Carbon 102 (2016) 437e454

Contents lists available at ScienceDirect

Carbon

journal homepage: www.elsevier.com/locate/carbon

Review article Diamond for neural interfacing: A review

* David J. Garrett a, , Wei Tong a, David A. Simpson a, Hamish Meffin b a School of Physics, The University of Melbourne, Victoria 3010, Australia b National Vision Research Institute, Department of Optometry and Vision Science University of Melbourne, Victoria 3010, Australia article info abstract

Article history: The variety of materials that fall in the diamond category is vast and increasing. From single, to micro to Received 9 December 2015 nano-crystalline, diamond can be highly electrically insulating or synthesized to possess varying degrees Received in revised form of conductivity. Diamond in all its forms is without peer in terms of chemical stability exhibits excellent 3 February 2016 biocompatibility and is famously mechanically robust. Diamond is however, a difficult material with Accepted 20 February 2016 which to fabricated devices owing to its extreme hardness, lack of ductility and weldability. New syn- Available online 26 February 2016 thesis and fabrication methods in recent decades have overcome some of these drawbacks and diamond has enjoyed a surge in interest as a biomedical material. In the field of neural interfaces a grand goal is permanent, high fidelity connections with neural populations. Diamond's longevity, biocompatibility and biochemical inertness make it a highly promising material with which to achieve this goal. This review covers recent uses of diamond in the three critical areas of neural interfacing: diamond as a growth substrate, as a neurochemical sensor electrode material and as a direct neural recording and stimulation material. Looking towards the future, the enticing prospect of nano-diamonds as optically active neural interfaces is reviewed. © 2016 Elsevier Ltd. All rights reserved.

Contents

1. Introduction ...... 438 2. Synthetic diamond ...... 438 2.1. Synthesis methods and products ...... 438 2.1.1. High-Pressure High-Temperature ...... 438 2.1.2. Chemical vapor deposition ...... 438 2.1.3. Characterization of diamond films ...... 439 2.2. Electrochemical properties of diamond and diamond films ...... 439 2.2.1. Boron-doped diamond ...... 439 2.2.2. Nitrogen-included ultrananocrystalline diamond (N-UNCD) ...... 441 3. Growth of neurons on diamond ...... 441 3.1. Effects of surface properties ...... 441 3.2. Necessity of adhesion promoter coatings ...... 443 4. Diamond electrode and diamond microelectrode array (MEA) fabrication methods ...... 444 5. Neurochemical sensors ...... 445 5.1. Electrochemical detection of neurotransmitters ...... 445 5.2. In vitro performance of diamond for electrochemical detection of neurotransmitters ...... 446 5.3. Neurochemical detection in living systems ...... 446 6. Diamond devices for neural stimulation or neural recording ...... 447 6.1. Neural recording ...... 449 6.2. Neural stimulation ...... 450 7. Future prospects ...... 450

* Corresponding author. E-mail address: [email protected] (D.J. Garrett). http://dx.doi.org/10.1016/j.carbon.2016.02.059 0008-6223/© 2016 Elsevier Ltd. All rights reserved. 438 D.J. Garrett et al. / Carbon 102 (2016) 437e454

8. Conclusions ...... 451 Acknowledgments ...... 452 References ...... 452

1. Introduction material has steadily increased. Initially interest was confined to diamond as a biocompatible growth substrate for various cell types Human fascination with diamonds goes back millennia owing to including neurons [20,21]. Invariably diamond is found to be non- the unique set of optical and physical properties that diamonds cytotoxic and in some cases has been shown to improve the possess and the extreme rarity of the naturally occurring gem stone viability of some cell types over those cultured or stored adjacent to [1]. The first synthetic (High-Pressure High-Temperature (HPHT)) more traditional materials. For instance, in one study, sperm stored diamonds were produced in the 1950s leading rapidly to com- in diamond coated petri dishes exhibit better viability than sperm mercial scale diamond manufacturing in the same decade. Cheap stored in conventional polystyrene petri dishes [22]. Recently synthetic diamonds led to an explosion in interest in the material, prototype implantable devices featuring predominantly diamond principally as an ultra-hard grit for cutting tools. Diamond cutting construction or at least diamond electrodes have been described tools are now produced by a wide variety of methods and are arising from the retinal prosthesis community [23,24]. Diamond inexpensive and common place [2]. Low pressure diamond growth has also received particular attention from the retinal prosthesis also has its origins in the 1950s [3]. Low pressure growth methods research community as a material for hermetic encapsulation of however did not gain real traction until the 1980s when Japanese implanted electronics [25]. scientists Seiichiro Matsumoto and Nobuo Sekata optimized a This review will cover research conducted during the previous number of chemical vapor deposition (CVD) methods to produce ten years featuring diamond as a neural growth substrate, diamond diamond [4e7]. Unlike HPHT diamonds, CVD diamond could be electrodes for electrochemical sensing of neurochemicals such as grown conformally on surfaces as a microcrystalline or nano- dopamine and diamond for direct detection of neuron activity or crystalline film. Such films found use as wear resistant coatings for electrically stimulating neural tissue. Special attention will be [8,9]. The very low chemical reactivity of diamond provides the focused towards devices where diamond is used substantially as a added advantage that such films are also highly effective at pre- construction material for neural interfacing. venting corrosion [10,11]. The first medical use of diamond was as a fi wear resistant coating for arti cial hip joint parts [11,12]. Anticor- 2. Synthetic diamond rosion coating of cardiovascular stents and dentistry components fi with diamond lms has also been investigated [11]. As part of the 2.1. Synthesis methods and products due diligence of developing diamond coated implants, the cyto- toxicity of diamond and diamond like coatings has been extensively 2.1.1. High-Pressure High-Temperature fi investigated. Invariably diamond and diamond like lms exhibit no The HPHT system is inspired by the process by which diamonds measurable cytotoxicity [11]. Similarly nano-diamonds, which are form naturally within the earth. Seed diamonds are placed in a attractive as drug delivery vehicles [13], are generally considered to chamber with a high-purity carbon source and a metal alloy sol- be biocompatible although the chemical surface termination has vent. The chamber is heated to a temperature around 1400 C, been shown to impact suitability for biomedical applications [14]. melting the metal solvent. Hydraulic pistons apply high pressure to Paget et al. for instance described no cytotoxicity or genotoxicity the chamber. Carbon from the high-purity source dissolves and e towards a range of human cell lines for COOH terminated nano- diffuses through the metal solvent. Dissolved carbon precipitates diamond [15] whereas Marcon et al. showed marked cytotoxicity of onto the diamond seed, as the pressure is released, adopting the NH2 terminated nanodiamonds towards human embryonic kidney crystal structure of the seed. HPHT diamonds treated by irradiation. cells [16]. Hence diamond-like coatings have been used in a variety Irradiation induces lattice defects in the diamond generating colour of areas, including food packaging and even processes to coat effects. fi plastic materials with diamond-like carbon lms [17]. Graphite is thermodynamically more stable than diamond but A parallel research path that arose as a result of the new ability the tetrahedral (or face centered cubic) stacking in diamond has a fi to form diamond lms was the intense study of the electronic higher atomic density, hence diamond is favored at very high properties of diamond. Pure diamond possesses a wide band gap pressure. HPHT techniques are also commonly used to treat natural and is therefore an electrical insulator. As CVD diamonds are grown diamonds in order to improve clarity or change color [26]. Dopants from a gas mixture, the inclusion of non-carbon impurities into such as boron can be included in the carbon source producing synthetic diamond is straight forward. The introduction of boron electrically conducting HPHT diamonds [27] and a variety of doping to generate electrically conducting forms of diamond treatments are available to generate coloured gem stones. For increased the potential uses of diamond dramatically [18]. Boron further reading on HPHT see the following reference. [28]. doped diamond (BDD) is a p-type semiconductor and can exhibit a wide range of conductivities depending on the synthesis method 2.1.2. Chemical vapor deposition and doping level. Superconductivity at low temperatures has even The CVD process, like HPHT, also employs diamond seeds upon been reported [19]. Recognition that conducting diamond could be which an activated carbon source precipitates, increasing the size of employed to interface with neural tissue swiftly followed. the crystal. In its simplest form, synthesis by CVD requires a dia- Compared to the general acceptance of noble metals such as plat- mond seed, a carbon-containing feed-gas and a method of acti- inum, the uptake of diamond in real world neuro-modulation de- vating the feed-gas mixture to generate reactive carbon and vices has been slow, perhaps due to challenges involved in hydrogen species. In all CVD processes the presence of atomic fabrication of devices from such an extremely hard material. In the hydrogen in the gas mixture is the factor that dictates the formation past decade however interest in diamond as a neural interface of diamond over graphitic forms of carbon [29,30]. Atomic D.J. Garrett et al. / Carbon 102 (2016) 437e454 439 hydrogen is capable of reducing the double bonds in graphitic The gas mixture in the diamond growth environment also dic- structures; thus, under the correct conditions, graphitic material is tates the chemical termination of the as produced diamond surface. preferentially etched and diamond growth can dominate [31]. Post processing by plasma or acid cleaning tend to radically alter Many different carbon sources have been used to grow CVD di- the surface termination. For instance boiling in acidic solutions, amonds, from the most mundane (e.g. methane) to the highly electrochemical oxidation or treatment with an oxygen plasma will exotic [32]. Analysis of the gases present in CVD chambers reveal terminate diamond with oxygen functionalities [40,42]. The rela- that the carbon decomposition species most commonly observed tive proportions of different oxygen functionalities (C]O, CeOH, are methane radicals, CH3 [30,33,34] and acetylene, C2H2 [30,34] COOH) are affected by the oxidation method employed [42]. The though many other carbon species have been identified [34]. majority of PCD films are grown in a hydrogen rich environment The two most common methods of synthesis by CVD are hot- and, as a result, the diamonds are hydrogen terminated from the filament CVD and plasma enhanced CVD [29]. Hot-filament em- reactor. Hydrogen terminated diamond surfaces are electrically ploys a thin, high melting temperature wire situated close to the conducting via an unusual surface transfer doping mechanism seed diamonds. A feed-gas mixture is introduced over the filament. [43,44]. Conversely oxygen terminated diamond exhibits immea- The filament is heated to 2000e2500 C simultaneously activating surably low surface conductivity [44]. Of greater significance in the gas mixture and heating the seed diamonds. In plasma terms or neuron interactions is the relative hydrophobicity and/or enhanced CVD an energy source such as microwave radiation is surface charge of the diamond. Oxygen terminated diamond is focused into the gas mixture to form a high energy plasma in close more hydrophilic than hydrogenated diamond films. In a biological proximity to diamond seed crystals. For detailed discussion of environment hydrophilic films tend to be a favored substrate for diamond synthesis by CVD see the following references. [34e36]. cellular attachment [45].

