Imidazole and Triazole Coordination Chemistry for Antifouling Coatings

Chalmers Publication Library (CPL) offers the possibility of retrieving research publications produced at Chalmers University of Technology. It covers all types of publications: articles, dissertations, licentiate theses, masters theses, conference papers, reports etc. Since 2006 it is the official tool for Chalmers official publication statistics. To ensure that Chalmers research results are disseminated as widely as possible, an Open Access Policy has been adopted. The CPL service is administrated and maintained by Chalmers Library.


Introduction
A major problem for the shipping industry is fouling of marine organisms, such as algae, mussels, and barnacles on ship hulls [1][2][3][4][5]. As the ship moves through the water, the hull is subjected to form drag and skin friction drag. Since fouling increases the surface roughness, the skin friction drag is increased, which subsequently results in increased fuel consumption and reduced manoeuvrability [3]. Traditional countermeasures against marine fouling are coatings containing antifouling toxicants that have had severe environmental consequences on marine ecosystems. The most wellknown examples are organometallic paint systems (comprising lead, arsenic, mercury, etc.) which met the market in the early 1960s. The notorious tributyltin self-polishing coating (TBT-SPC) was developed in the 1970s [1,6]. The reason for the success of the TBT-SPC paints is that the rate of release can be controlled on a molecular level. Generally, these paints are based on acrylic or methacrylic copolymers. The active substance (TBT) is covalently attached to the carboxylic group of the polymer, which is hydrolytically instable in slightly alkaline conditions (as in seawater) [1,4,6]. Toxic copper pigment is also included in the TBT-SPC paint systems, which in combination with TBT, provides protection against the entire spectrum of fouling organisms since some algae are tolerant to TBT but not to copper and vice versa [7]. As the use of TBT containing paints increased, marine life in the vicinity of harbours and marinas was severely affected [8,9]. The hazardous nature of TBT for the marine environment is ascribed to its weakening of fish immune systems, causing imposex and sterility in female gastropods (at such low concentration as 1 ng/L) as well as its bioaccumulative properties in mammals [1,4,6]. Starting with TBT ban for pleasure craft in France 1982, TBT-containing paint systems were successively phased out in the 1990s 2 Journal of Chemistry and finally banned by IMO (the International Maritime Organization) [10,11].
A number of tin-free alternatives have been developed due to the restrictive legislation regarding TBT paints. These can crudely be divided into biocide-release and biocide-free antifouling systems. Biocide-free approaches are dominated by paint systems called foul-release coatings. The antifouling mechanism lies in the low surface energy and smoothness of the coating. The binders in these coatings are silicone polymers, for example, PDMS (polydimethylsiloxane) and to some extent fluoropolymers, with both possessing very low surface energy.
In tin-free SPC paints, tin is replaced by other elements such as copper, zinc, and silicon. Even though the side group of a few tin-free SPC paint binders also possess biocidal properties, no alternative is nearly as effective as TBT [4,6]. To provide full and long-term fouling protection, it is now necessary to add other biocides, so called boosters, generally in combination with copper. The most frequently employed boosters are conventional agrochemicals, some of which, for example, irgarol, have been under scrutiny due to low rate of biodegradation as well as toxicity towards nontarget organisms [1,4,6,[10][11][12][13][14][15][16][17][18]. Since also the use of copper has been under debate recently, a number of other pigments are used such as Zn(II), Fe(III), and Ti(IV) oxides [1,3,4,6].
A generic problem with tin-free SPC paints is premature leakage of the booster biocides [3]. These are not anchored to the binder, as TBT is, and are hence subjected to free diffusion within the paint matrix. The immediate consequence of premature leakage is loss of antifouling effect prior to the hydrolysis-dependent lifetime of the coating.
A lot of research and development has been conducted on SPC paints from an environmental perspective. For instance, the antifouling substances secreted by marine organism, that is, natural biocides, have recently attracted attention from the scientific community. The antifouling mechanisms of these natural biocides are generally much more intricate and specific compared to those of conventional booster biocides [3,4,6]. Some examples of natural antifoulants are worth mentioning. Furanon, originating from red seaweed, is for instance 100 times more effective than TBT as antifoulant against barnacle [3]. Bufalin (a toad poison) is for example, considered to be 100 times more toxic than TBT and 6000 times more efficacious against barnacles [6]. Most substances are found in the phyla Porifera (sponges), Algae, Cnidaria (e.g., corals, sea anemones, and hydroids), Echinodermata (e.g., sea urchins), Bryozoa (moss animals), and bacteria [3,6]. Many antimicrobial substances are based on five-membered aromatic heterocycles such as Seanine and OIT (see Scheme 1). Moreover, triazole-based biocides represent a wide spectrum of in particular antifungals such as fluconazole and itraconazole (see Scheme 1) [19][20][21][22][23][24][25][26][27]. The imidazole fungicide clotrimazole (see Scheme 1) used as human and veterinary pharmaceutical has been found in the marine environment at concentrations affecting sterol metabolism in marine microalgal communities [28]. The interdisciplinary research program Marine Paint [2], in which our group has been a part, has since 2001 been investigating the substance medetomidine [29] (see Scheme 1) which is based on imidazole for antifouling purposes [30][31][32]. Medetomidine is exceptionally efficient against fouling barnacles [33], which is one of the most problematic foulants in the north Atlantic, yet, without seriously affecting nontarget organisms [34][35][36][37].

