panels. Adherent organisms attached to the backboard and nylon bolts holding the panels in position were generally removed; seaweed fronds were observed to have encroached onto the surface of test panels which may cause disturbance slime and algae on their surfaces. The PVC back-boards were removed from the water briefly and photographing was performed before reimmersion.
Figure 3.12: Examples of pMMA and Metamare formulation coated panels prior to the first immersion at NOCS pontoon (May 2010). The coatings on individual panels are binder +
Cu2O + FD (leftmost panels), binder + Cu2O (top centre panels), binder + FD (bottom centre panels) and blank binder only (rightmost panels). Each scheme was repeated on the PVC board’s front and back and at two depths.
In addition to the immersed panels, one panel of each type was prepared and retained in the laboratory for further testing and analysis. The post-immersion analysis with fluorescence microscopy was also carried out on these panels, which has aged an equal length of time but without any immersion.
3.6. LSCM stack image acquisition and binary processing
The overall objective of this work element was to visualise the binder distribution of additives in-situ. Having established excitation/emission data for the relevant compounds and binders, imaging of the distribution of biocides within the binder matrix is possible. A z- stack was performed at a step resolution of 1 – 1.5 µm, with extra images above and below the film to ensure full coverage. Each individual stack image should not be closer than the z- resolution of the OL; this ensures exclusivity of a given feature. This entails adjustment of the OL position relative to the specimen, to allow the focal plane to be adjusted within the sample. This was carried out for transparent specimens. As a certain portion of the biocidal additive is dispersed homogeneously within the binder film, the microscope was calibrated
91
Chapter 3 Materials and Methodology to eliminate the signal resulting from this part; the result is that the solid NP crystals show up within the film and that their distribution can be mapped.
Furthermore, quantitative analysis is possible by performing a binary transform on each individual image from the stack. A threshold minimum is applied to each image to remove noise and isolate the crystals. In practice, only a low threshold is required, especially if multiple accumulations are used to produce each image at the time of acquisition. After the binary transform has been applied, for any given pixel, the intensity is either 0 or 1 (Fig. 3.13). The ratio of signal to non-signal is easily calculated in the image analysis software (ImageJ) to determine the coverage of active substance in a given plane. The depth of each plane relative to the sample surface is calculated to allow the coverage to be plotted against depth. A top-down view and cross-section of a composite stack image is shown in Fig. 3.14.
Figure 3.13: Binary processing of a z-stack to quantify crystal distribution.
92
Chapter 3 Materials and Methodology
500 µm 80 µm
Figure 3.14: Top-down view of the 3D reconstruction of a pMMA/furan derivative film (left) on a microscope slide, with reconstructed cross-section (right).
A full characterisation of coatings was performed by LSCM before the dynamic immersion of panels was commenced. The coated panels to be immersed were imaged at three pre- defined points (Fig. 3.15).
right
middle
left
Figure 3.15: LSCM measurements are carried out at three locations on each panel.
By taking three widefield 1.5 mm2 image stacks (designated left, middle and right) it can be ensured that variation in film thickness or additive distribution is accommodated. The z- stacks were performed with a step resolution of between 1.0 and 1.5 µm.
93
Chapter 3 Materials and Methodology
3.7. Preparation of cross-sections for microscopy analysis
Cross-sections for fluorescence microscopy and optical microscopy analysis were prepared differently for each panel type:
1) Mild steel panels (dynamic immersion): After rotor immersion and panel retrieval, the coatings were separated from the steel substrate by cooling the panels with carbon dioxide at -78 °C in order to induce brittleness in the coating and primer, and mechanically fracturing the coating at either the primer-topcoat interface or the steel-primer interface. The latter is preferable as retaining the primer layer assists in correct sample orientation. The cryofracture method renders the paint coatings brittle and enables preservation of fragile leached layer structures. Paint chips from these coatings (of ~2 mm maximum width) were embedded in polymer- reinforced embedding wax (Paraplast XtraTM). Samples are allowed to cool and harden for a minimum of 4 hours at ambient conditions, and cross-sections were prepared by using a rotary microtome (Slee CUT 4062) equipped with ultra-sharp pathological Feather blades. The orientation of the paint chip was not permitted to deviate more than ~5˚ from vertical; in this way, the error in apparent thickness of entities measured in the cross-section resulting from this deviation was restricted to <1% (see Figure 3.16, Figure. 3.17).
2) PVC panels (static immersion): The panels were recovered from the pontoon after 10 months of immersion and a strip 2.5 cm wide was removed. The cross-sectioned strips were polished with five SiC paper grades (120 - 4000) to provide a smooth surface for optical microscopy and FM. Preferential polishing of either coating or PVC was not found to be an issue with any of the three binder types. Care was taken during sample handling and polishing to preserve the biofilm and weed on the panel surfaces.
The effect of incorrect orientation of the paint chip on the apparent thickness of structures (such as the leached layer, or overall thickness) was calculated (Figure 3.16, 3.17).
94
Chapter 3 Materials and Methodology
Figure 3.16: Representation of paint chip in various orientations, and the possible resulting distortion of measured features.
Figure 3.17: Percentage increase in apparent cross-section thickness resulting from deviation of the paint chip from vertical (i.e. 90˚ to cutting direction).
95
Chapter 3 Materials and Methodology
3.8. References
1. Claxton, N.S., Fellers, T.J., and Davidson, M.W., Laser Scanning Confocal Microscopy, 2005, Department of Optical Microscopy and Digital Imaging, National High Magnetic Field Laboratory, Florida State University. p. 37. 2. Göppert-Mayer, M., Uber Elementarakte mit zwei Quantensprungen. Annals of Physics, 1931. 5: p. 273-294. 3. Piston, D.W. and Davidson, M.W. Davidson, M. W. and The Florida State University. Specialized Microscopy Techniques. Multiphoton Fluorescence Microscopy. Molecular Expressions Microscopy Primer 2004 cited 2010 19/05; Available from: http://microscopy.fsu.edu/primer/techniques/fluorescence/multiphoton/multiphoto nintro.html. 4. Francolini, I., Norris, P., Piozzi, A., Donelli, G., and Stoodley, P., Usnic acid, a natural antimicrobial agent able to inhibit bacterial biofilm formation on polymer surfaces. Antimicrobial agents and chemotherapy, 2004. 48(11): p. 4360-4365. 5. Ingólfsdóttir, K., Usnic acid. Phytochemistry, 2002. 61(7): p. 729-736. 6. Cutler, H.G., Belson, N.A., Dawson, R., and Wright, D.A., Method for treating aquatic pests, 2000, Pharmacogenetics, Inc. 7. Dana, M.N. and Lerner, B.R., Black walnut toxicity, P.U.D.o. Horticulture, Editor 2001, Purdue University Cooperative Extension Service: West Lafayette, IN, USA. p. 1-2. 8. Randall, V.D. and Bragg, J.D., Effects of juglone (5'-hydroxy-l,4-naphthoquinone) on the algae Anabaenaflos aquae, Nostoc commune and Scenedesmus acuminatus. Proceedings: Arkansas Academy of Science, 1986: p. 52-55. 9. Kiil, S. and Yebra, D.M., Modelling the design and optimization of chemically active marine antifouling coatings, in Advances in marine antifouling coatings and technologies, D.M. Yebra and C. Hellio, Editors. 2009, Woodhead Publishing in Materials (CRC Press Ltd.). p. 393-421. 10. Kiil, S., Weinell, C.E., Yebra, D.M., Dam-Johansen, K., and Ka M. Ng, R.G., Marine biofouling protection: design of controlled release antifouling paints, in Computer Aided Chemical Engineering, 2007, Elsevier. p. 181-238. 11. Yebra, D.M., Kiil, S., and Dam-Johansen, K., Antifouling technology - past, present and future steps towards efficient and environmentally friendly antifouling coatings. Progress in Organic Coatings, 2004. 50: p. 75 -104. 12. Yebra, D.M., Kiil, S., and Dam-Johansen, K., Mathematical modeling of tin-free chemically active antifouling paint behavior. American Institute of Chemical Engineers' Journal, 2006. 52(5): p. 1926-1940. 13. Yebra, D.M., Kiil, S., Dam-Johansen, K., and Weinell, C., Reaction rate estimation of controlled-release antifouling paint binders: rosin-based systems. Progress in Organic Coatings, 2005. 53: p. 256-275. 14. Yebra, D.M., Kiil, S., Weinell, C., and Dam-Johansen, K., Supplementary material to 'Analysis of chemically-active antifouling paints by mathematical modelling'. Progress in Organic Coatings, 2006: p. 20. 15. Yebra, D.M., Kiil, S., Weinell, C.E., and Dam-Johansen, K., Effects of marine microbial biofilms on the biocide release rate from antifouling paints - a model-based analysis. Progress in Organic Coatings, 2006. 57(1): p. 56-66.
96
Chapter 3 Materials and Methodology
16. Merten, C., Barron, L.D., Hecht, L., and Johannessen, C., Determination of the helical screw sense and side-group chirality of a synthetic chiral polymer from Ramen optical activity. Angewandte Chemie International Edition, 2011. 50(42): p. 9973-9976. 17. Merten, C. and Hartwig, A., Structural Examination of Dissolved and Solid Helical Chiral Poly(trityl methacrylate) by VCD Spectroscopy. Macromolecules, 2010. 43(20): p. 8373-8378. 18. Ingólfsdóttir, K., Usnic Acid. Phytochemistry, 2002. 61: p. 729-736. 19. Salta, M., Wharton, J.A., Stoodley, P., Dennington, S.P., and Stokes, K.R., Antifouling performance against marine bacterial attachment for three natural products. Submitted, 2013. 20. Cussler, E.L., Diffusion: Mass Transfer in Fluid Systems. 3rd ed2007: Cambridge University Press. 655.
97
Chapter 4 Results and Discussion
4. Results and Discussion
4.1 Determination of excitation/emission spectra for binders and biocides
Excitation and emission spectra were determined first by a preliminary observation of the sample by simple light filter cubes. The three light filter cubes available were: a) UV; b) visible indigo; c) visible green, each covering a range of wavelengths corresponding to those parts of the spectrum. If any fluorescent behaviour was observed from samples under these filters, an indication could be taken as to which laser excitation wavelength to select. If no or very little fluorescent activity was observed, the sample would likely require selection of higher gain or widening of the pinhole aperture to increase signal.
4.1.1. pMMA pMMA was found to demonstrate no fluorescent character under light filters, 1PE or 2PE. The pMMA film displays no fluorescence even at extreme configurations of laser intensity and gain, so it is a convenient binder for examination of NP dispersion. The inert nature of pMMA is very conducive to studies of other copolymer components or additives, allowing resolution of complex structures within and avoiding ‘cross-talk’, which is excitation of multiple film components with a single excitation wavelength.
4.1.2. Usnic acid
Pure usnic acid was examined under three filter cubes (UV, visible blue, visible green) as a preliminary examination. A significant response was observed from all visible light cubes. Correspondingly, four 1PE laser lines covering the visible blue-green range were selected (458, 476, 488 and 514 nm; Ar/Kr laser). 458 nm is the lowest attainable wavelength with one photon laser lines. The crystals were easily discernable from the glass cover slip at all four excitation wavelengths. Lambda scans were performed at each of the four excitation wavelengths to ascertain emission spectra and optimum excitation wavelength for usnic acid (Fig. 4.1).
99
Chapter 4 Results and Discussion
Figure 4.1: Emission spectra for usnic acid for corresponding excitation wavelengths: 458 nm (violet circles); 476 nm (dark blue squares); 488 nm (light blue point-up triangle) and 514 nm (green point-down triangle). Arrows demonstrate excitation wavelengths.
The broadest and most intense emission spectrum was observed for the 514 nm excitation wavelength (Fig. 4.1). In practice, the crystals emit strongly enough for imaging purposes at any of the above wavelengths, as demonstrated in Fig. 4.2.
300 µm Figure 4.2: Usnic acid crystals on a microscope slide observed through LSCM. Excitation wavelength is 476 nm.
100
Chapter 4 Results and Discussion
The excitation wavelength was optimised by performing a series of incremental quick scans at different wavelengths, taking advantage of the tuneable multi-photon laser (Fig. 4.3). 920 nm excitation was selected as the optimal excitation wavelength.
Figure 4.3: Fluorescence images of a single usnic acid crystal, embedded in pMMA, across a range of excitation wavelengths. Detectors for all excitation wavelengths were set between 450 and 550 nm.
A series of λ scans (i.e. quantification of fluorescent emission across a variety of narrow wavelength windows, allowing precise definition of the emission spectrum) was performed at an excitation wavelength of 920 nm with increasingly small step resolutions to obtain
101
Chapter 4 Results and Discussion precisely the 2PE emission spectrum for usnic acid. It was found that the compound emits between 451 and 473 nm (Fig. 4.4).
2PE emission spectrum for usnic acid (excitation at 920 nm) 120
100
80
60
Intensity
40
20
0 430 440 450 460 470 480
Wavelength / nm Figure 4.4: 2PE emission spectrum for usnic acid at 920 nm excitation.
Figure 4.5: Emission spectra for juglone at 476 nm (dark blue circles), 488 (light blue squares) and 498 (green-blue triangles). Arrows demonstrate excitation wavelengths.
102
Chapter 4 Results and Discussion
At blue-green excitation wavelengths, juglone emits a broad peak at a slightly higher wavelength. The apex of the peak cannot be captured, as it is necessary to avoid the excitation wavelength which results in capture of reflected light. However, the intensity achieved is sufficiently high to allow signal detection without a wide detection range incurring background noise.
4.1.3. Copper (I) oxide
It was not expected that copper oxide would be detectable by fluorescence microscopy. However, when testing the pure copper oxide sample against excitation wavelengths used for other film components, it was observed that copper oxide can be stimulated with certain wavelengths. An emission spectrum has been determined for pure copper oxide for LSCM detection (Fig. 4.6). 2PE was required to observe any response from the copper (I) oxide, which did not respond to any single photon excitation wavelengths.
Emission spectrum for copper (I) oxide (excitation at 840 nm) 50
45
40
35
30
Intensity
25
20
15 350 400 450 500 550 600 650 700 750 Wavelength / nm Figure 4.6: Emission spectrum for copper (I) oxide at 840 nm excitation.
Selection of the peak between 550 and 650 nm allows for good resolution of copper oxide agglomerates and low background noise (Fig. 4.7).
103
Chapter 4 Results and Discussion
150 µm Figure 4.7: Copper (I) oxide agglomerates on a microscope slide.
When applied to films containing copper oxide, darkening of a small area of the film where illumination occurred was seen to develop, suggesting a photooxidation to copper (II) oxide by the laser light. This change occurred more rapidly in higher magnifications, perhaps resulting from the increased intensity of illumination over a smaller area. For this reason, LSCM was not employed as an alternative method for mapping copper distribution. Furthermore, the high density of the copper oxide (6.1 g cm-3) hinders penetration of the film, negating the confocal ability of the confocal fluorescence microscope.
4.1.4. Rosin
One of the principal binders of interest for the ACWS project is Metamare®, a controlled depletion polymer type binder based on pMMA and rosin. As rosin is the soluble component of the binder that imparts the pseudo-polishing behaviour, the ability to follow its degradation or depletion in the binder would be of great benefit. The Metamare® film responded to UV filter cubes; as the principal copolymer used in Metamare® is pMMA/BA, it was deduced that the widespread fluorescence was caused by the rosin. Despite the fluorescence observed from the filter cubes, the lowest possible 1PE wavelength achievable
104
Chapter 4 Results and Discussion
(458 nm) was not able to produce any activity. 2PE at 730 nm was found to produce significant fluorescence of the binder to distinguish it from the glass slide (Figs. 4.7 and 4.8).
Emission spectrum for rosin (from Metamare binder) at 730 nm excitation
40
30
20
Intensity
10
0 400 450 500 550 600 650 700 Wavelength / nm Figure 4.8: Emission spectrum for rosin at 730 nm excitation.
Some structure can be observed around the film edges when high resolution images are obtained. It is unclear if these are due to high local concentrations of rosin or if residual solvent is having an effect. In this particular sample, the Metamare® film was applied wet under a microscope slide. Microstructural anomalies resulting from rapid evaporation of the xylene around the edge of the film could also be responsible for the difference.
150 µm Figure 4.9: The edge of a Metamare film on a glass slide. Excitation wavelength is 730 nm.
105
Chapter 4 Results and Discussion
Pure rosin and abietic acid samples were obtained from chemical suppliers to confirm the analysis of rosin from the Metamare® binder. Brittle rosin crystals were analysed on a microscope slide (Fig. 4.10). These were found to respond to identical wavelengths as determined for the bulk Metamare product.
Figure 4.10: Rosin crystals viewed by fluorescence microscopy.
4.1.5. Poly(triphenyl methacrylate/butylacrylate)
Samples of pure pTrMA homopolymer were observed under light filters. All samples were found to respond to the UV light filter only. An emission spectrum for pTrMA was determined by using a 2PE wavelength corresponding to the UV excitation seen under light filter cubes. 730 nm was found to be a suitable excitation wavelength (Fig. 4.11).