2.1.3. Characterization of diamond films 2.2. Electrochemical properties of diamond and diamond films Synthetic diamonds are grown as standalone single crystals or as polycrystalline films or blocks. In general polycrystalline dia- 2.2.1. Boron-doped diamond mond materials are named according to the size of the diamonds When initially considering a new electrode material for any from which the material is constructed. The grades polycrystalline electrochemical application, three electrochemical material quali- diamond materials are summarized in Table 1. For further reading ties are typically evaluated, namely; corrosion resistance, back- see CVD Diamond for Electronic Devices and Sensors, edited by Ric- ground current and usable potential window [46]. Pure BDD cardo S. Sussmann [37]. exhibits extremely high corrosion resistance. Refluxing in a highly Beyond the diamond size, there are a number of other proper- corrosive solution such as HNO3:H2SO4 for instance serves to ties that are often cited when characterizing diamond films. One of remove graphitic impurities from microcrystalline diamond films the more critical of these is the nature of the grain boundaries that but has little effect on the integrity of the diamond grains [47,48]. bind the micro or nano diamonds into the matrix. Fig. 1 is a cartoon Background current during potentiodynamic electrochemical depiction of a generic PCD film indicating a diamond region and measurements is the principle source of noise, limiting the sensi- two grain boundaries regions between diamonds. tivity of an electrochemical sensor. The magnitude of the back- The carbon atoms in the diamond regions adopt tetrahedral, sp3 ground current is, in most cases, determined by the electrode/ geometry but the structure of the carbon in the grain boundaries is solution capacitance (better known as the electrochemical double complex and changes between film types. In high purity PCD films layer capacitance) of the material/solution combination. The with large diamond crystals the percentage of non-diamond (sp2/ magnitude of this capacitance depends on a number of factors sp3 amorphous) diamond in the grain boundaries is very low and including the graphitic (sp2) content of the diamond film, surface has little impact on the properties of the film. In UNCD films how- termination, doping level, crystal facet and roughness among ever the boundaries contribute a much higher proportion of the film others [48]. As-grown single crystal or high quality polycrystalline leading to higher sp2 content. When UNCD is nitrogen treated diamond films exhibit very low electrochemical double layer during growth (N-UNCD), the grain boundaries become thicker and capacitance values [47]. BDD (like other p-type semiconductors) more graphitic and sp2 content in the film can be as high as 65% exhibits a degree of voltage dependence of the electrochemical [38e41]. The electrical conductivity in N-UNCD films is thought to capacitance [49]. Minimum reported values occur when the dia- occur via the grain boundaries contrasting with traditional doping mond electrode is polarized at 0.1 V vs Ag/AgCl and are typically such as occurs with boron inclusions in diamond [38,41]. 1e2 mFcm 2 in magnitude [47], an order of magnitude lower than

Table 1 Grades and typical uses of diamond and polycrystalline diamond types.

Crystal size Appearance Common/proposed uses

Single Crystals Wide range Clear e coloured windows Gem stones Grit for grinding/cutting Low absorbance windows (x-ray, IR etc) Microcrystalline Diamond (MCD) films 0.5e50 mm Black - clear Implantable electrodes Implant encapsulation Cutting tools Heat dissipation Wear resistant coatings Nano Crystalline diamond (NCD) 10- 500 nm Black Wear resistance Anti-corrosion barrier films Barrier film (Biomedical) Ultra-nano crystalline diamond (UNCD) 5-10 nm Black Implantable electrodes Wear resistance Barrier film (Biomedical) 440 D.J. Garrett et al. / Carbon 102 (2016) 437e454

Fig. 1. Structure of a polycrystalline diamond film. values reported at freshly polished glassy carbon for instance purity BDD electrode (black line) and a BDD electrode with (25 mFcm 2). Fig. 2 shows MotteSchottky plots taken from the considerable sp2 (or non-diamond carbon) impurities (purple line) (111) and (100) faces of single-crystal boron-doped HPHT diamond. in a 0.1 M KNO3 solution [50]. Both types of diamond exhibit a wide The slope of a MotteSchottky plot relates to the acceptor con- region where no current passes between the electrode and the centration in the semiconductor and, in this instance, illustrates the solution. The purple trace contains additional features due to preferential uptake of boron into the (100) facet of diamond during oxidation of sp2 carbon at z 1.4 V and reduction of dissolved ox- growth (2 1019 cm 2 for (111) vs 4.8 1021 cm 2 for (100) [46,49]) ygen (ORR) between 0.8 and 1.5 V. The large increases in current and illustrates the effect that dopant concentration can have on magnitude occurring at the extremes of the CV are due to oxidation overall double layer capacitance. The very low capacitance values at (splitting) of water in the positive (anodic) direction and reduction ca. 0.1 V for BDD may be explained by the development of a space- of water in the negative (cathodic) direction. Cathodic water charge region within the diamond, depleting the concentration of splitting and dissolved oxygen reduction are both inner sphere carriers at the surface [47]. The absence of sp2 carbon at the surface electrochemical reactions [51] requiring a close interaction be- of high-purity diamond and consequently oxygen functionalities tween the respective molecules and electro catalytic sites on the may also be a contributing factor [47]. electrode. Defects in the diamond caused by regions of non- The low capacitance of BDD is an attractive material quality for diamond carbon serve as electrocatalytic sites hence reduction of obtaining low-noise electrodes, but equally attractive is the very water is more facile and dissolved oxygen reduction is observed. In wide electrochemical window that diamond exhibits in aqueous both cases however there is still a wide region (z3 V) where almost electrolyte solutions. This wide electrochemical window permits no current is measured, one of the widest electrochemical windows detection of analytes across a wide range of voltages without the known. This has safety implications for neural stimulation and interference of faradaic (electrochemical reactions) background implications for detection of neurochemicals by oxidation where current due to electrochemical oxidation and/or reduction of water. additional current from oxidation of water acts as a source of noise. Fig. 3 shows cyclic voltammograms (CVs) conducted using a high-

Fig. 3. Cyclic voltammograms of high purity BDD (black trace) and BDD with sp2 Fig. 2. MotteSchottky plots for individual (111) (1) and (100) (2) faces of the single impurities (purple trace), illustrating the wide limits for water oxidation and reduction 2 crystal recorded in H2SO4 solution. Potential vs Ag/AgCl [46]. and the additional catalytic activity brought about by the presence of sp impurities [50]. (A colour version of this figure can be viewed online.) D.J. Garrett et al. / Carbon 102 (2016) 437e454 441

2.2.2. Nitrogen-included ultrananocrystalline diamond (N-UNCD) 1000 C during growth of their films compared to z800 C for N-UNCD is a material constructed of very small (z2e5 nm) Pleskov, a factor that may also have altered the nature of the films. diamond crystals interspersed in a non-diamond carbon matrix. The two results are an excellent illustration of the wide variety of The presence of nitrogen in the CVD growth chamber increases the properties that can occur in CVD grown N-UNCD and the need for crystallinity of the non-diamond regions leading to graphitic elec- careful control of growth conditions in order to achieve repeatable trical conductivity through the network of grain boundaries. The material properties. electrical and electrochemical properties of the material are strongly influenced by the amount of nitrogen present during 3. Growth of neurons on diamond growth. Considering that N-UNCD films can contain as much as 50% non-diamond carbon [40,52] the electrochemical properties of N- Due to its chemical and biochemical inertness, diamond is UNCD are surprisingly diamond-like. generally considered as a biocompatible material, meaning that it is Fig. 4 shows CV recorded from N-UNCD samples grown with a chemically non-cytotoxic when in contact with biological cells. This high (e) and low (c) N2 concentration in the CVD plasma. CVs were makes diamond a material of interest for coating of medical de- recorded before (black trace) and after (dashed trace) electro- vices, building artificial organs, and as a growth support for bio- chemical oxidation by polarization above 2.5 V vs Ag/AgCl in an logical cells. Diamond has been used as a substrates for cultivation aqueous electrolyte [40]. Both the CVs exhibit the wide water of a variety of cell types including; neurons [55] fibroblasts [56], window associated with diamond electrodes despite their esti- osteoblasts [57] and many other cell lines [58]. In all cases diamond mated (by x-ray photoelectron spectroscopy) z50% non sp3 carbon exhibited no measurable cytotoxicity and, in some cases, appears to content. Anodisation however resulted in a dramatic increase in promote cell adhesion and proliferation over conventional mate- electrochemical capacitance of the high-N2 nanodiamond film from rials such as glass or tissue culture polystyrene. 25 mFcm 2 to a value of 160 mFcm 2. Platinum for instance, the Whether acting as a neurochemical sensor or as a recording/ most commonly used stimulation material typically exhibits stimulating electrode, a healthy, non-toxic interface with the local 50e100 mFcm2 [53]. In this work the authors were targeting neurons is clearly critical if the electrode is to operate successfully, neural stimulation where high electrochemical capacitance is in particular over extended periods. Therefore, the subject of dia- desired. The authors attribute the startling rise in capacitance to an mond as a growth substrate is important in the context of this re- increase in surface roughness and to a greater propensity of the view. The assumption being that materials capable of promoting high-N2 diamond to form carbon-oxygen functionalities at the cultured neuron adhesion and proliferation, extended neurite surface upon anodization. Fig. 5 illustrates the high nanoscale (a,b) growth and perhaps formation of neurite networks are more likely and microscale (c,d) roughness of the films grown in this experi- to maintain their electrical and chemical function in an in vivo ment and the change in structure from a granular structure at 5% N2 setting. (Fig. 5 (a)) to a grassy appearance (Fig. 5 (b)) at 20% N2. These results stand in contrast to those obtained by Pleskov et al. 3.1. Effects of surface properties [54] where a maximum of 5.1 mFcm 2 was measured on N-UNCD films grown with 25% N2 in the CVD plasma. Comparison of the The surface properties of diamond vary according to different synthesis methods between these two papers reveal that the most fabrication methods. Post-processing steps such as sterilization can notable differences in the preparation of the films was the growth alter the chemical composition of the surface. Therefore, it is time and the substrate temperature. The shorter growth time and necessary to compare the growth of neurons on diamond films careful seeding described by Pleskov et al. resulted in films that with different surface characteristics. The relative sp2/sp3 content is were considerably smoother than those of Garrett et al., a factor a variable that has been shown to affect the viability and growth of that would contribute directly to the lower capacitance values some cell types on diamond like carbon (DLC) [59] including my- measured. Garrett et al. report substrate temperature in excess of eloblasts, HEK293 and fibroblasts. This specific question however has not to date been investigated for neurons. The effects chemical surface termination and morphology on neural adhesion, prolifer- ation have been attracted scientific interest with conflicting results. The as-grown diamond films are normally hydrogen terminated due to the growth environment. Hydrogen termination however is unstable in atmosphere. Oxygen termination can be attained easily by several different techniques such as acid boiling, oxygen plasma, UV and ozone irradiation. Rezek et al. developed an atomic force microscope technique to measure adhesion forces of proteins on surfaces [60] and applied that technique to show that epithelial cells exhibited changes in morphology and improved adhesion on oxygen terminated NCD compared with H terminated [61]. Ariano et al. [62] compared neuron growth on oxygen and hydrogen terminated homoepitaxial diamond films finding no significant difference, either in morphology or in survival ratios between the ciliary ganglion neurons grown on two types of diamond surface and control cultures (plastic dishes), provided that a mixture of