Antifoulant Immobilization.
Booster biocides, including the azole-based molecules already mentioned, are typically small molecules which are molecularly dispersed in the antifouling paint [3,51]. As a consequence, the biocides leach out from the coating within a couple of weeks (diffusive release from thin coatings ∼100 m) [52]. This results in a premature loss of fouling protection which is a severe problem since the desired lifetime of a marine coating is ∼5 years [3,4]. However, when the biocides are immobilized on macromolecules or colloids (with negligible diffusion coefficients given there large sizes), the effective diffusion coefficient will be reduced by the magnitude of the stability constant for immobilization [53]. We have explored the possibility to immobilize the very efficient antifoulant medetomidine on polymers and nanoparticles by exploiting the strong and specific coordination chemistry of the molecule's imidazole moiety. This is treated in Section 5.

Triggered
Release and Imidazole-Cu 2+ Coordination. The approach discussed in the previous section is obviously only applicable for azole-based biocides/antifoulants. Most booster biocides do not contain a strong ligand moiety for coordination chemistry. We have therefore developed more generic release systems based on polymeric poly(methyl methacrylate) microcapsules [52][53][54][55][56][57][58][59][60] with some novel advances involving ultrathin polyelectrolyte multilayers for sustained release [53,[60][61][62][63][64][65][66]. However, in order to maintain protection against fouling, it is important that the biocide concentration reaches high enough concentrations at the coating surface (see Scheme 2). Consequently, it would be beneficial if the release of the antifouling substance could be triggered by some physical or chemical parameters at the hull-water interface. We have explored the possibility to achieve triggered release by incorporating salt in the otherwise rather hydrophobic microcapsule shell using azole coordination chemistry. The salt will initiate swelling of the shell by increasing the free volume of the polymer [ Poly(1-vinylimidazole-co-methylmethacrylate)

Monomeric imidazoles and five-membered nitrogen heterocyclic antifouling biocides
] ] ] ] Scheme 1: Molecular structure, names, and abbreviations of polymers used for azole coordination and low molecular weight imidazole containing substances and five-membered nitrogen heterocyclic biocides investigated in this review. Potential nitrogen coordinating borderline ligands (imidazole, triazole) or hard ligands (amines), oxygen coordinating hard ligands, and sulphur coordinating soft ligands are marked with blue, purple, red, and yellow, respectively. Weak ligands such as organic halides have not been marked.
trigger the release from the microcapsule shell as a function of contact with water or moisture (see Scheme 2). The functionalization of polymers with azoles (mainly imidazole) is treated in Section 3, and the material properties and water sorption with respect to the presence of metal ions (in particular Cu 2+ and Zn 2+ ) in the polymer matrix are treated in Section 4.
The azole coordination has also been used to trigger the release of medetomidine. This is treated in Section 5.1.

Towards Cu 2+ /Cu + Catalysis in Novel Biocide-Free Marine
Coatings. Copper is a potent antifouling biocide for in particular marine organisms [67,68]. Yet, copper is also a trace element and constitutes an essential part of many enzymes where redox reactions are the central function. However, excess copper concentrations are lethal for the organism since vital biological processes (disruption of cell homeostasis such as pH balance, membrane potential, and osmosis) are affected via the inactivation of enzymes and by the precipitation cytoplasmic proteins into metal-protein aggregates [69,70]. Many studies are pointing towards the fact that copper crosses the biological membranes as Cu + which is subsequently the biologically most active oxidation state [69]. However, Cu + is unstable in an aqueous environment and disproportionates to Cu 2+ and Cu 0 [71,72]. Moreover, only free ions of Cu 2+ and Cu + are considered to be bioavailable [73][74][75]. A lot of recent focus in the catalysis and microbiology community is addressing Cu + stabilization by altering the Cu 2+ /Cu + redox potential via the proper choice of ligand environment, both with respect to chemistry and geometry (see Section 2) [76][77][78][79]. Here, the ligand environment in copper proteins, including both the Cu(I) [77][78][79] and the Cu(II) [79][80][81][82][83][84][85][86][87][88][89][90][91][92][93][94] oxidation statse, is a continuous source of inspiration.
Encouraged by our findings concerning azole-Cu 2+ coordination mentioned in the previous Section and recent advances in Cu 2+ /Cu + stabilization, we have developed a generic concept of biocide-free antifouling coatings. By incorporating azole-type ligands in the antifouling coating, the pool of free and sequestered Cu 2+ in the marine environment may be absorbed. It is important that the interaction is strong enough since a large fraction of the dissolved copper in the sea is associated with organic ligands. Typically, weak ligands are considered to have stability constants of the order 10 9 and strong ligands of the order 10 13 [73][74][75]. However, this concept is still feasible given the fact that azole-type ligands have stability constants for copper coordination ranging between 10 14 -10 20 depending on chelate properties and chemistry as discussed in Sections 2 and 6 and the references within. Furthermore, regarding the antifouling coating, the presence of strongly Cu + stabilizing ligands, for example, nitriles and sulphur-containing ligands, may catalyse the reduction of Cu 2+ to Cu + . This will produce high concentrations of Cu 2+ and Cu + , both of which are efficient antifouling agents Journal of Chemistry 5 [67,68], at the coating interface without any net addition of biocides to the environment (see Scheme 3). The azole specificity towards Cu 2+ is treated in Section 4.1, and the Cu + stabilization including our on-going work with the triazole ligand is briefly discussed in Section 6.2.

