Controlling Grafting from Aryldiazonium Salts: A Review of Methods for the Preparation of MonolayersTony Breton A C and Alison J. Downard B C
A MOLTECH-Anjou, Université d’Angers, UMR CNRS 6200, 2 Boulevard Lavoisier, 49045 Angers, France.
B MacDiarmid Institute for Advanced Materials and Nanotechnology, Department of Chemistry, University of Canterbury, Private Bag 4800, Christchurch 8140, New Zealand.
Tony Breton was born in La Rochelle (France) in 1977. He received his Ph.D. in organic electrosynthesis from the University of Poitiers in 2004. After a one-year post-doctoral research study on the electrochemical modification of surfaces with Professor Daniel Belanger (Montreal - Canada), he joined the MOLTECH-Anjou laboratory at the University of Angers (France) where he is currently Assistant Professor. His current work concerns the elaboration and characterisation of functionalised conductive surfaces with the aim of establishing controlled structure–property relationships.
Alison Downard is Professor of Chemistry at the University of Canterbury, New Zealand. She has an Honorary Doctorate from the Université de Rennes, France, and is a Fellow of the Royal Society of New Zealand. Alison gained her Ph.D. at the University of Otago, New Zealand, and undertook post-doctoral research at the University of Southampton with Professor Derek Pletcher and at UNC-Chapel Hill with Professor T. J. Meyer before taking up her position at the University of Canterbury. She has broad ranging interests in electrochemistry; her current major focus is on electrochemistry for surface modification.
Australian Journal of Chemistry 70(9) 960-972 https://doi.org/10.1071/CH17262
Submitted: 15 May 2017 Accepted: 6 July 2017 Published: 9 August 2017
Surface modification by grafting from aryldiazonium salts has been widely studied and applied to many substrates as a simple and versatile method for preparing strongly adherent organic coatings. Unless special precautions or conditions are used, the usual film structure is a loosely packed disordered multilayer; however, over the past decade, attention has been paid to establishing strategies for grafting just a monolayer of modifiers to the surface. To date, four general approaches to monolayer preparation have emerged: use of aryldiazonium ions with cleavable protection groups; use of aryldiazonium ions with steric constraints; grafting in the presence of a radical scavenger; and grafting from ionic liquids. This review describes these approaches, illustrates some of their applications, and highlights the advantages and disadvantages of each.
Grafting surface films from aryldiazonium salt precursors has become a widely used approach for the preparation of modified surfaces.[1,2] The stability of the resulting surface attachment, the wide substrate compatibility of the general methods, and the chemical diversity of the modifiers are the most important drivers behind the popularity of this grafting strategy. Various mechanisms have been detected or proposed for generation of the active modifier species; the dominant mechanism depends on the aryldiazonium derivative and the grafting medium. Aqueous acid or acetonitrile solution are the most commonly employed media, and under these conditions, at conducting substrates, film formation has been proposed to proceed, at least primarily, as shown in Fig. 1. The figure shows that aryl radicals are generated by reduction of the aryldiazonium ion: reduction can be carried out electrochemically or, for some substrates, by the substrate acting as the reducing agent. The latter is usually described as ‘spontaneous grafting’. Fig. 1 shows a second key feature of grafting from aryldiazonium salts: the most common outcome is formation of a loosely packed disordered multilayer film, typically less than ~8 nm in thickness. Multilayers form because as radicals are generated, they can either react with the substrate or with already attached modifiers. A second pathway may also lead to multilayer film growth: direct reaction of the aryldiazonium ion with the surface-attached aryl groups (or with the surface). This reaction accounts for the common observation of azo linkages within multilayer films. We have recently shown that the importance of this pathway depends on the activating effect of the para substituent towards electrophilic substitution.
A multilayer structure may be advantageous for some potential applications of these grafted films, and in other cases, the thickness of the film may be unimportant. However, when a layer with a well-defined structure is required, a monolayer has obvious advantages. Starting from a monolayer, further coupling reactions can be used to build up, in a controlled manner, a functional interface for the targeted application. Another advantage of thin layers, and in the limit, monolayers, is that they allow faster electron transfer between the electrode and solution-based redox centres than do thicker multilayers. This can be an important factor in applications such as electrochemical sensors and electrocatalysts.
To our knowledge, Pedersen, Daasbjerg, and coworkers were the first to report a deliberate strategy for limiting the growth of multilayers when grafting from aryldiazonium salt solutions. A decade ago, they demonstrated the ‘formation–degradation’ approach, which gave films with close to monolayer thickness.[5,6] In that work, they also recognised the importance of generating a layer with chemically reactive terminal groups; this ground-breaking research is described in the following sections. Also noteworthy in this regard is the report of Bartlett, Kilburn, and coworkers describing the use of tert-butyloxycarbonyl (Boc)-protected aminomethylbenzenediazonium salt. After grafting, deprotection, and post-functionalisation by amide coupling with an anthraquinone derivative, electrochemically measured surface concentrations were assumed to correspond to monolayer coverages. However, more recent work with the same diazonium ion has shown that multilayer coatings are formed (see below).
An important aspect of research into monolayer preparation from aryldiazonium salts is characterisation of the layers, and in particular, demonstration that monolayers rather than multilayers have been formed. When the grafted aryl group is electroactive, its surface concentration can be easily determined from the charge associated with its oxidation and/or reduction. Surface concentration can also be estimated from X-ray photoelectron spectroscopy (XPS) and quartz crystal microbalance (QCM) measurements. However, there are several factors that prevent surface concentration being a reliable indication of monolayer formation. First, considering the mechanism for film growth shown in Fig. 1, close-packed monolayers are not expected to be formed by this grafting method, except when steric constraints automatically limit grafting to a monolayer (see below). The high reactivity of radicals and the strong bonding of aryl groups to the surface prevent self-organisation, and as oligomeric structures form, shielding will prevent further attack of radicals at the substrate surface. (To our knowledge, there is only one well-characterised example of short-range self-organisation of modifiers during electrografting of aryldiazonium salts; see following sections.) In fact, from experiment, typical surface coverages appear to be in the range of ~25 to 40 % of a close-packed layer. These values have been estimated based respectively on the number of electroactive modifier groups in a monolayer-thick slice of a multilayer film on carbon, and on the fraction of a gold surface bound to aryl groups (and therefore not able to be oxidised) in a multilayer aryl film on gold. Clearly a ‘high’ experimental surface concentration (relative to the coverage for an ideal close-packed monolayer) can exclude the possibility of monolayer formation, but the surface concentration that might correspond to a monolayer, which has unknown packing density, cannot be determined. A second complicating factor is that a measured surface concentration that corresponds to 25–40 % of an ideal close-packed layer of the modifier may correspond to very sparse attachment of oligomeric structures on the surface, rather than individual modifiers. It is worth noting that monolayers with low surface coverages may be more accurately described as ‘sub-monolayers’; however, ‘monolayers’ will be used in the present report to indicate all layers in which individual modifiers (rather than oligomers) are immobilised on the surface.