proteic adhesion molecules (poly-D-lysine and laminin) was applied before the cell seeding for cell anchoring. Similar results were re- ported on NCD with different surface termination. Ariano et al. [63] Fig. 4. CVs conducted in pH 7.4 phosphate buffered saline with 0.1 M KCl supporting used GT-17 cells without exogenous adhesion molecules and rat electrolyte. N-UNCD grown in the presence of 5% N (a) and 20% N (b) before (black 2 2 hippocampal neurons were cultivated by Ojovan et al. [64] with trace) and after (dashed trace) electrochemical oxidation by polarization above 1.6 V vs fi Ag Ag/Cl in an aqueous electrolyte [40]. poly-DL-ornithine and laminin coatings, both showing no signi cant difference on two types of NCD surfaces with respect to cell 442 D.J. Garrett et al. / Carbon 102 (2016) 437e454

Fig. 5. Scanning electron microscope images of N-UNCD films grown with 5% (a) and 20% (bed) N2 in the CVD plasma [40]. adhesion and proliferation. Neuroblastoma cells also grew equally biomedical applications. Later, human fetal stem cells were also well on unmodified (hydrogen terminated) and oxygen terminated shown to attach better onto NCD films with oxidized surfaces diamond-like carbon surfaces [65]. Bendali et al. concluded overall without applying additional coatings instead of hydrogen termi- that a protein coating was required to promote healthy growth of nated surfaces [69]. retinal neurons on diamond but noted that results were much more Incorporation of elements other than carbon and hydrogen is consistent when oxygen terminated diamond was used over H commonly used to alter diamond conductivity for fabricating terminated. The authors cite the lack of stability of H terminated biomedical electronic devices. Higher survival rates of primary diamond as a possible reasons for the result [66]. cortical neurons on phosphorus doped than un-doped diamond- In contrast, other publications report better attachment of like carbon with poly-D-lysine was reported, suggesting that the neurons on oxygen terminated surfaces. Oxygen terminated phosphorus-doped diamond-like carbon can direct neuronal diamond-like carbon supported rat cortical neuron growth much growth [65]. Though elemental boron can be mildly toxic, boron better than the unaltered surfaces [65]. Poly-D-lysine coating was when used as a dopant in diamond is locked in the diamond lattice applied to promote cell adhesion by altering the surface charge and therefore does not affect biocompatibility [70]. Nistor et al. [71] from negative to positive (cells are negatively charged). Therefore cultured pluripotent stem-cell-derived human neurons on dia- the different neuron growth may be associated with different sur- mond and found that compared with un-doped diamond, boron- face hydrophilicity, which leads to different degrees of poly-D-lysine doped diamond has no adverse effect on cell survival, neurite for- adsorption on the surface. Rat cortical neuron attachment was mation and the apoptosis levels of cells. When nitrogen is added found to be better on N-UNCD films oxidized by hydrogen peroxide into gas mixture during UNCD deposition to increase the conduc- sterilization than those on the as-grown surfaces with hydrogen tivity, the number of surviving rat cortical neurons also increased termination after 24-h cell culture without any additional coating, and their neurite length extended [72]. The different neural inter- which may also be associated with their different protein adsorp- action with UNCD and N-UNCD may be the result of different grain tion capacities (unpublished). Chen et al. [67] found that neural shapes and sizes, which further influence the protein adsorption on stem cells proliferated well without additional coating on both the diamond surfaces. hydrogen and oxygen terminated UNCD but slightly better on Another factor that affects neuronal growth is surface hydrogen terminated surfaces. Neuron stem cells on oxygen- morphology. Both pluripostent stem-cell-derived human neurons terminated surfaces demonstrated a preference to differentiate [71] and primary rat cortical neurons [24] showed preference to into oligodendrocytes, while those on hydrogen-terminated UNCD diamond with small diamond microcyrstals. Babchenko et al. [69] tended to differentiate into neurons. The increased levels of compared NCD films with flat, porous or nano-rods structures. neuronal differentiation are due to adsorbing fibronectin from When seeded without laminin coating, the neural stem cells spread medium [68]. Therefore, controlling the diamond surface proper- over flat diamond surface homogeneously but assembled into ties has the potential to differentiate neural stem cells for different bundles on the porous structures, and only neurospheres were D.J. Garrett et al. / Carbon 102 (2016) 437e454 443 found on the diamond nanorods, suggesting it is an inappropriate similar topography and roughness, solid NCD films do not promote condition for neural stem cells attachment. Piret et al. [73] cultured neuronal adhesion whereas films of isolated nanodiamonds do [74], mouse spinal cord and hippocampal cell cultures on flat and 3D- despite the fact that the roughness of these films are below the nanostructured boron-doped diamond substrates. At 8 days in vitro, roughness window for promotion of neuronal adhesion on silicon cells were found to attach and extend neurites on both substrate [75]. The conclusion Edginton et al. draw is that the important in- types and no significant difference was observed. teractions are subcellular (nanoscacle) interactions and possibly the size, shape and surface charge of the nanodiamonds helps offer fi 3.2. Necessity of adhesion promoter coatings good adhesion for proteins such as vitronectin and bronectin expressed by the growing neurons. are that Cultivation on homo- There is a wealth of evidence that neurons have the ability to epitaxial diamond was published in 2004 [55] describing dissoci- sense and react to the chemical and mechanical properties of bio- ated cortical murine neurons seeded onto a laminin patterned materials. In the natural system, neurons respond to proteins in the diamond surface. Neurons extended neurites along the laminin extra cellular matrix of adjacent neurons and function according to patterns, indicating their preference to laminin-covered surface these interactions. Typically, during neuronal cell culture, examples over bare homoepitaxial diamond. Chick cilary ganglion neurons of these proteins such as fibronectin, vitronectin, lysine and laminin [62] and GT1-7 cell lines [63] were also found unable to attach onto are adsorbed to the culture substrate in order to facilitate healthy homoepitaxial diamond without any coating. neuronal adhesion, proliferation and neurite outgrowth. There The conflicting results about the requirement of promotion layer exists however some conflicting evidence regarding whether have been focused on polycrystalline diamond, including micro- additional promoting layers such as poly-D-lysine, poly-DL-orni- crystalline (MCD), NCD and UNCD. Nistor et al. reported that for thine or laminin are necessary for neuron attachment and prolif- long-term survival of human pluripotent cells on MCD, it is eration on diamond films. Similar debates exist over whether necessary to supplement the culture medium with laminin once a neurons can survive and form functional networks on bare dia- week, even though laminin was coated onto the diamond surface mond surfaces. Table 2 is a summary of results for neurons grown before cell seeding [71]. Rat hippocampal neurons grown on bare on diamond with/or without the use of adhesion promoter NCD revealed poor adhesion and cell clustering [64]. Tong et al. proteins. used UNCD as a substrate for rat cortical neuron growth. Although In essence, as with most cells, neurons prefer surfaces with the neurons attached and extended neurites on UNCD surfaces features mimicking their biological environment such as the after the initial 24-h incubation [24,72], they did not survive for extracellular matrix they are native to. Adsorption of extracellular more than 7 days if no additional coating layer was applied. proteins on substrate materials is determined by properties such as In contrast, the studies of Chen et al. found that mouse neural surface charge, surface functional groups, surface roughness and stem cells did attach, proliferate and differentiate on bare UNCD surface curvature. For neurons grown on homoepitaxial diamond surfaces for at least 9 days [67,68]. Human fetal neural stem cells for instance, an adhesion promotion layer appears to be essential. were also found to survive on bare NCD films up to 8-days after cell There has been no report describing neuron survival on homo- seeding [69]. For biocompatibility studies, the required addition of epitaxial diamond without a coating. This may well be a result of neurotrophic factors to their culture could prevent neuronal the very low roughness of homoepitaxial diamond. Again however degeneration, therefore the cell proliferation of neural stem cells the role of roughness on diamond is difficult to understand. In their could mask a material toxicity [66]. Cockroach neurons [76] and 2013 article, Edginton et al. point out that, despite possessing GT1-7 cell lines [63] attached well on UNCD and NCD without a

Table 2 Publication list on neuron growth on diamond without additional coating.