Theoretical Considerations of Imidazole and Triazole Coordination Chemistry
In this work, the coordination of azole ligands (mainly imidazole) by Cu 2+ , Cu + , and Zn 2+ ions has been investigated as presented in Scheme 4. It is therefore pertinent to describe the fundamentals for the strong interaction between the previously mentioned species. In the following subsections, the azole coordination chemistry will be briefly reviewed in terms of coordination geometry, molecular orbitals and electronic structure, covalency, aromaticity, and so forth.

HSAB Principles and -Interaction.
The strong metalligand bonding between the metal ions Zn 2+ , Cu + (diamagnetic d 10 full d-shell configuration), Cu 2+ (paramagnetic d 9 configuration), and the imidazole or triazole ligand (see Scheme 5) may be qualitatively explained by HSAB (Hard Soft Acid Base) principles [95,98,99]. To begin with, only the electron lone pair donation from the base (ligand) to the acid (metal), -bonding, will be considered. Hard bases preferentially bond to hard acids via ionic bonding, whereas soft bases favourably bond to soft acids via covalent bonding [98,99]. In the ligand field depiction (see Scheme 7) this implies that the energy levels of the metal (M) and ligand (L) orbitals (the d -and -levels) will be similar [100,101]. The ensuing M-L molecular orbitals will have almost equivalent proportions of metal and ligand character, respectively, and the electrons spend their time at the metal and ligand equally [100]. Ionic bonding denotes large orbital energy differences with essentially no electron sharing [98][99][100].
Hard acids and bases are small, hard to polarize with low and high electronegativity, respectively [98,99]. They possess large HOMO-LUMO (Highest Occupied Molecular Orbital, Lowest Unoccupied Molecular Orbital) gaps, favouring ionic bonding [100]. Soft acids and bases are large and polarizable with intermediate electronegativity and with small HOMO-LUMO gaps [98,99]. Note that Cu + is considered a soft acid and that triazole is softer than imidazole. A high oxidation state usually renders the transition metal hard, whereas the soft has low oxidation state (≤2) [100]. With regard to bases, unsaturation usually makes them softer [99]. For regulardonors, the coordinate bond may be viewed as polar covalent, especially with regards to borderline acids and bases (see below) [99,100]. The selectivity of different metal-ligand pairs in polymers is treated in Section 4.1.

-Interaction.
-bonding is another important interaction (besides -bonding) in transition metal complexes (see Schemes 5 and 7) [100]. Ligands with more than one lone pair, labelled -donors or -bases (see Schemes 5 and 7), for example, Cl − and RO − , may donate additional electron density to a metal d orbital ( -acid) of proper symmetry (see Scheme 7, -bond) [100]. This interaction is typically considered hard and destabilizes the metal d-orbitals but will be favourable for hard, electropositive (especially d 0 ) transition metals [100]. Certain ligands, labelled -acceptor or -acid have low-lying vacant LUMOs of proper symmetry with regard to the metal d -orbitals ( -base) susceptible for metal to ligand -bonding (back donation), an interaction typically considered soft. This interaction may be favourable since it alleviates the metal of some electron density in the coordination compound. For unsaturated ligands, the antibonding * -orbital typically accepts electron from the metal d -orbitals, which is favourable for electron-rich metals in low oxidation states with high energy d-orbitals that are basic in character [100].

Coordination Geometry.
Imidazole is coordinated by Cu 2+ in an essentially square planar geometry Cu(Im) 4 2+ ⋅ X − 2 , similar to pyridine (see Schemes 6(a) and 6(c)), with the imidazole rings occupies coplanar positions with respect to Cu(II) [95]. This favours -interaction between the imidazole ring orbitals and metal d orbitals and locates the electron hole in a distinctly separated orbital (see Scheme 6(a)) [79,95]. In the case of Zn 2+ or Cu + , tetrahedral geometry is frequently encountered (common for d 10 metals [101]), ). Octahedral geometry (see Scheme 5(c)), Zn(Im) 6 2+ ⋅ X − 2 , is also possible for Zn(II) but not for Cu(I). The coordination is less dependent on electronic factors for Zn(II) and more on steric factors in contrast to Cu 2+ complexes [95,101]. As an example, Zn 2+ with Cl − as counterion coordinates unsubstituted imidazole in octahedral geometry of ℎ symmetry [Zn(Im) 6 ] 2+ (see Section 6.1) [38]. However, the addition of a methyl group at position 1 or 4 on the imidazole ring, which gives a slight steric hindrance, results in the formation of pseudotetrahedral complexes [Zn(Me-Im) 2 Cl 2 ] of 2V symmetry, even at large excess of imidazole ligands [38]. It should be noted that the nature of the counterion has an effect on the type of complex formed and that octahedral ( ℎ symmetry) and tetrahedral ( symmetry) complexes of Zn 2+ and methylated imidazoles have been reported [95].
The coordination geometries for triazole ligands are more complicated. Since triazole is bidentate, with the possibility to bind to two metal centers, the coordination usually results in polymeric (1D, 2D or 3D) structures [102][103][104][105]. This is a popular method to produce metal-organic frameworks which is also possible for imidazole ligands at sufficiently high pH [38,95]. However, this is outside the scope of this review and will not be further treated. Mononuclear coordinate complexes comprising triazole ligands do exist provided that the triazole is substituted with sufficiently bulky groups. In these situations, the triazole displays the same type of coordination geometries as the imidazole ligand [102][103][104] (see Scheme 6).  Scheme 4: Structure and numbering of the imidazole 1,2,4-triazole and 1,2,3-triazole. N 1 is labelled pyrrole nitrogen and N 3 (imidazole) and N 2 /N 4 (triazole) are labelled pyridine nitrogen. Note that N 1 is also sp 2 hybridized and the free electron pair is donated to the ring -orbital system in order to obtain aromaticity. The lone electron pair of N 3 is, however, free rendering of this nitrogen basic and the predominant site for coordination [95].