Emission spectrum for pTrMA (excitation at 730 nm)
40
30
20
Intensity
10
0 400 450 500 550 600 650 700 Wavelength / nm Figure 4.11: Emission spectrum for pTrMA at 730 nm excitation.
106
Chapter 4 Results and Discussion
Good resolution of pTrMA is achieved by detecting either between 430-520 nm or 600-700 nm. The selection of which detection range to use should be determined by analysis of background noise at the time of testing, to determine whether other film components or additives also emit in one or both of these limits.
4.1.6. Summary of emission data
The emission spectrum for each substance and the corresponding recommended detection ranges are summarised below (Fig. 3.20). All substances tested except pMMA were found to have an activity, rendering them imageable by this technique.
TrMA (2PE)
Cu2O (2PE)
Juglone (1PE)
Rosin (2PE)
Usnic Acid (1PE)
Usnic Acid (2PE)
O (2PE) O
2
TrMA(2PE) (2PE) Rosin
UA (2PE) UA
UA (1PE) UA
JGL (1PE) JGL
Cu
400 500 600 700 800 900 Wavelength / nm
Figure 4.12: Summary of excitation/emission spectrum data determined for each component. Vertical lines correspond to excitation wavelengths; horizontal lines correspond to the recommended detection ranges.
4.2. Roughness/thickness assessment of primer and topcoat
On visual inspection of the primed panels for erosion analysis, two topography scales were evident. The largest of these was an orange peel-like dimpling effect with a periodicity between 2 or 3 mm, with a fine scale graininess within this dimension. The roughness and
107
Chapter 4 Results and Discussion waviness of the surface can be decoupled within the microscope’s software package (Fig. 4.13). Primer average and peak-to-trough roughnesses were determined using the Alicona InfiniteFocus microscope. Sample images and models of the primer surface are shown below (Figs. 4.14 and 4.15)
a
b c
d
Fig. 4.13: Decoupled primer roughness (a, b) and waviness (c, d) provided by optical profilometry. Scale bars 1 mm.
108
Chapter 4 Results and Discussion
300 µm
Figure 4.14: Optical profilometer microscope image of a small section of primer.
20x 541 µm
714 µm
Figure 4.15: Alicona 3D image of a 741 mm x 514 mm primer surface area overlaid with false colour contouring. Total peak-to-trough roughness = 17 µm.
The objective lens selected for imaging for the LSCM microscope was a 10x magnification lens, producing a 1.5 mm x 1.5 mm specimen zone. The vertical (z-plane) resolution generated by this lens is 2.5 µm. This compromise between specimen area size and vertical resolution was chosen as being appropriate to paint coating thickness and scale of application. A widefield approach such as this allows for maximum compensation of heterogeneity in additive distribution, position of leach fronts, and film thickness. The peak-
109
Chapter 4 Results and Discussion to-trough roughness for primed panels was measured using a range of magnifications, to capture a range of specimen area sizes comparable to that of the LSCM. The rectangular nature of the Alicona microscope view field renders direct comparison impossible. The range of magnifications tested and the corresponding specimen area sizes are detailed below in Table 4.1. Further images of the primer at each magnification can be found in Appendix A.
Instrument Magnification x(mm) y(mm) Area Alicona 50x 0.286 0.217 (mm0.06 2) Alicona 20x 0.713 0.54 0.39 Alicona 10x 1.43 1.08 1.54 LSCM 10x 1.5 1.5 2.25 Alicona 5x 2.84 2.16 6.13 Alicona 2.5x 5.67 4.29 24.3
Table 4.1: Specimen area sizes for various magnifications across the different instruments.
For each magnification, 15-20 measurements were carried out on three primed panels. The mean and standard deviation for each lens was calculated and plotted in Fig. 4.16. The error bars were calculated as 3σ. The grey lines show that the predicted range of primer roughness for the specimen size employed for the LSCM study is 23 – 44 µm.
110
Chapter 4 Results and Discussion
Primer roughness versus specimen area size 70
m 60
2.5x 50
40 5x
30 10x 20x 20 50x
Peak-to-trough roughness / / roughness Peak-to-trough 10
0 0.01 0.1 1 2.25 10 100 2 Specimen area size / mm Figure 4.16. Specimen area size against measured peak-to-trough roughness. Error bars are given as 3σ. Linear regression is represented by the black dashed line and 99% confidence levels represented by the red dashed line. The grey lines represent predicted range of the peak-to-trough roughness over the working area for LSCM.
The surface roughness was also assessed for the topcoats, over the 10x magnification area (1.43 x 1.08 mm2). The reflectivity of all blank topcoats made focussing difficult, resulting in poor imaging capacity and the inability to calculate roughness. The copper (I) oxide bearing coatings were visibly smooth, but demonstrated some surface defects on the millimetre scale and waviness on a centimetre scale. The defects (mostly bubbles and craters) resulted in a rougher surface, with a peak-to-trough roughness in the order of up to 20 µm. The p(TrMA/BA) films also had visible surface defects, combined with a waviness resulting from the coating method. Airless spray would be recommended as a preferable coating method if less volatile solvents were to be employed in future formulations.
4.1.1. Implications of primer and topcoat roughnesses for LSCM imaging
Primer and topcoat roughnesses have repercussions for the LSCM imaging of the both the surface of the topcoat, and the bottom part of the topcoat, which is intermeshed with the primer layer. For example, if a 1 µm step resolution is used for slices of the optical stack, then the signal will decrease over the last 20 to 50 slides of the stack as primer occupies an
111
Chapter 4 Results and Discussion increasingly large area of the xy plane. The effects of primer roughness on LSCM measurement are illustrated in Fig. 4.17.
Figure 4.17: Exaggerated representation of film structure demonstrating the effect of primer and film roughness heterogeneities on z-stack scanning.
Owing to the fine z step resolution and the good optical sectioning that can be achieved with the LSCM, the asperities of the coating surface and primer can impact the signal. Three regions in the coating can be delineated as follows:
Region A: Rapidly increasing signal, dependent on concentration of active substance. If active material is present in the very surface of the topcoat, then the tops of asperities become visible before the ‘bulk’ of the film, resulting in increasing signal. This signal increase is dependent upon the peak-to-trough roughness of the top coat over the
112
Chapter 4 Results and Discussion zone of interest, and the presence of active substance within this zone. Surface defects such as bubble pits from solvent evaporation and waviness from uneven coating application can cause considerable surface roughness on the widefield scale of LSCM.
Region B: Signal dependent on concentration of active substance. The optical plane is occupied entirely by the topcoat layer, so any location in the xy plane bearing active material will result in a signal.
Region C: Slowly decreasing signal, dependent on concentration of active substance. The penetration of active into the optical plane is limited by the extent of the primer’s peak-to-trough roughness. The primer layer may have significant roughness compared to the average thickness of the coat, so can partially occupy a large number of slides on the stack. In fact, primers are often designed and applied with a high roughness in mind, to enhance mechanical adhesion of the topcoat.
Region B will contain the majority of slides in thick coatings. For example, in a 150 µm thick coating with 10 µm peak-to-trough surface roughness and 30 µm peak-to-trough primer roughness, using 1 µm step resolution, the ratio of Regions A:B:C should be 10:110:30. Regions A and C must be taken into account in interpretation of quantifications of additive distribution. However in thinly coated films with high primer and/or surface roughness, Regions A and C can dominate the z-stack.
4.2. Pre-immersion usnic acid distribution
4.2.1. Assessment of vertical distribution of crystals with LSCM
Although the majority of this work focussed on the depletion of the homogenously dispersed portion of the biocide additive, the nature of crystal distribution is likely to have an effect on binder properties and efficacy. Using the z-stack acquisition and binary processing methods described above, the distribution of the furan derivative crystals was determined in the three binder types in the pre-determined locations on the panels (Fig. 4.18), before immersion was carried out. The peak-to-trough roughness of the primer layer is illustrated by horizontal lines corresponding to each stack.
113
Chapter 4 Results and Discussion
4.2.1.1. Usnic acid crystal distribution in pMMA
Figure 4.18: Usnic acid crystal distribution in three locations within pMMA film. The coloured lines delineate the start of the primer interface with the topcoat.
Figure 4.19: Top down composite view of crystals in the pMMA film (1.5 mm x 1.5 mm area).
The needle-like usnic acid crystals in the coating are large, up to more than 200 µm in length (Fig. 4.19). The distribution of usnic acid for the pMMA/FD coating was non homogeneous, with a decline in concentration from the coating surface until a peak is reached at the primer layer (Fig. 4.17). Disparities in the depth profile between individual stacks are due to variation in film thickness across the panel. The peak location in furan distribution is attributed to a settling mechanism, which is schematically illustrated in Fig. 4.20.
114
Chapter 4 Results and Discussion
Figure 4.20: Proposed settling mechanism for usnic acid crystals in the binder.
The proposed mechanism for usnic acid crystal dispersion in the pMMA binder involves four stages:
A: The film is applied to the substrate. Some of the FD is dissolved in the polymer solution, whereas some is already crystallised in the mixture (yellow crystals). B: The crystals begin to settle. Chloroform rapidly evaporates due to the high volatility of the solvent and the large surface area, thus film viscosity begins to increase and thickness begins to decrease. Consequently, crystals begin to precipitate out throughout the film (green crystals). C: As more solvent is lost, previously formed crystals continue to settle, resulting in an increasing concentration gradient towards the bottom of the drying film. New crystals continue to nucleate in the top of the film. D: viscosity becomes high enough to prevent further settling. As the last solvent evaporates, new crystals form within the whole of the film. As a result, most of the crystals have settled to the bottom of the film, but some crystals are present throughout the entire layer thickness. This inhomogeneous distribution is likely to have implications for AF efficacy in the field, due to the low concentration of additive at the surface of the coating.
The large size of the crystals also has implications for the efficacy of the binder. Replenishment of the dispersed usnic acid in the binder may be reduced if crystals have a low surface area.
115
Chapter 4 Results and Discussion
4.2.1.2. Usnic acid crystal distribution in Metamare
Figure 4.21: Usnic acid crystal distribution in three locations within Metamare film. The coloured lines delineate the start of the primer interface with the topcoat.
Figure 4.22: Top down composite view of crystals in the Metamare film (1.5 mm x 1.5 mm area).
The distribution of FD in Metamare was more homogeneous than for the other two binder types (c.f. Figs 4.18 and 4.23). Although the additive is at its highest concentration at the primer-topcoat interface, this concentration peak is only slightly larger than elsewhere in the films (Fig 4.21). Over the entire film (Region B), the concentration is relatively homogeneous. Furthermore, small fluctuations in additive concentration are apparent throughout the stacks, and these do not correspond particularly well between stacks. This
116
Chapter 4 Results and Discussion suggests a greater degree of spatial heterogeneity, both vertically and horizontally. Despite the low surface concentration and slight increase at the primer, it appears that the settling phenomenon observed in the other formulations does not occur to such an extent in this binder. The decline in the surface layer is partially attributed to the roughness of the sample surfaces – the ‘Region A’ section of the vertical structure (see Fig. 4.16). The signal decreases into the primer layer (Region C).
Metamare has an exceptionally long drying time compared to the other two binder systems, in excess of a month (Section 3.4). Although settling of non-dissolved FD crystals in the solution was also observed to occur in the preparation formulation beakers after long periods, it is possible that enough viscosity was lost by initial solvent evaporation to prevent settling of the crystals. Another possibility is that the FD’s three phenol groups engage in hydrogen bonding with the rosin’s carboxylic acid group, permitting the dissolved portion of the FD to remain in solution longer.
The crystals formed in the Metamare coating (Fig. 4.21) are comparatively smaller than those formed in the pMMA coating (Fig. 4.18).
4.2.1.3. Usnic acid crystal distribution in p(TrMA/BA)
Figure 4.23: Usnic acid crystal distribution in three locations within p(TrMA/BA) film. The coloured lines delineate the start of the primer interface with the topcoat.
117
Chapter 4 Results and Discussion
Figure 4.24: Top down composite view of crystals in the p(TrMA/BA) film (1.5 mm x 1.5 mm area).
Similarly to the pMMA binder, a distinctive settling behaviour was observed in this binder. At the top of the film, no fluorescent activity from the additive was observed (Fig. 4.23). In fact, imaging of dust particles on the film surface was necessary to locate the film surface. Owing to the lack of active NP at the top of the binder, this is likely to present a suboptimal distribution in terms of AF efficacy. The efficacy would need to rely solely on the diffusion of active substance through the binder and the release of crystals through mechanical leaching and chemical polishing of the copolymer.
Crystal size and coverage in the xy-plane (Fig. 4.24) is similar to that observed for the Metamare binder. Coverage in the xy-plane is more complete than for pMMA owing to the smaller size of the crystals, although the vertical distribution is inhomogeneous (Fig 4.23). A settling mechanism similar to that observed for the NP additive in pMMA is again likely to be responsible for the observed distribution.
4.2.2. Assessment of vertical distribution of usnic acid with fluorescence microscopy
Fluorescence microscopy (FM) was employed for aged coatings retrieved from immersion because of the detritus that settled on the film surface during rotor testing, and the heavy fouling that formed in the sea. In order to determine the quantities of usnic acid lost during immersion, the same procedure was carried out on cross-sectioned aged panels that had been preserved at ambient conditions for the same length of time. This allows for any change within the coating that may occur over time to be accounted for within the
118
Chapter 4 Results and Discussion immersed samples. FM was employed to analyse the vertical distribution of additive in unimmersed CDP, pMMA and p(TrMA/BA) binders (Fig. 4.25). In addition, blank laboratory- aged samples of each binder were assessed in the same manner to determine background fluorescence intensity (FI) levels for each medium (Fig. 4.25).
Fig. 4.25: Average fluorescence intensity measurements for 10 month aged, unimmersed pMMA (blue, short dashed), CDP (black, solid) and p(TrMA/BA) (red, long dashed) coatings containing usnic acid (left) and for blank binders (right). Boundaries represent 90% confidence intervals. FITC filter block was employed for all measurements.
The blank pMMA and CDP binder samples demonstrate very low levels of fluorescence (100 units) under a FITC (fluorescein isothiocyanate) filter block. However, p(TrMA/BA) does exhibit some autofluorescence up to around 4-500 units, which probably results from the aromatic structures of the triphenyl group. The data for binders with NP additive demonstrate an elevated FI compared to blanks even for short exposure times (80 ms). As discussed earlier, high variance is observed for these samples owing to the presence of crystalline usnic acid within the cross-sections. FI for p(TrMA/BA)/FD and CDP/FD are of the same order (Fig. 4.25). Higher FI for pMMA is observed in comparison to the other binder types; this may correspond to a higher solubility of usnic acid in chloroform, used as a solvent for the pMMA formulations. The higher solubility in this solvent corresponds to a higher ratio of dispersed usnic acid to crystalline usnic acid for this formulation, as a result of better dispersion in solution.
An observation for all samples aged on PVC substrates is a decrease in FI towards the substrate. This phenomenon was investigated by measuring line profiles through the coatings, which demonstrates a gradient of FI in the substrate in all PVC panels, resulting
119
Chapter 4 Results and Discussion from diffusion of the active substance into the substrate (Fig. 4.26). This was not observed for any cross-sections from the rotor panel coatings, where an epoxy tie-coat is employed between topcoat and substrate, or for recently wax-embedded samples, implying a time- dependant and substrate-dependant effect. The penetration of usnic acid reaches as far as 50 µm into the substrate. This indicates a diffusion of biocide into the substrate for PVC panels that could have important implications for long-term immersion schemes using this or similar material. The epoxy primer employed for steel panels was examined after accelerated ageing. The primer did not exhibit any FI concentration gradient and remained inert after 10 months of immersion. Use of an epoxy tiecoat is recommended for future immersion work.
Fig. 4.26: Fluorescence image of p(TrMA/BA)/FD coating on PVC substrate (FITC filter). FI profile associated with the solid line (perpendicular to coating) is demonstrated below.
120
Chapter 4 Results and Discussion
4.2.3. Summary of pre-immersion usnic acid distribution
An ‘approach’ period of increasing FI is observed as the z-stack scanning approaches the sample surface, owing to surface defects. Three stacks were performed for each sample type in three locations, and good agreement was shown between data sets. Vertical distribution of crystals within the matrix was generally shifted towards the interface with the primer layer in all the pMMA and p(TrMA/BA) systems. This is attributed to settling during the drying phase. Vertical distribution of crystals in the Metamare binder was more homogeneous, with a consistent crystal concentration in the film ‘bulk’, decreasing in the surface asperities and intermeshing primer region. The dispersed FD was imaged in all binder types, before immersion of the samples. The blank binders are largely inert, other than p(TrMA/BA), which possesses some inherent fluorescence. Dissolved FD was homogeneously distributed in all binder types. A diffusion of FD into the substrate was observed on coated PVC panels. This is a significant finding for long-term immersion trials. When using an epoxy primer, no diffusion occurred.
4.3. Results of panel immersion tests
4.3.1. Static immersion (NOCS pontoon)
The coated panels were immersed at the NOCS pontoon in May 2010. Monthly visits to assess the fouling state of the panels were carried out (Fig. 4.27). ‘Front’ refers to the sunlit side of the panel, whilst ‘Back’ refers to the shaded rear side of the panel.