A commonly used method for monitoring film growth from aryldiazonium salts is to record cyclic voltammograms of solution-based redox probes at the modified electrode. When the surface film is sufficiently thick and dense, it blocks access of the redox probe to the surface, resulting in a clearly observed decrease in the rate of electron transfer. For some methods of monolayer preparation, redox probe measurements may give useful qualitative information, for example, by demonstrating that a film becomes less blocking after a deprotection step.[8,12,13] However, such measurements cannot confirm that a film is indeed a monolayer because the blocking properties of the monolayer will depend on its density and thickness (i.e. the ‘height’ of the modifier). To illustrate this point: sparse monolayers of phenyl derivatives were found to have no detectable impact on redox probe voltammetry[12,13] whereas monolayers of derivatised calixarenes were highly blocking.
A more reliable indicator of monolayer formation is the thickness (height) of the layer. Atomic force microscopy (AFM) is most commonly used to directly measure film thickness, including the thickness of monolayers of modifiers with heights less than 1 nm. The AFM method involves grafting the layer to a substrate with a very low surface roughness (typically sub-0.5 nm), using an AFM tip to scratch away a section of film, and then profiling across the scratch. Pyrolysed photoresist film (PPF), which is a glassy carbon-like material with a typical roughness of <0.5 nm, is the commonly used substrate. Ellipsometry can also be used to measure the thickness of monolayer films on metal and silicon surfaces.[17,18]
In the following sections, the current four main methods for preparing monolayers from aryldiazonium salts are reviewed. We have included examples in which a deliberate strategy has been used to generate monolayers, and where there is firm evidence for monolayer formation, rather than simply an assumption that a monolayer is the grafting product.
Monolayers Prepared Using Aryldiazonium Derivatives with Cleavable Links
As mentioned above, Pedersen and Daasbjerg were the first to report a strategy aimed at generating monolayer films from aryldiazonium salt grafting. Fig. 2 shows their ‘formation–degradation’ approach involving use of a diaryl disulfide diazonium derivative with a cleavable linkage in the para position, which, after cleavage, gives a reactive layer suitable for further on-surface chemistry. As illustrated in Fig. 2, it was expected that owing to steric constraint, ring A of the diaryl disulfides should be less susceptible to radical attack than ring B, and so cleavage of the disulfide linkage should remove most, if not all, of the multilayer film. XPS measurements confirmed this expectation and measurement of film thickness by AFM depth profiling showed that when X = Cl (Fig. 2), layer thickness decreased from 3 ± 1 to 1.5 ± 0.5 nm after cleavage of the disulfide bond. The calculated height of thiophenolate is 0.6 nm, suggesting that more than a monolayer remained on the surface; however, the AFM measurement was performed on glassy carbon, which is rough relative to the layer thickness, and so firm conclusions could not be reached. The measured surface concentration of thiophenolate groups was 4 × 10-10 mol cm-2, which is in the range expected for a monolayer.
In further elegant work, Pedersen and Daasbjerg reported the grafting and characterisation of a diazonium salt with a para substituent with an extended chain terminated with a bulky alkyl hydrazone group (Fig. 3). Once an initial layer was grafted to the surface, electrostatic repulsion between the cationic ammonium terminus and diazonium ions in solution was expected to decrease the amount of film growth, and the bulky hydrazone group was also expected to protect surface-bound aryl groups from radical attack. After grafting to a polished glassy carbon electrode, removal of the hydrazone group by acid hydrolysis gave a layer of benzaldehyde groups, with a film thickness and surface concentration very similar to those found for the thiophenolate layer described above. The authors concluded that the film comprised 1–2 layers of benzaldehyde groups; however, given the special design features of this diazonium derivative, the build-up of oligomeric structures should be highly disfavoured and so it would be interesting to measure the surface concentration and layer thickness using a smoother substrate (e.g. PPF).
The principle of using a bulky protecting group that, when removed, leaves a monolayer of chemically reactive aryl groups has been very successfully implemented by Leroux, Hapiot, and coworkers in their ‘protection–deprotection’ strategy shown in Fig. 4. They prepared a para-substituted ethynyl aryldiazonium salt, protected by a triisopropylsilyl (TIPS) group and showed that after deprotection, an ethynyl-terminated monolayer remains, which reacts with azide derivatives via the CuI-catalysed Huisgen 1,3-dipolar cycloaddition pathway (a click reaction). In follow-up work, the researchers demonstrated that trimethylsilyl (TMS) and triethylsilyl (TES) protecting groups also gave (after deprotection) monolayers, with the density of the monolayer depending, in the expected way, on the size of the protecting group. TIPS has also been used as a protecting group for preparation of monolayers of benzylic hydroxyl groups, which were shown to be readily transformed into other reactive functionalities.
The reliability of monolayer formation, and convenience (and assumed high yield) of post-functionalisation of the layer by reaction with azides has seen silyl-protected ethynyl-derivatised aryldiazonium salts applied in several areas. For example, an amperometric hydrogen peroxide biosensor was fabricated by clicking azide-derivatised horseradish peroxidase (HRP) to an ethynyl-terminated monolayer on screen-printed carbon electrodes. The immobilised HRP maintained its redox activity, showing high rates of electrocatalytic hydrogen peroxide reduction and electron transfer with the electrode. In another biosensing example, ethynylphenyl-modified boron-doped diamond electrodes were prepared by grafting and deprotection of the corresponding TIPS-protected aryldiazonium salt, and an azide-derivatised sialic acid-mimic peptide was coupled to the layer. Highly sensitive detection of the influenza virus was achieved using impedance spectroscopy at the modified electrodes.