Comparison with Other Nitrogen
Consequently, these borderline acid-base pairs form favourable bonds with bond properties which are intermediate to those of soft and hard acid-base pairs [98][99][100]. Imidazole is generally considered to be a moderate -donor and a weak -acceptor, with -donor and -acceptor abilities in between saturated amines such as ammonia (NH 3 ) and unsaturated amines such as pyridine [95]. This positions imidazole above oxygen donors and below ammonia and pyridine in the spectrochemical series [95,106]. In rare cases, -donor abilities from the imidazole-ring -orbital have been reported [107,108].
The series of nitrogen donor ligands with increasing basicity follows the order NH 3 > imidazole > pyridine, whereas the order of increasing -acceptor ability is the reversed [95]. For Cu(II) and Zn(II), the bond strength follows the order imidazole > NH 3 > pyridine, reflecting a dominating importance of -bonding but also a significant contribution from -bonding [95].

Functionalization of Polymers
For the antifouling coatings containing azole ligands as described in Sections 1.2 and 1.3, it is necessary to develop suitable synthetic routes in order to functionalize polymers with azole ligands. All polymers and low molecular weight imidazole containing substances used in this work are presented in Scheme 1. We have investigated both N-and Csubstituted azoles. The attachment position of the polymer backbone (the substituent) on the azole ring has important consequences for the coordination chemistry as described in Section 6.1.

Imidazole Functionalization.
A popular technique to functionalize polymers is to graft functional groups onto an active site on the polymer backbone [111][112][113][114][115][116]. It is often easier to polymerize a smaller monomer with an active site than a large bulky monomer with the functional group already attached. This type of functionalization is sometimes called postpolymerization modification [115,116].
An epoxide route was used in order to functionalize ethyl(hydroxyethyl)cellulose (EHEC) with imidazole groups [49]. In short, the procedure involved 4 steps: protection of Scheme 5: Coordinate bonding modes (left) and molecular orbitals (right) envisaged for azole-Cu(II) in square planar geometry as obtained from DFT calculations [38]. The bonding principles are generic and applied also for Zn(II) and Cu(I) complexes.   [96]. (a) With regard to cupric complexes, there is possibility for weakly bound counterions or coordinating solvent molecules in axial position and the geometry is then labelled tetragonal (distorted octahedral symmetry). However, the ligand field splitting will be similar in tetragonal and square planar symmetry if one accounts for the important 4s-3dz 2 mixing, which effectively lowers the metal dz 2 orbital in both symmetries [97]. (b) The ligand field splitting in tetrahedral complexes is the inverse of an octahedral one but with some mixing of the d and p-orbitals. The situation with -interaction is, however, more complicated since the -interacting ligand orbitals may form bothand -bonds, and the effect on the ligand field splitting will be a reflection of both bonding modes [96,97]. the pyrrole nitrogen, reduction of the aldehyde to epoxide, epoxide opening and attachment to the polymer backbone, and finally removal of the protective group (see Scheme 8). The modified polymer and its interaction with Cu 2+ are evaluated in Section 4.2.2.
We have also developed a synthesis procedure to attach the biological amino acid histidine to a polymer backbone as well as to a polymerizable monomer using maleimide chemistry [48]. The synthetic routes comprised both grafting onto by maleimide bond formation on a maleic anhydride polymer (see Scheme 8) and conventional radical polymerization (see Scheme 8). The successful grafting of histidine onto P(E-alt-MAh) (poly(ethylene-alt-maleicanhydride)) was confirmed by the formation of the maleimide bond. The DS (degree of Im-EHEC (see Figure 11)

