Nanodiamond Surface Redox Chemistry: Influence of Physicochemical Properties on Catalytic Processes

Modification of an electrode with an immobilised layer of nanodiamond is found to significantly enhance the recorded currents for reversible oxidation of ferrocene methanol (FcMeOH). Current enhancement is related to nanodiamond diameter, with enhancement increasing in the order 1000 nm < 250 nm < 100 nm < 10 nm < 5 nm. We attribute the current enhancement to two catalytic processes: i) electron transfer between the solution redox species and redox-active groups on the nanodiamond surface; ii) electron transfer mediated by FcMeOH + adsorbed onto the nanodiamond surface. The first process is pH dependent as it depends on nanodiamond surface functionalities for which electron transfer is coupled to proton transfer. The adsorption-mediated process is observed most readily at slow scan rates and is due to self-exchange between adsorbed FcMeOH + and FcMeOH in solution. FcMeOH + has a strong electrostatic affinity for the nanodiamond surface, as confirmed by in situ infrared (IR) experiments.


Introduction
In recent years electrochemical experiments have been undertaken using electrodes composed of, or modied with, undoped nanodiamond materials. 1Given that bulk diamond is intrinsically insulating, with a band gap of 5.47 eV, it may seem unlikely that undoped diamond can perform as an electrode material.However, unsaturated bonding at the diamond surface and the presence of oxygen functionalities can give rise to surprisingly rich redox chemistry. 2,3For diamond nanomaterials the large surface atom to bulk atom ratio allows surface properties to dominate, hence nanodiamond is able to participate in electrochemical processes despite its insulating nature.2][3][4][5] Most commonly used are detonation nanodiamond powders that have been used to construct cavity electrodes, [5][6][7] sintered into pellet electrodes 8 or dropcoated from solution onto electrodes. 2,3,9Detonation nanodiamond is synthesised by detonation of carbon-based explosives in the absence of oxygen, resulting in a carbonaceous soot from which the nanodiamond powder can be extracted and puried. 10The resulting powder is composed of relatively monodisperse diamond nanoparticles, with average diameters of 5 to 10 nm, as shown in the TEM image in Fig. 1a.Detonation nanodiamond exhibits an array of surface carbon-oxygen functionalities (including ketones, esters, alcohols and acids) that have been characterised using FTIR 11,12 and XPS. 13,14lectrodes modied with an immobilised surface layer of detonation nanodiamond show an enhanced, catalytic electrochemical response towards various analytes such as nitrite, 7 oxygen 15 and haemoglobin, 16 as well as common redox probes like Ru(NH 3 ) 6 3+/2+ and Fe(CN) 6 3À/4À . 2,3,9 is proposed that surface functionalities, defect sites and unsaturated bonding give rise to surface electronic states with energies within the band gap of undoped diamond. 33]9 This mechanism is illustrated schematically in Fig. 1b.Briey, a species such as FcMeOH can undergo oxidation at an electrode to produce FcMeOH + .In the presence of nanodiamond, the FcMeOH + may then be reduced by redox active functionalities on the nanodiamond surface, resulting in regeneration of FcMeOH.This continual regeneration of FcMeOH in the electrode diffusion layer results in an enhanced catalytic current.In situ IR studies of 5 nm diameter nanodiamond in the presence of redox probes have implicated surface quinoneand hydroquinone-like moieties in this redox mediation. 11Catalytic current enhancements at nanodiamond-modied electrodes, via the mechanism shown in Fig. 1b, have previously been reported for charged species such as Ru(NH 3 ) 6 3+/2+ and Fe(CN) 6 3À/4À . 3,9[3]9 However, as the charge on the nanodiamond surface changes with the protonation state of the surface functionalities, electrostatic interactions between the surface and redox probe must be taken into account when carrying out analysis.
In this paper we focus on nanodiamond materials of different size and morphology and their interaction with the outer-sphere and uncharged redox probe FcMeOH.We nd that electrodes modied with nanodiamond exhibit signicantly enhanced catalytic currents for the reversible oxidation of this neutral species compared to charged redox species.We investigate the electrochemical response when changing the diameter of the nanodiamond particles used and the pH and ionic strength of the electrolyte solution.Analysing the effect of these physicochemical parameters, alongside factors such as redox species concentration and voltammetric scan rate, provides further insight into the mechanisms of current enhancement and nanodiamond surface properties.