The use of the silyl-ethynylphenyl protection–deprotection strategy for the preparation of catalytic electrodes for the oxygen reduction reaction (ORR) has been demonstrated in clever work by Liu’s group.[23–25] In their first example, an iron porphyrin was covalently attached to carbon nanotubes (CNTs) via an imidazole ligand, which was clicked onto a monolayer of ethynylphenyl groups. The molecular assembly exhibited exceptional ORR activity and stability in both acidic and basic media. In later work, Liu’s group coordinated copper complexes to CNTs and to reduced graphene oxide (rGO), after first synthesising N-donor ligands on the carbon substrate. Bidentate, mononucleating ligands were synthesised by clicking azidopyridine to the ethynylphenyl monolayer, giving triazole and pyridine donor groups (Fig. 5).
The ORR activity and stability of the covalently immobilised mononuclear copper complexes were superior to those of physisorbed analogues but the activity was lower than that for the similar covalently coupled multinuclear copper complexes. For the latter complexes, Liu and coworkers synthesised dinucleating tetradentate N-donor ligands on the rGO substrate by a similar route to above, but beginning with grafting of the diazonium ion formed in situ from 6-((triisopropylsilyl)ethynyl) pyridin-3-amine. As shown in Fig. 6, this gave surface-immobilised ligands with two triazole and two pyridinic donors. The dinuclear complexes had high ORR activity and excellent stability, and high selectivity towards the desirable four-electron reduction pathway.
The application of this protection–deprotection strategy has also been explored in the context of molecular electronics. Yamamoto and Einaga clicked an azido derivative of azobenzene onto an ethynyl-terminated monolayer on heavily boron-doped diamond and found that photoinduced isomerism of azobenzene modulated the superconducting critical temperature. In this application, the low density of azobenzene groups that results from removal of the bulky protecting groups might be assumed to aid in the isomerisation reaction. Similarly, the low density of tethers and of the tetrathiafulvalene molecules clicked to the tethers allowed the surface assembly of charge-transfer complexes with tetracyanoquinodimethane derivatives. In contrast, McCreery and coworkers found that electronic junctions prepared by clicking azidoferrocene onto ethynyl-terminated monolayers grafted to PPF showed (undesirable) short-circuit responses between the PPF and a top layer of e-beam carbon or gold. The authors attributed the short-circuiting to the pinholes that remain after the bulky protecting groups are removed. Similar responses were obtained when either TIPS or the smaller TMS were used as the initial protecting group. In this context, we have recently examined, experimentally and by computation, the surface coverage of monolayers prepared by several protection–deprotection strategies (see below for these other strategies). The study highlighted the low surface coverage of tether groups after removal of bulky protecting groups: for ethynyl-terminated layers, the calculated coverage was 0.2–0.3 of a close-packed monolayer if the phenyl groups are assumed to be freely rotating, and close to 0.1 if no rotation is assumed.
Ethynylphenyl monolayers are clearly very convenient for further on-surface chemistry via the 1,3-dipolar Huisgen cycloaddition click reaction; Sonogashira and Glaser cross-coupling reactions have also been demonstrated at these surfaces. However, to expand the types of chemistry that can be used for post-functionalisation of monolayers, and to allow the possibility of forming mixed monolayers with two (or more) tethers with orthogonal reactivity, we have developed protection–deprotection strategies for preparing monolayers with para-carboxy and -amine substituents. Fig. 7 shows the 9-fluorenylmethyl (Fm)-protected carboxybenzene diazonium ion, which after grafting and deprotection gave a well-characterised monolayer. Post-functionalisation with amine derivatives was conveniently carried out using oxalyl chloride to convert the carboxylate functionality to the more reactive acyl chloride. The measured surface concentration of electroactive ferrocene and nitrophenyl groups coupled to the layer suggested the reactions proceeded with high yields.
A similar strategy was adopted to prepare amine-terminated monolayers. In that work, both fluorenylmethyloxycarbonyl (Fmoc) and the smaller Boc protecting group were used to protect aminobenzene and aminomethylbenzene diazonium salts (Fig. 8). After grafting and deprotection, film thickness measurements by AFM depth profiling indicated that both of the Fmoc-protected diazonium salts and the Boc-protected aminobenzene diazonium salt yielded monolayers, whereas multilayers were present after deprotection of the Boc-protected aminomethylphenyl layer. This finding was attributed to the relatively small size and flexibility of the para substituent of the latter diazonium salt. Electroactive carboxylic acid derivatives were coupled to the layers using both oxalyl chloride and N, N,N′,N′-tetramethyl-O-(1H-benzotriazol-1-yl)uronium hexafluorophosphate (HBTU) as activating agents for amide bond formation. The measured surface concentrations of the coupled redox centres were consistent with the oxalyl chloride-activated carboxylic acid derivatives coupling to both the amine groups and the glassy carbon surface, whereas use of HBTU resulted in immobilisation on the amine-terminated tether groups only.
A very different approach was taken by Page and Wildgoose, who synthesised a calixarene-tetradiazonium salt bearing tert-butyldimethylsilyl-protected ethynyl groups on the upper (narrower) rim (Fig. 9 shows the amine precursor to the diazonium salt). After grafting to glassy carbon, the ethynyl groups were deprotected and an azidoferrocene derivative was clicked to the surface. From the electrochemically measured surface concentration of ferrocene groups, and with the assumption that all ethynyl groups had reacted with azidomethylferrocene, the grafted calixarenes were described as a ‘complete’ monolayer. Although this calculation appeared not to take account of surface roughness, it is clear that a reasonably dense layer was grafted and that the terminal ethynyl groups were at least largely reactive towards click coupling. This strategy, which expands on the work of Lagrost and coworkers[14,32,33] (described in the next section), has the advantages of generating a monolayer of modifiers with multiple anchoring points, which should lead to high stability, and gives a layer of reactive tethers which can be conveniently functionalised using click chemistry.
All of the protection–deprotection approaches described above have the advantages of reliably generating monolayers (when an appropriate protecting group is used) and of generating reactive surface tethers for straightforward, follow-up on-surface chemistry. With the exception of the last example, removal of the bulky protecting groups must necessarily leave monolayers with low surface coverages, allowing other modifiers to be immobilised, giving mixed monolayers.[29,34] The disadvantage of the general approach is the synthetic effort required to prepare the protected aryldiazonium salts.
Monolayers Prepared by Relying on Steric Constraints of the Aryldiazonium Ion
Knowing that the main mechanism responsible for the growth of the polyaryl layer proceeds via radical attack on the free positions of the grafted aryl rings (Fig. 1), Pinson and Podvorica were the first to speculate that the presence of bulky substituents on the aryl ring should prevent multilayer growth. Electrochemical reduction of 3,5-bis-tert-butylbenzenediazonium ion (3,5-TBD) on copper, silicon, and gold by chronoamperometry led to the formation of films with thicknesses between 1 and 1.6 nm (measured by ellipsometry), very different to the 16-nm thick layer obtained using 4-methylbenzenediazonium ion (Fig. 10). Assuming that aromatic rings are tilted 40° to the surface normal, as shown by previous density functional theory (DFT) calculations, this result was in good agreement with the presence of a monolayer on the substrate. The steric hindrance of the tert-butyl substituents appeared to be the key factor as the use of trifluoromethyl groups, which are less bulky, did not limit growth to a monolayer.