4-step grafting with protective group
Scheme 8: Imidazole functionalization of polymers and monomers [48,49] and monomers used for AIBN initiated radical polymerization [39,47,48]. substitution) ranging from 80-90% was estimated from NMR spectra and elemental analysis. In conclusion, imide bond formation between amine and anhydride is an efficient way of grafting histidine onto a polymer backbone [48]. Radical or anionic polymerization of maleimide monomers is an established synthesis procedure, thoroughly investigated [120][121][122][123][124][125][126][127][128]. Since the maleimide group is strongly electron withdrawing, effectively reducing the electron density at the maleic double bond, it is custom to copolymerize it with for example, styrene or other monomers containing electron donating groups. Such a copolymerization results in rather alternating copolymers [120]. This copolymerization technique is also common for polymerization of maleic anhydride, which is not possible to homopolymerize [120]. However, it is still possible to homopolymerize maleimides, even though the reaction rate is very slow [125]. The reaction rate in this case is additionally decreased since the histidine side group renders the monomer large and bulky [122,129]. Both routes were successful, but the maleimide route facilitates both grafting onto and radical polymerization. Furthermore, the maleimide route is conveniently carried out in one step without the necessity of protective groups [48].
However, triazole synthesis is most synonymous with the copper catalyzed alkyne-azide huisgen cycloaddition (CuAAC) reaction [117] (see Scheme 9). This reaction is a well-known example of a class of chemical reactions called "click reactions" defined for the first time by Kolb et al. [133]. Click reactions are characterized by high quantitative yields and tolerance towards the use of functional groups. The reactions are also relatively insensitive to the solvent system. Consequently, these characteristics make click reactions a very good tool for polymer synthesis and modifications.
It is obvious that the CuAAC reaction can be used for triazole functionalization of azide or alkyne containing polymers which has been investigated by us [50]. However, the CuAAC reaction can also be used to synthesize triazole-based vinyl monomers. Thibault synthesized vinyl containing triazole monomers (R 1 = vinyl in Scheme 9) which were polymerized via reversible addition chain-transfer (RAFT) reaction [134]. Using this methodology, a variety of monodisperse polymers carrying different functional group were synthesized [134].

Properties of Coordinated Polymers
The coordination of polymers carrying ligand sites by metal ions has a profound effect on the polymeric material [135][136][137][138][139], which includes properties and functions such as optical, electrical, rheological, molecular self-assembly, and solubility properties [137][138][139][140][141]. The basis for the majority of properties and functions relies on the specificity and strength of coordination for a given set of metal ion and ligand. Therefore, regarding the use of these materials in antifouling coatings (triggered release or Cu 2+ sorption/redox reaction), it is important to understand the details of the coordinate interactions on both a microscopic and a macroscopic levels.
In our group, the coordinate interaction of the materials has been characterized using EPR, vibrational spectroscopy (far-FTIR and mid-FTIR), ab initio, and DFT (Density Functional Theory) calculations. The material properties have been characterized using mainly DSC (differential scanning calorimetry) and vibrational spectroscopy.

Bond Strength and Selectivity.
Regarding the PVM copolymers discussed in Section 3.1, the polymer contains two possible metal-coordinating ligands: the imidazole and the ester groups (see Scheme 1). The carbonyl group is regarded as a hard ligand with the oxygen atom as the ligating group, although with a theoretical possibility of acting as a soft -bond donor [100]. However, the metal ions exclusively coordinate the imidazole ligand of the polymer [47]. This is evident since only imidazole bands in FTIR are altered upon coordination (see Figure 1) [39,47]. Regarding copper complexes, EPR affirms this fact since complexation of PMMA results in no signal (see Figure 2(c)). The EPR results concern only Cu 2+ complexes. Zn 2+ is diamagnetic and hence EPR silent. Furthermore, the environment appears to be square planar (or tetragonal) due to the resemblance between the EPR spectra of square planar [Cu(Me(1)-Im) 4 ] 2+ and PVM-4-Cu 2+ . This result is quite spectacular since the enforced coordination symmetry results in a high entropic penalty for the polymer chain. It should be noted that the EPR parameters and spectrum of PVM-4-Cu 2+ are especially similar to imidazole methylated at position 1 (Me(1)-Im) coordinated by Cu 2+ (see Figure 2) [38,47].

Coordinate Crosslinks.
The coordinate bonds between the imidazole ligand and the metal ions generate crosslinks between the polymer chains [47]. The crosslinks cause the glass transition temperature, , to increase since the chain segments in proximity of a crosslink are stiffened due to the reduced rotational freedom (see Figure 3) [47,145]. The stiffness depends on the strength of the crosslink. Hence, the increase of will depend on both the number of effective crosslinks as well as the strength of the crosslinking bond [47,135,136,146]. Regarding the polymers, the increase of is more pronounced for PVM-44 than PVM-4 due to the larger number of possible crosslinks. However, when the effect of the particular metal ion is inspected, it is observed that Cu 2+ raises more than Zn 2+ for PVM-4 while the opposite holds for PVM-44. This may appear contradictory but is explained by coordination bond strength and coordination number. The bond strength between imidazole and Cu 2+ which is stronger than the bond strength between imidazole and Zn 2+ which is most important for PVM-4 with a low number of imidazole ligands. However, the total number of possible To be more precise, the coordination complexes are separated by less than 20Å (see Figure 2(b2)) which is indicated by the EPR half field signal (see Figure 2(b2)). crosslinks may be higher for Zn 2+ complexes since this metal allows for octahedral geometry (see Section 2.3), whereas Cu 2+ is restricted to square planar geometry. This explains why PVM-44 with a high number of imidazole ligands coordinated to Zn 2+ acquires a higher than PVM-44 coordinated to Cu 2+ despite a weaker coordinate bond.

Hygroscopicity and Metal
Coordination. The hygroscopicity of a PVM polymer containing a metal ion salt is highly dependent on the fraction of imidazole containing monomers. In addition, the water molecules can crosslink two imidazole groups of a PVM polymer in a similar manner as the metal cation. The hydrogen-bond interaction is too weak to alter the thermal properties of the polymer, yet, the change of the polarization of the water molecules is evident by the blue shift of the (HOH) water bending vibration (see Figure 4). However, when metal ions are introduced, this population of bridged water molecules disappears and is replaced by bridging metal ions (see Figure 4) [39,47]. For PVM-4, with a low fraction of imidazole ligands, the introduction of metal ions results in a significant increase of solitary water as indicated by the area of the V(OH) water stretching vibrations [39]. However, for PVM-44 with a high fraction of imidazole ligands, the metal ion coordination results in a decrease of the amount of sorbed water [39]. Note that this polymer is already rather polar due to the high fraction of N-VIm monomers. Consequently, the triggered release effect discussed in Section 1.2 is more applicable for hydrophobic polymers which are functionalized with crosslinking metal ions.