Equipment
All electrochemical experiments were carried out using a mAutolab PGSTAT potentiostat (Eco Chemie, Utrecht, Netherlands) running GPES (v4.9), or a CHI910B Scanning Electrochemical Microscope/bipotentiostat (CH Instruments, Inc. Austin, Texas, USA) running CHI (v6.28).Electrochemistry was performed using a three electrode setup.The working electrode was a boron-doped diamond disk (d ¼ 3.0 mm), press tted into a PEEK body (Windsor Scientic, Slough, United Kingdom).The counter electrode was fabricated in-house from platinum wire and sheet, cut into ags of approximately 4 cm 2 .A stable reference potential was achieved by using an Ag/AgCl (sat.KCl) RE-5B reference half-cell (BASi, West Lafayette, America).
A Carbolite MTF 1200 horizontal tube furnace was used to pre-treat and clean all nanodiamond samples.The temperature control unit and associated electronics were designed and built in house.Mid-infrared spectra were recorded (100 scans) in attenuated total reectance (ATR) mode using a Bruker Tensor 27 Fourier transform infrared (FTIR) spectrophotometer (Bruker, Coventry, United Kingdom) tted with a room temperature DLaTGS detector at 4 cm À1 resolution.The Platinum ATR accessory consisted of a diamond prism operating with one reection.Background spectra were collected prior to each experiment (100 scans).Transmission electron micrograph (TEM) images were recorded using a Jeol JEM 2100 TEM with a 200 kV accelerating voltage using a LaB6 lament.All nanoparticles were deposited from the same solutions used for modifying electrode surfaces.Carbon coated copper TEM grids were used as the nanoparticle support.
All nanodiamond samples were treated in a furnace at 425 C for 4 h (including temperature ramp times, which were set at 10 C min À1 ) in air.This was to remove any sp 2 graphitic carbon from the surface and to maximise the number of oxygen terminating functional groups. 17Nanodiamond suspensions were prepared by adding nanodiamond to puried water, centrifuging twice (13 200 rpm for 10 min) and briey sonicating to ensure large aggregates were broken up.The mass per volume concentrations of the various nanodiamond solutions were: 5 nm ND 1.50 mg ml À1 ; 10 nm ND 1.50 mg ml À1 ; 100 nm ND 1.55 mg ml À1 ; 250 nm ND 1.60 mg ml À1 and 1000 nm ND 1.7 mg ml À1 .

Preparation of nanodiamond-modied electrodes
Nanodiamond lms were drop coated onto clean, dry electrodes by pipetting the various nanodiamond suspensions (2.5 ml) onto the electrode surface.The droplet was positioned and manoeuvred to ensure complete coverage of the electrode's surface.The nanodiamond coated electrode was le to air dry for 20 min and checked under an optical microscope to ensure even and complete coverage.