Diamond Cell type Results Ref author year

Homoepitaxial Murine cortical neurons Neurons extended neurites along laminin patterns [55] Specht 2004 diamonda Homoepitaxial Chick cilary ganglion cells No cell attached [62] Ariano 2005 diamond Homoepitaxial GT1-7 cell line No cell attached [63] Ariano 2009 diamond MCDb Human pluripotent cells For long term culture (up to 150 days), the medium needs to be supplemented with laminin once a [71] Nistor 2015 week NCD GT1-7 cell line Cells adhered and functional viability was confirmed [63] Ariano 2009 NCD Rat hippocampal neurons No cell attached at 2 DIV [77] Thalhammer 2010 NCD Rat hippocampal neurons Cells attached and extended neurites on both flat and nanostructured surfaces but formed clusters [73] Edgington 2013 NCD Rat hippocampal neurons Cells revealed poor adhesion and cell clustering [64] Ojovan 2014 NCD Human fetal neural stem Cells could survived but better with laminin coating [69] Babchenko cells 2013 NCD Mixed rat retinal cells and Glial cells only survived on coated surfaces but neurons survived even better on bare diamond surface [66] Bendali 2014 purified ganglion cells UNCD Murine neural stem cells Cells attached, proliferated, and differentiated on diamond for at least 9 days [67] Chen 2009 UNCD Cockroach neurons Cells attached well on diamond surface at 2 DIV [76] Voss 2012 UNCD Rat cortical neurons Cells attached and extended neurites at 1 DIV but could not survive at 7 DIV [24,72] Ganesan 2014 Tong 2014 nanodiamond Rat hippocampal neurons Nanodiamond coating promoted neuronal attachment with respect to neurite outgrowth, direct [77] Thalhammer interaction with the substrate, long-term survival and functional properties 2010

Note. a Surface coated with laminin patterns before cell seeding. b Surface coated with poly-DL-ornithine and laminin before cell seeding. 444 D.J. Garrett et al. / Carbon 102 (2016) 437e454

Fig. 6. (a) electrochemically sharpened platinum wire shown before (upper) and after (lower) CVD depiction of boron doped PDC. (b) Is a higher magnification SEM image of the PCD film. coating. Primary mammalian neurons are more demanding than pertinent here is the work of Edgington et al. who found that those from other animals and cell lines. Therefore, they may not be nanodiamond coatings on glass universally promoted murine sufficient for biocompatibility assessment of materials for mammal hippocampal neuron adhesion regardless of surface functionality implantations. and with or with adhesion promoter proteins [74]. A comparison Primary rat hippocampal neurons were shown to attach and with a continuous diamond film was not conducted in this instance. extend neurites on boron-doped NCD with nanostructured surfaces and the functional viability of the attached neurons at 14 days 4. Diamond electrode and diamond microelectrode array in vitro was confirmed from neural recording [73]. However, the (MEA) fabrication methods cells were found to form clusters on the surface, indicating that cells may prefer to attach onto themselves rather than the diamond Fabrication of flat films of diamond is relatively straight forward, surface. Bendali et al. [67,68] reported interesting results on NCD however for most neural interfacing applications microelectrodes using adult retinal cell cultures from rats. The use of a protein or microelectrode arrays are desired. Diamond, by virtue of its coating increased cell survival, particularly for glial cells, but the extreme hardness, lack of ductility and chemical reactivity, its biopolar neurons grew in direct contact with bare diamond. Using inability to melt and the high temperatures present during CVD purified adult retinal ganglion cells, the neurons survived even synthesis is a challenging material with which to fabricate devices. better on NCD without any protein coating at 6 DIV and a protein Methods to produce single microelectrode and MEAs have however pattern on diamond cannot lead to ordered neuron growth. Thal- been developed and have been reviewed elsewhere [78,79]. For the hammer et al. [77] reported that nanodiamond powder coating purposes of this review a small number of examples will be chosen could promote rat hippocampal neuron growth on many different to illustrate the methods most commonly employed for neural substrates, with a potency similar to the effect of a standard protein interfacing. coating. In Thalhammer's work however the cells did not survive on Standalone diamond electrodes are commonly fabricated by an NCD film in spite of the film having a similar surface topography CVD growth of diamond on a high melting temperature metal to the nanodiamond powder film. The difference may be due to substrate. An example from Park et al. [80] is depicted in Fig. 6 their different protein adsorption capacities of the nanodiamond showing boron-doped PCD deposited onto an electrochemically powder and NCD film. The nanodiamond film may adsorb certain sharpened platinum wire. Tungsten wire is also a common sub- proteins more efficiently than NCD which in turn could promote strate for diamond electrodes of this type [81]. the neuron attachment and neurite outgrowth on the surface. Also Multi electrode arrays are most commonly formed by removing

Scheme 1. Generic method of fabricating a diamond MEAs via dry etching of a continuous diamond film. D.J. Garrett et al. / Carbon 102 (2016) 437e454 445

Fig. 7. Example of a BDD MEA developed by Chan et al. [85]. (A colour version of this figure can be viewed online.) material from a continuous film of conducting diamond to create are chemicals that transmit signals between nerve cells or between discrete islands or channels of diamond following variations of the nerve and other specialized cells such as muscle cells at the generic scheme depicted in Scheme 1 [73,82e84]. Fig. 7 shows an neuromuscular junction. Neurotransmitters are released at synap- example from Chan et al. [85] of a device fabricated by this method. tic sites on axons where they diffuse across the synaptic cleft and An alternative to removal of diamond by dry etching was are detected by receptors on dendrites extending from the target demonstrated by Ganesan et al. [24] who produced an N-UNCD cell. Thus the expression of neurotransmitters is an essential part of microelectrode array in which individual electrodes were isolated normal neural function and hence anomalies in neurotransmitter by removing sections of the film with laser milling apparatus. In function are linked to wide range of diseases. The biogenic amines, this instance the authors used insulating PCD as the substrate. The catacholamines, dopamine, norepinephrine (noradrelaline) and PCD was laser milled with a series of vias upon which the N-UNCD epinephrine (adrenaline) along with serotonin are the most com- film was deposited. The vias acted as hermetic electrical feed- mon neurotransmitter targets for electrochemical detection. The throughs for the electrode array so that the array was suitable for three catacholamines, all derived from the precursor 3,4- direct bonding to microelectronics. The same group has previously dihydroxyphenylalanine (DOPA) and are shown in Scheme 2 proposed a method to generate penetrating diamond electrodes by along with serotonin [91], another common target for electro- growing the diamond film in an etched silicone template to mold chemical detection. diamond spikes. Both electrode types are depicted in Fig. 8 [86].An Dopamine is the most abundant of the catacholamines and, like alternative laser ablation method was previously demonstrated by noradrenaline and adrenaline, it is intimately involved in modula- Pagels et al. [87] In This method a standalone film of BDD was laser tion of many aspects of brain function among other crucial physi- ablated through a mask generating an array of BDD columns. ological roles. Accurate measurement of dopamine, in particular, is Insulating PCD was grown over the structure and a final polishing highly useful because dopamine expression is linked to neurolog- step revealed the BDD columns penetrating through the PCD, ical diseases such as schizophrenia and attention deficit hyperac- essentially the inverse of Ganesan et al.’s method. tive disorder (ADHD) and is implicated in most diseases of Another notable alternative fabrication method is that addiction [91]. Serotonin, or 5-hydroxytryptamine is a dicyclic employed by Bergonzo et al. who, rather than etching the diamond neurotransmistter biologically derived from tryptophan. Among film itself, pattern the diamond seed film prior to CVD. The result is numerous physiological roles serotonin levels are most commonly a patterned diamond film directly from the CVD chamber linked to depression, hence the success of the serotonin reuptake [23,88e90]. inhibitor (SRI) antidepressant drugs. All of these molecules can be detected electrochemically via oxidation of the molecule to their respective corresponding ketones according to Scheme 3. 5. Neurochemical sensors The reactions are generally electrochemically reversible. A persistent difficulty however is the susceptibility of the oxidized 5.1. Electrochemical detection of neurotransmitters forms of these molecules to nucleophilic attack by adjacent mole- cules. The resultant dimers are insoluble and form films on the The wide electrochemical window that diamond possesses electrode surface rapidly reducing the sensitivity of the electrodes. coupled with low background current has made diamond an One of a number of possible byproducts formed by reaction of attractive material for neurochemical detection. Neurotransmitters 446 D.J. Garrett et al. / Carbon 102 (2016) 437e454