In a second study, published shortly after, Pinson and Podvorica investigated the role of the steric hindrance induced by one or two alkyl substituents on the benzenediazonium ion on the grafting efficiency on copper. It was first shown that the presence of a methyl substituent in both positions 2 and 6, or one ethyl substituent in position 2 of the benzene ring completely prevent grafting to the surface. This observation was explained in the light of DFT calculations that gave a theoretical decrease of the modifier surface binding energy of up to 50 % because of the geometrical constraints introduced by the surface. In the case of monosubstitutions 2-, 3-, and 4-methyl, 2-methoxy, and 4-tert-butyl, and disubstitutions 3,5- and 2,4-dimethyl and 3,5-bis(trifluoromethyl), multilayer growth was invariably obtained. These results demonstrate that positions 3, 4, and 5 on the grafted phenyl rings are all sites for radical attack and that for the derivatives above, steric hindrance is too low to have an impact. Interestingly, the grafting of pentafluorobenzene diazonium ion also led to a multilayered film, a finding that was attributed to homolytic aromatic substitution of fluorine, leading to the formation of dimers, compatible with the layer growth. The only diazonium ion salt able to generate strictly monolayered films was 3,5-TBD thanks to its two bulky substituents that prevent radical attack in position 4.
The ability of 3,5-TBD to generate monolayers was exploited in very elegant work by Lvasenko, De Feyter, and coworkers. Raman spectroscopy was used on both highly ordered pyrolytic graphite (HOPG) and Cu-supported graphene (synthesised by chemical vapour deposition (CVD)) to evidence the difference in grafting between this aryldiazonium ion and the well-known 4-nitrobenzenediazonium ion. The intensity ratio of D to G bands, characteristic of an sp2-to-sp3 hybridisation after grafting, was found to be much higher for 3,5-TBD, corresponding to a higher functionalisation density on the surface. Those results were confirmed by scanning tunnelling microscope (STM) images where only scattered aggregates of polymerised nitrophenyl groups were visible whereas 3,5-tert-butylphenyl clustered groups were densely distributed (Fig. 11). As polyaryl growth was found to occur at a very low concentration of 4-nitrobenzenediazonium ion, the authors concluded that radical attack at surface-attached nitrophenyl groups must be kinetically very favourable. In contrast, grafting from 3,5-TBD, for which the tert-butyl moieties prevent further radical attachment to already grafted species, should enable radicals to continue to graft directly to the surface, resulting in high-density grafting.
Working on HOPG, Ivasenko and De Feyter demonstrated patterning by use of the STM tip to remove grafted 3,5-tert-butylphenyl groups from a densely grafted monolayer, and in subsequent work, the technique was extended to the preparation of nanocorrals on HOPG. Examination of self-assembly of 10,12-pentacosadiynoic acid within nanocorrals of different sizes, shapes, and orientations led to insights into factors that control the nucleation and growth processes involved in self-assembly.
From the above discussion, it is clear that the inherent ability of 3,5-TBD to yield monolayers has advanced our understanding of film growth pathways, and the dense monolayers formed by this diazonium ion have been used very creatively in fundamental studies of other systems. Other work aimed at using diazonium salt derivatives with the inherent ability to generate monolayers has yielded layers that present reactive terminal groups for post-functionalisation reactions. This work is described in the following.
Lagrost and coworkers were the first to report the synthesis and grafting of calixarenediazonium salts. They grafted the diazonium ions prepared from the amine precursors shown in Fig. 12 to gold and carbon electrodes and thoroughly characterised the resulting layers. Through AFM measurements on PPF and ellipsometry measurements on gold, they confirmed that the grafted layers were monolayers; electrochemical assessment of the blocking properties of the layers suggested that they were densely packed. The formation of densely packed monolayers was attributed to the position of the diazonium moieties on the calixarene rim, which orient the structure on the surface, and to the methylene groups in the narrow rim, which prevent radical attack at the phenyl ring meta to the diazonium functionality. Amide coupling reactions with ferrocenemethylamine were used to attach ferrocene groups to monolayers of 3 and 4 (Fig. 12) and electrochemical measurement of the surface concentration of ferrocene coupled to surface 3 indicated a ‘rather’ densely packed monolayer. Interestingly, the surface concentration of ferrocene coupled to surface 4 was only twice as high, suggesting only two ferrocene groups could be coupled to each calixarene. Nevertheless, as the authors point out, in addition to the presumed stability advantage of multiple anchor points, the approach paves the way for preparation of functionalised surfaces with controlled spatial relationships between coupled species.
Continuing with their interest in monolayer-modified surfaces with controlled composition, Lagrost, Reinaud, and Jabin examined the formation of mixed monolayers on gold substrates by one-pot grafting from mixtures of calixarene tetra-diazonium salts. Using mostly the diazonium ions prepared from derivatives 2 and 4 (Fig. 12), they compared the ratio of each calixarene on the surface with that of the diazonium salts in the deposition solution and found reasonable agreement. This was not unexpected because the reduction potentials of all four diazonium ions were shown to be very similar. This is a very useful consequence of the common core structure of the diazonium ions, and allows easy tuning of monolayer composition. Post-functionalisation of mixed monolayers via amide bond formation with ferrocenepentylamine revealed that dilution of modifier 4 ~1 : 10 with modifier 2 resulted in an increased number of ferrocene groups coupled, per modifier 4. This beneficial effect on reactivity of dilution of the active site is frequently observed with a variety of surface modifiers, and is generally attributed to a decrease in steric crowding.