Coordination-Mediated Multilayer
Assembly. The strong coordinate cross-links discussed previously can be used to assemble polymer multilayers. Polymer multilayers are most synonymous with the polyelectrolyte layer-by-layer technique which has received a lot of attention during the last decade [64]. However, coordination chemistry is a feasible tool to assemble multilayers with anisotropic properties as provided by the restricted coordination geometry of a given metal ion. Hedin et al. studied the bilayer formation of the imidazole functionalized biopolymer Im-EHEC (see Scheme 1) using QCM-D [49]. Although Im-EHEC was functionalized with only 1% imidazole moieties per monomer unit, the introduction of copper resulted in a slight crosslinking of the adsorbed polymer layer, and the subsequent addition of Im-EHEC solution gave an additional adsorbed monolayer of Im-EHEC (see Figure 5).

Antifoulant Immobilization and Controlled Release
In the following subsections, ways to immobilize and control the release of antifouling agents using coordinate interactions, with a special focus on medetomidine, will be presented. The immobilization of medetomidine has been evaluated using HPLC, NMR, and vibrational spectroscopy as well as monitored in situ in model systems using QCM-D and SPR. The controlled release has been monitored from real coatings using liquid scintillation spectrometry ( 14 C − probed release substances) as well as from model surfaces using the surface sensitive techniques QCM-D and SPR.

Polymer Coordination.
Triazole derivatives represent a wide class of antimicrobial substances (see Scheme 1) [19][20][21][22][23][24][25][26][27] as mentioned in the Introduction. Their use in antifouling coatings has been claimed in a number of patents [147][148][149][150]. As an example, Tanaka and coworkers immobilized antifoulants based on triazole, thiadiazole, and benzotriazole in resins containing a high fraction of carboxylate groups [147]. In a similar manner, Shtykova et al. studied the immobilization of the imidazole containing medetomidine (see Scheme 1) on an alkyd resin using NMR diffusometry [44]. It was found that the self-diffusion coefficient of medetomidine was decreased by more than one order of magnitude due to immobilization of medetomidine on the alkyd resin. It was proposed that the mechanism of immobilization was hydrogen bonding between the protonated imidazolium ion of medetomidine and deprotonated carboxylate groups of the alkyd resin [44].
This concept was further developed by Handa et al. to encompass polymers with very strong Lewis acid sites. Medetomidine was immobilized and released from the sulphonated polymer SPSEBS (see Scheme 1, the nonsulphonated polymer PSEBS was used as a reference) as monitored in situ using QCM-D [41]. In the solvent xylene, large amounts of medetomidine were adsorbed almost irreversibly, which was ascribed to a coulombic charge-transfer reaction in the hydrophobic solvent. The interaction was much weaker in artificial seawater with less adsorbed medetomidine and a pronounced release upon rinsing [41]. This system could consequently be used as a triggered release system (see Scheme 2) as discussed in Sections 1.1 and 1.2. Thickness (m * 10 6 ) Figure 5: Bilayer formation of Im-EHEC (see Scheme 1) with CuCl 2 (measured using QCM-D and modeled using a Voight-based viscoelastic model). The second layer is thin and corresponds to a monolayer. Reprinted from [49].