Characterisation of nanodiamond materials using FTIR and TEM
Fig. 2 shows representative FTIR spectra and TEM micrographs for two of the nanodiamond samples used in this study: 5 nm detonation nanodiamond and 100 nm microcrystalline nanodiamond (produced by mechanical grinding of synthetic diamond).Two different detonation nanodiamond materials were studied and examination with TEM (Fig. 1a, Fig. 2b) revealed the powder to be composed of discrete spherical and relatively monodisperse particles; one sample with mean particle diameter 5 nm and another with a slightly larger particle diameter closer to 10 nm.In both samples the individual nanoparticles aggregate together to form larger agglomerates.The surface of all of the nanodiamond samples used is highly oxidised due to the pretreatment procedure of heating in air, described in section 2.2 above.The FTIR spectrum of the 5 nm nanodiamond is complex, with a great number of different functionalities contributing to the vibrational spectrum.These functionalities have been characterised and discussed in detail previously, 11,12 and include alcohols and carbonyl species: 3393 cm À1 (nOH), 2926 cm À1 (nCH 2 ), 2847 cm À1 (nCH), 1729 cm À1 (nC]O) and 1634 cm À1 (nOH Abs. ).The convoluted set of peaks between 1000-1500 cm À1 contains a mix of bending and stretching modes of various C-C, C-H and C-O containing functionalities.Of most interest to the present study is the presence of redox-active functionalities such as unsaturated ketone or 'quinone-like' moieties that can participate in redox cycling. 11he 100, 250 and 1000 nm samples have pronounced crystal facets, consistent with nanodiamond produced through mechanical grinding of High Temperature High Pressure synthesis diamond. 18Analysis of the size distribution of the 100 nm nanodiamond sample gives a mean particle size of 98 nm with a standard deviation of 35 nm; the modal average size is ca.100-120 nm.
The crystalline nanodiamond samples all show a large absorption peak at ca. 1100 cm À1 in the IR spectra, attributable to nitrogen centres at defect sites within the bulk material. 19The IR spectra of these larger nanodiamonds also show absorption bands attributed to surface functionalities at 3394 cm À1 (nOH), 2923 cm À1 (nCH 2 ), 2855 cm À1 (nCH), 1790 cm À1 (nC]O), 1631 cm À1 (nOH Abs. ) and 1103 cm À1 (nN-V À ), demonstrating the presence of similar surface functionalities to those on the detonation nanodiamond particles.Although the nanodiamond samples were produced by different synthetic methods, making direct comparison difficult, the oxidative pre-treatment results in similar surface oxygen functionalities, allowing us to anticipate similar electrochemical properties.

Cyclic voltammetry (CV) investigations of nanodiamond-modied electrodes
3.2.1The effect of nanodiamond particle diameter on the voltammetric response of FcMeOH.Fig. 3a shows the cyclic voltammograms (CV) of 1 mM FcMeOH at a clean boron doped diamond (BDD) electrode (black line) and a BDD electrode modied with 100 nm nanodiamond particles (red/blue lines).Both the The 1000 nm nanodiamond has a negligible effect on the electrochemical response, the CV being essentially identical to that at the clean BDD.Enhanced currents are observed at the other electrodes, with peak currents increasing in the order 250 nm < 100 nm < 10 nm < 5 nm.Oxidation and reduction peaks are enhanced equally in each case.The degree of current enhancement at the 10 nm and 5 nm nanodiamond-modied electrodes is quite dramatic.This has been quantied in Fig. 3c, where a 'charge enhancement ratio' has been calculated for each nanodiamond diameter.This is dened as the charge passed during oxidation or reduction at the nanodiamond-modied electrode divided by that passed at the clean BDD electrode under the same conditions, i.e.Q ND-BDD /Q BDD .Charge passed during the oxidation of FcMeOH or reduction of FcMeOH + is increased by a factor of ten when the electrode is modied with a layer of 5 nm or 10 nm nanodiamond.A four-fold enhancement is noted for the 100 nm particles, falling to two-fold for the 250 nm nanodiamond, and no enhancement is seen for an electrode modied with 1000 nm particles.
In comparison to previous studies using the Fe(CN) 6 3À/4À redox probe, 9 the current enhancements for FcMeOH are considerably higher.This is shown in Fig. 3d, where the charge enhancement ratio for Fe(CN) 6 4À oxidation is compared to that for FcMeOH oxidation at electrodes modied with different diameter nanodiamond.The data shows the same relationship between charge enhancement and nanodiamond diameter for both redox species, however, the magnitude of enhancement shows a strong dependence on the chemical nature of the redox species, with FcMeOH currents being signicantly more enhanced.enhanced at the 100 nm nanodiamond-modied electrodes over all pH values, but the enhancement is pH dependent with I p increasing as the pH is lowered.As for Fig. 3a, DE p for the nanodiamond-modied electrode is $30 mV over the whole pH range.I p vs. n analysis suggests mixed diffusion controlled and adsorption behaviour.
Each of the CVs at the nanodiamond-modied electrodes exhibits a non-zero 'off-set' reduction current at 0 V.At this potential there should be no species in the solution available to be reduced (only FcMeOH is present and this can only be oxidised).We have reported this phenomenon previously for 5 nm nanodiamond-modied electrodes in solutions containing Fe(CN) 6