Taking a different pathway to grafting calixarene tetra-diazonium salts, Mattiuzzi, Lagrost, and coworkers utilised the spontaneous formation of aryl radicals through decomposition of aryldiazonium salts in basic solution to graft monolayers to gold and polymer substrates. The layer grafted to gold from a sodium hydroxide solution (pH 13) of the diazonium salt from calixarene 4 (Fig. 12) was examined by AFM and ellipsometry and confirmed to have monolayer thickness. Using the same methodology, calixarenes 2 and 4 (Fig. 12) were successfully grafted to propylene substrates, and 2 also to polyethylene terephthalate and polystyrene. AFM imaging clearly showed the formation of surface layers on polypropylene and XPS confirmed the presence of modifier 2; however, AFM depth profiling measurements could not be performed owing to the softness of the substrate. Nevertheless, all results were consistent with the presence of very thin modifying layers. Both amine and alcohol derivatives were successfully coupled to the layer of 4 on polypropylene, and hence, this work demonstrates a simple and effective route to preparation of well-defined and controlled reactive layers on non-conducting substrates for tailoring of surface properties and functionality.
One final example of a diazonium ion that generates only a monolayer on electrografting is a heteroleptic polypyridine RuII complex reported by Lemercier, Lacroix, and coworkers. Fig. 13 shows the amine precursor to this complex (compound 1), and that of the parent complex, 2. AFM measurements of layers electrografted from the diazonium salt of 1 gave thicknesses consistent with monolayer formation whereas the parent diazonium compound forms multilayers. Restriction of film growth to a monolayer with the diazonium derivative of 1 was attributed to steric effects. STM imaging of monolayers on HOPG clearly showed ordering of the molecules, a phenomenon suggested to arise by a process in which a grafted molecule acts as a seeding agent, inducing ordering as subsequent modifiers approach the surface.
As is the case for monolayers prepared through the protection–deprotection strategy, the use of modifiers with steric constraints appears to be a reliable method for preparing monolayers with non-reactive and reactive terminal groups. The commercial availability of the precursor amine to 3,5-TBD should see further use of this modifier for fundamental studies. However, for research aimed at building up functional interfaces from well-defined monolayers with reactive terminal groups, the disadvantage of non-trivial synthetic procedures applies. Strategies that avoid this disadvantage are described in the following sections.
Monolayers Prepared by Grafting in the Presence of Radical Scavengers
As an experimentally simple approach to monolayers, we have proposed to limit film growth using 2,2-diphenyl-1-picrylhydrazyl (DPPH) as a radical scavenger during the grafting process on glassy carbon and PPF. An increasing concentration of the scavenger can be added to the deposition solution to trap the aryl radical and to control the thickness of the organic layer. For the electrografting of nitrobenzenediazonium salt, well-defined monolayers can be obtained in the presence of an excess of DPPH (Fig. 14).
The mechanism involved in this radical control approach is proposed to be purely diffusion-controlled and is based on the concentration equilibrium between the aryldiazonium ion and the scavenger. The observation that grafting cannot be fully suppressed, even at high scavenger concentrations, is most likely due to the instantaneous reaction of the aryl radicals produced at the substrate/solution interface. We assume that the use of a radical scavenger leads to homogeneously modified surfaces thanks to two phenomena: planar diffusion prevents dendritic oligomer formation because of the drastic decrease in aryl radical concentration moving away from the surface due to the presence of scavenger, and prolonged electrolysis or multicycle voltammetry allows pinholes in the layer to be filled, leading to a compact layer, even at less reactive carbon sites.
Recently, it has been shown that an efficient control of film growth via this strategy is, in fact, dependent on the para substituent of the aryldiazonium salt. Electron-donating groups favour uncontrolled secondary grafting mechanisms and strict monolayer functionalisation can only be obtained by the use of electron-withdrawing para substituents (Fig. 15).
The method, initially developed under electrografting conditions, has been extended to spontaneous grafting from acetonitrile solutions where the amount of scavenger required to generate monolayers on glassy carbon was found to be one-tenth that required during electrografting. Under those conditions, an estimated 50 % of modifiers were immobilised on the surface via azo links (rather than direct C–C coupling of aryl groups), reflecting the competition between the two main film-forming mechanisms (Fig. 1). The spontaneous grafting pathway was successfully applied to the modification of activated carbon powder for supercapacitor development. Modification from nitrobenzenediazonium salt solution using radical control led to an increased proportion of electroactive material in the film and a conservation of the porosity.
In very recent work, a monolayer of nitrophenyl groups electrografted from a solution of nitrobenzenediazonium salt in the presence of DPPH was subsequently reduced to give a reactive amine-terminated monolayer. After successful post-functionalisation with the 2,2,6,6-tetramethylpiperidin-1-yl)oxyl (TEMPO) redox species, the advantages afforded by thin layers, in terms of electron transfer rate and interfacial reactivity, were clearly demonstrated. The same strategy was used by Creus and coworkers to prepare amine-terminated monolayer films on HOPG. The modified surfaces maintained good electrical conductivity, allowing electron transfer reactions to take place. Gold electrodeposition was found to be strongly favoured on the monolayer-modified surfaces compared with pristine HOPG, offering the possibility of technological applications incorporating patterned carbon surfaces. The same group also used the radical control strategy to produce a precisely controlled sub-monolayer coverage of nitrophenyl groups on HOPG for investigation of the effect of the modifier on the electrical properties of the system. Combined with DFT calculations, scanning probe microscopy measurements confirmed an almost perpendicular orientation of the nitrophenyl group with a dipole moment directed towards the surface. Furthermore, the work function of HOPG was found to increase after modification with nitrophenyl groups.
The experimental simplicity of adding a commercially available radical scavenger to the grafting solution is the major advantage of this approach to monolayer preparation. The disadvantage is the limited chemical diversity of monolayers that can be accessed. Nevertheless, as the examples described above illustrate, the ability to prepare a monolayer of reactive aminophenyl groups, via grafting and subsequent reduction of a nitrophenyl monolayer, is a very useful application of the strategy.
Monolayers Prepared Using Ionic Liquids as the Grafting Medium
The development in the early 2000s of ionic liquids (ILs) as solvents for electrochemical reactions led to investigations of their use for grafting from aryldiazonium salts. Initially demonstrated with CNTs as the substrate in imidazolium-based IL, the feasibility of the grafting reaction was extended to glassy carbon using hydrophobic or hydrophilic imidazolium-based ILs. Ghilane and Randriamahazaka were the first to report the lowering of surface coverages compared with those obtained in classical solvents using nitrobenzenediazonium salt and suggested the possibility of controlling the layer thickness by electrografting in ILs. The decrease of diffusion coefficients and aryl radical trapping by the acidic protons of ILs were both invoked to explain their observations. Shortly after, the same authors published a detailed study dealing with the effect of the solvent viscosity on the electrografting efficiency of nitrobenzenediazonium salt on glassy carbon by the use of several ILs. They found a clear dependence of the surface coverage of nitrophenyl groups on the viscosity of the IL, demonstrating the impact of diffusion on the grafting kinetics. Moreover, by adding a co-solvent to the most viscous IL, a fine modulation of the surface concentration was obtained (Fig. 16, left) and, interestingly a limiting value of ~2 × 10-10 mol cm-2 was reached for viscosity higher than 100 cP (1 cP = 1 mPa s). AFM measurements on modified gold surfaces confirmed the decrease of the layer thickness down to 1 nm, corresponding to ~1–2 layers (Fig. 16, right).