Fig. 8. Example of a flat N-UNCD high density MEA isolated by laser milling (Left). Templated diamond spikes produced by Ganesan et al. (Right) [24,86]. (A colour version of this figure can be viewed online.) serotonin with its corresponding ketone is shown in Fig. 9. Similar and determined (in agreement with Güell et al. [93] and Patel et al. reactions occur with the catacholamines dopamine, and the [96]) that the diamonds principle advantage is due to its lack of adrenalines. Evaluation of electrodes for electrochemical detection fouling. In this work the authors report enhanced sensitivity by of neurotransmitters therefore usually includes standard sensitivity modifying the surface of the BDD with carboxy functionalities via a assessment and also evaluation of the electrode material's ability to photochemical reaction with a carboxy terminated alkene. Building avoid fouling due to formation of films on the electrode surface. on earlier work [97], Tryk et al. [98] also demonstrate the impor- tance of surface termination for the detection of dopamine at BDD e 5.2. In vitro performance of diamond for electrochemical detection electrodes. Several papers [99 101] demonstrate that N-UNCD also of neurotransmitters is highly sensitive for detection dopamine in the presence of ascorbic acid. Surface termination on diamond electrodes can also fi Güell et al. [93] compared the performance of smooth glassy be used to covalently link molecules which can impart speci c fi carbon and BDD with disordered carbon nanotube (CNT) films for bene ts. For instance covalent attachment of Tyrosinase was use to electrochemical detection of serotonin. They found that BDD was enhance BDD for electrochemical detection of small molecules more sensitive than glassy carbon but less sensitive than CNT films. including dopamine [102]. A summary of neurotransmitter detec- Detection limits of 500 nM, 2 mM and 10 nM were established tion results where a detection limit has been calculated is included respectively. BDD electrodes however were much slower to foul in table. In a unique experiment Dankerl et al. demonstrated that a fi than CNT films. Fig. 10 shows 10 sequential cyclic voltammograms diamond eld effect transistor based upon the hydrogen surface conducted in a 10 mm serotonin solution on a BDD electrode (a) and conductivity could be employed to construct a sensor for acetyl- a CNT film electrode (b). The peak due to oxidation of serotonin choline [103]. Table 3 shows a summary of published detection occurs at z0.45 V for BDD and z0.38 V on CNTs indicating some limits towards serotonin and dopamine. degree of catalysis of the reaction on the CNT electrode. The peak magnitude on the BDD electrode however reduces by 65% over the 5.3. Neurochemical detection in living systems 10 scans compared with 90% for the CNT film electrode attesting to the lower propensity of the BDD electrode to become fouled. The The unparalleled chemical stability and very high biocompati- authors improved on this result by extending the cyclic voltam- bility of diamond materials make them an obvious target for mogram to 0.6 V during the detection scans. Fig. 10 (c) shows fabrication of long lasting implantable sensor electrodes. Several sequences of CVs conducted in 10 mM serotonin on BDD (upper) and groups have demonstrated efficacy of diamond electrodes in living CNT film (lower) including a cathodic excursion to 0.6 V for each systems. Gosso et al. demonstrated that an array of nine BDD CV. The degree of fouling is much reduced on the CNT film and electrodes confined within a disc-shaped area of 20 mm diameter practically eliminated on BDD. were effective in mapping exocytosis of catacholamines across the Lete et al. employed BDD microelectrodes and established a membrane of single mouse and bovine chromaffin cells. Expression detection limit of 44 nM for dopamine. Importantly they conducted of catacholamines from the chromaffin cells was elicited either by this study in the presence of the common interfering molecule electrical or chemical stimulation. The nine BDD electrodes were ascorbic acid [94]. The low detection limit was facilitated by a peak- polarized at þ0.8 V vs Ag/AgCl and the current on each electrode current voltage separation of 410 mV between the two analytes. simultaneously monitored. Exocytosis of catacholamines in the Kondo et al. [95] investigated the electrochemical detection of a vicinity of an electrode resulted in a corresponding spike in recor- number of catecholamines on BDD in the presence of ascorbic acid ded current due to oxidation of the neurotransmitters. This work

Table 3 Summary of published detection limits of various diamond preparations towards serotonin and dopamine.

Electrode Target analyte Limit of detection (nM) Mixed analyte Ref

BDD ST 500 No [93] BDD DA 44 AA [104] COOH-BDD DA 100 AA [95] N-UNCD DA 360 AA [99] O-BDD DA 50 AA [97] Tyrosinase-BDD DA 1300 AA [102] D.J. Garrett et al. / Carbon 102 (2016) 437e454 447

vasoconstriction method and demonstrated that the BDD elec- trodes, grown on a sharpened platinum wire exhibited significantly better antifouling performance than carbon fibre electrodes. Car- bon fibre electrodes suffered from a 30% response attenuation over a 7 h period compared with just 8% for the BDD electrodes [80]. Patel and Swain et al. demonstrated in 2008 that a BDD electrode could also be used to detect nitric oxide released by neurons and circular muscle in guinea pig ileum [107]. The BDD electrode out performed carbon fibre in terms of signal amplitude in that study. The research also stands as an example of the utility of micro- electrodes for electrochemical detection of species other than the regular catecholamine targets. The most complete example to date of the utility of diamond for neurochemical detection in vivo is that of Yoshimi et al. [108] later highlighted by Trouillon et al. [109] In this work the effectiveness of BDD electrodes for detection of dopamine in vivo was established in a mouse model using previously established methods [110]. The study employed side-by-side carbon fibre and BDD microelectrodes electrodes inserted into the mouse striatum. The striatum is an area of the brain associated with the reward system receiving glutamine and dopamine inputs during activity that is perceived as pleasur- able. In the anaesthetized mouse model, dopamine expression was evoked by electrical stimulation with one of the electrodes whilst concurrently recording dopamine oxidation current with the other. Administration the drug nomefesine (a dopamine reuptake inhib- Scheme 2. Structure of catacholamines derived from DOPA and serotonin. itor) increased the amplitude of the current detected on the recording electrode thus identifying dopamine as the principle catecholamine responsible for the signal [111]. This additional experiment adds considerable strength to the work as ampero- metric detection is inherently a non-specific technique. The high- light of the paper however was the use of BDD electrodes to record dopamine expression in the stratum of two awake behaving Japa- nese monkeys. A chamber was attached to each monkey's skull permitting access to the cortex above the striatum. The monkeys were conditioned to an orange juice reward using a Pavlovian conditioning paradigm where a light flash could be used to forecast the juice reward. BDD electrodes were inserted into the striatum and amperometric recordings taken during presentation of a sequence of light flash, no light flash, OJ reward, no OJ reward stimuli. Fig. 11 (c) shows the collated current traces recorded over 67 paired trials (paired ¼ both light flash and juice reward) in one of Scheme 3. Electrochemical oxidation/reduction pathways of catecholamines and the monkeys. The sharp peak in oxidation current immediately serotonin. following the light flash indicates a strong Pavlovian response to the impending juice reward followed by a smaller current peak following the juice reward itself. Only single current peaks were builds on the earlier efforts of Kisler et al. where nitrogen doped observed when a juice reward was administered without a pre- diamond-like carbon electrodes were compared with indium tin ceding light flash. The study demonstrates that BDD is an effective oxide (ITO) and gold electrodes for simultaneous amperometric dopamine detection material and in all cases the authors report and fluorescence imaging of exocytotic events in chromaffin cells superior sensitivity exhibited by the BDD electrodes over more [105]. traditional carbon fibres. Examples of whole tissue detection of neurotransmitters with diamond has been dominated by Greg M. Swain and co-workers. In 6. Diamond devices for neural stimulation or neural 2006 Park and Swain et al. demonstrated living system detection of recording norepinephrine in two separate studies [80,106]. In the second of these published studies, sections from the small intestine of rats All nerves transmit signals to one another by rapidly exchanging þ þ were removed and perfused to maintain. A BDD recording electrode ions such as sodium (Na ) and Calcium (Ca2 ) across their cell þ þ was positioned touching a blood vessel within the section. Vaso- membranes. In a neuron's resting state, Na and K are pumped constriction of the blood vessel was initiated by stimulation of across the membrane by specialized proteins generating a trans- perivascular nerves on the blood vessel with a second electrode. membrane electrical potential. During a natural nerve depolariza- Concurrently norepinephrine released by the perivascular nerves tion, a stimulus (i.e. a neurotransmitters arriving at a synapse) may during vasoconstriction was detected by oxidation of the cate- cause a change in the permeability of a region of the cell mem- cholamine at the BDD electrode [106]. Fig. 11 shows photographs of brane. Ions pass into the cell altering the transmembrane electrical a vessel before and after vasoconstriction and the corresponding potential in that region. If the transmembrane potential changes current spike recorded at the BDD electrode due to release of sufficiently, voltage gated ion channels will open en-mass, causing norepinephrine. The earlier study introduced the rat an axon potential travelling out from the initial segment of the axon 448 D.J. Garrett et al. / Carbon 102 (2016) 437e454

Fig. 9. One of the principle dimers formed by a nucleophilic substitution reaction between serotonin and its corresponding ketone. Also shown is a sequence of AFM images recorded by Patel et al. [92] illustrating the change in topography of a BDD electrode due to electrochemically generated film formation upon oxidation of serotonin at the electrode. The line profiles shown correspond to the dashed line indicated in panel (a). (A colour version of this figure can be viewed online.)

Fig. 10. Sequences of 10 cyclic voltammograms conducted in 10 mm serotonin solution conducted by Güell et al. [93] The electrodes used in each case were (a) BDD, (b) CNT film, (c- upper) BDD with 0.6 V cathodic excursion (c-lower) CNT film with 0.5 V cathodic excursion. (A colour version of this figure can be viewed online.) and along the axons and dendrites of the cell. An action potential the electromagnetic field around the electrode can induce a cor- arriving at a synapse triggers neurotransmitters to be released and responding flux of ions in the adjacent tissue. If the flux of ions diffuse across the synaptic cleft to signal the next neuron in the results in a sufficiently large change in the transmembrane po- chain. The sudden flux of ions occurring during a neuron depolar- tential of a local neuron, an axon potential can occur and be ization causes a small perturbation in the electric field around the transmitted along a nerve. Though the two aims are clearly inter- neuron. This perturbation can induce a corresponding perturbation related, neural recording and stimulation electrodes require very in a nearby electrode which, when measured, manifests as a different properties [53]. momentary change in electrode voltage. Conversely if a small packet of charge is delivered to the tip of an electrode, the change in D.J. Garrett et al. / Carbon 102 (2016) 437e454 449

Fig. 11. (a) Images taken and traces reproduced from [104] showing microscope images of a blood vessel before and after vasoconstriction induced by the stimulation electrode. Also shown (b) is the measured vessel width and corresponding current recorded at the BDD microelectrode due to oxidation of norepinephrine expressed for perivascular nerves on the blood vessel. (c) Shows the oxidation current due to dopamine release recorded at a BDD microelectrode following a light flash induced Pavlovian response and an orange juice reward response from the striatum of an awake, behaving Japanese monkey. (A colour version of this figure can be viewed online.)