4À
. 9 It is attributed in the present case to reduction of FcMeOH + species within the diffusion layer at the electrode surface.FcMeOH + is proposed to be generated spontaneously by the oxidation of FcMeOH at the nanodiamond surface (Fig. 4c).This oxidation takes place through electron exchange between nanodiamond surface functional groups, which are reduced concomitantly with the generation of FcMeOH + .The magnitude of the off-set current is pH dependent as shown in Fig. 4b.At pH 5, signicant initial reduction currents are observed, but these become negligible at pH 8 and above.
3.2.3.The effect of solution ionic strength on the CV response of FcMeOH at nanodiamond-modied electrodes.The CVs for 1 mM FcMeOH at a 100 nm nanodiamond-modied BDD electrode in electrolytes of varying ionic strength are shown in Fig. 5.The ionic strength was varied by altering the concentration of the KCl electrolyte across ve orders of magnitude.As the ionic strength is decreased, a pre-peak emerges from the oxidation wave, which shis to more negative potentials with decreasing ionic strength (increasing the pre-peak to main-peak separation).The height of the main oxidation wave is decreased with decreasing ionic strength, while that of the reduction peak increases.In CV experiments the emergence of a pre-peak to an oxidation peak is usually taken as evidence of outer-sphere adsorption of the oxidation product at the electrode surface.In this case, the oxidation product FcMeOH + is likely to experience a positive electrostatic interaction with the negatively charged nanodiamond surface. 20,21Decreasing the ionic strength of the medium will increase the extent of adsorption, as the length to which the electric eld of a surface can be felt is inversely proportional to the concentration of electrolyte (as described by DLVO theory). 22,23This is consistent with the observations in Fig. 5b, where the adsorption pre-peak can be seen to become more prominent as ionic strength is decreased and hence more FcMeOH + adsorbs onto the nanodiamond surface.
3.2.4.The effect of CV scan rate on the CV response of FcMeOH at nanodiamond-modied electrodes.The studies described in this section were carried out using 5 nm nanodiamond modied electrodes, as these showed the most extreme current enhancements.CVs were carried out in 1 mM and 10 mM FcMeOH over scan rates of 1-500 mV s À1 (shown in Fig. 6, where currents have been normalised for clarity).Below scan rates of 10 mV s À1 , two oxidation waves are clearly observed; one wave is positioned at the same E p as the diffusion-controlled oxidation peak at the clean BDD electrode, and the other is a pre-peak to this main wave.A pre-peak to the reduction wave is also observed.As the scan rate is increased the two waves merge to form one broad peak located midway between the pre-peak and main peak positions.At scan rates above ca.100 mV s À1 only one peak is observed, at the position expected for a simple diffusion-controlled mechanism.However, at very low concentrations of FcMeOH (1 mM), at the faster scan rates, this peak becomes a steady state current response, indicating that oxidation is no longer taking place under diffusion-control.