Ghilane and Randriamahazaka pointed out that it is not easy to understand why diffusion effects should directly influence the amount of film electrografted to the electrode. They noted that it might even be expected that the amount of film should increase as solution viscosity increases owing to slower diffusion of aryl radicals away from the electrode. The authors concluded that the mechanism of electrografting in ILs is more complex than the simple generation of radicals at the electrode followed by coupling to the surface or to already attached groups, and they proposed several possible factors that might influence grafting in ILs, and be different, or additional, to those operating in lower-viscosity media.
The use of IL solvent has been extended by Bélanger and coworkers to the electrografting of in situ-generated aryldiazonium ions on carbon. A protic IL, able to promote the diazotization reaction (the conversion of the amine precursor into the diazonium ion) by proton exchange, was used in the presence of 4-nitro- and 4-chloroaniline to lead to the grafting of the corresponding aryl derivatives. A voltammetric study of the nitrophenyl-modified electrode revealed a surface coverage of 9 × 10-10 mol cm-2. The chemical composition, determined by XPS, was found to be very similar to those previously reported for electrodes modified from aqueous and acetonitrile media, showing that no major change occurs in the modification process when an IL is used. AFM measurements of films grafted to PPF gave a film thickness of ~1.5 nm (1–2 layers), in good agreement with voltammetric data. Importantly, it was found that an increase of the modification time did not cause a detectable increase of the thickness of the deposited film, allowing a good control of the grafting under stationary conditions. The authors postulated that the relatively large ions of the protic IL at the electrode surface could inhibit the growth of the grafted layer and explain this self-limitation.
Very recently, Lagrost and coworkers reported work focussed on the functionalisation of glassy carbon and CNTs using in situ-generated 4-nitrobenzene diazonium ion in a Brønsted acidic IL. Their results were consistent with those of Bélanger described above, demonstrating self-limiting film growth, whatever the electrolysis duration. In addition, an electrochemical study combined with AFM measurements showed that this procedure allows the formation of a more compact layer than that obtained by grafting from acetonitrile or aqueous acidic solutions.
The main advantage of the use of ILs to control the grafting resides in the versatility of the approach that is not limited by synthetic aspects or (presumably) the particular aryldiazonium derivative. However, even though the increase of the IL viscosity appears to be an easy route to efficiently decrease the thickness of grafted layers, it is not clear that it can strictly prevent the grafting of aryl radicals onto already attached groups, and thus whether it affords true monolayers. Deeper understanding of the grafting process in ILs is required to reveal the scope of the approach in this regard.
To date, there is no experimentally simple, reliable, versatile, and widely applicable method for grafting monolayers from aryldiazonium salts. When the requirement for a significant amount of organic synthesis does not present a drawback, the protection–deprotection methods and use of diazonium salts with steric constraints are currently the most reliable and versatile methods for forming monolayers of reactive tethers. Together, they afford pathways for generating monolayers with ethynyl, hydroxyl, carboxyl, and amino-terminal groups that allow straightforward post-functionalisation of the monolayers with a variety of functional groups. However, grafting a monolayer of nitrophenyl groups via the radical trap approach cannot be surpassed for sheer simplicity and convenience. Conversion of the layer into reactive aminophenyl groups is also very simple, and hence this strategy provides easy access to a platform for further on-surface chemistry. As highlighted above, monolayers can be reliably prepared through grafting from 3,5-TBD or the diazonium ion of the heteroleptic RuII complex of Fig. 13; however, these layers have, to date, had limited application. Finally, grafting using ILs as the solvent requires further study to establish whether limitation of film growth to a monolayer, or near-monolayer, is a general result that applies to a range of aryldiazonium salts, and indeed whether monolayers can be reliably prepared in this way.
Conflicts of Interest
The authors declare no conflicts of interest.
References D. Bélanger, J. Pinson, Chem. Soc. Rev. 2011, 40, 3995.
| CrossRef |
 M. Delamar, R. Hitmi, J. Pinson, J. M. Savéant, J. Am. Chem. Soc. 1992, 114, 5883.
| CrossRef | 1:CAS:528:DyaK38XktFCqu7o%3D&md5=3bc493d79aa329914227188b221f5a92CAS |
 P. Doppelt, G. Hallais, J. Pinson, F. Podvorica, S. Verneyre, Chem. Mater. 2007, 19, 4570.
| CrossRef | 1:CAS:528:DC%2BD2sXos1Sksrg%3D&md5=5780ed33e21d218dd3acaa5324e8aa0fCAS |
 T. Menanteau, M. Dias, E. Levillain, A. J. Downard, T. Breton, J. Phys. Chem. C 2016, 120, 4423.
| CrossRef | 1:CAS:528:DC%2BC28XitFers78%3D&md5=8ce30643b275c097e90540d716d5c0f9CAS |
 L. T. Nielsen, K. H. Vase, M. D. Dong, F. Besenbacher, S. U. Pedersen, K. Daasbjerg, J. Am. Chem. Soc. 2007, 129, 1888.
| CrossRef | 1:CAS:528:DC%2BD2sXot1ahtg%3D%3D&md5=68644be64c1adf3db47d2d4de540bbdaCAS |
 K. Malmos, M. D. Dong, S. Pillai, P. Kingshott, F. Besenbacher, S. U. Pedersen, K. Daasbjerg, J. Am. Chem. Soc. 2009, 131, 4928.
| CrossRef | 1:CAS:528:DC%2BD1MXjtFKnur0%3D&md5=b3d3d6261bf14361ef9ec208c76e52adCAS |
 J. M. Chretien, M. A. Ghanem, P. N. Bartlett, J. D. Kilburn, Chem. – Eur. J. 2008, 14, 2548.
| CrossRef | 1:CAS:528:DC%2BD1cXltVejtbs%3D&md5=178eb5513561d029028968f481cb2b36CAS |
 L. Lee, Y. R. Leroux, P. Hapiot, A. J. Downard, Langmuir 2015, 31, 5071.
| CrossRef | 1:CAS:528:DC%2BC2MXmsFagur4%3D&md5=1352eccf2996b5967e44615465cf105fCAS |
 V. Q. Nguyen, X. Sun, F. Lafolet, J.-F. Audibert, F. Miomandre, G. Lemercier, F. Loiseau, J.-C. Lacroix, J. Am. Chem. Soc. 2016, 138, 9381.