121
Chapter 4 Results and Discussion
4.3.1.1. pMMA coatings – front
pMMA coatings – front at surface pMMA coatings – front at 1m depth
May 2010
June 2010
July 2010
122
Chapter 4 Results and Discussion
August 2010
September 2010
March 2011
Figure 4.27a: Fouling progression of pMMA panels (front) over 10 months from May to March
The front face of the backboard is significantly fouled by the first measurement after 1 month. There is a proliferation of green algal growth (possibly Ulva sp.) on the PVC backboard. The panels are considerably less fouled, although there is limited brown/green
123
Chapter 4 Results and Discussion algal fouling on the blank and FD-only panels (~10% coverage). By August, the blank pMMA panel is severely fouled (>50%) and the pMMA/FD panel is >90% fouled with diatomaceous slime and brown/green algae. The pMMA/Cu2O and pMMA/Cu2O/FD panels are considerably fouled by slime only, possibly Amphora sp. By September, all panels are significantly fouled, including the copper-containing panels.
4.3.1.2. pMMA coatings – back
pMMA coatings – back at surface pMMA coatings – back at 1m depth May 2010
June 2010
124
Chapter 4 Results and Discussion
July 2010
August 2010
September 2010
125
Chapter 4 Results and Discussion
March 2011
Figure 4.27b: Fouling progression of pMMA panels (back) over 10 months from May to March
The back face of the backboard is less fouled after the first month than the front side, and early hard foulers including spirobids and barnacles are more prevalent. The quantity and size of these increases in the next month (July), accompanied by the appearance of ascidians and hydroids and a few tubewoms. In addition, brown slime has covered most of the remaining backboard area, and has begun to foul the surface panels (~20%) although the panels immersed at 1 m depth remain mostly free of fouling (<5% slime coverage). By August and September, the blank pMMA panel and pMMA/FD panel at each depth are completely fouled, and the backboard is overgrown with sea squirts, particularly at 1 m depth, necessitating cleaning every month to ensure that the growth does not encroach directly on the panel surfaces. The copper-pigmented panels resisted the majority of fouling up to the end of the experiment, with 30-40% brown slime coverage and no hard fouler or ascidian fouling.
126
Chapter 4 Results and Discussion
4.3.1.3. Metamare coatings – front
Metamare May 2010 coatings – front at surface Metamare coatings – front at 1m depth
May 2010
June 2010
July 2010
127
Chapter 4 Results and Discussion
August 2010
September 2010
March 2011
Figure 4.27c: Fouling progression of Metamare panels (front) over 10 months from May to March
In the first month of immersion, green algae dominate the backboard fouling community. The blank and CDP/FD panels are less fouled than the backboard at both depths (5 and 10% coverage respectively), whilst the pigmented panels also resist fouling at this early stage (5% coverage). However, by July (one month later) green algal material (probably Ulva sp.) has entirely covered the backboard and black MM panel, and most of the CDP/FD panel (70%).
128
Chapter 4 Results and Discussion
The pigmented panels still resist the adhesion of the algal spores and exhibit about the same levels of coverage as the previous month (5%). In July, the pigmented panels continue to resist fouling well while other species begin to appear on the other panels and backboard, including brown weed and tubeworms. The pigmented panels have a light coverage (10%) of thin brown slime. However, by September, the pigmented panels also become more heavily covered by the brown slime (>50%), and they are covered with slime by March 2011. Despite the eventual slime fouling on the copper (I) oxide-containing panels, they continued to resist all other fouler types.
4.3.1.4. Metamare coatings – back
Metamare coatings – back at surface May 2010 Metamare coatings – back at 1m depth
May 2010
June 2010
129
Chapter 4 Results and Discussion
July 2010
August 2010
September 2010
130
Chapter 4 Results and Discussion
March 2011
Figure 4.27d: Fouling progression of Metamare panels (back) over 10 months from May to March
One month into the trials, the small fouling community comprises barnacles, spirobids and brown algal slime, as observed on the shaded side of the other board. At this point, the panels containing copper (I) oxide are free of fouling at both depths, whilst the blank and CDP/FD panels have a thin slime layer at both depths. There is evidence of fish grazing on the surfaces of these panels. By July 2010, both of these panels (CDP/FD and MM Blank) are entirely covered by fouling, as is the backboard. The fouling community consists of ascidians, hydroids, barnacles, spirobids and thick brown diatomaceous slime. The two pigmented panels, CDP/Cu2O and CDP/Cu2O/FD, are resisting fouling well with a very limited amount of brown slime attaching to the surface (1-2% or less). They continue to resist fouling until the end of the trial in March 2011, with <5% fouling on all panels except the
CDP/Cu2O panel at 1 m depth, which is beginning to accumulate more slime (20%). Notably, the surface CDP/Cu2O/FD panel remains totally free of fouling throughout the trial. This is discussed more in the next section. The findings of the static immersion in terms of fouling coverage are summarised in Table 4.2.
131
Chapter 4 Results and Discussion
pMMA pMMA/FD pMMA/Cu2O pMMA/Cu2O/FD F B F B F B F B 1 month 2 months 3 months 4 months 10 months
MM MM/FD MM/Cu2O MM/Cu2O/FD F B F B F B F B 1 month 2 months 3 months 4 months 10 months
No fouling Light slime fouling Moderate slime fouling Heavy slime fouling only Heavy slime coverage + limited hard fouling, sea squirts or weed Heavy hard fouling, sea squirts and/or weed Rapid hard fouling colonisation but low slime
Table 4.2: Fouling progression summary.
4.3.1.5. Static immersion discussion
Over the first two months of immersion, both micro and macrofouling were observed on the panel front and rears (Fig. 4.27). Slime and green alga populated the front of panels, with the backs being extensively colonised by slime but lacking macroalgae. Small, hard spirobids were observed on the front and particularly on the rear of both panels, and these also colonised the NP-containing surfaces rapidly with no apparent deterrence.
The panels suffered an aggressive fouling period over July owing to the good weather conditions. Extensive coverage of green algae is noted on the front of both panel schemes at the July assessment and thereafter, which is followed by large brown alga specimens. The shaded sides of the panels are dominated by barnacles, bryozoans, hydroids, tunicates and ascidians. The ascidians in particular seem to favour the niches around the panels and
132
Chapter 4 Results and Discussion proliferate rapidly following attachment.
They outcompeted tubeworms in the zones around the panels, despite tubeworms being much more prolific in the blank space between panel schemes (Fig. 4.28). Once hard fouling is well established, a continual presence of fouling was observed even through the winter months until next March when the panels were removed. It was noted that pigmented pMMA panels inhibited fouling during the first 1 month of static immersion on the panel fronts and 3 months on the rear. Front panels were freely fouled by green algae after the first month, although hard fouling was less frequently observed on these panels than on other formulations and controls.
Rear panels still demonstrated a reasonable performance after the first 3 months, though some slime attachment was noted, as well as three large tubeworms on one of the rear panels after 3 months. Interestingly, the fouling state of the tubeworm-fouled panel did not further deteriorate throughout the remainder of the immersion trial. The non-pigmented pMMA/FD Figure 4.28: The back of a board, coatings were comparable to the blanks on both demonstrating proliferation of sea the front and rear sides. Despite the favourable squirts around panels. Note that bioassay results obtained within the ACWS project, tubeworms dominate the area the FD biocide on its own was not observed to between panel schemes. have any efficacy against settling organisms on front or back.
MM coatings demonstrated a good level of performance. The non-pigmented CDP/FD coatings repelled fouling at the early stage, but by at the 2 month stage they were covered ~95% (rear) and 70% (front) by macroalgae and hydroids, and outperformed blank controls
133
Chapter 4 Results and Discussion in this early stage, implying some degree of efficacy of FD on its own. Post-immersion analysis demonstrated a greater content of dispersed FD remaining in the rear panels than the front ones. This was an unexpected result give the heavier fouling observed on the front-attached panels. CDP/Cu2O offered a good performance on front panels for the first 3 months, and on rear panels for the duration of the 10 month immersion experiment, with only brown slime adhering on the rear panels. Combinatorial CDP/Cu2O/FD coatings displayed a comparable performance to CDP/Cu2O on the lit side, but excellent resistance to fouling on the rear-side panels, compared to pigment only and NP only formulations. One of the two rear panels remained completely unfouled throughout the entire year (Fig 4.27), and its performance was favourable even compared to commercial paint controls. Copper- containing surfaces offered good fouling resistance on the backs of the panels. However, they were heavily fouled by slime and the green alga species on the panel fronts, despite resisting attachment of larger organisms. It was noted that crabs often grazed the copper- containing panel surfaces (e.g. Metamare back/surface September 2010). The combinatorial copper oxide/NP-bearing panels were the best performers, particularly in the Metamare coatings. These were only fouled by small amounts of slime after 5 months, whereas the pigmented coatings without NP were heavily slimed. This is particularly interesting in light of erosion and post-immersion FM analysis data, as later discussed.
Coatings containing FD without pigment fouled more rapidly than blanks in the pMMA coatings, but less quickly than blanks in the Metamare coatings. Little influence of sample depth was noted, but significant differences were observed between the boards’ fronts and backs. This was primarily demonstrated by the proliferation of green algae on the front only, and by the colonisation of the panel back by tubeworms and ascidians.
The efficacy of the FD biocide in the immersion trials was limited to its performance as a co- biocide with copper oxide. Coatings with usnic acid only were generally incomparable from the blanks. The exception was in the case of combinatorial CDP/Cu2O/FD coatings, which outperformed Cu2O only ones and demonstrated excellent AF potential. These results are fully discussed later in tandem with the analysis of usnic acid and copper oxide distribution in the eroded coatings by fluorescence and optical microscopy, respectively (see Section 5.5). After 10 months of exposure to natural conditions, copper oxide depleted leached layers on CDP/Cu2O/FD panels were measured to be 17 μm on average for front-mounted panels (σ = 4.31 µm; Fig 4.29) and universally considerably thicker on the rear side panels (31 μm, σ = 7.86 µm; Fig 4.29). Leached layers were much less developed for panels without
134
Chapter 4 Results and Discussion
FD; mean copper oxide leached layer thickness was 18 μm for rear panels (σ = 5.56 μm) and leached layer thickness was negligible for front-mounted CDP/Cu2O panels, indicating easier copper oxide depletion in the combinatorial coatings. An interesting observation was the gradual discolouration of CDP panels from red to white, corresponding with the formation of a leached layer and loss of copper oxide. The faster discolouration of the combinatorial panels compared the pigment-only ones corroborates their thicker leached layer development.
Figure 4.29: Cross-sections of CDP/Cu2O/FD panels from rear (left) and front (right) of the exposed board. Scale bars = 50 μm.
4.3.1.6. Static immersion summary
Usnic acid only demonstrated efficacy when present as a co-biocide in the case of
CDP/Cu2O/FD, which had superb performance on the shaded side. This was attributed to In other combinations and on most front panels, the effect of FD as a co-biocide was small or negligible. A poor AF effect for usnic acid was concluded, particularly given the high loading (10%) that was employed for this study. Copper (I) oxide demonstrated potent broad AF efficacy in both pMMA and Metamare binders. As expected, the performance of pigmented Metamare was superior to that of pigmented pMMA. The Metamare binder bearing both copper (I) oxide and FD performed excellently on the shaded side of the panels. This finding is further discussed in the context of fluorescence microscopy results in the following sections (see Section 5.5).
4.3.2. Dynamic immersion (TNO, Den Helder)
Dynamic immersion tests using the TNO rotor system were carried out simultaneously with the static immersions. After the 24-hour static conditioning period, all four coatings of Metamare (blank) and Metamare/FD had completely detached from the primer and had to
135
Chapter 4 Results and Discussion be removed. No further delaminations were observed. Measurements of film thickness were performed on a fortnightly basis at TNO (as described in section 3.5.1.). The cumulative change in thickness compared to the pre-immersed films was compared for each point.
4.3.2.1. pMMA coatings
Figure 4.30: Thickness change of pMMA and reference epoxy coatings with time: a) pMMA
Blank; b) pMMA/FD; c) pMMA/Cu2O; d) pMMA/Cu2O/FD; e) reference non-polishing epoxy
136
Chapter 4 Results and Discussion coating.
Fluctuations of roughly ±4 µm in thickness can be observed in all coatings, including the reference non-polishing epoxy binder, which can be considered to result from the natural changes in temperature and ambient humidity as the panel thicknesses are gauged out the water (Fig. 4.30). However, a statistically significant polishing behaviour was determined for the blank binder and the pigmented binders, although for both binders bearing NP the furan derivative appeared to have a retarding effect on erosion rate. This could be explained by an increase in the hydrophobicity of the coatings, which was confirmed by comparison of the water droplet contact angle on pMMA coatings without additive (contact angle = 67.5°) and with additive (contact angle = 89°). A concomitant increase in Persoz hardness was observed by pendulum testing (10% usnic acid content increased the relative hardness of pMMA film by 22%, from 183 to 223 (units are seconds passed for oscillations to fall from 12° to 4°). Overall decrease in coating thickness was estimated by performing a linear regression over the test period, normalized against the changes exhibited by the control epoxy reference coating, and then calculating the gradient of the resulting line. The calculated polishing rates are shown below (Table 4.3) within example cross-sections from post-immersion testing (Fig. 4.31)
Thickness decrease / µm month-1 Coating Panel 1 Panel 2 Average pMMA Blank 1.59 1.62 1.61
pMMA / Cu2O 2.37 2.46 2.42 pMMA / FD 0.21 1.05 0.63
pMMA / Cu2O / FD 1.76 2.27 2.02
Table 4.3: Calculated polishing rates for pMMA-based coatings.
Expected polishing rates for commercial SPCs on this apparatus are generally in the order of 5-6 µm / month (Bakker, M., Pers. Comm.). This is around 3 times the polishing rate that would be expected for the same coatings in service. pMMA/Cu2O and pMMA/Cu2O/FD coatings demonstrated a statistically significant ‘polishing’ behaviour. Some reduction in thickness was also observed for the blank pMMA binder. After 7 days, thickness had increased by around 1-2 µm compared to pre-immersion, possibly as a result of swelling
137
Chapter 4 Results and Discussion
(Fig. 4.30). From then after, thickness decreased linearly and constantly (Table 4.3). Thickness decreases in the order of 1 µm/month are effectively negligible.
Figure 4.31: Cross-sections from pMMA/Cu2O/FD (top) and pMMA/FD (bottom) coatings after 6 months of rotor immersion.
The thickness reduction observed for the pigmented coatings could be due to a smoothing effect as copper oxide pigment leached out from the coatings’ surfaces (Fig. 4.32), and as the consideration surface defects were smoothened by the flow in the tank. The addition of a high volume content of copper oxide pigment was found to have softened pMMA coatings by around 15%, which would make them more susceptible to erosion (Appendix C). The surfaces of these panels were studied with optical microscopy and heterogeneity in the distribution of the copper pigment in the film surface was apparent. Although some smoothing might be expected to occur shortly after immersion, consistent thickness decrease is unexpected. In the blank binder no pigment is present to leach, providing channels and thereby allowing erosion.
pMMA should be non-hydrolysable in water at standard operating conditions, as CH3 is a highly unreactive leaving group and the C-O bond is highly stable and has no dipole. Thus pMMA was not expected to present any polishing behaviour, as it is considered to be inert
138
Chapter 4 Results and Discussion under the operating conditions. However, the possibility of hydrolysis was investigated to ascertain whether the polishing could be attributed to this phenomenon or otherwise.
Figure 4.32: Surface of pMMA/Cu2O coating, with pale copper-depleted areas clearly visible (solid circle) and surface debris from rotor immersion (dotted circle). Scale bar 50 μm.
4.3.2.2. Metamare coatings
Figure 4.33: Thickness change of Metamare coatings with time: CDP/Cu2O (left) and
CDP/Cu2O/FD (right).
As previously mentioned, the blank and FD-only Metamare coatings detached from the primer within the first few hours of conditioning. However, the remaining Metamare coatings (containing pigment and NP/pigment) demonstrated a significant initial polishing rate similar to that expected for commercial SPCs (Fig. 4.33). The initial polishing phase
139
Chapter 4 Results and Discussion resulted in a loss of about 10 µm within the month for both coatings, before a stabilisation for the remainder of the test. At 3 months, the mean leached layer thickness (measured from 5 cross-sections) for the combinatorial CDP/Cu2O/FD panel was ~23 ± 5 μm thick, which increased to ~56 ± 5 μm after 6 months (Fig. 4.34). This behaviour is expected for a CDP type coating 1, 2 3.
Figure 4.34: Cross-section of wax-embedded CDP/Cu2O/FD from the rotor exposure trials. Note the attached primer layer that remains adhered to the coating in the second image. Scale bar 100 μm.
The Metamare coating containing both the furan derivative and Cu2O pigment polished slightly faster than the Metamare coating with pigment only (Fig. 4.33). This result is contrary to that observed for the pMMA coatings. The loss of the dispersed FD crystals by erosion or dissolution, in the same manner as rosin, may have provided pathways for copper oxide to diffuse out into the water, thereby increasing the overall speed of binder degradation. Further LSCM analysis of these eroded panels will help to study this hypothesis further by comparing the leach fronts of the various components.