In situ ATR-IR spectroscopy of 5 nm nanodiamond in the presence of FcMeOH
The electrochemical data is strongly suggestive of adsorption, of the oxidation product FcMeOH + in particular, at the nanodiamond surface.In situ ATR-FTIR spectroscopy experiments were therefore performed to monitor changes in the IR absorption at the nanodiamond/solution interface in solutions containing FcMeOH or FcMeOH + .The experiments were performed by drop-casting a 5 nm nanodiamond lm onto the ATR crystal.An enclosed liquid cell was clamped over the nanodiamond modied ATR crystal and gently lled with water to solvate the nanodiamond layer.Aer allowing time to equilibrate, a small volume of FcMeOH or FcMeOH + solution was gently introduced to the cell, and the FTIR spectrum between 3500-500 cm À1 was recorded over 120 min.Fig. 7 shows the IR difference spectra obtained aer 120 min of nanodiamond lm immersion in the FcMeOH FcMeOH + solution, along with a spectrum of the solid FcMeOH for comparison.
In the absence of nanodiamond, there is negligible change over 120 min in the IR spectra of the ATR crystal in FcMeOH solution.With the nanodiamond lm, the presence of FcMeOH produces increases in infrared absorption at ca. 3200-3300 cm À1 (nOH) and 1650 cm À1 (dOH), i.e. predominantly in regions associated with OH vibrational modes.There are a greater number of changes observed for the nanodiamond lm exposed to FcMeOH + solution.These occur in regions associated with OH, C-H and C-O stretches and vibrations: 2925 cm À1 (nCH 2 , asym.), 2845 cm À1 (nCH 2 , asym.), 1648 cm À1 (dOH), 1576-1393 cm À1 (d/uCH 2 ) and 1044 (nC-C-O), showing many features in common with the solid FcMeOH spectrum. 24

Mechanism of the catalytic current enhancement
With the exception of 1000 nm nanodiamond, modication of the BDD electrode with an immobilised layer of nanodiamond results in enhanced CV currents for the reversible oxidation of FcMeOH.Catalytic current enhancements have previously been reported for the redox probes Fe(CN) 6 3À/4À , IrCl 6 2À/3À and Ru(NH 3 ) 6 2+/3+ , among others; 1-3,6,9 however, currents achieved here for the neutral FcMeOH species are considerably higher than those observed for charged redox molecules.The current enhancement is partly attributed to a catalytic feedback mechanism, as shown in Fig. 1b.In this scheme, FcMeOH undergoes oxidation at the BDD electrode to produce FcMeOH + .FcMeOH is then regenerated rapidly by the reduction of FcMeOH + at the nanodiamond interface.On the CV return scan the opposite process is proposed to take place; FcMeOH generated at the electrode is immediately oxidised back to FcMeOH + aer donating an electron to the nanodiamond surface.For this mechanism to be feasible, the oxidation and reduction of nanodiamond surface functionalities must be at similar potentials to the standard potential of the solution redox probe.Electrode-immobilised 5 nm detonation nanodiamond has been shown to exhibit reversible redox peaks in this potential range, which shi potential with changing solution pH. 9 In addition, in situ IR studies have provided evidence that 'quinone-like' surface moieties on the nanodiamond are involved in the redox cycling. 11Although the 100 nm and 250 nm diamond particles used in this study are synthesised by different methods to the 5 nm nanodiamond and possess different morphologies, the observed current enhancements suggest similar surface species on each that can contribute to a catalytic reaction mechanism.All of the samples were oxidised prior to use and among the oxygen functionalities present it is likely that 'quinone-like' and 'hydroquinone-like' species are present on each to support the catalytic cycle in Fig. 1b.

Relationship between the particle size and the current enhancement
In this study an inverse relationship between particle size and observed current enhancement was observed for both the oxidation of FcMeOH and reduction of FcMeOH + in the presence of 5, 10, 100 and 250 nm diameter nanodiamonds.The extent of the enhancement observed, quantied as the 'charge enhancement ratio', depends on the availability of nanodiamond surface functionalities to support the feedback mechanism.The concentration of these functionalities will depend on the nanodiamond particle size as well as the surface density of these groups.This latter factor will vary with oxidation pre-treatment, as well as the type of exposed crystal facets, edge sites and the defect density.A log-log plot of charge enhancement ratio with nanodiamond diameter produces a straight line, suggesting that a monomial power law relates these two properties.A straightforward relationship between the surface area to volume ratio of the particles and the charge enhancement ratio (s) should result from a s ¼ 3r À1 t to the data (where r is the particle radius).However, a more complex relationship is found.Fractal scaling could be one explanation for the unusual relationship obtained, and it should also be noted that the 100 nm and 250 nm samples are not spherical, like the 5 and 10 nm samples.Most likely, the density of redox active sites on the different diamond samples differs.The highly active detonation nanodiamond surface is very defective and has much in common with amorphous carbons; it exhibits an incredibly diverse and dense coverage of functionalities.In contrast, the 100 nm and 250 nm samples are more 'diamondlike' in their character and thus the surface defect density and degree of sp 2 bonding is much lower, resulting in fewer active sites.The 1000 nm diamond does not support the catalytic current enhancement (although neither does it block electrode activity).It is proposed that there are insufficient surface groups present on this material to produce a detectable catalytic current.The same mechanism is likely to take place at the surface of the 1000 nm diamond, however, the surface atom to bulk atom ratio is much smaller, so currents generated via this mechanism will be negligible in comparison to the diffusion controlled currents.