| CrossRef | 1:CAS:528:DC%2BC28XhtFyitLjI&md5=69404a012f54ee4dc238fee8be5c079bCAS |
 P. A. Brooksby, A. J. Downard, Langmuir 2004, 20, 5038.
| CrossRef | 1:CAS:528:DC%2BD2cXjvVCktro%3D&md5=deb6769896a6187b897280b66262e6baCAS |
 M. G. Paulik, P. A. Brooksby, A. D. Abell, A. J. Downard, J. Phys. Chem. C 2007, 111, 7808.
| CrossRef | 1:CAS:528:DC%2BD2sXkvFSjt7k%3D&md5=4685816659051c8ec110feabaef7ca36CAS |
 Y. R. Leroux, H. Fei, J. M. Noel, C. Roux, P. Hapiot, J. Am. Chem. Soc. 2010, 132, 14039.
| CrossRef | 1:CAS:528:DC%2BC3cXhtFygtrfI&md5=6d363934ec87b021717be0c5516b2bb4CAS |
 L. Lee, H. F. Ma, P. A. Brooksby, S. A. Brown, Y. R. Leroux, P. Hapiot, A. J. Downard, Langmuir 2014, 30, 7104.
| CrossRef | 1:CAS:528:DC%2BC2cXovVSjs78%3D&md5=20d25b9e8d8213fbc9e3417dfaee9111CAS |
 A. Mattiuzzi, I. Jabin, C. Mangeney, C. Roux, O. Reinaud, L. Santos, J. F. Bergamini, P. Hapiot, C. Lagrost, Nat. Commun. 2012, 3, 1130.
 F. Anariba, S. H. DuVall, R. L. McCreery, Anal. Chem. 2003, 75, 3837.
| CrossRef | 1:CAS:528:DC%2BD3sXks12nur4%3D&md5=2f7edfc68cd98926c73f8e1a783aad31CAS |
 S. Ranganathan, R. L. McCreery, Anal. Chem. 2001, 73, 893.
| CrossRef | 1:CAS:528:DC%2BD3MXjvV2ntw%3D%3D&md5=6d9d5ca24af49c18e75ef445f2f2b661CAS |
 C. Combellas, F. Kanoufi, J. Pinson, F. I. Podvorica, J. Am. Chem. Soc. 2008, 130, 8576.
| CrossRef | 1:CAS:528:DC%2BD1cXnt1Witrg%3D&md5=a17f6173ab023c398fd5a4189c046b08CAS |
 C. Combellas, D. E. Jiang, F. Kanoufi, J. Pinson, F. I. Podvorica, Langmuir 2009, 25, 286.
| CrossRef | 1:CAS:528:DC%2BD1cXhsVOmtr7L&md5=ca2ee61d8778af15ebfe8849f165da5aCAS |
 Y. R. Leroux, P. Hapiot, Chem. Mater. 2013, 25, 489.
| CrossRef | 1:CAS:528:DC%2BC3sXhtFWrurc%3D&md5=e446d73294bf0cc6b8c1d8c3f60f8546CAS |
 W. J. Liu, T. D. Tilley, Langmuir 2015, 31, 1189.
| CrossRef | 1:CAS:528:DC%2BC2MXmt12l&md5=1bdb60c324132af6f7157f271ffda8a7CAS |
 A. Hayat, J. L. Marty, A. E. Radi, Electroanalysis 2012, 24, 1446.
| CrossRef | 1:CAS:528:DC%2BC38XmslKntb8%3D&md5=fa17c0a7e585e07c400a29324d64774fCAS |
 T. Matsubara, M. Ujie, T. Yamamoto, M. Akahori, Y. Einaga, T. Sato, Proc. Natl. Acad. Sci. USA 2016, 113, 8981.
| CrossRef | 1:CAS:528:DC%2BC28Xht1ehurbP&md5=0488dfd516b3c1813ade0b7345955554CAS |
 P. J. Wei, G. Q. Yu, Y. Naruta, J. G. Liu, Angew. Chem. Int. Ed. 2014, 53, 6659.
| CrossRef | 1:CAS:528:DC%2BC2cXot1SitLc%3D&md5=0c2da3767c30e57abd540f6ccd910513CAS |
 R. C. Wang, T. L. Yin, P. J. Wei, J. G. Liu, RSC Adv. 2015, 5, 66487.
| CrossRef | 1:CAS:528:DC%2BC2MXht1Gmu7jP&md5=703dd3af0f4969b15187925a26f8e605CAS |
 Y.-T. Xi, P.-J. Wei, R.-C. Wang, J.-G. Liu, Chem. Commun. 2015, 7455.
| CrossRef | 1:CAS:528:DC%2BC2MXltFWltLs%3D&md5=b96c5f03bab62cf1a95eabf931e4654bCAS |
 K. Natsui, T. Yamamoto, M. Akahori, Y. Einaga, ACS Appl. Mater. Interfaces 2015, 7, 887.
| CrossRef | 1:CAS:528:DC%2BC2cXitVKqsLrJ&md5=9f3f979ab6afaf7196177f3c07a19440CAS |
 J. Jalkh, Y. R. Leroux, A. Vacher, D. Lorcy, P. Hapiot, C. Lagrost, J. Phys. Chem. C 2016, 120, 28021.
| CrossRef | 1:CAS:528:DC%2BC28XhvVGqtrzP&md5=9b415e3ad148a377038aa09737ce8f9cCAS |
 S. Y. Sayed, A. Bayat, M. Kondratenko, Y. Leroux, P. Hapiot, R. L. McCreery, J. Am. Chem. Soc. 2013, 135, 12972.
| CrossRef | 1:CAS:528:DC%2BC3sXht1Ojur3E&md5=4b950bb066c53ae1d3468e2257bbb629CAS |
 L. Lee, N. R. Gunby, D. L. Crittenden, A. J. Downard, Langmuir 2016, 32, 2626.
| CrossRef | 1:CAS:528:DC%2BC28Xjtlentrw%3D&md5=d5e0cbe07cc49fa17a323aaed7fcf764CAS |
 A. A. S. Gietter, R. C. Pupillo, G. P. A. Yap, T. P. Beebe, J. Rosenthal, D. A. Watson, Chem. Sci. 2013, 4, 437.
| CrossRef | 1:CAS:528:DC%2BC38XhslKktbfF&md5=3203ba399935338cdb85461ed86f96efCAS |
 J. P. Buttress, D. P. Day, J. M. Courtney, E. J. Lawrence, D. L. Hughes, R. J. Blagg, A. Crossley, S. E. Matthews, C. Redshaw, P. C. B. Page, G. G. Wildgoose, Langmuir 2016, 32, 7806.