The good polishing rate exhibited by these formulations is promising and helps to explain the good performance of CDP/Cu2O/FD in field trials (see Section 4.3.1). However, the poor topcoat-primer adhesion could be problematic for future formulations.
140
Chapter 4 Results and Discussion
4.3.2.3 p(TrMA/BA) coatings
Figure 4.35: Thickness change of p(TrMA/BA) copolymer coatings with time: a) p(TrMA/BA)
Blank; b) p(TrMA/BA/FD; c) p(TrMA/BA)/Cu2O and d) p(TrMA/BA)/Cu2O/FD. p(TrMA/BA) was synthesised at the University of Southampton as an experimental SPC which in principle should have a polishing mechanism analogous to that of the TBT- copolymer. However, no significant polishing behaviour was observed during the erosion tests (Fig. 4.35; Table 4.4), The rotor data demonstrate that none of the p(TrMA/BA) coatings exhibited either polishing or erosion, with no statistically significant thickness
2 decrease over the 3 month period (i.e. for p(TrMA/BA) / Cu2O, r = 0.28). Example cross- sections from post-immersion testing are shown below (Fig. 4.36).
141
Chapter 4 Results and Discussion
Thickness decrease / µm month-1 Coating Panel 1 Panel 2 Average TrMA Blank 0.97 0.73 0.85
TrMA / Cu2O 1.13 0.21 0.67 TrMA / FD 0.58 1.97 1.28
TrMA / Cu2O / FD 0.22 0.11 0.16 Reference Epoxy 0.58 n/a 0.58
Table 4.4: Calculated thickness change rates for p(TrMA/BA)-based coatings.
Fig. 4.36: Cross-sections from p(TrMA/BA)/Cu2O (top) and p(TrMA/BA)/FD coatings after 6 months of rotor immersion.
All thickness decreases were non-significant other than for the FD-only binder, and were less than those observed for both pMMA and Metamare-based binders. Although the thickness of the p(TrMA/BA) / FD coating does decrease over time according to calculation from the linear regression alone (Table 4.4), the error is so large that there is no certainty in this result, and the coating thickness could in fact not have decreased at all. The large error
142
Chapter 4 Results and Discussion for these coatings stems from a large discrepancy between adjacent points’ thickness changes, with some points increasing in thickness while others decrease. This could result from operator error in locating the exact same point each time, as the surfaces of the p(TrMA/BA) panels were quite uneven with many small scale defects such as bubbles and pits.
However, these erosion tests were conducted with first generation p(TrMA/BA) copolymers whose polishing capabilities had not been tested in any other way. Normal procedures for selection of products for erosion testing rely on down-selection of a variety of different formulations before the most promising are selected for further, more detailed experimental analysis. For this reason, erosion testing is a lengthy and time-consuming process and is therefore inappropriate for down-selection of large groups of materials, and the first attempts at polymer blends should not be expected to give optimal results. In order to examine the lack of polishing, FTIR analysis of the retrieved panel was carried out.
4.3.2.3.1. FTIR analysis of the p(TrMA/BA) panel surface
FTIR analysis of the retrieved panel surface (Fig 4.37) confirmed hydrolysis of the trityl group within the film, despite a lack of polishing. These peaks diminished in scale after aggressive cleaning, implying that the hydrolysis is confined to the surface few µm of the film only. FTIR data demonstrate the presence of trityl alcohol in the film surface, demonstrating that although some saponification may have occurred, the leaving group is retained in the film structure, further contributing to a low wettability. It is possible that the brown film of material that adheres to all surfaces in the rotor system also contributes to the trapping of the eroded monomer.
143
Chapter 4 Results and Discussion
Figure 4.37: FTIR reflectance spectrum for eroded p(TrMA/BA) coating, before cleaning.
After erosion, several differences can be observed (Fig. 4.38).The increased noise compared to the original spectrum for p(TrMA/BA) (Fig. 4.37) is attributed to the film of unknown matter on the surface that develops on all surfaces in the rotor. Aside from the increased signal noise, the several peaks that demonstrate the presence of carboxylic acid were noted. Furthermore, a medium alcohol peak (3700 cm-1) was observed. These demonstrate the hydrolysis of the trityl monomer within the film to form trityl alcohol and leave a carboxylic acid group on the resin. The large peak at 1040 cm-1 is usually attributed to C-O from primary rather than tertiary alcohol, but is a known characteristic of trityl alcohol spectrum (as obtained from a spectral database), as a result of its unusual structure. The peak immediately left of this one is also attributed to the trityl alcohol and is similarly observed on its known spectrum. A notable peak at about 1600 cm-1 is attributed to carboxylates, likely sodium salts of the carboxylic acid group.
An attempt was made to remove the surface film by incrementally more aggressive cleaning techniques (Fig. 4.38). Rinsing with distilled water at low and then high pressures made no difference to the observed spectra. Gentle wiping with tissue paper followed by a further rinse resulted in a much cleaner spectrum with less noise; in particular, the broad peak
144
Chapter 4 Results and Discussion demonstrating H-bonding between 2000 – 3000 cm-1 was much diminished, as the surface film was visibly removed. Many of the peaks corresponding to carboxylic acid are also observed to decrease in size relative to the other peaks, suggesting a limitation of hydrolysis to the surface layer only, in contrast to the polymer bulk. A further stage of more aggressive scouring was carried out, which further reduced all the peak sizes.
The appearance of three peaks at roughly 1920, 1850 and 1690 cm-1 was noted for all the immersed panels. These are observed neither on the non-immersed samples, nor the known spectra for pMMA, trityl alcohol or simple organic compounds featuring carboxylic acid. Their relative intensity is not affected by the cleaning of panels. These are attributed to organic carbon impurities in the film. A
B
Figure 4.38: FTIR reflectance spectrum for eroded p(TrMA/BA) coating, after cleaning with tissue paper (A) and scourer (B). Receding peaks compared to the un-cleaned panel marked by black arrows.
These results demonstrate that although the saponification of the ester group took place, the desired polishing of the film did not occur. Considered as a possible reason for this was the high weight content of butylacrylate as a co-monomer, which acted to prevent erosion of the depleted copolymer by preventing ingress of water by virtue of its hydrophobicity. The wettability of this film is much lower than pMMA, owing to the large, hydrophobic trityl and butyl side groups (see section 3.2). However, the high water uptake of this polymer
145
Chapter 4 Results and Discussion would suggest that this is not a viable reason for non-polishing. Most likely, the depleted polymer backbone retains enough strength to avoid polishing. Although the copolymer is a 50:50 wt.% blend, the actual ratio of repeating units is 2.5625:1 (due to the large size of the triphenyl group).
4.3.2.4 Dynamic erosion summary
Pigmented pMMA coatings demonstrated a slow but statistically significant linear thickness decrease. As no such decrease was observed for non-pigmented pMMA samples, this is attributed to copper oxide depletion and degradation of the skeletal binder matrix. The CDP/FD and CDP blank coatings all detached from the epoxy primer in the 24 hour static period of the dynamic testing. The remaining pigmented Metamare coatings demonstrated a rapid erosion of roughly 10 µm in the first month followed by stabilisation of the binder, with no further statistically significant erosion occurring. Copper oxide depletion continued throughout this period, as evidenced by increasing leached layer thicknesses throughout this period. This in indicative of a failure of the weakened leached layer structure to polish, even in the aggressive rotor regime. The p(TrMA/BA) coatings did demonstrate some hydrolysis, but failed to polish. This was attributed to the very high butylacrylate content relative to TrMA monomer.
4.4. Post-immersion natural product distribution
4.4.1. pMMA coatings
4.4.1.1. Non-pigmented coatings
The novel fluorescence technique was firstly applied to the pMMA/FD films in order to ascertain the extent of leaching within the cross-section (Fig. 4.39). Optical imaging of the cross-sectioned rotor-immersed film revealed the presence of crystals remaining within the film despite the aggressive leaching regime (Fig. 4.31). The unimmersed sample prepared simultaneously with the rotor samples demonstrated a high FI relative to the blank binder. The unimmersed binder as well as the rotor samples all demonstrated a high FI variance
146
Chapter 4 Results and Discussion with respect to film position resulting from crystals in the cross-section, the presence of which was not observed to diminish over time. After 6 months of immersion, the FI significantly decreased compared to the unimmersed sample, especially at the film surface. Estimation of areas beneath the mean curves demonstrates a 35% additive loss after 6 months. After 3 months, the profile was very similar, with 27% lost on average (data not shown for clarity reasons). The lack of thickness decrease coupled with the slow leaching of the biocide could explain the fact the AF performance was equivalent to the blank panel under pontoon immersion conditions (see Section 4.3.1).
Figure 4.39: a): Optical cross-section image of pMMA/FD film after 6 months of rotor immersion; b): fluorescence microscopy image of unimmersed pMMA/FD film cross-section; c): fluorescence microscopy image of pMMA/FD film cross-section after 6 months of rotor immersion. The fluorescent signal from the mean of three cross-sectioned samples for unimmersed (solid line, black) 6 month eroded (dotted line, grey) pMMA/FD coatings is shown right. Shaded areas represent 90% prediction intervals for the mean lines.
The pMMA/FD coatings from the pontoon immersion were also cross-sectioned and imaged in the same manner (Fig. 4.40). A large decrease occurred in FI over the 10 month period with an average FI reduction of 61% in the front-mounted panels and 25% in the back- mounted panels. An unexpected observation was the slight decrease in FI towards the bottom of the film, i.e. the PVC substrate, observed for all samples. When FM inspection was carried out, it was observed that a gradient of FI was present at the interface, corresponding with a diffusion of FD into the substrate throroughout the test period; this
147
Chapter 4 Results and Discussion was also observed for other binder types on PVC (see previous section), but never for the rotor immersion panels which used an epoxy primer tiecoat.
Figure 4.40: a) Fluorescence cross-section of pMMA/FD coating before-immersion, b): after 10 months of pontoon immersion from the shaded back side of the board, c): after 10 months of pontoon immersion from the lit front side of the board. Scale bars 200 μm. The average signal from line profiles associated with three cross-sectioned samples of unimmersed control (solid line, black), rear-mounted pMMA/FD panels (long dashed line, red), front-mounted panels (short dashed line, grey) and blank pMMA panel (solid line, blue) is shown (right) including 90% prediction intervals.
4.4.1.2. Pigmented coatings
The FI was also assessed in the cross-sections of pMMA/Cu2O/FD and pMMA/Cu2O coatings after 3 and 6 months of erosion using a rhodamine filter block (Fig. 4.40) which offered a good contrast for all film components. The pMMA/Cu2O exhibited a slightly higher FI than that of the blank film (Fig. 4.41), but which was much less than the pigmented FD- containing film. Signal maxima were expected to correspond to the presence of FD crystals in contact with the imaged surface, or gaps between copper oxide agglomerates. The average signal is markedly higher for the cross-sections from the FD coating than for the non-FD coating due to loci of high signal. The result of these numerous areas of high FI is that high variance of line profiles is observed for these coatings. The profiles demonstrate no FI gradient within the film, and the 3 month aged samples have no statiscally significant difference from the 6 month samples (p > 0.1). The observation of thickness decrease for
148
Chapter 4 Results and Discussion these samples - coupled with the lack of leached layers with respect to both copper (I) oxide and usnic acid - suggests that breakdown of the depleted film occurs in a fairly synchronous manner with the copper oxide depletion. As no separate FD leached layer is observed, the leaching of FD appears to be impaired by the copper front, and that water permeation below this front is hindered. This is in good agreement with other studies that have been carried out for copper leaching from coatings2, 4, 5
Figure 4.41: Fluorescence cross-sections of pMMA/Cu2O (a) and pMMA/Cu2O/FD (b) films after 10 months of static immersion. Optical microscope image of the pMMA/Cu2O/FD cross-section is also shown (c), with an attached thin algal film on the surface (~10 μm in thickness). Scale bars 200 μm. The average signal from line profiles associated with three cross-sectioned samples of pMMA/Cu2O/FD (long dashed line, grey), pMMA/Cu2O (short dashed line, red) and blank pMMA panel (solid line, blue) is demonstrated (right) including 90% confidence intervals.
Unlike the pMMA/FD coatings, there was no discernible difference between front- and back-mounted panels. Given the lack of leaching of FD beyond the copper front, and the good matching between copper depletion and binder degradation, this is an expected result.
149
Chapter 4 Results and Discussion
4.4.2. Metamare coatings
4.4.2.1 Non-pigmented coatings
The fluorescence technique was also applied to multiple cross-sections from statically immersed CDP panels containing only FD, without pigment (i.e. CDP/FD). Unimmersed controls demonstrated a much higher fluorescence intensity thoroughout the film thickness compared to blanks (Fig. 4.42). Fluorescence intensity in the cross-section varied as a result of visible crystals that provided local maxima of high signal. No significant trend in fluorescence intensity was observed with film position, implying an even distribution of crystals and dissolved additive in the bulk. 10 month pontoon-immersed panels demonstrated an interesting trend, with a different behaviour observed for panels mounted on the boards’ front and back sides, as noted for pMMA/FD coatings. In both schemes, a decrease was observed in the film surface relative to the unimmersed films. However, in the case of the front mounted samples an average of 53% less FI was detected in cross-sections throughout the entirety of the film thickness. A lower variance was also remarked for these samples, which may correspond to a decreased occurrence of crystals in the cross-section, which could result from their dissolution into the binder bulk as concentration decreases in the surrounding polymer. The average additive loss from the rear-mounted panels was estimated at 12.2% overall, although the only significant decrease from the unimmersed film occurred in the surface 30 µm.
150
Chapter 4 Results and Discussion
Figure 4.42: CDP/FD cross-sections from unimmersed control (top left) and rear-mounted (middle left) and front-mounted (bottom right) after 10 months of static immersion. Mean fluorescence intensity profiles for two front (short dashed, grey) and two back panels (medium dashed line, red) are demonstrated relative to the non-FD bearing blank binder (long dashed, blue) with 90% confidence intervals. The fluorescence intensity relating to the unimmersed control sample is denoted by the solid black line. Scale bars 200 μm.
As noted for pMMA/FD coatings, the release of FD was significantly faster on the front- mounted panels. This could have resulted from hydrodynamic variation between both sides, or from the different fouling community on the front. It should be noted that none of these non-pigmented coatings offered a good AF performance regardless of FD leaching rate.
4.4.2.2. Cu2O Pigmented coatings
As previously described, examination of rotor immersed and statistically immersed panels revealed the development of leached layers resulting from asynchronous copper oxide depletion and binder erosion rates. To recap, leached layers formed more rapidly in the rotor tests. At 3 months, the mean leached layer thickness (measured from 5 cross-sections) for the combinatorial CDP/Cu2O/FD panel was ~23 ± 5 μm thick, increasing to ~56 ± 5 μm after 6 months. The rate of thickness decrease for rotor samples during the first month was rapid, before stabilisation; this behaviour is expected for a CDP type coating1, 6, 7. Thickness loss for CDP/Cu2O panels was slightly greater than for CDP/Cu2O/FD panels in the rotor trial,
151
Chapter 4 Results and Discussion although leached layer thickness was of a similar order.
FM was employed to assess the binder distribution of FD within the statically-aged
CDP/Cu2O and CDP/Cu2O/FD coatings using the rhodamine filter block (Fig. 4.43). Fluorescent signal associated with the organic compound allows the distribution to be assessed relative to the non-pigmented, blank binder and pigmented binder without FD. The blank erodible binder demonstrates a degree of autofluorescence as a result of the high rosin content. The attribution of this fluorescence to rosin in the binder was confirmed by separate analysis of a pure rosin sample under the same excitation wavelength as described in Section 3. The inherent signal from the rosin within the binder is diminished by the addition of inert copper. Signal is uniformly increased in the cross-section when FD is present in the binder. The signal also demonstrates a large variance in the 90% confidence intervals. This results from exposed solid crystals in the cross-section, which are randomly distributed in the films, which have a much higher fluorescence intensity than the surrounding film area, rather than any inherent inaccuracy in the method, similarly to the pMMA films where crystals of usnic acid were dispersed in the pigmented matrix. Mean fluorescence intensity measurements from the immersed CDP/Cu2O/FD panels (Fig. 4.43) reveal a defined structure, with a depleted signal in the surface layer equivalent to the blank binder. The fluorescence intensity gradually increases to a depth of 20 - 40 μm, whereafter the intensity is stable at a much higher level than the blank binder as a result of high FD content. The layer where the usnic acid content is lower has a thickness corresponding to the copper oxide leached front. Below the leached front, the signal derived from the binder is significantly higher than the other two formulations as a result of the FD content. This again demonstrates that leaching of FD is impaired below the copper leach front, but is readily removed from the copper oxide leached layer, likely as a result of its highly porous structure. Analysis of cross-sections from the front-immersed panels showed similar results, with a shallower FD leached layer corresponding to the reduced copper oxide leached layer also noted for these samples. The presence of the skeletal leached layer is also highlighted by using a different filter block with increased gain; this image (Fig. 4.43d) demonstrates the copper and FD front relative to this depleted leached layer.