pH dependence of the current enhancement
Current enhancement with 100 nm nanodiamond-modied electrodes was found to increase in the order pH 9 < pH 8 < pH 7 < pH 6 < pH 5. The redox response of FcMeOH at the BDD electrode is independent of pH, hence the ability of the nanodiamond to contribute to the catalytic cyclic must be the proton dependent step.5 nm nanodiamond exhibits reversible redox peaks in voltammetry that shi by ca.60 mV to more negative potentials with each increase in pH unit.This is typical of a coupled reaction of one proton per electron transferred and consistent with a quinone/hydroquinone redox transformation, which in aqueous solution is a 2H + /2e À reaction. 9In the present study, we propose that similar surface species are present on the 100 nm nanodiamond and that the redox potentials of these functionalities are similarly pH dependent.The pH 5 response suggests the potentials of the surface species are coincident with the potential of the FcMeOH/FcMeOH + couple at this pH.This match allows the most efficient electron transfer to take place between the nanodiamond and the solution species and hence the current enhancement is greater.As the pH is increased, the potential of the surface functionalities moves towards more negative potentials (according to Nernstian analysis of the proton dependence of the process), while that of the FcMeOH/FcMeOH + couple is unchanged.As the potentials of the solution and surface species are not so well matched, the driving force for the catalytic process is decreased and the current enhancement is not so pronounced.Hence smaller current enhancements are noted at pH 9.

pH dependence of the magnitude of the off-set current
Further evidence for pH dependent electron transfer between the nanodiamond surface groups and the FcMeOH species comes from inspection of the magnitude of the off-set reduction current at 0 V in Fig. 4. As described above, we attribute this current to the reduction of FcMeOH + present in the diffusion layer.The FcMeOH + is proposed to be generated by the spontaneous oxidation of FcMeOH at the nanodiamond (ND) surface in the absence of an applied potential: FcMeOH + ND ox./ FcMeOH + + ND red.
For such a process to be thermodynamically feasible, the nanodiamond surface functionalities must undergo reduction at a higher potential than the potential of the FcMeOH/FcMeOH + couple.The magnitude of the reduction current for the FcMeOH + species suggests this is the case at pH < 8 with the largest driving force for the process occurring at the lower pH values.At pH 8 and above the reduction currents are very small, indicating that electron transfer between the FcMeOH and the nanodiamond is negligible.This is consistent with the proton dependence of the potential of the surface functionalities.As pH increases, the potential shis to more negative values and by pH 8 is proposed to be found below that of the FcMeOH/ FcMeOH + couple, hence spontaneous electron transfer does not take place.

Ionic strength dependence of the voltammetric response
At low ionic strengths, a pre-peak to both the main oxidation and reduction waves becomes evident.This, along with the observed increased width of the voltammetric peaks in the presence of nanodiamond, as well as the decreased peak to peak separation (ca.30 mV), is diagnostic of adsorption of both FcMeOH and FcMeOH + at the electrode surface.As no evidence for adsorption is observed at the clean BDD electrode it is assumed that there is a strong affinity of the FcMeOH and FcMeOH + species for the nanodiamond surface.The positively charged FcMeOH + can clearly interact via electrostatic interactions with the negatively charged surface oxygen functionalities.Some current enhancement for the Ru(NH 3 ) 6 3+ species at the 5 nm nanodiamond modied electrodes has previously been attributed to such an interaction.The strong adsorption of the neutral FcMeOH is more surprising as the nanodiamond surface is considered quite hydrophilic and less likely to interact with a largely hydrophobic molecule.However, the adsorption of hydrophobic species to oxidised nanodiamond has previously been reported and in fact exploited for the delivery of water-insoluble drugs. 25The sp 2 bonding present at the nanodiamond surface provides ample opportunity for hydrophobic and p-p stacking interactions with the aromatic pentacyclo-moieties of the FcMeOH molecule.