6.1. Neural recording recording [115]. Though high purity diamond is electrically insu- lating, hydrogen termination imparts limited conductivity to the In order to induce a voltage change in a recording electrode the diamond surface by a surface transfer doping mechanism [43]. firing neuron must first overcome the electrochemical capacitance Ariano et al. successfully recorded high quality action potentials and also the electrical impedance of the electrode solution inter- from mouse hypothalamic GT1-7 cells cultured on the diamond. A face. The relationship between recording quality and the various particularly interesting feature of this work is that the diamond is parameters available for electrodes include size, shape, material high purity single crystal and therefore optically transparent. etc. is highly complex [112]. The impedance of the electrode solu- Therefore the technology could be easily coupled with optical tion interface and the size of the electrode dictate the volume of techniques. Using a different technology again Dankerl et al. tissue that can be sampled. High impedance, small electrodes for fabricated a diamond transistor array which they employed to re- instance sample only the closest firing neurons. Hence high cord the membrane potential of HL-1 cardio myocytes and human impedance, small electrodes are more likely to detect the firing of embryonic kidney cells (HEK293) [103,116]. single neurons, typically referred to as single unit recording. Low Halpern et al., building on earlier work [117], demonstrated a impedance, large electrodes are more easily affected by the wider diamond electrode for neural recording in the unusual choice of a electrical environment and therefore tend to record from a larger Sea Hare (Aplysia Californica), a slug like marine animal [81]. The population of neurons. Such recordings are typically called local animal has an accessible nerve (the buccal mass) which is associ- field potentials indicating activity of a region of neural tissue as ated with visually identifiable behaviors associated with grasping opposed to individual cells. The majority of neural recording with food. The electrode consisted of a hook-like stainless steel wire diamond has been conducted with small electrodes fabricated from coated with BDD grown by hot filament CVD. The hook electrode BDD. was inserted into the buccal mass and recordings taken during Neural recordings from cultured neurons or other electrogenic normal behavior. The authors hypothesize and indeed show cells is a common method for assessing the properties of new (though only in a single animal) that the BDD electrode had better electrode materials [113]. Examples on diamond however are rare. signal to noise ratio (average spike height/random noise magni- Maybeck et al. elegantly demonstrated this method with an array of tude) and higher resistance to fouling than a stainless steel control. 64 boron doped NCD electrodes coated with HL-1 cardio myocytes Stainless steel is not favored as a recording electrode due to its lack [114]. They concluded that BNCD electrodes functioned as well as of stability compared to more typically used metals such as plat- gold electrodes in terms of signal magnitude. In a very different inum however the results from this experiment were promising approach Ariano et al. demonstrated that high purity, single crystal, [81]. Chan et al. not only used their custom designed probe shown hydrogen terminated diamond could be employed for neural in Fig. 7 for neurochemical detection but also employed it to record 450 D.J. Garrett et al. / Carbon 102 (2016) 437e454 directly from the auditory cortex of guinea pig [85]. In that instance feedthrough and high density electrode array [24] and that the however a relatively poor signal to noise ratio (z2) was reported. material can be safely employed to stimulate retinal ganglion cells The authors speculate that the recorded neurons/electrode distance [123]. The retinal ganglion cells in that case were in rat retina, and/or high impedance interconnects may have been the cause of excised and perfused. Single diamond stimulating electrodes were the poor result. The result also illustrates one of the principle dif- placed on the surface of the retina and a patch clamp recording ficulties involved in characterizing electrodes in vivo, namely it is electrode was attached to an adjacent ganglion cell to measure the usually impossible to know the distance from the electrode to the effectiveness of stimulation. Three different electrodes were used in recorded neuron. The recent study of Piret et al. however used a 12 experiments yielding a stimulation threshold range between 19 diamond MEA to record activity in an excised perfused mouse hind and 164 mCcm 2, all well below the safe charge injection limit. In brain [73]. In that study they compared as-grown BDD with that work the capacitance of the materials when formed into single nanostructured BDD finding that the nanostructured variety had microelectrodes was improved to 300 mFcm 2 and the authors greatly increased capacitance (3 mF cm 2), decrease impedance, report no electrochemical changes the electrode performance after reduced noise level and was effective at both stimulation and continual pulsing at 50 Hz for one week even at the upper charge recording of neurons in the mouse hind brain sample. Increasing injection limit. Shivdasani et al. the following year, conducted a the effective surface area of electrodes has been previously pilot in vivo study showing that single 120 120 mm square N- demonstrated as a highly effective method of modifying the elec- UNCD electrodes in an array, attached retinally, could elicit a trochemical and hence recording and stimulation properties of response in the visual cortex of a cat with charge injections as low neural interface electrodes [118]. The role of capacitance in as 30 mCcm2. Some electrodes however required over recording however is not so well studied. In Piret et al.'s work the 400 mCcm2 to elicit a response, well over the safe limit of high capacitance acted as a high frequency filter, minimizing the 300 mCcm 2. The BVA diamond array was attached to the inner impact of thermal noise. Thus electrode interface capacitance is a surface of the retina (epi-retinal), a notoriously difficult position to further parameter the warrants consideration when optimizing achieve both close contact and damage free retinal stimulation. A recording electrodes [119]. magnetic attachment method has been proposed by BVA but safe close attachment remains challenging [124]. Close proximity of the 6.2. Neural stimulation electrode to the target neurons is pivotal in achieving low thresh- olds for stimulation therefore improvements in the attachment Diamond, prima facie, is not a promising material for neural strategy could improve in vivo stimulation thresholds. stimulation. Simply put, like other carbon materials [120], its The DREAMS project, utilised an array of BDD microelectrodes inherent electrochemical capacitance is too low. Effective neural transferred to a flexible substrate [23] and have successfully uti- stimulation requires that a sufficient amount of charge is delivered lized their general strategy to fabricate a range of MEA types to the tip of the electrode to affect stimulation of neurons without [89,125]. In early work the capacitance and stimulation perfor- the voltage on the electrode becoming extreme enough to cause mance was not discussed however the expected high biocompati- damage to those same neurons. Extreme positive or negative bility of the diamond array is confirmed by extended subretinal voltages cause damaging electrochemical reactions to occur such as implantation in rat. The subretinal implantation location has some water splitting or irreversible oxidation/reduction of biomolecules. significant advantages including securing the device between the High capacitance forms of diamond have however been reported. retina and the choroid and providing close contact between the Like other carbon materials such as carbon nanotubes or graphene, electrode array and the retina [73,84]. In collaborative work arising high capacitance materials can be constructed by forming micro or from the DREAMS project however, a high surface area form of BDD nanostructured aggregates with very high real surface area. Boron produced by growing the diamond film over a field of carbon doped diamond when very heavily doped can exhibit high elec- nanotubes (Fig. 13) resulted in a capacitance (measured by cyclic trochemical capacitance values as demonstrated by Watanabe et al. voltammetry) of 3 mF cm 2 [73], the highest yet reported for dia- [121] In that work it appears that non diamond inclusions in the mond. As described in section 6.1,10mm diameter nanostructured material may be responsible for the increased capacitance. BDD electrodes in an array format were effective in both stimu- The principle interest in diamond as a stimulation material has lating and recording from cultured cardiomyocytes. come from the retinal implant community. Specifically from the Bionic Vision Australia (BVA) consortium, the MEDINAS project, 7. Future prospects funded by the French National Research Agency, and the Diamond to Retina Artificial Microinterface Structures (DREAMS) project Beyond the electronic and electro-chemical techniques funded by the European Commission. The BVA strategy was to described extensively in this review, new optical approaches using fabricate a rigid diamond array capable of connecting directly to intrinsic defect centers in diamond have been proposed as a plat- and hermetically encapsulating control electronics, thus negating form for multi-magnetometer arrays (MMAs). These non-invasive the need for a large cable connecting to the array from a remote high resolution MMAs present an attractive proposition to image controller. the magnetic fields arising from the transmembrane potentials The BVA device utilizes N-UNCD as the conducting form of generated by neuronal activity [126]. The technology is based on a diamond. Similar to heavily doped PCD the N-UNCD characterized particular defect center in diamond, known as the negatively- by Garrett et al. [40] (Fig. 12) contained as much as 50% non- charged nitrogen-vacancy center (NV) [127]. This defect can be diamond carbon according to XPS analysis. Interestingly, the ma- engineered into the diamond lattice through the introduction of terial did not exhibit high capacitance until a strong electro- nitrogen in the gas phase of CVD growth or alternatively via ni- chemical oxidation of the surface of the electrode in aqueous trogen ion implantation. These fabrication approaches have led to electrolyte was conducted. They showed that this oxidation step the development of 2D magnetic imaging arrays which have been resulted functionalization of the surface with oxygen, a process that used to image and reconstruct the stray magnetic fields from cur- has been shown to drastically increase the electrochemical capac- rent carrying wires [128] and magnetic nanoparticles within living itance of carbon nanotubes [122]. The maximum capacitance re- organisms [129]. The detection technology for 2D MMA imaging of ported in this first publication was 154 mFcm 2. The authors go on neuronal activity is based on optically detected magnetic resonance to demonstrate that this materials can be used as a hermetic (ODMR) of the NV magnetic sub levels. Using wide-field D.J. Garrett et al. / Carbon 102 (2016) 437e454 451

Fig. 12. Structure of the BVA retinal stimulation implant showing the stimulator chip, diamond electrode array and hermetic feedthrough and a diamond box to close the capsule and protect the electronics (From Lichter et al. [25]). (A colour version of this figure can be viewed online.)