152
Chapter 4 Results and Discussion
Figure 4.43: Fluorescence cross-sections of various CDP films after 10 months of pontoon immersion: a) Blank CDP (rhodamine), b) CDP/Cu2O (rhodamine), c) CDP/Cu2O/FD
(rhodamine), d) CDP/Cu2O/FD (GFP) demonstrating leached layer not visible with rhodamine filter block. All scale bars 200 μm. The average signal from line profiles associated with three samples of each binder type are shown (right) including 90% prediction intervals:
Blank CDP binder (short dashed line, blue), CDP/Cu2O (long dashed line, red), 10 month immersed CDP/Cu2O/FD (solid line) and unimmersed CDP/Cu2O/FD (dotted line, yellow).
4.4.3. p(TrMA/BA) coatings cross-sections of the novel copolymer coatings from rotor immersion were also prepared in order to characterise leaching behaviour of usnic acid in this binder type. Owing to the small quantities of copolymer available, the coatings were only submitted for rotor erosion, without static seawater immersion taking place. As observed in other binder formulations, fluorescence intensity in the cross-section varied as a result of a few cross-sectioned crystals that provided local maxima of high signal. Analysis of p(TrMA/BA)/FD after 3 months and 6 months of rotor immersion demonstrates a progressive and rapid leaching of FD from the binder (Fig. 4.44). Progressive dissolution of the FD crystals throughout the study resulted in a decreasing variance in the fluorescence intensity with respect to depth for the 3 month and 6 month intervals. Fluorescence intensity is much depleted in the three month sample with 85.6% of the additive lost, and is in line with blank values in the surface 20 μm,
153
Chapter 4 Results and Discussion suggesting a total depletion of the additive at the film surface. In the 6 month sample, concentrations are in line with blank samples throughout the entire film, although small loci of high signal are observed towards the bottom of the film in the cross-section, corresponding to a few remaining crystals of FD within the depleted binder. To all intents and purposes the biocide can be considered completely depleted at this point. For the purposes of comparison with modelling work later on, a value of 100% depletion was used for the 6 month aged p(TrMA/BA) samples.
Figure 4.44: p(TrMA/BA)/FD coating – un-immersed (a), 3 month erosion scheme (b) and 6 month erosion scheme (c). The mean signal derived from multiple line profiles across three cross-sections from each panel are shown right – un-immersed (solid line, grey) 3 months (long dashed line, red) and 6 months (dotted line, cyan).The fluorescence intensity from blank p(TrMA/BA) is denoted by the short dashed line (black). Scale bars are 100 μm.
The FI profiles taken from rotor immersed p(TrMA/BA)/Cu/FD films could not be compared with an unimmersed control, as a control for this combination was not prepared. However no statistically significant difference could be observed for FI profiles from either the 6 month aged coating or the intermediate 3 month one (Fig. 4.45). FI levels in both cross- sections were low owing to the high copper oxide content, with high variance (as generally observed for pigmented coatings)
154
Chapter 4 Results and Discussion
1400 6 months 1200 3 months
1000
800
600
Fluorescence Intensity Fluorescence 400
200
0 0 20 40 60 80 100 120 140
Distance from film surface / m
Figure 4.45: Fluorescence intensity in rotor-aged p(TrMA/BA)/Cu2O/FD coatings; after 3 months of immersion (solid red line, 90% confidence intervals shown by dashed red lines) and 6 months of immersion (solid black line, 90% confidence intervals shown by dashed black lines).
Despite the lack of a control, the lack of significant change between the 3 and 6 month eroded panels implies the absence of FD leaching below the surface of the film (i.e. the copper front in this case).
4.4.4. Summary of post-immersion usnic acid distribution
Fluorescence microscopy was successfully used to image NP distribution in binder cross-sections, even for pigmented binders. Faster depletion of additive occurred for front-mounted panels in both pMMA (after 10 months, 25% lost from back vs. 61% lost from front) and Metamare coatings (after 10 months, 12% lost from back vs. 53% from front). Faster depletion of additive from pMMA than for the CDP, Metamare, was a surprising result. Given that the Metamare also developed very thick leached layers without polishing, and the separation from the primer very rapidly in the rotor trials
155
Chapter 4 Results and Discussion
(even before rotation commenced), it would appear that the formulation of this binder is not very well tailored to use as an antifouling coating. The failure to polish might be alleviated by addition of a higher rosin content or reduction or elimination of the butylacrylate. Usnic acid leaching from rotor trials occurred preferentially in the surface, with a more typical Fickian-type diffusion profile. Leaching from pMMA in immersion trials occurred uniformly throughout the samples. This could have been partly due to a simultaneous diffusion of pMMA into the substrate as observed for non-immersed, 10 month aged samples (see Section 3.2.2). In pigmented samples, leaching of FD never occurred below the copper (I) oxide front; i.e. the pigment leaching rate is the limiting factor in controlling overall leaching for the usnic acid biocide. This was the case in all binder types and is a useful confirmation of other authors’ observations. FD leached extremely rapidly from the p(TrMA/BA)/FD coating, despite its inability to polish, with 86% depletion after only three months and 100% depletion at 6 months. This is further discussed in the next section.
4.5. Overall discussion
4.5.1. Erosion, copper oxide distribution and leaching
Copper leaching was observed only in the erodible CDP and pMMA binders. In the p(TrMA/BA) binders, no polishing or leached layer development occurred. Even with no polishing occurring, reaction of seawater with the exposed particles4, 8 occurred at a rate insufficient to provide a visible leached layer (≥1 μm) even over the 6 month dynamic erosion period, or 10 months of static exposure, with 30 vol.% Cu2O present. Although this was not the desired result for the copolymer, FTIR analysis of the sample surfaces confirmed a degree of hydrolysis and the presence of triphenyl alcohol as an expected product of saponification of the ester. It is possible that hydrolysis did occur in-situ and that the remaining binder was too hydrophobic to ‘polish’ as a result of the dominant butyl copolymer, which composes about 72% of the units, or that hydrolysis occurred ‘in pot’ before coating or immersion as a result of a lack of stability of the copolymer formulation. In p(TrMA/BA) binders, copper leaching did not occur either, with no leached layer observed during microscopy examination of the samples post-immersion.
156
Chapter 4 Results and Discussion
In the pMMA coatings, a thickness decrease occurred, but no leached layers developed, implying a coupling of the two effects; i.e. loss of copper (I) oxide leading to degradation of the depleted binder. The release of the large copper oxide particles is intrinsically linked to the breakdown of the containing medium. In the Metamare (CDP) binder, a rapid thickness decrease occurred in rotor trials. However, the initial phase of binder erosion lasted only a month; copper oxide continued to be lost thereafter, resulting the formation of leached layers well within 3 months of rotor exposure, and the generation of increasingly thick leached layers after 6 months of exposure. The increasing leached layer of CDP panels was also in evidence during pontoon trials, where the coatings grew paler throughout time. Optical microscopy inspection of these coatings after their removal confirmed significant leached layer development.
The thickness decrease of the CDP coatings was actually of a similar order to that exhibited by the pMMA coatings. Given the much larger leached layers observed in the former, it may be inferred that copper oxide release was significantly accelerated by the presence of rosin. The failure of the Metamare binder to degrade after the first month, even in aggressive accelerated conditions, should be of concern to its manufacturer. Unfortunately, thickness decrease was not confirmed in the statically immersed pMMA coatings. The leached layer was not observed in these samples, but it seems reasonable to assume that a similar mechanism of copper oxide loss and binder erosion was in effect, as observed for rotor- tested pMMA coatings, in light of the good AF performance of pigmented panels that could not be explained without copper leaching. Although the pigmented pMMA panels demonstrated some positive inhibitory behaviour, a superior performance of the CDP coatings compared to the pMMA coatings was noted. This corroborates the thicker leached layers observed in the former, corresponding to increased copper oxide delivery rate.
An interesting observation is the much greater thickness of copper oxide leached layers on panels on the rear, dark side of the boards (31 μm after 10 months) compared to the front ones (17 μm), an observation that was remarkably uniform among all CDP samples. The difference between front and back panels could result from a more aggressive hydrodynamic regime on one panel side. There are two possible scenarios: faster copper (I) oxide leaching from the boards on the rear side, or faster erosion of the leached layer on the front-mounted panels. Although the latter seems more plausible, insufficient data is available on the physical surroundings of the pontoon to attribute the discrepancy to any hydrodynamic effect. Nevertheless it was noted that the operation of large ferry vessels was
157
Chapter 4 Results and Discussion observed in the harbour from time to time. Their operation resulted in extremely aggressive currents around the periphery of the harbour that inclined towards the front sides of the mounted boards. The force of these currents was enough to push the boards to near- horizontal and dislodge some loose fouling, so may have had an effect on the leached layer thickness on the front-mounted panels.
AF performance of the rear panels was highlighted, as the rear CDP/Cu2O/FD panels resisted fouling over the 10 month period. It is worth pointing out that the optical and fluorescence microscopy demonstrated that this binder type was the only one where synchronous FD and Cu2O leaching was confirmed. The enhanced performance is not attributed to the effect of FD, however; as minimal inhibitory effects were observed for FD in any of the other panels without pigment. It is more likely that the addition of a further ‘leachable’ agent allows for a more rapid and complete copper (I) oxide release. This would also explain the thicker leached layer observed for these samples.
The copper oxide leached layers formed in the CDP coatings by the rotor method were roughly four times as thick as those formed in an equivalent length of static exposure; for example, after 6 months of rotor exposure, roughly 56 μm of leached layer had formed uniformly across the combinatorial panel’s surface, whereas after 10 months of static immersion, only 31 μm had formed on the unfouled rear panels, with as little as 17 μm of leached layer on the heavily fouled front panels. As a crude estimation, the data suggest that the rotor approach (at 17 knots and 25 oC) was somewhere from four to seven times more aggressive that the equivalent length static exposure in terms of inducing pigment leaching, depending on the formulation. Rotor testing remains a crucial tool in advanced testing of selected promising formulations.
4.5.2. Natural product distribution and leaching
LSCM and FM were employed to characterise FD crystal distribution and dispersed FD distribution, respectively. The quantity of FD crystals in each mixture varied slightly, as determined by the solubility limit of the compound in each solvent. As a result of the fine optical exclusivity of the LSCM method, pre-determination and decoupling of binder and topcoat roughness scales is necessary to assess the distribution of crystals in the z-plane. In clear resins, this technique is carried out non-destructively and provides a large area specimen size. For discrete objects such as crystals or microcapsules, this is far more rapid
158
Chapter 4 Results and Discussion to carry out than a suitably large number of cross-sections. Crystals were well dispersed in the CDP binder, but their distribution was skewed towards the bottom of the pMMA and p(TrMA/BA) films, possibly as a result of the slightly lower viscosity of these mixtures. These crystals were also prominent in coating cross-sections produced for FM analysis. Analysis of FI was carried out by assessing line profiles at pre-set intervals across the cross-section images. Where one of these line profiles bisected a crystal in the cross-section, an FI peak occurred. The result of these peaks is a high variance for sample sets. In nearly all post- immersion analyses, variance of the sample sets decreased over time with average FI. Decreasing FI indicates leaching of the additive, and the decreasing variance indicated a lower frequency of crystals in the binder matrix as the solid-state FD replenished the surrounding matrix, a well-established phenomenon in analogous drug delivery mechanisms9. The use of fluorescence microscopy to carry out analyses of biocide distribution represents a novel step in post-immersion analysis, and is a good candidate for extrapolation to mainstream biocides, many of which possess potential fluorophores, as a part of future work.
FD release from all three binders was observed in rotor and static testing. The amount of additive released from each binder varied significantly with p(TrMA/BA) being the only binder to be virtually depleted over the testing period; the rotor-tested panel removed at the 3 month interval contained only 14% of the additive present in the equivalent non- immersed coating, and after 6 month the FI had depleted to base levels for the polymer medium, with the exception of a few discrete crystals. As a result of failure at the topcoat- primer interface, the CDP blank and CDP/FD formulations were lost early in the rotor experiment. However the pontoon-immersion equivalent CDP formulations exhibited a decrease in FI throughout the static immersion experiment. After 6 months, less than half of the original amount remained (47%) across the entire film. FI decreased more rapidly in the surface of the film. In the case of a non-polishing, non-erodible binder, assuming no diffusion at the binder/substrate interface, one would expect an inverted Fickian-style curve for additive concentration relative to film position to develop as leaching occurs more rapidly in the surface, down the concentration gradient (i.e. the seawater in this case). In reality, many processes are superimposed upon the diffusion of an additive on the molecular scale, such as diffusion into the substrate, and particularly water ingress within the coating resulting in swelling and plasticisation of the polymer chains. The latter is likely to be affected greatly by the degree and nature pigmentation and development of leached layer front. Diffusion of FD in acrylic coatings is expected to be very slow, as a result of a
159
Chapter 4 Results and Discussion combination of the low swelling capacity of hydrophobic acrylic resins, the closeness of the side groups of the polymer chains, the large size of the penetrant molecule – which is considerably larger than the pMMA monomer - and its capacity for molecular interaction with the ester groups of the acrylic monomer. pMMA demonstrated a similar degree of leaching to the CDP, with roughly 40% remaining after 10 months of static immersion. However, the FI profile throughout the coating bulk was very different in the case of pontoon-immersed samples; counterintuitively, there is no decrease of concentration at the surface compared to the bulk, and there is debatably even a slight decrease towards the substrate observed in all samples, although it is not statistically significant. In fact, diffusion of usnic acid into the substrate was noted, and should be subject to further investigation, being a potentially important aspect of biocide activity neglected in existing literature. Although not covered in the remit of this thesis, the same phenomenon was observed for coatings where juglone was incorporated into topcoats directly in contact with PVC substrate; in this instance, the diffusion of juglone into the substrate was even noticeable with optical microscopy, owing to the deep brown colouring of this compound.
As discussed, seawater penetration below the copper oxide front appears to be impeded, representing an efficient mechanism for limiting co-biocide release. Previous authors have commented on the limitation of water intrusion by the copper oxide front 4, 10. Although some water inevitably enters below this layer, the hydrophobicity and magnitude of the pigment particles seems sufficient to limit plasticization of the polymer below this front, preventing diffusion of the large FD penetrant. In binders containing FD without pigment, leaching of the additive was observed to occur at a faster rate. The observation of this free leaching in the binders without Cu2O lends further support to the influence of pigment leaching in controlling the leach rate of co-biocides. For example, pMMA/Cu2O/FD coatings demonstrated an erosion (thickness decrease), but no leaching beyond the surface with respect to FD content. However, leaching of FD did occur in the equivalent non-pigmented coating. This is attributed to the differing mechanisms governing the dissolution of each molecule; copper oxide relies on seawater interaction and the formation of soluble species at the pigment-water interface. Diffusion of the additive is likely facilitated when seawater is present in the surrounding matrix, so is hindered below the pigment front. For example, when formulated with pigment, leaching from the p(TrMA/BA) binder was totally inhibited for the FD biocides and for the copper oxide, as no degradation of the film surface took place to allow copper release. Without pigment, however, the FD biocide was able to leach from the film, in this case very rapidly. As a last example, the leached layer of copper (I)
160
Chapter 4 Results and Discussion oxide corresponding exactly to the usnic acid leached layer in the combinatorial CDP coatings, with a rapid transition from blank binder FI levels to control FI levels across the leached layer interface. This binder offered excellent AF performance for the duration of the 10 month immersion trial as a result of the synchronous leaching of FD and copper (I) oxide. The impairment of FD leaching below the copper front was also notable in the p(TrMA/BA) formulations, where FD leaching occurred extremely fast from the non-pigmented binder, but was not significantly reduced when paired with a high volume of copper (I) oxide pigment. Despite the lack of a control coating for comparison with the latter formulation, there was no statistically significant decrease in FI between 3 and 6 months of immersion.
Figure 4.46: Calculated Hansen solubility parameters for pMMA, TrMA and usnic acid.
The discrepancy in FD leaching behaviour between the p(TrMA/BA) and traditional acrylic binders is of great interest – FD leaching from p(TrMA/BA) occurred significantly faster than from either of the other binders. The most striking difference between the two materials is the difference in water uptake: the copolymer possesses a much higher water uptake capacity than pMMA (10% water volume fraction at saturation at ambient conditions, compared to 1% for the other binders; see section 3.2), a surprising result given the hydrophobicity of the aromatic side groups. It is probable that water uptake is of great importance to determining the potential loss rate of contained compounds - the polymer
161
Chapter 4 Results and Discussion network is slightly plasticised as water uptake occurs, allowing for a facilitating effect of diffusion in the polymer; the diffusive system is therefore a very complex, interactive one. The intrusion of water into the binder matrix is likely to have an even greater effect for highly water-soluble compounds, as may be observed for certain fractions of crude algal extracts or NPs that may exhibit AF behaviour11-14. Further exploration of this effect was the focus of the modelling work carried out in the next section.