| CrossRef | 1:CAS:528:DC%2BC28XhtFOktLfL&md5=b4fe2ccac8d2a78fc2c314f16fb4101bCAS |
 L. Santos, A. Mattiuzzi, I. Jabin, N. Vandencasteele, F. Reniers, O. Reinaud, P. Hapiot, S. Lhenry, Y. Leroux, C. Lagrost, J. Phys. Chem. C 2014, 118, 15919.
| CrossRef | 1:CAS:528:DC%2BC2cXhtFSmsLfO&md5=5fb4a22c7607a571a6779944c4505b05CAS |
 L. Troian-Gautier, D. E. Martinez-Tong, J. Hubert, F. Reniers, M. Sferrazza, A. Mattiuzzi, C. Lagrost, I. Jabin, J. Phys. Chem. C 2016, 120, 22936.
| CrossRef | 1:CAS:528:DC%2BC28XhsFWqtr3P&md5=a6f76ec62b0d2b9081063e5cc374945fCAS |
 L. Lee, P. A. Brooksby, Y. R. Leroux, P. Hapiot, A. J. Downard, Langmuir 2013, 29, 3133.
| CrossRef | 1:CAS:528:DC%2BC3sXitlCgtrg%3D&md5=88ea85d3286b657b767097a4bd60abbcCAS |
 J. Greenwood, T. H. Phan, Y. Fujita, Z. Li, O. Lvasenko, W. Vanderlinden, H. Van Gorp, W. Frederickx, G. Lu, K. Tahara, Y. Tobe, H. Uji-i, S. F. L. Mertens, S. De Feyter, ACS Nano 2015, 9, 5520.
| CrossRef | 1:CAS:528:DC%2BC2MXmvVynsL0%3D&md5=a2dd213cbcfea96fbb2c17530ce09b20CAS |
 L. Verstraete, J. Greenwood, B. E. Hirsch, S. De Feyter, ACS Nano 2016, 10, 10706.
| CrossRef | 1:CAS:528:DC%2BC28Xhs1OnsL7K&md5=a4d881fd598f1e69b5a9e41c9f3045a0CAS |
 T. Menanteau, E. Levillain, T. Breton, Chem. Mater. 2013, 25, 2905.
| CrossRef | 1:CAS:528:DC%2BC3sXhtVChu7fL&md5=212c2f5996eb1590251df6f6a39d7104CAS |
 T. Breton, E. Levillain, T. Menanteau, A. J. Downard, Phys. Chem. Chem. Phys. 2015, 17, 13137.
 T. Menanteau, E. Levillain, T. Breton, Langmuir 2014, 30, 7913.
| CrossRef | 1:CAS:528:DC%2BC2cXpslygsL0%3D&md5=57acaec1b5ef441cf4eb40c2c1adb8acCAS |
 T. Menanteau, C. Benoît, T. Breton, C. Cougnon, Electrochem. Commun. 2016, 63, 70.
| CrossRef | 1:CAS:528:DC%2BC28XjvF2nsg%3D%3D&md5=e1cab90deafa8eec91d406cebe52bf09CAS |
 T. Menanteau, S. Dabos-Seignon, E. Levillain, T. Breton, ChemElectroChem 2017, 4, 278.
| CrossRef | 1:CAS:528:DC%2BC28XitFCmsb7K&md5=5af6d838c5216f8d913844078fd62a3cCAS |
 M. C. R. Gonzalez, A. G. Orive, R. C. Salvarezza, A. H. Creus, Phys. Chem. Chem. Phys. 2016, 18, 1953.
| CrossRef | 1:CAS:528:DC%2BC2MXitVWjsrnJ&md5=ad36ea2af8b06dfd96cfd741046e2e84CAS |
 M. C. R. Gonzalez, P. Carro, L. Vazquez, A. H. Creus, Phys. Chem. Chem. Phys. 2016, 18, 29218.
| CrossRef | 1:CAS:528:DC%2BC28XhsF2mtbbL&md5=beedf0311e39e58f5eac953952096849CAS |
 B. K. Price, J. L. Hudson, J. M. Tour, J. Am. Chem. Soc. 2005, 127, 14867.
| CrossRef | 1:CAS:528:DC%2BD2MXhtVGqs7fP&md5=78e3c98bea57dc059a5c4d7affddac7eCAS |
 P. Actis, G. Caulliez, G. Shul, M. Opallo, M. Mermoux, B. Marcus, R. Boukherroub, S. Szunerits, Langmuir 2008, 24, 6327.
| CrossRef | 1:CAS:528:DC%2BD1cXlvFSnsLw%3D&md5=288727b1de8e991b66a23421781e1e5aCAS |
 J. Ghilane, P. Martin, O. Fontaine, J.-C. Lacroix, H. Randriamahazaka, Electrochem. Commun. 2008, 10, 1060.
| CrossRef | 1:CAS:528:DC%2BD1cXnvVGnurg%3D&md5=9d0291b4acb3ae4bae08e018fb4ca00aCAS |
 O. Fontaine, J. Ghilane, P. Martin, J.-C. Lacroix, H. Randriamahazaka, Langmuir 2010, 26, 18542.
| CrossRef | 1:CAS:528:DC%2BC3cXhsVSisr7J&md5=9a64733669edb90102f255a55a2ec05fCAS |
 G. Shul, C. A. C. Ruiz, D. Rochefort, P. A. Brooksby, D. Bélanger, Electrochim. Acta 2013, 106, 378.
| CrossRef | 1:CAS:528:DC%2BC3sXhtFCisrvE&md5=47efbcd7c31372122fac7ec5402edcb3CAS |
 J. Carvalho?Padilha, J.-M. Noël, J.-F. Bergamini, J. Rault-Berthelot, C. Lagrost, ChemElectroChem 2016, 3, 572.
| CrossRef | 1:CAS:528:DC%2BC28Xot1Chuw%3D%3D&md5=d6b6670e1d1477b7427513beb58602efCAS |