Fig. 13. Nanostructured BDD film described by Piret et al. [73] exhibits electrochemical capacitance of 3 mF cm 2. microscopy the ODMR from diffraction limited sensing volumes ms [131]. From the theoretical modelling nT/ms is within reach (0.4 0.4 1 mm) can be imaged with millisecond timing over provided the nitrogen impurity levels can be well controlled and wide fields of view of 100 100 mm [130]. More recent reports have more importantly the conversion efficiency of N to NV improves extended the field of view out to mm2 at the expense of spatial from between 1 and 5% up to 50%. Such improvements may bring to resolution [131]. Since the magnetic sensors are embedded in the life a fundamentally new view of neuronal activity and network diamond substrate total internal reflection microscopy (TIRF) can dynamics. be used to excite and report changes in the local magnetic field. This eliminates any photo-toxicity effects from the excitation light 8. Conclusions ensuring the method is non-invasive. Optical imaging provides fast and parallel signal acquisition and offers the potential to be incor- The set of materials available in the diamond stable offer an porated with other optical techniques such as opto-genetics which unparalleled level of choice in terms of combinations and variations could see the development of a fully integrated 2D solution for in electrical and optical properties. The very high biocompatibility optical stimulation and recording. The advantages of magnetic and chemical stability of diamond materials has always positioned based detection include no local charge screening as is the case them at least at the periphery of the neuromodulation research with electrical recording and potentially higher spatial resolution field. With modern synthesis and processing techniques however as the extracellular magnetic signal falls away much faster than the diamond materials have enjoyed a surge in popularity and the first complementary electrical signal [126]. In this regard it is critical to devices substantially fabricated from diamond have appeared. The bring the NV magnetic sensors as close as possible to the source of mechanical hardness, lack of ductility of diamond and difficulties the magnetic field. This can be achieved by culturing the neuronal connecting diamond to traditional electronics will always present network directly on the diamond MMA or as discussed in Section fabrication challenges. Recent advances however indicate that as 3.2 coating the homoepitaxial diamond within a thin adhesion machining methods continue to improve, diamond materials may layer of laminin or PDL to promote growth. From the physical increase in importance in the neural interfacing field and in the modelling of a hippocampal CA1 pyramidal neuron the magnetic biomedical research area generally. As a growth substrate diamond field generated from a single axon is expected to be of order tens of is clearly non-cytotoxic but claims that forms of diamond actively nanoTelsa/ms [126], with the majority of the magnetic signal enhance neural growth are presently debated. In most cases arising from intracellular ion currents. The current state-of-the-art treatment of diamond with an adhesion promoter protein is rec- 2D NV-based magnetometer has a DC magnetic sensitivity of 30 mT/ ommended to assist with long term survival of cultured neurons. 452 D.J. Garrett et al. / Carbon 102 (2016) 437e454

Diamond electrodes for electrical recording and neurotransmitter prosthesis, Biomaterials 35 (3) (2014) 908e915. detection are notable for their low noise performance but in [25] S.G. Lichter, M.C. Escudie, A.D. Stacey, K. Ganesan, K. Fox, A. Ahnood, et al., Hermetic diamond capsules for biomedical implants enabled by gold active particular for their ability to resist fouling. Diamond electrodes for braze alloys, Biomaterials 53 (2015) 464e474. neural stimulation are feasible where a process to increase elec- [26] T.W. Overton, J.E. Shigley, A history of diamond treatments, Gems Gemol. 44 e trochemical capacitance is included. Beyond electrical interroga- (1) (2008) 32 55. fl [27] V.S. Bormashov, S.A. Tarelkin, S.G. Buga, M.S. Kuznetsov, S.A. Terentiev, tion of neural tissue, optical recording of magnetic uctuations in A.N. Semenov, et al., Electrical properties of the high quality boron-doped neural tissue utilizing NV defects in nanodiamonds is an intriguing synthetic single-crystal diamonds grown by the temperature gradient prospect and a promising research area for the future. method, Diam. Relat. Mat. 35 (2013) 19e23. [28] S. Nad, Y.J. Gu, J. Asmussen, Growth strategies for large and high quality single crystal diamond substrates, Diam. Relat. Mat. 60 (2015) 26e34. Acknowledgments [29] M. Werner, R. Locher, Growth and application of undoped and doped dia- mond films, Rep. Prog. Phys. 61 (12) (1998) 1665e1710. [30] J.J. Gracio, Q.H. Fan, J.C. Madaleno, Diamond growth by chemical vapour DJG is supported by an Australian Research Council (ARC) DECRA deposition, J. Phys. D-Appl Phys. 43 (37) (2010) 22. grant DE130100922 and a National Health and Medical Research [31] J.E. Butler, A.V. Sumant, The CVD of nanodiamond materials, Chem. Vap. Council (NHMRC) Project Grant APP1101717. Depos. 14 (7e8) (2008) 145e160. [32] J. Morales, L.M. Apatiga, V.M. Castano, Growth of diamond films from tequila, Rev. Adv. Mater Sci. 21 (2) (2009) 134e138. References [33] F.G. Celii, J.E. Butler, Direct monitoring of ch3 in a filament-assisted diamond chemical vapor-deposition reactor, J. Appl. Phys. 71 (6) (1992) 2877e2883. [1] P.W. May, Diamond thin films: a 21st-century material, Philos. Trans. R. Soc. [34] J.E. Butler, Y.A. Mankelevich, A. Cheesman, J. Ma, M.N.R. Ashfold, Under- Lond Ser. A-Math Phys. Eng. Sci. 358 (1766) (2000) 473e495. standing the chemical vapor deposition of diamond: recent progress, J. Phys- [2] J. Konstanty, Sintered diamond tools: trends, challenges and prospects, Condes Matter 21 (36) (2009) 20. Powder Metall. 56 (3) (2013) 184e188. [35] M. Schwander, K. Partes, A review of diamond synthesis by CVD processes, [3] R.M. Hazen, The Diamond Makers, Cambridge University Press, Cambridge, Diam. Relat. Mat. 20 (9) (2011) 1287e1301. 1999. [36] R.S. Balmer, J.R. Brandon, S.L. Clewes, H.K. Dhillon, J.M. Dodson, I. Friel, et al., [4] M. Kamo, Y. Sato, S. Matsumoto, N. Setaka, Diamond synthesis from gas- Chemical vapour deposition synthetic diamond: materials, technology and phase in microwave plasma, J. Cryst. Growth 62 (3) (1983) 642e644. applications, J. Phys-Condes Matter 21 (36) (2009) 23. [5] Y. Matsui, S. Matsumoto, N. Setaka, TEM electron-energy loss spectroscope [37] CVD Diamond for Electronic Devices and Sensors, Wiley, 2009. study of the diamond particles prepared by the chemical vapor-deposition [38] J. Birrell, J.A. Carlisle, O. Auciello, D.M. Gruen, J.M. Gibson, Morphology and from methane, J. Mater Sci. Lett. 2 (9) (1983) 532e534. electronic structure in nitrogen-doped ultrananocrystalline diamond, Appl. [6] K. Kitahama, K. Hirata, H. Nakamatsu, S. Kawai, N. Fujimori, T. Imai, et al., Phys. Lett. 81 (12) (2002) 2235e2237. Synthesis of diamond by laser-induced chemical vapor-deposition, Appl. [39] J. Birrell, J.E. Gerbi, O. Auciello, J.M. Gibson, D.M. Gruen, J.A. Carlisle, Bonding Phys. Lett. 49 (11) (1986) 634e635. structure in nitrogen doped ultrananocrystalline diamond, J. Appl. Phys. 93 [7] S. Matsumoto, Chemical vapor-deposition of diamond in rf glow-discharge, (9) (2003) 5606e5612. J. Mater Sci. Lett. 4 (5) (1985) 600e602. [40] D.J. Garrett, K. Ganesan, A. Stacey, K. Fox, H. Meffin, S. Prawer, Ultra-nano- [8] M. Murakawa, S. Takeuchi, Applications of CVD diamond to wear resistant crystalline diamond electrodes: optimization towards neural stimulation components, J. Jpn. Soc. Tribol. 41 (9) (1996) 748e753. applications, J. Neural Eng. (1) (2012) 9. [9] M.S. Wong, R. Meilunas, T.P. Ong, R.P.H. Chang, Tribological properties of [41] E.J. Correa, Y. Wu, J.G. Wen, R. Chandrasekharan, M.A. Shannon, Electrical diamond films grown by plasma-enhanced chemical vapor-deposition, Appl. conduction in undoped ultrananocrystalline diamond thin films and its Phys. Lett. 54 (20) (1989) 2006e2008. dependence on chemical composition and crystalline structure, J. Appl. Phys. [10] F.A. Almeida, M.M. Maru, M. Shabani, F.J. Oliveira, R.F. Silva, C.A. Achete, 102 (11) (2007) 10. Enhancing the tribological performance under biodiesel lubrication using [42] M. Wang, N. Simon, C. Decorse-Pascanut, M. Bouttemy, A. Etcheberry, M.S. Li, CVD diamond coated parts, Wear 302 (1e2) (2013) 1370e1377. et al., Comparison of the chemical composition of boron-doped diamond [11] R.K. Roy, K.R. Lee, Biomedical applications of diamond-like carbon coatings: a surfaces upon different oxidation processes, Electrochimica Acta 54 (24) review, J. Biomed. Mater Res. Part B 83B (1) (2007) 72e84. (2009) 5818e5824. [12] L. Mattei, F. Di Puccio, B. Piccigallo, E. Ciulli, Lubrication and wear modelling [43] P. Strobel, M. Riedel, J. Ristein, L. Ley, Surface transfer doping of diamond, of artificial hip joints: a review, Tribol. Int. 44 (5) (2011) 532e549. Nature 430 (6998) (2004) 439e441. [13] D. Passeri, F. Rinaldi, C. Ingallina, M. Carafa, M. Rossi, M.L. Terranova, et al., [44] M.C. Rossi, F. Spaziani, S. Salvatori, G. Conte, Electronic properties of Biomedical applications of nanodiamonds: an overview, J. Nanosci. Nano- hydrogen and oxygen terminated surfaces of polycrystalline diamond films, technol. 15 (2) (2015) 972e988. Phys. Status Solidi a Appl. Res. 199 (1) (2003) 71e76. [14] J.C. Arnault, Surface modifications of nanodiamonds and current issues for [45] S. RS, CVD Diamond for Electronic Devices and Sensors, Wiley, 2009. their biomedical applications, in: N. Yang (Ed.), Novel Aspects of Diamond: [46] Y.V. Pleskov, Electrochemistry of diamond: a review, Russ. J. Electrochem 38 from Growth to Applications, Springer-Verlag Berlin, Berlin, 2015, pp. (12) (2002) 1275e1291. 85e122. [47] G.M. Swain, The use of cvd diamond thin-films in electrochemical systems, [15] V. Paget, J.A. Sergent, R. Grall, S. Altmeyer-Morel, H.A. Girard, T. Petit, et al., Adv. Mater 6 (5) (1994) 388e392. Carboxylated nanodiamonds are neither cytotoxic nor genotoxic on liver, [48] Y.V. Pleskov, Part II Electrochemistry Fo Diamond Films in Synthetic Dia- kidney, intestine and lung human cell lines, Nanotoxicology 8 (2014) 46e56. mond Films: Preparation, Electrochemistry, Characterization, and Applica- [16] L. Marcon, F. Riquet, D. Vicogne, S. Szunerits, J.F. Bodart, R. Boukherroub, tions, Wiley & Sons, Hoboken, New Jersey, 2011. Cellular and in vivo toxicity of functionalized nanodiamond in Xenopus [49] Y.V. Pleskov, Chapter 6 in Comprehensive Treatise of Electrochemistry, embryos, J. Mater. Chem. 20 (37) (2010) 8064e8069. Plenum Press, New York, 1980. [17] T. Hasebe, A. Hotta, H. Kodama, A. Kamijo, K. Takahashi, T. Suzuki, Recent [50] J.V. Macpherson, A practical guide to using boron doped diamond in elec- advances in diamond-like carbon films in the medical and food packing trochemical research, Phys. Chem. Chem. Phys. 17 (5) (2015) 2935e2949. fields, New Diam. Front. Carbon Technol. 17 (6) (2007) 263e279. [51] A.N. Correia, S.A.S. Machado, Hydrogen evolution on electrodeposited Ni and [18] W.B. Wilson, Evidence for hopping transport in boron-doped diamond, Phys. Hg ultramicroelectrodes, Electrochimica Acta 43 (3e4) (1998) 367e373. Rev. 127 (5) (1962) 1549. [52] S. Al-Riyami, S. Ohmagari, T. Yoshitake, X-ray photoemission spectroscopy of [19] E.A. Ekimov, V.A. Sidorov, E.D. Bauer, N.N. Mel'nik, N.J. Curro, J.D. Thompson, nitrogen-doped UNCD/a-C: H films prepared by pulsed laser deposition, et al., Superconductivity in diamond, Nature 428 (6982) (2004) 542e545. Diam. Relat. Mat. 19 (5e6) (2010) 510e513. [20] M. Amaral, P.S. Gomes, M.A. Lopes, J.D. Santos, R.F. Silva, M.H. Fernandes, [53] S.F. Cogan, Neural Stimulation and Recording Electrodes. Annual Review of Nanocrystalline diamond as a coating for joint implants: Cytotoxicity and Biomedical Engineering, Annual Reviews, Palo Alto, 2008, pp. 275e309. biocompatibility assessment, J. Nanomater (2008) 9. [54] Y.V. Pleskov, M.D. Krotova, V.G. Ralchenko, A.V. Saveliev, A.D. Bozhko, Elec- [21] M. Amaral, A.G. Dias, P.S. Gomes, M.A. Lopes, R.F. Silva, J.D. Santos, et al., trochemical behavior of nitrogenerated nanocrystalline diamond electrodes, Nanocrystalline diamond: In vitro biocompatibility assessment by MG63 and Russ. J. Electrochem 43 (7) (2007) 827e836. human bone marrow cells cultures, J. Biomed. Mater Res. Part A 87A (1) [55] C.G. Specht, O.A. Williams, R.B. Jackman, R. Schoepfer, Ordered growth of (2008) 91e99. neurons on diamond, Biomaterials 25 (18) (2004) 4073e4078. [22] A.P. Sommer, D. Zhu, F. Gagsteiger, H.J. Fecht, Sperm performance better on [56] B. Shi, Q.L. Jin, L.H. Chen, O. Auciello, Fundamentals of ultrananocrystalline diamond than on polystyrene, MRS Proc. (2013) 1511 (mrsf12-1511-ee07- diamond (UNCD) thin films as biomaterials for developmental biology: 05). embryonic fibroblasts growth on the surface of (UNCD) films, Diam. Relat. [23] P. Bergonzo, A. Bongrain, E. Scorsone, A. Bendali, L. Rousseau, G. Lissorgues, Mat. 18 (2e3) (2009) 596e600. et al., 3D shaped mechanically flexible diamond microelectrode arrays for [57] B. Shi, Q. Jin, L. Chen, A.S. Woods, A.J. Schultz, O. Auciello, Cell growth on eye implant applications: The MEDINAS project, Irbm 32 (2) (2011) 91e94. different types of ultrananocrystalline diamond thin films, J. Funct. Biomater. [24] K. Ganesan, D.J. Garrett, A. Ahnood, M.N. Shivdasani, W. Tong, A.M. Turnley, 3 (3) (2012) 588. et al., An all-diamond, hermetic electrical feedthrough array for a retinal [58] P. Bajaj, D. Akin, A. Gupta, D. Sherman, B. Shi, O. Auciello, et al., D.J. Garrett et al. / Carbon 102 (2016) 437e454 453