Contribution of the adsorption processes to the enhanced current response
The currents for FcMeOH oxidation at the nanodiamond-modied electrodes are much greater than observed previously for the Fe(CN) 6 4À species.As the potential of the two species is similar, the same redox interactions with the nanodiamond surface groups are proposed.The difference in current enhancement is attributed to the extent to which the redox molecule can approach closely to the surface of the nanodiamond and the resulting efficiency of electron tunnelling between the two.The very negatively charged Fe(CN) 6 4À is unlikely to be able to approach closely to the surface due to electrostatic repulsion, hence in the absence of any other factors it is not surprising that FcMeOH shows a higher current response that Fe(CN) 6 4À .However, the adsorption behaviour of FcMeOH and FcMeOH + is clearly a contributor to the very signicant and unexpectedly large current enhancements we observe.As stated above, FcMeOH undergoes spontaneous oxidation at the nanodiamond surface to generate FcMeOH + .This positively charged species will interact very strongly with the surface and become adsorbed.The strong interaction is illustrated by the affinity of the FcMeOH + for the nanodiamond surface, as found in the IR studies in Fig. 7.The contribution of the Gibbs energy of adsorption can shi the redox potentials of an adsorbed species relative to the solution species and results in the observation of pre-peaks in the CV. 26,27Here we propose that the surface adsorbed FcMeOH + also participates in the catalytic feedback mechanism, as shown in Fig. 8. Solution FcMeOH undergoes electron exchange with adsorbed FcMeOH + to generate solution FcMeOH + and adsorbed FcMeOH.The adsorbed FcMeOH can be re-oxidised to adsorbed FcMeOH + by the underlying BDD electrode, or by a redox process with the nanodiamond surface groups.Thus, in addition to the catalytic currents generated by the mechanism shown in Fig. 1b, the adsorbed species also mediate a catalytic process.The accumulative inuence of both processes explains why currents are so signicantly enhanced for this solution species compared to others previously studied.

Scan rate dependence of the voltammetric response
The scan rate dependence of the response illustrates further the different processes contributing to the catalytic currents.At slow scan rates, the pre-peak for the adsorption mediated process is very distinct from the main peak during the oxidation process.The two catalytic processes take place at distinct potentials; rst via the adsorbed FcMeOH + species and then via the nanodiamond surface groups, indicating that adsorbed FcMeOH + is easier to reduce than the nanodiamond surface under these conditions.The factor determining the rate of the adsorption-mediated mechanism is the availability of adsorbed FcMeOH + to oxidise the solution FcMeOH.Assuming the rate of regeneration of adsorbed FcMeOH + by reduction at the BDD electrode or nanodiamond surface is fast enough, the limit to this reaction is simply how much FcMeOH + is adsorbed at the nanodiamond interface.The adsorption-mediated process is so distinctly observed at slower scan rates because the ux of FcMeOH is smaller under these conditions, and hence mass transport of reactant can keep up with the rate of regeneration of the available FcMeOH + .At faster scan rates, the higher ux of solution FcMeOH leads to diffusion-controlled currents overwhelming the adsorption response and the pre-peak merges with the main diffusion-controlled peak.At higher scan rates and low concentrations a steady-state current begins to emerge (Fig. 6a).In this regime, we assume that catalysis takes place largely by the nanodiamond-mediated mechanism shown in Fig. 1b.The steady state current shows that the rate of FcMeOH + generation at the BDD electrode is exactly matched by the rate of FcMeOH + reduction by the nanodiamond surface functionalities.At slow scan rates a peak response is obtained for this process, as a thick diffusion layer containing FcMeOH + builds up and there are insufficient nanodiamond surface functionalities to maintain the catalytic cycle.At faster scan rates the diffusion layer is much thinner, so there is less FcMeOH + for the nanodiamond to reduce, hence the rate of electron transfer from the nanodiamond can keep up with the rate of FcMeOH + generation, leading to a steady state current response.