Metal Ion and Metal Oxide Nanoparticle Coordination.
The hydrogen bonds mentioned in the previous section are relatively strong interactions; yet, the strength of coordinate metal-ligand bonds discussed in Section 2 is significantly higher. Therefore, the Nydén group explored the feasibility of using transition metal coordination for medetomidine immobilization. Fant et al. investigated the adsorption and controlled release of medetomidine from PVM polymers (see Scheme 1, PVM-44) coordinated by Cu 2+ or Zn 2+ using QCM-D [40]. Cu 2+ was adsorbed in much higher amounts than Zn 2+ by the PVM polymer which is in line with the stability constants. Moreover, the metal ion adsorption resulted in a pronounced swelling of the PVM polymer by water which corroborates its feasibility to be used as a material for triggered release applications as discussed in Sections 1.2 and 4.2.1. The subsequent addition of medetomidine resulted in significant adsorption for both PVM coordinated by Cu 2+ and Zn 2+ (see Figure 6). Note that the resulting complex contains imidazole ligands from both the PVM polymer and medetomidine. However, the interaction of medetomidine with Zn 2+ was more reversible than with Cu 2+ and resulted in a much faster release (see Figure 6). Thus, the release profile may be tailored by coordinating the PVM polymer with mixtures of Zn 2+ and Cu 2+ [40].
In addition to polymers coordinated by metal ions, metal oxide nanoparticles are feasible vehicles for antifoulant immobilization using coordination chemistry [45,46]. Shtykova and coworkers studied the immobilization of some common booster biocides including medetomidine and Seanine (see Scheme 1) on different types of metal oxide nanoparticles (see Figure 7). It was apparent that the booster biocides containing ligand moieties (medetomidine, irgarol, diuron, and Seanine) gave the strongest adsorption. Note that the isothiazolinone Seanine (see Scheme 1) has the possibility to bind via hard interaction (carbonyl oxygen) or via soft interaction (sulphur free electron pair). Regarding the nitrogen which is of pyrrole type, interaction at this position is only possible at very high pH for isothiazolinones which Time (min) Figure 6: Medetomidine absorption (measured using QCM-D and SPR) to a Cu 2+ doped, Zn 2+ doped, and metal free 21 nm thick PVM film. It is apparent that metal coordination significantly enhances medetomidine adsorption and that the coordination to Cu 2+ is stronger and less reversible. Reprinted from [40]. The combination of QCM-D and SPR measurements can be used to estimate the amount of sorbed solvent in an adsorbed film. The acoustic QCM-D technique probes the entire adsorbed viscoelastic layer, including bound solvent. In contrast, the SPR method measures changes in refractive index at the surface and consequently takes only the mass of the adsorbent into account. The difference in adsorbed mass from QCM-D and SPR measurements can therefore be related to the amount of solvent.
are unsubstituted at position 2. However, the affinity of medetomidine for the metal oxide particles was substantially higher than that of any other biocide. This indicates that the imidazole group of medetomidine is key for its successful immobilization (see Figure 7(a)). The order of increasing interaction strength with the imidazole ligand was slightly different for the metal oxides compared with the corresponding metal ions (see Figure 7(b), and ZnO displays a stronger interaction with medetomidine than CuO) [42]. This is reasonable since the metal centers in the oxides are much more electron-rich compared to the free metal ions. The immobilization of medetomidine on ZnO nanoparticles (which are common pigments in antifouling marine paints) resulted in a significant decrease in the release rate from a varnish coating as compared with medetomidine molecularly dispersed in the same coating (see Figure 7(c)) [42].

Effect of Ligand Substituent Position and Ligand Chemistry
The coordinate interaction as described in Section 2 is not only dependent on the type of ligand used but also on the substituents attached at the azole ring. This has been investigated by us in detail using vibrational spectroscopy, EPR, ab initio, and DFT calculations and is presented in Section 6.1.
In Section 6.2, these results are discussed with respect to Cu 2+ /Cu + stabilization as presented in Section 1.3.

"Natural" versus Synthetic Ligands.
When inspecting the polymers discussed in Section 3 (see Scheme 1), it is observed that the imidazole ligand is attached to the polymer backbone at either position 1 or 4, and this holds for in principle any arbitrary imidazole-containing polymer. Moreover, regarding imidazole from biological sources, such as histidine, the imidazole ring is always substituted at position 4. This is in contrast to synthetic imidazole-containing polymers where the imidazole ring is usually attached to the backbone at position 1, that is, via the pyrrole nitrogen. Subsequently, it may be speculated why nature has chosen this particular position for substituent attachment. To investigate this, the methylated imidazole model compounds Me(1)-Im and Me(4)-Im (see Schemes 1 and 10) were analysed with respect to similarities and differences in coordination strength, coordination geometry, covalency, and so forth [38]. Both compounds display profound covalence in thebond ( 2 = 71% according to DFT calculations) which is also evident by the hyperfine splitting in the EPR spectrum (see Figure 8) [38]. An even better indication of the high degree of covalence is the -interaction coefficient (EPR; 2 ≈ [0.65, 0.69], ab initio; 2 ≈ [0.51, 0.60]) [38,151]. Small differences as an effect of position regarding the coordinate interaction are evident since the Me-C 4 bond is slightly more polarizable than the Me-N 1 bond. As a consequence, Me(4)-Im is a stronger base than both Me(1)-Im and Im since the methyl substituent is an electron-donating group (see Scheme 10) [38]. However, the stability constant is lower for the corresponding Me(4)-Im complex compared to Me(1)-Im and Im [38,152] which suggest a significant -acceptor interaction in addition to the -donor interaction in the coordinate complex [38] as discussed in Section 2. However, much larger effects were obtained by changing the type of substituent as was apparent from ab initio calculation. By systematically changing the position of the substituent as well as the substituting group, it was observed that Δ (the energy of complexation) and (M-L) (metal-ligand asymmetric stretch vibration) were more affected by the chemistry of the group than its actual position.
The coordination geometries for Me(1)-Im and Me(4)-Im are more or less identical. For instance, the steric hindrance of the methyl group upon coordination is the same for both methylated imidazoles (see Schemes 1 and 10 and Figure 9) since Me(4)-Im actually coordinates as Me(5)-Im [38,154] due to tautomerism and steric restrictions. As a consequence, the coordination geometries for the methylated imidazole ligands are very similar, both for Cu 2+ and Zn 2+ complexes, in contrast to unsubstituted imidazole (see Figure 9). However, an important difference between Me(1)-and Me(4)-Im is the reduced "degrees of freedom" for Me(1)-Im. Me(4)-Im, equivalent to Im or triazoles, contains two potential sites for coordination or protonation. Consequently, Me(4)-Im and Im may, as a function of ligand metal ratio and pH, form coordination polymers which is restricted to Me(1)-Im [38]. Reprinted from [42].
If the polymeric imidazole ligands used in this work are compared, that is, 1-VIm and histidine, the position of the substituent in the imidazole ring is not the most important factor (see Scheme 11). The chemical nature of the substituent, more precisely its electron withdrawing or electron donating abilities, is more important with respect to the coordinate interaction. However, the position of the ring substituent is important when it comes to bidentate coordination, the possibility to conduct protons between adjacent imidazole ligands and ring polarization due to coordinating nucleophiles [38].  6.2. Cu 2+ /Cu 1+ Stabilization. As mentioned in the introduction, Cu + is highly unstable in water and disproportionates to Cu 2+ and Cu 0 . This is a consequence of the hydrate complex where water, which is a hard base (see Section 2.1), strongly stabilizes Cu 2+ (border line acid) over Cu + (soft acid). However, it is possible to alter the redox potential in favor of Cu + by changing the ligand environment [100,158]. In Figure 10, 4 mM CuCl 2 has been dissolved in water/glycerol (hard ligands) or acetonitrile (MeCN, soft ligand) [38]. The area of the EPR signal is proportional to the concentration of Cu(II) ions. The MeCN system displays a smaller area compared to the water/glycerol system which indicates reduction of Cu(II) to the EPR silent Cu(I). It is well known that nitriles strongly stabilize Cu + in favour of Cu 2+ [38,47,76,[159][160][161]. However, soft ligands such as nitriles prevent the catalytic ability of Cu 2+ /Cu + by stabilizing the Cu(I) oxidation state too well [76]. In order to find a proper ligand environment, scientists are once again finding inspiration in biological systems, in particular copper containing enzymes [76,79]. Regarding the efficient catalytic activity and the electron transfer capability found in type 1 blue copper proteins, the ligand set of the copper complex contains both Cu 2+ and Cu + stabilizing ligands (histidine and cysteine/methionine resp.) [79,162]. In addition, the coordination geometry is forced into a distorted tetrahedral symmetry which is an intermediate geometry between square planar/tetragonal (Cu 2+ coordination) and tetrahedral (Cu + coordination, see Scheme 6) [79]. This has been suggested to facilitate the ease of electron transfer and oxidation/reduction of the copper ion [79,162].