Intermolecular attractions between the dispersed biocide and the polymer will also reduce its effective diffusivity. The chemical dissimilarity of somewhat polar FD with the completely nonpolar trityl monomer may also have been a contributing factor to its more rapid depletion, whereas a degree of hydrogen bonding is possible between FD and the ester group in the acrylate polymer or carboxylic acid moiety of the rosin, owing to its hydroxyl groups. Hansen solubility parameters15, 16 were calculated for trityl monomer, methyl methacrylate and the FD molecule (Fig 4.46). Calculated values for MMA were in very good agreement with measured literature values16. The estimations demonstrate that the three structures lack any particular affinity, but indicate that the TrMA is extremely chemically dissimilar from FD, and should undergo weaker molecular interactions with the additive than the MMA monomer. The possibility of molecular interactions between FD and MMA is expected to slightly slow the overall diffusion of FD in this binder17, 18, although this phenomenon is discussed in more depth in the modelling section (Section 5).
4.6. References
1. Yebra, D.M., Kiil, S., and Dam-Johansen, K., Antifouling technology – past, present and future steps towards efficient and environmentally friendly antifouling coatings. Progress in Organic Coatings, 2004. 50: p. 75-104. 2. Yebra, D.M., Kiil, S., Dam-Johansen, K., and Weinell, C., Reaction rate estimation of controlled-release antifouling paint binders: rosin-based systems. Progress in Organic Coatings, 2005. 53: p. 256-275. 3. Berglin, M. and Elwing, H., Erosion of a model rosin-based marine antifouling paint binder as studied with quartz crystal microbalance with dissipation monitoring (QCM- D) and ellipsometry. Progress in Organic Coatings, 2008. 61(1): p. 83-88. 4. Yebra, D.M., Kiil, S., and Dam-Johansen, K., Mathematical modeling of tin-free chemically active antifouling paint behavior. American Institute of Chemical Engineers' Journal, 2006. 52(5): p. 1926-1940. 5. Yebra, D.M., Kiil, S., Weinell, C., and Dam-Johansen, K., Supplementary material to 'Analysis of chemically-active antifouling paints by mathematical modelling'. Progress
162
Chapter 4 Results and Discussion
in Organic Coatings, 2006: p. 20. 6. Berglin, M. and Elwing, H., Erosion of a model rosin-based marine antifouling paint binder as studied with quartz crystal microbalance with dissipation monitoring (QCM- D) and ellipsometry. Progress in Organic Coatings, 2008. 61: p. 83-88. 7. Yebra, D.M., Kiil, S., Dam-Johansen, K., and Weinell, C.E., Reaction rate estimation of controlled-release antifouling paint binders: Rosin-based systems. Progress in Organic Coatings, 2005. 53: p. 256-275. 8. Yebra, D.M., Kiil, S., Weinell, C.E., and Dam-Johansen, K., Supplementary material to 'Analysis of chemically-active antifouling paints by mathematical modelling'. 2006: p. 20 pp. 9. Cussler, E.L., Diffusion: Mass Transfer in Fluid Systems. 3rd ed2007: Cambridge University Press. 655. 10. Kiil, S. and Yebra, D.M., Modelling the design and optimization of chemically active marine antifouling coatings, in Advances in marine antifouling coatings and technologies, D.M. Yebra and C. Hellio, Editors. 2009, Woodhead Publishing in Materials (CRC Press Ltd.). p. 393-421. 11. Blunt, J.W., Copp, B.R., Munro, M.H.G., Northcote, P.T., and Prinsep, M., Marine natural products. Natural Product Reports, 2004. 21: p. 1-49. 12. Blunt, J.W., Copp, B.R., Munro, M.H.G., Northcote, P.T., and Prinsep, M., Marine natural products. Natural Product Reports, 2006. 23: p. 26-78. 13. Hellio, C., Berge, J.-P., Beaupoil, C., Le Gal, Y., and Bourguognon, N., Screening of marine algal extracts for anti-settlement activities against microalgae and macroalgae. Biofouling, 2002. 18(3): p. 205-215. 14. Hellio, C., De La Broise, D., Dufossé, L., Le Gal, Y., and Bourguognon, N., Inhibition of marine bacteria by extracts of microalgae: potential use for environmentally friendly antifouling paints. Marine Environmental Research, 2001. 52: p. 231-247. 15. Hansen, C.M., 50 Years with solubility parameters - past and future. Progress in Organic Coatings, 2004. 51: p. 77-84. 16. Hansen, C.M., Hansen solubility parameters: a user's handbook. 2nd ed2007, Boca Raton, Florida: CRC Press. 224. 17. Sakelleriou, P. and Kapadia, K., Diffusion of organic molecules through epoxy/acrylic copolymer films. European Polymer Journal, 1996. 32(5): p. 601-604. 18. Tonge, M.P. and Gilbert, R.G., Testing models for penetrant diffusion in glassy polymers. Polymer, 2001. 42: p. 501-513.
163
Chapter 5 Mathematical Modelling of NP Diffusion
5. Mathematical Modelling of NP Diffusion
5.1. Background
5.1.1. The challenge of modelling marine paint systems
The ability to accurately predict coating lifetimes is a holy grail for coating manufacturers. The development of formulations requires constant tweaking based on results from natural immersion tests at a variety of sites, as well as accelerated erosion tests. Depending on observations made, modifications to the Tg may be required, for example, resulting in a change in type or relative quantities of monomer. The biocide may turn out to have a plasticising effect on the binder and release too rapidly, so encapsulation of the biocide is necessary in order to mitigate these phenomena. However a change in one parameter can be deleterious for overall coating performance, owing to the complex coupling of processes within the binder. To quote Yebra et al.,1
‘The main difficulty in developing efficient AF products is that the coupling of the main paint processes is so marked that typically only very few formulations among a large number of possibilities result in adequate paint polishing and biocide leaching simultaneously. The substitution of one major paint component may cause a dramatic misbalance in the performance, which often takes years of research and development to solve.’
As discussed in previous chapters, a reasonable understanding has been attained in three key areas, thanks to the efforts of the previous authors: the kinetics of rosin degradation in CDP binders, simulation of TBT release rate from SPCs, and the mechanisms of copper oxide depletion via reaction with seawater1-10. Even despite the complex and complete representation of redox chemistry and speciation of major ions involved in the reaction pathways of copper oxide and its derivatives in a CDP binder, the resulting models tend to be reasonably specific to the binders within individual studies, with extrapolation of parameters in more general terms proving an elusive goal. For example, Yebra et al.1 commented that mechanistic assumptions made on dissolution kinetics of copper (I) oxide would probably not hold for other binder chemistries and wettabilities.
165
Chapter 5 Mathematical Modelling of NP Diffusion
5.1.2. Modelling biocide release
Conspicuous by its absence is modelling of the release rate of conventional biocides. Many similarities remain with the analysis of copper oxide and TBT leaching, in that the degradation of the binder is likely to play a key role in allowing for increased water ingress. However, the mechanisms of biocide depletion are likely to be significantly different. Yebra et al.1 considered an impermeable binder in which diffusion occurs within tortuous pathways generated by conversion of the binder, and where the dissolution of copper particles occurs exclusively at the pigment front (Fig. 5.1).
Figure 5.1: Paint binder pores created by rosin and pigment dissolution (adapted from Yebra et al.1).
The case for organic biocides is different, as these will be evenly dispersed in the polymer matrix. The formation of pores will be determined by the dissolution of pigment and rosin, if present. However the natural uptake of water in the polymer matrix itself should not be neglected; absorption of water into polymer binders is highly variable, depending on
166
Chapter 5 Mathematical Modelling of NP Diffusion wettability of the binder and its dispersed additives, free volume in the matrix in the form of cavities, and volume fraction of insoluble particulates (i.e. pigments). Water uptake in polymer binders will manifest in the form of swelling. Swelling for the present binder systems was determined in a previous chapter (see Materials and Methodology). Loss of dispersed material into this swelling water fraction is also inevitable, and in hard non- pigmented binders may represent the major source of depletion, as erosion of the polymer is unlikely. Ternery (polymer + penetrant + diluent) systems have been considered in the literature in the case of rubbery regimes, but not in the case of glassy ones. Quantification of this loss would be a useful tool in the understanding of the effect of binder water uptake and rate of diffusion of a given additive in the ‘base’ binder, before addition of pigments and other elements. The movement of the biocide in the diluent is not free; it will be restricted by the polymer chains, whose packing will be affected by the degree of swelling (i.e. Fig 5.2). Exploration of this phenomenon is the goal of the present chapter. The principal objective is the comparison of theoretical and experimental data for leaching rates of the furan derivative usnic acid biocide (FD), which represents a novel body of work owing to some of the previous limitations on biocide detection that have been resolved by use of fluorescence microscopy techniques.
A B C
Figure 5.2: In heavily congested networks where the diluent comprises a low volume of the overall system, the properties of the network dominate diffusivity. Here, the yellow ball represents the migrant, blue the diluent, and black the polymer or protein network. A: free diffusion of a migrant in a liquid; B: diffusion in a network containing a low volume content of fibres or barriers; C: diffusion in a network containing a high volume content of fibres or barriers.
167
Chapter 5 Mathematical Modelling of NP Diffusion
5.1.2. Selection of a model, and its limitations
As discussed in more detail in Section 2, estimation of diffusivity in polymers is a classically challenging case as a result of the heterogeneity of the polymer medium and its variable state11. Literature consideration of diffusion focuses more towards polymer systems above
12-15 Tg and those papers considering diffusivity in glassy systems are driven more towards the migration of very small penetrants and self-diffusion of monomers16-18. For the purpose of the present study, a free-volume based physical model (Gray-Weale et al. 1997) was selected for the estimation of diffusion coefficients19, 20.
The model predicts diffusion coefficients by representing the penetrant molecule and side groups on the polymer backbone as spherical entities. The geometry of the reactant state and transition state cavities are represented by spheres (with four interacting groups) and cylinders (three interacting groups), respectively. These values for the number of interacting groups are drawn from the authors’ own computer simulations of molecular interaction. However, they were observed for a wide range of polymer-penetrant couples, and are hence considered safe to use in this context. The modelling of each group or penetrant entity as spherical unified atoms greatly simplifies calculation. The cavity jump is schematically represented in Fig. 5.3. The spaces between side groups on adjacent chains are repsented as cavities, and the critical energy required for movement of a penetrant molecule between cavities must be calculated. The activation energy is determined by calculating the difference in energy between the reactant state (cavity) and transition state, where the penetrant must deform the side groups to its width and overcome repulsive forces owing to the side groups’ proximity. The frequency of jumps (time) can be then calculated and the jump length (distance) can be obtained from geometric measurements, allowing for determination of the diffusion coefficient (cm2 s-1).
168
Chapter 5 Mathematical Modelling of NP Diffusion
Fig 5.3: Reactant-transition-reactant cavity jump demonstrating the number of interacting units (4-3-4) within each stage. Note that owing to limitations in the model, only one side group type can be considered at a time.
The nature of the model is such that the diffusion coefficient must be calculated on a per side group basis (i.e. see Fig 5.3). Therefore, for pMMA, the calculation of the penetrant- side group interactions must be performed twice, once for the acrylate side group and once for the methyl group pendant to the backbone. The geometric mean of the diffusion coefficients obtained for each side group is taken the find the effective diffusion coefficient for the polymer. In the case of copolymers (i.e. >1 type of repeating unit,) the mean is weighted relative to the more frequent unit; e.g. for the p(TrMA/BA) copolymer 50:50 wt.% blend, the ratio of butyl monomer units to trityl monomer units is 2.5625:1.
The limitations of this model are the following:
i) intermolecular forces are neglected;
169
Chapter 5 Mathematical Modelling of NP Diffusion
ii) some of the parameters employed are based on the authors’ own computer simulations of molecular dynamics, including some that are molecule specific;
iii) the penetrants whose diffusivities are calculated in the present study are somewhat larger than recommended for use in this model;
iv) there is no possibility of including the effect of water uptake in the model.
However, the geometric parameters of the model allow for adaptation to the particular scenario of diffusion in a partially aqueous environment, negating iv). The neglecting of intermolecular forces (i.e. i)) is likely to be problematic for highly polar compounds in binders featuring PEG or amine moieties, for example. In many situations, a slight overprediction of diffusion coefficient could be expected. Intermolecular forces could be taken into account in future iterations of the work with some study of components’ solubility parameters and calculation of the strengths of hydrogen bonds and van der Waals forces that must be overcome. In the case of the present work, assessment of the HSPs for each component demonstrated a low potential affinity (see Section 2 and 3). As a result, the neglecting of intermolecular attraction is not likely to result in error for the current molecules. For ii), many of the parameters affected are low sensitivity. In some instances, values were chosen to minimize potential error. For iii), no model describing the migration of large penetrants in glassy polymers was located in the present available literature. However, one of the key reasons for error in the case of large molecules is the increased possibility of deviation from a sphere. Chemicalize simulations (a free online service for calculations of ionisation and geometry of molecular structures; http://www.chemicalize.org/) were carried out for usnic acid and (for reference purposes) juglone molecules (Fig. 5.4). Usnic acid demonstrated significant deviation from a sphere (elongation along the ring-furan-ring axis), although the juglone molecule did not.
170
Chapter 5 Mathematical Modelling of NP Diffusion
Figure 5.4: Chemicalize simulations of molecular geometry for usnic acid (left) and juglone (right).
The chemicalise simulations were also used to validate calculations for usnic acid molecular geometry in the Gray-Weale model later.
5.1.3. Description and modification of the Gray-Weale model
The output of the model is the diffusion coefficient of the selected penetrant (i.e. usnic acid) in the chosen polymer, D (cm2 s-1), as a function of cavity jump length and frequency:
2 (Eq. 5.1)
where
− ( ) (Eq. 5.2)
and
+ 2 (Eq. 5.3)
RS Where kB is the Boltzmann constant, h is the Planck constant, E0 is the critical energy, ri is the radius of the reactant state cavity, T is the temperature (K) and σsg is the Lennard-Jones diameter of the side group. Variation of T allows for inclusion of the elevated temperature
171
Chapter 5 Mathematical Modelling of NP Diffusion conditions of the rotor immersion regime. Q†/Q is the ratio of the partition functions of the penetrant in the Lennard-Jones force field of the polymer, in the deformation of the side groups, and the penetrant’s internal motions, and takes a value between 0 and 1. As these values were obtained by mechanistic simulations that were unavailable, a value of 0.025 was specifically selected for Q†/Q to minimize the possible error. Despite the best efforts, this assumption should be borne in mind as the most likely source of error in the calculations presented here, representing about an order of magnitude at the extreme values (i.e. 0 and 1) of Q†/Q. However, the median value of simulations carried out by the authors in Gray-Weale et al. (1997) was found to be 0.05, which gives a close value to the one selected here. The critical energy E0 is derived from the difference between the minima of the sum of deformation potentials and Lennard-Jones potentials (both kJ mol-1):
( + ) ( + ) (Eq. 5.4)
The geometry of the reactant state (RS) and transition state (TS) cavities are represented by spheres (with four interacting groups) and cylinders (three interacting groups), respectively. The modelling of each group or penetrant entity as spherical unified atoms greatly simplifies calculation. The cavity jump is schematically represented in Fig. 5.3.
Determination of the Lennard-Jones potentials for the interacting side groups and penetrant in the transition and reactant states are given by:
2 4nε [( ) ( ) ] (Eq. 5.5)
Where n is the number of interacting side groups (4 in reactant state, 3 in the transition state), ϵ is the potential well depth, r is the distance between the centre of each side group and that of the penetrant. The Lennard-Jones equation (Eq. 5.5) describes the interactive potential for a given distance between two molecules; the 12th power term encompasses the Pauli repulsion at very short ranges, whereas the 6th power term describes weaker interactive forces at greater distances, tending to 0. Both σ and ϵ are derived from their values for the interacting side group A and penetrant B in the following manner:
√ (Eq. 5.6)
+ (Eq. 5.7) 2
Estimation of well depth ϵ for the penetrant molecule is determined by the following 19:
172
Chapter 5 Mathematical Modelling of NP Diffusion
2 (Eq. 5.8)
Many of the required side group well depth values were estimated within the paper and were used herein. Calculation of the Lennard-Jones diameter, σ, for the penetrant molecule and side groups is crucial for estimation of distances, and is determined in an additive manner of molecular volumes as described in 19, 20:
4 √∑ (Eq. 5.9)
Where the contributions of individual atoms (Å3) are given by the following:
C +8.9 H +2.2 -O (ester) +5.5 -O (OH) +6 =O +4.5 C-C-C-C-C-C (ring) -9
Table 5.1: contribution of atoms to overall Lennard-Jones diameter.
Deformation potentials, i.e. the energy required to deform the side groups by a given distance in order to allow passage of the penetrant molecule, are calculated by the following:
2 2 (Eq. 5.10)
2 (Eq. 5.11)
TS RS Where Ln = σsg, ri is the radius of the transition state cavity, ri is the radius of the reactant state cavity, and µ (shear modulus) is equal to
(Eq. 5.12) 2 +
Where E is Young’s modulus and v is Poisson’s ratio for the polymer. Values for pMMA are given in Tonge and Gilbert (2001).