Conclusions
Modication of a BDD electrode with an immobilised layer of nanodiamond was found to enhance currents for the reversible oxidation of FcMeOH.Current enhancement is inversely related to nanodiamond diameter, with enhancement increasing in the order 1000 nm < 250 nm < 100 nm < 10 nm < 5 nm.In comparison to previous studies using negatively charged redox probes, we nd that currents are much larger for the neutral species FcMeOH under the same experimental conditions.We attribute the current enhancement to two catalytic processes: i) electron transfer between the solution redox species and redox active groups on the nanodiamond surface; ii) electron transfer mediated by FcMeOH + adsorbed onto the nanodiamond surface.The rst process is pH dependent as it is believed that 'quinone-like' and 'hydroquinone-like' species on the nanodiamond surface mediate the redox process and thus it is coupled to proton transfer.The FcMeOH/FcMeOH + redox cycle is more efficiently catalysed than the Fe(CN) 6 4À/3À transformation reported previously, as the neutral and positive redox species are attracted rather than repelled by the negatively charged nanodiamond surface.The adsorption mediated process is observed most readily at slow scan rates and is due to electron self-exchange between adsorbed FcMeOH + and FcMeOH in solution.FcMeOH + has a strong affinity for the nanodiamond surface, as conrmed by in situ ATR-FTIR experiments.
This study provides insight into the interaction of solution species with diamond and carbon surfaces under different physicochemical conditions.It also reveals mechanistic details of catalytic processes associated with the surfaces of nanodiamond materials.Catalytic currents within electrochemistry are of interest as they can allow energy intensive reactions to proceed at lower overpotentials (or allow a greater rate of reaction at the same overpotential).Such catalysts have applications in important sectors such as synthesis, energy conversion and molecular detection.Due to the particular mechanism of the nanodiamond electrocatalysis, nanodiamonds could nd use within sensing devices, for example allowing detection of analytes at very low concentrations.

Fig. 1
Fig. 1 (a) TEM micrograph of 10 nm diameter nanodiamond.(b) Illustration of the nanodiamond feedback mechanism.The green arrows depict the direction of electron transfer.
3.2.2.The effect of the solution pH on the CV response of FcMeOH at nanodiamond-modied electrodes.CVs for 1 mM FcMeOH at clean BDD and 100 nm nanodiamond-modied BDD electrodes in different solutions of pH 5-9 are shown in Fig. 4a.At the BDD electrode the CV response of FcMeOH is independent of pH, producing voltammograms with DE p ca. 60 mV, consistent with oneelectron, reversible, diffusion-controlled electrochemistry.The currents are

Fig. 4
Fig. 4 (a) Cyclic voltammograms (1 st scan) of FcMeOH (1 mM) in aqueous solutions of K 2 HPO 4 -KH 2 PO 4 (0.1 M) in various ratios to produce pH 5-9 electrolytes, at the clean BDD electrode (black) and nanodiamond modified BDD electrode (red), recorded at 5 mV s À1 .(b) Initial off-set currents measured with different electrolytes of pH 5-9 (data collected from the 1 st scan of the experiments displayed in (a)).(c) Illustration of FcMeOH oxidation performed by the nanodiamond.The green arrow depicts the direction of electron transfer.

Fig. 8
Fig. 8 Illustration of nanodiamond surface adsorbed FcMeOH + and its associated redox reaction.Re-oxidation of the FcMeOH product can then occur by either (1) the underlying electrode (when at sufficient potential) or (2) reaction with nanodiamond surface functionalities.The green arrows depict the direction of electron transfer.