Journal of Chemistry
We are currently investigating triazole-based ligands for Cu + stabilization [50] as described in Section 1.3 with respect to antifouling properties. The polymers are synthesized by a grafting onto approach as described for imidazoles [48] in Section 3.1 using the CuAAC click reaction (see Section 3.2). The triazole ligand have, in accordance with imidazole, a very high affinity for Cu 2+ . Yet, triazole ligands may also stabilize Cu + . Moreover, the coordination geometry can be tuned by the proper spatial distribution of ligands of different chemical nature. By tailoring the ligand environment and the subsequent coordination geometry in triazole containing polymers, our intention is to prepare antifouling hydrophilic polymers [163], inspired by copper enzymes, which may absorb Cu 2+ from the sea as well as facilitate the Cu 2+ → Cu + reduction.

Conclusion and Outlook
Azole coordination chemistry shows great potential with respect to antifouling coatings for a number of reasons. Many antifouling agents contain azole moieties. The imidazole containing biocide medetomidine is promising antifouling agent with respect to protection against fouling of barnacles. The fact that the azole moiety is a ligand can be used to immobilize these biocides on macromolecules or nanoparticle and thus control their release (both sustained and triggered release are possible). Moreover, the polymeric materials used for the purpose of controlling the release of antifouling agents can be given new properties via transition  (1)-Im (green) and Me(4)-Im (blue). (b) Derivative of the EPR signal with inset (b1) displaying the fine structure at = +3/2 [38]. The fine structure indicates four equivalent ligands. However, some signals appear to be doublets which may be explained by the Cu 2+ isotope mixture [79,153]. metal coordination of pendant azole ligands in the polymer. However, the most promising antifouling approach lies in the very strong affinity of azoles for Cu 2+ which is a potent antifouling biocide. An antifouling coating containing azole ligands can subsequently absorb large amounts of naturally abundant copper and obtain antifouling protection without any net release of biocide. Moreover, the Cu 2+ → Cu + redox potential may be altered towards reduction to Cu + , which is the most bioactive oxidation state. This can be achieved by tailoring the ligand chemistry where triazole ligands are promising for these purposes. We believe that the potential of this bioinspired approach has not been fully explored. A lot of inspiration can be found in nature, in particular copper enzymes. These enzymes are optimised through evolution for copper absorption as well as copper reduction by using a proper choice of ligand chemistry and ligand geometry.

Conflict of Interests
The research in this work has been financed by the foundation for environmental strategic research, MISTRA, which is administrated by the Swedish state. In the paper, the brand names of two antifouling biocides are mentioned; Selectope and Seanine. It is hereby stated that neither of the authors . R 1 and R 2 are defined as in Scheme 1. The coordination geometry for Cu 2+ complexes is tetragonal D4 h symmetry for the methylated imidazoles, equivalent to unsubstituted imidazole. The imidazole ligands are almost perpendicular to the coordination axis and the chlorides are weakly bound axially [38,[154][155][156]. However, regarding Zn 2+ complexes, unsubstituted imidazole forms octahedral complexes [38,155] [Zn(Im) 6 ] 2+ , whereas the methylated imidazoles form tetrahedral C2v complexes [38,156,157] [Zn(Me-Im) 2 Cl 2 ] despite the large excess of imidazole ligand. is in any financial relations with the previously mentioned commercial identities. Therefore, no conflict of interests exists.