Deformation potentials and Lennard-Jones potentials were calculated for transition and reactant states and the difference in minima of these functions was determined for
173
Chapter 5 Mathematical Modelling of NP Diffusion
calculation of critical activation energy E0. E0 must be calculated separately for each side group (i.e. individual treatment of the methyl group and acrylic group for pMMA) to determine two diffusion coefficients; the ‘true’ coefficient was obtained by taking the logarithmic mean of the two values.
As a first attempt to simulate the effect of film swelling and plasticisation of the polymer chains, the radii of reactant and transition state cavity terms in the jump length and deformation potentials were allowed to vary with respect to the volume of a sphere and cylinder, respectively, for a given volume increase within the film. This represents the separation of chains and the lower critical energies required to achieve deformation of the side groups. The reactant state or transition state cavity radius at a given volume
RS TS percentage increase are delineated by rɸ and rɸ respectively, where
3( + ) √ (Eq. 5.13)
+ √ (Eq. 5.14)
RS TS Where V0 and V0 are the initial volumes of the reactant state cavities, derived from initial RS TS ri and ri values, and φ is the volume increase of the saturated binder resulting from water swelling expressed as a percentage, and h0 is the initial height of the transition state RS TS cavity. At φ = 0, these values are equivalent to initial ri and ri values. These terms represent the increasing volume of the reactant and transition cavities, and result in a reduction in the deformation potentials and hence overall critical energy. The effect of
RS TS varying rɸ and rɸ to represent swelling of the polymer was tested across a range of swelling values, corresponding to calculated swelling values for the binder systems
RS TS employed. These two terms replace ri and ri respectively in the original equations, namely in the jump length term and deformation potential estimates.
5.2. Determination of diffusion coefficients for usnic acid in polymers
The Lennard-Jones diameter was calculated according to the atom contribution method for the penetrant usnic acid molecule and for each side group of the various binders:
3 V1 = 18 x 8.9 (carbons) (Å )
174
Chapter 5 Mathematical Modelling of NP Diffusion
3 V2 = 16 x 2.2 (hydrogens) (Å )
3 V3 = 3 x 4.5 (double-bonded oxygen) (Å )
3 V4 = 1 x 5.5 (single-bonded oxygen) (Å )
3 V5 = 3 x 6.0 (single-bonded oxygen) (Å )
3 V6 = -9 x 2 (6-membered ring) (Å )
3 ΣVi = 214.4 (Å )
4 √∑ (Eq. 5.9)
Methacrylate side group
3 V1 = 2 x 8.9 (carbons) (Å )
3 V2 = 3 x 2.2 (hydrogens) (Å )
3 V3 = 1 x 4.5 (double bonded oxygen) (Å )
3 V4 = 1 x 5.5 (single bonded oxygen) (Å )
3 ΣVi = 34.4 (Å )
4 2
Methyl side group
3 V1 = 1 x 8.9 (carbons) (Å )
3 V2 = 3 x 2.2 (hydrogens) (Å )
3 ΣVi = 15.5 (Å )
Butylacrylate side group
3 V1 = 5 x 8.9 (carbons) (Å )
3 V2 = 9 x 2.2 (hydrogens) (Å )
175
Chapter 5 Mathematical Modelling of NP Diffusion
3 V3 = 1 x 4.5 (double bonded oxygen) (Å )
3 V4 = 1 x 5.5 (single bonded oxygen) (Å )
ΣVi = 74.3 (Å3)
Triphenyl side group
3 V1 = 20 x 8.9 (carbons) (Å )
3 V2 = 16 x 2.2 (hydrogens) (Å )
3 V3 = 1 x 4.5 (double bonded oxygen) (Å )
3 V4 = 1 x 5.5 (single bonded oxygen) (Å )
3 V6 = -9 x 3 (6-membered ring) (Å )
ΣVi = 196.2 (Å3)
42
Calculation of the large penetrant molecule by the method given in Gilbert and Smith as implemented in 20 gives a value of 8.68 Å. These values are in good agreement with predictions from Chemicalize simulations of molecular geometry. σ values for CH3, CO2CH3, BA and TrMA side groups were calculated at 3.56 Å, 4.72 Å, 6.1 Å and 8.43 Å, respectively.
The well depth ε/kB, as given above, was estimated to be 1050 K for usnic acid, 740 K for TrMA, 500 K for the methacrylate side group, and 70 K for the methyl side group.
Lennard-Jones potentials and deformation potentials were calculated using these values across a range of interaction distances. For example, for calculation of deformation potentials and an interaction distance of 4.87 Å and a swelling value of zero:
2 (Eq. 5.11)
2 2 4 2
4
176
Chapter 5 Mathematical Modelling of NP Diffusion
3 24
2 2 (Eq. 5.10)
2 2 4
2 24
3 4 4
The unusual unit PaÅ3 unit can be converted to joules by the following:
1 Å = 0.1 nm
Å3 = 1 x 10-3 nm3
= 1 x 10-30 m3
1 Pa = 1 N m-2
3 −2 −3 3
3 −3
3 −3
The overall transition and reactant state functions are determined by adding the deformation potential and LJ potentials at each swelling value, and the critical energy for the transition is determined by obtaining the difference in the minima of these functions. The jump frequency and distance, and hence D are obtained from this value as earlier discussed. Representative figures of the Lennard-Jones potentials and deformation potentials for the interaction of usnic acid with CH3 and CO2CH3 in the case of pMMA are given below (Fig. 5.5). The critical energy is found by the difference between the minima of each pair of functions, as described in Eq. 5.4).
177
Chapter 5 Mathematical Modelling of NP Diffusion
Figure 5.5: Lennard-Jones and deformation potentials for CH3 and CO2CH3 side groups in the case of pMMA (0% swelling).
Values for D for every side group present in the investigated polymers were established using the modifed Gray-Weale model19 as a function of various temperature values, corresponding to pontoon and rotor immersion, and swelling values corresponding to measured water uptake values for each polymer. These values are summarised below in Table 5.2.
178
Chapter 5 Mathematical Modelling of NP Diffusion
T / ˚C DCH3 DCO2C4H9 DCO2CH3 DC6H63C2O2 ɸ 1 10 1 10 1 10
10 1.585 x 10-12 5.931 x 10-11 2.641 x 10-31 1.894 x 10-29 1.626 x 10-21 3.298 x 10-43
20 2.992 x 10-12 9.909 x 10-11 2.325 x 10-30 5.503 x 10-28 6.274 x 10-21 7.497 x 10-42
25 4.046 x 10-12 1.264 x 10-10 6.531 x 10-30 1.411 x 10-27 1.191 x 10-20 3.304 x 10-41
Table 5.2: calculated diffusion coefficients for usnic acid relative to each side group type at relevant swelling values and temperature values.
A modified diffusion coefficient can be obtained for usnic acid in pMMA and other binders, based on the variable transition and reactant state geometries in function of swelling, by combining the geometric mean of each side group relative to its abundance in the polymer. For example, estimation of diffusion coefficient by this method yields a value of 3.077 x 10- 17 cm2 s-1 for usnic acid in pMMA at 10 ˚C. For a value of 1% swelling (obtained from gravimetric measurements, see Section 3.2) using variable cavity geometries, we obtain a modified D value of 5.08 x 10-17 cm2 s-1. Full sensitivity analyses of the model to various parameters were carried out in the Tonge and Gilbert paper on the Gray-Weale et al.
19 RS TS model . Given that ri and ri are quite sensitive parameters, it was expected that the introduced φ parameter for swelling would have a significant effect on diffusivity. To carry out a sensitivity assessment, D was calculated across a wide range of feasible swelling values for pMMA (Fig. 5.6). The relationship of D with φ is complex owing to the influence of the latter in more than one area within the equations.
179
Chapter 5 Mathematical Modelling of NP Diffusion
Figure 5.6: Effect of the geometric swelling parameter (expressed as a percentage volume increase) on the calculated diffusion coefficient for pMMA.
It can be observed for low values for φ up to about 7% that changes to overall D are within an order of magnitude of the original D value. At about 11-12% volume increase, diffusivity of FD in pMMA increases by almost two orders of magnitude. Thus for highly water- absorbent polymers, such as the p(TrMA/BA) in this study, rapid release rate might be expected.
The application of the D values determined for each side group to each binder type (where panels were immersed, for later comparison) gives the following overall diffusion coefficients (Table 5.3):
Diffusion Coefficient / cm2 s-1
Pontoon Rotor
pMMA 5.08 x 10-17 2.2 x 10-16 CDP 1.44 x 10-17 - p(TRMA/BA) - 2.47 x 10-19
Table 5.3: Overall calculated diffusion coefficients for each coating type at 10 °C (pontoon immersion) and 25 °C (rotor immersion).
Despite the increase in diffusivity resulting from the introduced swelling parameter, which has a large effect on Dp(TrMA/BA), the predicted diffusion coefficient for the latter is the
180
Chapter 5 Mathematical Modelling of NP Diffusion slowest. On temperature difference alone (neglecting hydrodynamicity), the rotor testing is predicted to result in release rate of usnic acid that is about an order of magnitude faster than than static testing. Because rosin degradation is not accounted for in the present model, the reduced release rate of usnic acid from the Metamare (CDP) coating is accounted for by the presence of butylacrylate as a copolymer (80:20), which requires slightly higher critical energy for deformation. In section 5.5, the agreement or disagreement of these results with assessed immersed coatings will be fully discussed.
5.3. Determination of diffusion coefficients for usnic acid in seawater
An estimate of seawater diffusion of usnic acid was obtained much more simply than the polymer model. It was estimated for 20 oC, using the Wilke-Chang correlation21 for bulk solid diffusion in liquids, using seawater physical parameters derived from the NPL online database22. The equation predicts:
0.6 (Eq. 5.15) Vm
Where D is the diffusion coefficient of the solute in the solvent (cm2 s-1), M is the molecular weight of the solvent (daltons), T is temperature (K), µ is the solvent viscosity (centipoises),
3 -1 Vm is the molar volume of the solute (cm mol ), and s is a corrective parameter describing the capacity of the solvent for self-interaction, and is 2.6 for water. The diffusion coefficient was estimated to be 5.936 x 10-6 cm2 s-1.
5.4. Integration of diffusion coefficients into a simple mathematical model
The obtained values for usnic acid diffusion in a variety of polymer binder types allow for estimation of coating lifetime by integration into a basic diffusion model based on Fick’s second law, which states:
(Eq. 5.16)
Where D is the diffusion coefficient of the diffusion body in the medium (converted to µm2 s-1), z is the position in the medium (µm, where 0 is the interface), t is time (s), and c is the concentration of a chemical the diffusing body at time t. MATLAB code (Appendix A) was created to allow simulation of this equation with varying D, t and zmax values, in order to calculate c for the given parameters.
181
Chapter 5 Mathematical Modelling of NP Diffusion
At a given t value, the profile of c can be established for a range of z values relative to the interface. Likewise, calculation of time taken for the concentration in the coating to fall below a certain threshold value can be achieved. For example, using the established diffusion coefficients for the pMMA at 10 °C (pontoon immersion) and 25 °C (rotor immersion), the relative concentration of usnic acid across multiple timesteps is possible (Fig 5.7 and 5.8), where 1 is the starting concentration, assuming concentration- independent diffusion (i.e. no usnic acid molecule self-interaction).
Figure 5.7: Modelled FD depletion from a static pMMA film 100 µm in thickness, at 10 ˚C.
182
Chapter 5 Mathematical Modelling of NP Diffusion
Figure 5.8: Modelled FD depletion from a static pMMA film 100 µm in thickness, at 25 ˚C.
By performing this kind of modelling work, it is possible to calculate the point at which the effective concentration of additive falls below a certain threshold level. This might be useful for biocides with well-known efficacy, as the effective lifetime of the coating may be limited by that biocide’s level in the binder. However in the present study, the data obtained for each binder were used for comparison with aged samples retrieved from the static and dynamic immersion studies as discussed in the next section.
5.5. Comparison of modelling and experimental data
5.5.1. Determination of ‘release coefficients’ for experimental data
The integration of diffusion coefficients from a physicochemical model into a mathematical model allows us to estimate effective lifespan for a coating. Furthermore, at any given timeframe, the remaining biocide content within the model binder can be determined. In Section 4, fluorescence microscopy was applied to retrieved panels from rotor and pontoon immersion regimes to ascertain the distribution of usnic acid within the binder. Given that
183
Chapter 5 Mathematical Modelling of NP Diffusion the total biocide content is now known in the binder, and the immersion time required to deplete to that level, it is possible to reverse engineer a diffusion coefficient for the experimentally observed data, with some caveats: the diffusion coefficient will not be a true representation of diffusion, because it will also include losses due to hydrodynamic processes, convection etc. For this reason, experimentally derived release terms are hereafter referred to as a ‘release coefficient’ to differentiate them from true diffusion coefficients, as determined by modelling work.
As discussed in the previous section, front- and back-mounted panels demonstrated very different leaching rates with respect to usnic acid (Table 5.4). For the purposes of this study, a separate release coefficient was determined for front and back, and in addition for the average value (i.e. average %FD remaining after t for front panels + average %FD remaining after t for back panels/2). In determination of the release coefficients for rotor samples, it was notable for the pMMA rotor samples that the amount lost at each available time step (27% after 3 months, 35% after 6 months) corresponds very closely to a perfect exponential loss rate; i.e. using either value to calculate the release coefficient gives a very similar value. For the rotor samples of p(TrMA/BA) the 3 month value was used (i.e. 86% lost).
Mathematical Modelling Experimental
Diffusion Coefficient / cm2s-1 Release Coefficient / cm2s-1 Pontoon Pontoon Rotor Rotor Average Front Back
pMMA 5.08 x 10-17 2.2 x 10-16 5.52 x 10-17 1.11 x 10-16 1.9 x 10-17 6.22 x 10-17
CDP 1.44 x 10-17 - 3.25 x 10-17 8.35 x 10-17 4.78 x 10-18 -
p(TRMA/BA) - 2.47 x 10-19 - - - 8.8 x 10-16
Table 5.4: Estimated diffusion coefficients from modelling results, and experimentally derived release coefficients.
5.5.2. Discussion
The modelled data provided an excellent correlation for loss rate for usnic acid from pMMA in the static immersion trials. However, agreement for the estimated diffusion coefficient for 25 ˚C compared poorly with observed samples from rotor trials, however, with over an
184
Chapter 5 Mathematical Modelling of NP Diffusion order of magnitude of discrepancy. Surprisingly, the modelled loss rate was faster than the actual loss rate, despite the former not taking into account the hydrodynamically accelerated aspect of rotor testing. It can be observed that for the pMMA coating samples, the calculated depletion rate of the FD additive in the rotor system was only about 13% greater than statically immersed samples. This is substantially less than would be predicted from the estimated Arrhenius plots of similar materials for the same 15 ˚C shift20, whereby the diffusion coefficient is increased by around an order of magnitude. It is possible that the low swelling of the pMMA, coupled with a lack of erodible components such as rosin or copper oxide, inhibited significant loss of additive by formation of channels and subsequent convection/hydraulic effects, effectively negating the effect of increased ‘service’ speed. In that case, one might expect a significant leaching rate increase in the case of the erodible binder. Unfortunately this was not verifiable, as all CDP samples detached from their primer within 24 hours of immersion.
The calculated average release coefficient for the pontoon-immersed pMMA samples was also influenced greatly by anomalously fast FD depletion from all four front-mounted panels; just taking the rear-mounted (shaded side) panels into account, a release coefficient of 1.9 x 10-17 cm2s-1 is obtained for the pontoon-immersed pMMA. Using this release coefficient corresponds to a biocide depletion rate which is 225% more rapid in the rotor system, which is in line with similar comparisons between static and dynamic ageing regimes23, 24, as well as observations on the increase in copper leaching rate in the previous section, although the correlation with the estimated D value from the model is then less favourable. As previously discussed, a higher release rate of additive was also observed for front-mounted CDP panels compared to the rear ones. The variance of release coefficient, if assumed to be of an exponentially decreasing nature, varied by around an order of magnitude for identical formulations on the front and back of the boards. This highlights the difficulties in obtaining reproducible results for immersion trials, even between samples of a similar formulation, especially between panels on the shaded back face of the boards and those on the front.
The average release coefficient for CDP showed surprisingly good agreement with the calculated diffusion coefficient, despite the fact that rosin degradation was not taken into account in the model. It is possible that the rosin degradation rate has a minimal effect on the release rate of dispersed biocides; the development of wide-open ‘pores’ in the binder matrix, resulting from rosin dissolution, is intended to increase the internal surface area, creating channels for the release of large copper (I) oxide particles. The mechanism by
185
Chapter 5 Mathematical Modelling of NP Diffusion which the biocide is delivered is governed on the nano-scale by its diffusion between polymer chains. In the polymer matrix surrounding pores and copper particles, the same mechanisms are likely to dominate regardless of channels allowing copper (I) oxide transport. This analysis also corroborates the similar FD release rate observed for pMMA and CDP binder, and the fact that observed copper (I) oxide depletion rates were very different.
Estimation of diffusion coefficient for the p(TrMA/BA) copolymer proved to be highly inaccurate. Even allowing for a significant effect of swelling on the cavity geometries, the trityl side group is so large that the critical energy required to laterally deform the side group results in a diffusion coefficient in the order of 10-40 cm2 s-1 or slower (Table 5.2), skewing the overall calculated coefficient considerably even though the group is present in a 1:2.5625 ratio with butylacrylate. The model is probably not representative of such a large side group, which may not be deformed in the usual manner to allow cavity jumps. In fact, the rapid biocide loss observed experimentally may result from a different mechanism, and the conformation of copolymer chains may be highly unusual as a result of the unusually large side group; indeed a helical conformation of the TrMA homopolymer is well documented25, 26. A similar phenomenon in the 50:50 copolymer could explain its surprisingly high water uptake for such hydrophobic material and could provide a more straightforward route for diffusive loss of dissolved biocide by means of the internal helix. This could be established by computational modelling of the copolymer chains and determination of the steric effects or the bulk side groups.
5.6. Summary of modelling work and results
A modification to the original literature model allowed cavity geometries (and corresponding jump length) to change as a function of volume increase of the whole film, using gravimetrically determined water uptake values. This allowed prediction of diffusion of FD in each binder type based on side group calculations.
Calculated diffusion coefficients were integrated to a mathematical model of Fick’s second law, allowing prediction of the FD concentration profile at any given timestep.
186
Chapter 5 Mathematical Modelling of NP Diffusion
The diffusion equation was employed to determine ‘release coefficients’ from experimental immersion panels, encompassing diffusion, convection and all other possible methods of additive loss.
Estimated diffusion coefficients for pMMA demonstrated an excellent agreement with release coefficients.
Diffusion coefficients for CDP showed also showed a good agreement with release coefficients, despite neglecting rosin release in the model. In fact, in light of the results herein, rosin degradation is not suspected to be a key factor in biocide release for dispersed biocides, as they do not rely on the formation of channels for reaction with seawater and significant release.
Diffusion coefficients for p(TrMA/BA) were about two and a half orders of magnitude slower than experimentally derived values. p(TrMA/BA)/FD demonstrated by far the most rapid biocide leaching in immersion trials, but had the slowest estimated diffusion coefficient. The trityl group is probably too abnormal to be accurately represented by the model, and the polymer is likely to have an unusual conformation that invalidates the model that was used herein.
The model overpredicted the increase in biocide depletion resulting from rotor immersion, which was not as significant as expected compared to static immersion.
5.7. References
1. Yebra, D.M., Kiil, S., and Dam-Johansen, K., Mathematical modeling of tin-free chemically active antifouling paint behavior. American Institute of Chemical Engineers' Journal, 2006. 52(5): p. 1926-1940. 2. Kiil, S., Weinell, C.E., Yebra, D.M., Dam-Johansen, K., and Ka M. Ng, R.G., Marine biofouling protection: design of controlled release antifouling paints, in Computer Aided Chemical Engineering, 2007, Elsevier. p. 181-238. 3. Yebra, D.M., Kiil, S., Dam-Johansen, K., and Weinell, C., Reaction rate estimation of controlled-release antifouling paint binders: rosin-based systems. Progress in Organic Coatings, 2005. 53: p. 256-275. 4. Yebra, D.M., Kiil, S., Weinell, C., and Dam-Johansen, K., Supplementary material to 'Analysis of chemically-active antifouling paints by mathematical modelling'. Progress in Organic Coatings, 2006: p. 20.
187
Chapter 5 Mathematical Modelling of NP Diffusion
5. Yebra, D.M., Kiil, S., Weinell, C.E., and Dam-Johansen, K., Dissolution rate measurements of sea water soluble pigments for antifouling paints: ZnO. Progress in Organic Coatings, 2006. 56(4): p. 327-337. 6. Yebra, D.M., Kiil, S., Weinell, C.E., and Dam-Johansen, K., Effects of marine microbial biofilms on the biocide release rate from antifouling paints - a model-based analysis. Progress in Organic Coatings, 2006. 57(1): p. 56-66. 7. Kiil, S., Dam-Johansen, K., Weinell, C.E., Pedersen, M.S. & Codolar, S.A., Estimation of polishing and leaching behaviour of antifouling paints using mathematical modelling: a literature review. Biofouling, 2003. 19 (supplement). p. 37-43. 8. Kiil, S., Dam-Johansen, K., Weinell, C.E., and Pedersen, M.S., Seawater-soluble pigments and their potential use in self-polishing antifouling paints: simulation- based screening tool. Progress in Organic Coatings, 2002. 45(4): p. 423-434. 9. Kiil, S., Weinell, C.E., Pedersen, M.S., and Dam-Johansen, K., Mathematical modelling of a self-polishing antifouling paint exposed to seawater: a parameter study. Chemical Engineering Research and Design, 2002. 80(1): p. 45-52. 10. Kiil, S., Weinell, C.E., Pedersen, M.S., and Dam-Johansen, K., Analysis of self- polishing antifouling paints using rotary experiments and mathematical modeling. Industrial & Engineering Chemistry Research, 2001. 40(18): p. 3906-3920. 11. Budd, P.M., McKeown, N.B., and Fritsch, D., Free volume and intrinsic microporosity in polymers. Journal of Materials Chemistry, 2005. 15: p. 1977-1986. 12. Coughlin, C.S., Mauritz, K.A., and Storey, R.F., A general free volume based theory for the diffusion of large molecules in amorphous polymers above Tg. 3. Theoretical conformational analysis of molecular shape. Macromolecules, 1990. 23(12): p. 3187-3192. 13. Coughlin, C.S., Mauritz, K.A., and Storey, R.F., A general free volume based theory for the diffusion of large molecules in amorphous polymers above Tg. 4. Polymer- penetrant interactions. Macromolecules, 1991. 24(7): p. 1526-1534. 14. Mauritz, K.A. and Storey, R.F., A general free volume based theory for the diffusion of large molecules in amorphous polymers above Tg. 2. Molecular shape dependance. Macromolecules, 1990. 23(7): p. 2033-2038. 15. Mauritz, K.A., Storey, R.F., and George, S.E., A general free volume based theory for the diffusion of large molecules in amorphous polymers above Tg. 1. Application to di-n-alkyl phthalates in PVC. Macromolecules, 1990. 23(2): p. 441-450. 16. Chen, W.-C., Lee, S.-J., and Ho, B.-C., Diffusion coefficients of acrylic monomers in poly(methyl methacrylate). Journal of Polymer Research, 1998. 5(3): p. 187-191. 17. Greenfield, M.L. and Theodorou, D.N., Geometric analysis of diffusion pathways in glassy and melt atactic polypropylene. Macromolecules, 1999. 26: p. 5461-5472. 18. Reis, R.A., Nobrega, R., Oliveira, J.V., and Tavares, F.W., Self- and mutual diffusion coefficient equation for pure fluids, liquid mixtures and polymeric solutions. Chemical Engineering Science, 2005. 60: p. 4581-4592. 19. Gray-Weale, A.A., Henchman, R.H., Gilbert, R.G., Greenfield, M.L., and Theodorou, D.N., Transition-state theory model for the diffusion coefficients of small penetrants in glassy polymers. Macromolecules, 1997. 30(30): p. 7296-7306. 20. Tonge, M.P. and Gilbert, R.G., Testing models for penetrant diffusion in glassy polymers. Polymer, 2001. 42: p. 501-513.
188
Chapter 5 Mathematical Modelling of NP Diffusion
21. Wilke, C.R. and Chang, P., Correlation of diffusion coefficients in dilute solutions. American Institute of Chemical Engineers' Journal, 1955. 1(2): p. 264-270. 22. Bullard, E. National Physical Laboratory. Kaye & Laby: Tables of Physical and Chemical Constants. 2.7.9. Physical Properties of Sea Water. 2011. 23. Yonehara, Y., Yamashita, H., Kawamura, C., and Itoh, K., A new antifouling paint based on a zinc acrylate copolymer. Progress in Organic Coatings, 2001. 42: p. 150- 158. 24. Valkirs, A.O., Seligman, P.F., Haslbeck, E., and Caso, J.S., Measurement of copper release rates from antifouling paint under laboratory and in situ conditions: implications for loading estimation to marine water bodies. Marine Pollution Bulletin, 2003. 46(6): p. 763-779. 25. Merten, C., Barron, L.D., Hecht, L., and Johannessen, C., Determination of the helical screw sense and side-group chirality of a synthetic chiral polymer from Ramen optical activity. Angewandte Chemie International Edition, 2011. 50(42): p. 9973-9976. 26. Merten, C. and Hartwig, A., Structural Examination of Dissolved and Solid Helical Chiral Poly(trityl methacrylate) by VCD Spectroscopy. Macromolecules, 2010. 43(20): p. 8373-8378.
189
Chapter 6 Conclusions
6. Conclusions
The principal aim of this work was to develop a method of analysing the distribution of organic biocides within a binder system. Laser scanning confocal microscopy and epifluorescence microscopy ahave been developed as complementary techniques for the imaging of specific NPs within thin coatings. The capacity of the LSCM to carry out non- destructive, quantitative assessment of additive crystal distribution in non-pigmented media allowed new insights into antifouling performance. Fluorescence microscopy applied to a series of eroded film cross-sections allows visualisation of additive leaching progression, complementing optical microscopy or SEM-EDX assessments of pigment leach fronts. This allows quantitative assessment of the progression of leaching additives, even in pigmented binder systems. These novel applications of LSCM and FM have been employed to assess the distribution of the main biocide of interest in the ACWS project, usnic acid, in a series of three binders; pMMA, a commercial CDP and a novel copolymer, p(TrMA/BA). A complete test matrix of binders was prepared, with and without usnic acid and copper (I) oxide, as a biocidal pigment. The test matrix of panels was immersed in static conditions and accelerated dynamic conditions for 10 and 6 months respectively. These coatings were fully analysed before and after immersion with a full suite of optical and fluorescence microscopy techniques to assess change in copper oxide and FD distribution. Although no antifouling effect was concluded for this specific NP during the immersion study, its inclusion resulted in increased copper leaching from the binder, which was confirmed by the development of thicker leached layers and favourable antifouling performance, with panels on the shaded side deterring fouling for the entire 10 month period. The synchronous depletion of both biocides (FD and copper (I) oxide) was noted on microscopy investigation of the cross-sections. An enhanced leaching rate of copper was noted in the CDP binders bearing both biocides, resulting in the observed antifouling efficacy. Depletion of copper occurred from pMMA and CDP binders during static and dynamic erosion tests; copper oxide leaching from the CDP binders was more rapid, but the binder failed to polish after the first month of rotor immersion. The pMMA binder degradation occurred at the same rate as the copper oxide depletion in a linear manner, resulting in the lack of a leached layer. Some antifouling efficacy was observed for pMMA panels during static immersion as a result of this copper depletion.
191
Chapter 6 Conclusions
NP distribution was visualised in-binder using fluorescence microscopy, demonstrating leaching of FD from the binders during immersion. The pMMA/FD coating slost 27% after 3 months and 35% after 6 months in rotor tests, and (on average) 43% over the 10 month static immersion period. The CDP/FD coatings lost (on average) 33% of the FD over the 10 month immersion period. Despite the much faster copper depletion noted for CDP, the leach rate of FD was similar or even slower for this binder; it was concluded that rosin degradation has a negligible effect for the release of dispersed biocide, as the mechanism by which it increases copper (I) oxide dissolution rate (i.e. generation of channels and pores) is not important on a molecular scale of diffusion. Another significant observation was the total inhibition of FD leaching below the pigment front, demonstrating that pigment depletion is the limiting factor in these combinatorial coatings. This confirms other authors’ suggestions that this may be the case as a result of the large size and hydrophobicity of the pigment. The p(TrMA/BA)/FD coatings demonstrated the most rapid leaching rate, with 86% depletion after only 3 months of rotor immersion. This binder also failed to polish, although FTIR confirmed that some hydrolysis had taken place. This is attributed to the high occurrence of non-saponifiable butylacrylate groups.
An existing model for transition state theory-based estimations of diffusion coefficient for small penetrants in glassy polymeric systems was adapted by adding a swelling term to make existing parameters for cavity geometry variable as a function of film volume increase, as calculated from values for water uptake in each binder (1% for pMMA and CDP, 10% for p(TrMA/BA)). This modified model was applied to determine diffusion coefficients at the operation temperature for static and dynamic immersions, in order to determine the accuracy of predictions using such a model. The diffusion coefficients were integrated into a mathematical MATLAB model to simulate the depletion of FD from each binder type over time. Estimations of ‘release coefficients’ were made by using the fluorescence profiles for eroded coatings to reverse-engineer diffusion coefficient values that also include losses due to convection and other mechanisms. Agreement with pMMA leach rates was very good, and was also good for CDP binders. As the model did not take rosin degradation into account, the close agreement is further proof that rosin degradation is not a key factor in biocide diffusion. Agreement of the model with experimental results for FD leaching from p(TrMA/BA)/FD was very poor, with about 2.5 orders of magnitude of discrepancy. The abnormal size of the trityl side group may invalidate the use of the model in this situation, and furthermore, the conformation of the polymer is likely to be markedly different from the other binders because of the known inclination of the bulky side group to form unusual
192
Chapter 6 Conclusions structures. The formation of helical chains would also serve to explain the very high water uptake of this polymer, which was not predicted based on the hydrophobicity of the film. The mechanism of rapid diffusion may have relied on unusual chain conformation in this binder. In light of these results, water uptake appears to be the dominant mechanism governing biocide diffusion in this binder.
6.1 Further work
A continuation of work should focus primarily on use of a smaller, more compact biocide with known efficacy, possibly using a commercial biocide such as DCOIT. Simultaneous modelling should be carried out, but could be strengthened using use of atomistic simulations for more accurate determination of initial cavity geometry in the same manner as carried out in the original paper, as well as for Q†/Q, which was not able to be estimated in this study, instead having a value selected to minimize the discrepancy based on the original work. Using a matrix of acrylic monomers with a variety of side-lengths, i.e. from methyl and butyl up to stearyl methacrylate, would allow for a more thorough investigation of the effects of water uptake on diffusion rates. In addition to pontoon studies, methods of carrying out laboratory exposure-based diffusion tests could be investigated. UV-HPLC would be a possible method of measuring concentrations in solution in such an experiment, supporting the use of fluorescence microscopy for analysis of cross-sections.
In addition, rounding out the model by incorporating copper oxide presence in the binder, as well as the dissolution processes, would increase the applicability of the model to commercial products, which are unlikely to feature a lone biocide with no pigmentation. This could be achieved by incorporating the mechanisms of copper oxide dissolution from first principles, or by modelling the movement of the copper leach front as determined from experimental data.
Lastly, quantification of intermolecular interaction (biocide-biocide and biocide-polymer) could be an important addition to the model. This would lead to some degree of concentration-dependent diffusion in the case of the former, and an increased accuracy of the estimated diffusion for the latter, as the sum of dipole forces and hydrogen bonding forces, etc., would need to be taken into account in estimation of critical energy for the
193
Chapter 6 Conclusions cavity jump. This might be achieved by means of using HSP values or atomistic simulations of molecular interaction and energies.
194
Appendix A: Diffusion Code for MATLAB
% Euler's implicit scheme for solving the diffusion equation % dC/dt = D * d^2C/dz^2 where D is the diffusion coefficient,and C(t) is the % concentration of a chemical at time t. % % Use dC/dt = C(t+dt) - C(t) / dt % d^2C/dz^2 = C(t+dt) - 2 *C(t) + C(t-dt) / dt^2
% Begin by defining a vector C which has entries giving the concentration % of C at each point along the z-axis. %% xx=100; C=zeros(xx+1,1); nz=length(C); dz=1/nz; C(1)=1; % lhs bc Cnew=C; %% E=*enter diffusion coefficient here*; % E = input('Enter the diffusion coefficient in cm^2/s:'), inputting from model results or experimentally derived results; D=E*1e8; % Diffusion coefficient in um^2/s
% Define the time steps for the problem dt=100; % time-step t=*enter time here*; % total time to integrate for, in s steps=t/dt; % number of time-steps needed
% The Courant number is a measure of stability, need D * dt / dz^2 < 1
% Integrate the problem forwards in time, and print out the temperature
195
ic=0; for i=1:steps %intermediate %% for n=2:xx Cnew(n) = (C(n) + D*dt*(C(n+1) + C(n-1) - 2*C(n))/(dz^2)); end %% %rhs bc Cnew(xx+1)=C(xx);
C=Cnew;
ic=ic+1; if rem(ic,100) == 0 fprintf('Value for C at z=1 is %2.2f at time- step %6i\n', C(101), i); else end end
C*enter time here*UA=Cnew; %create time-stamped results file
Diff=1-Cnew;
% Plots the results whilst retaining previous graphs for comparison hold on z=0:100/xx:100; plot(z,Diff) axis(0 100 0 1) title('Concentration for D = ',num2str(E),' cm^2/s after t = ', num2str(t), ' seconds.') xlabel('z um') ylabel('Concentration')
196