Ultrananocrystalline diamond film as an optimal cell interface for biomedical neuro-chemical and -electrical recording in neural prosthesis, applications, Biomed. Microdevices 9 (6) (2007) 787e794. J. Microelectromech Syst. 18 (3) (2009) 511e521. [59] F.Z. Cui, D.J. Li, A review of investigations on biocompatibility of diamond- [86] K. Ganesan, A. Stacey, H. Meffin, S. Lichter, U. Greferath, E.L. Fletcher, et al., like carbon and carbon nitride films, Surf. Coat. Technol. 131 (1e3) (2000) Diamond penetrating electrode array for epi-retinal prosthesis, in: 2010 481e487. Annual International Conference of the Ieee Engineering in Medicine and [60] B. Rezek, E. Ukraintsev, A. Kromka, Optimizing atomic force microscopy for Biology Society (Embc), 2010, pp. 6757e6760. characterization of diamond-protein interfaces, Nanoscale Res. Lett. 6 (2011) [87] M. Pagels, C.E. Hall, N.S. Lawrence, A. Meredith, T.G.J. Jones, H.P. Godfried, et 10. al., All-diamond microelectrode array device, Anal. Chem. 77 (11) (2005) [61] B. Rezek, E. Ukraintsev, M. Kratka, A. Taylor, F. Fendrych, V. Mandys, 3705e3708. Epithelial cell morphology and adhesion on diamond films deposited and [88] M. Bonnauron, S. Saada, C. Mer, C. Gesset, O.A. Williams, L. Rousseau, et al., chemically modified by plasma processes, Biointerphases 9 (3) (2014) 8. Transparent diamond-on-glass micro-electrode arrays for ex-vivo neuronal [62] P. Ariano, P. Baldelli, E. Carbone, A. Gilardino, A. Lo Giudice, D. Lovisolo, et al., study, Phys. Status Solidi A Appl. Mat. 205 (9) (2008) 2126e2129. Cellular adhesion and neuronal excitability on functionalised diamond sur- [89] A. Bonnauron, S. Saada, L. Rousseau, G. Lissorgues, C. Mer, P. Bergonzo, High faces, Diam. Relat. Mat. 14 (3e7) (2005) 669e674. aspect ratio diamond microelectrode array for neuronal activity measure- [63] P. Ariano, O. Budnyk, S. Dalmazzo, D. Lovisolo, C. Manfredotti, P. Rivolo, et al., ments, Diam. Relat. Mat. 17 (7e10) (2008) 1399e1404. On diamond surface properties and interactions with neurons, Eur. Phys. J. E [90] R. Kiran, L. Rousseau, G. Lissorgues, E. Scorsone, A. Bongrain, B. Yvert, et al., 30 (2) (2009) 149e156. Multichannel boron doped nanocrystalline diamond ultramicroelectrode [64] S.M. Ojovan, M. McDonald, N. Rabieh, N. Shmuel, H. Erez, M. Nesladek, et al., arrays: design, fabrication and characterization, Sensors 12 (6) (2012) Nanocrystalline diamond surfaces for adhesion and growth of primary 7669e7681. neurons, conflicting results and rational explanation, Front. Neuro- [91] K. Jackowska, P. Krysinski, New trends in the electrochemical sensing of engineering (2014) 7. dopamine, Anal. Bioanal. Chem. 405 (11) (2013) 3753e3771. [65] S. Kelly, E.M. Regan, J.B. Uney, A.D. Dick, J.P. McGeehan, E.J. Mayer, et al., [92] A.N. Patel, P.R. Unwin, J.V. Macpherson, Investigation of film formation Patterned growth of neuronal cells on modified diamond-like carbon sub- properties during electrochemical oxidation of serotonin (5-HT) at poly- strates, Biomaterials 29 (17) (2008) 2573e2580. crystalline boron doped diamond, Phys. Chem. Chem. Phys. 15 (41) (2013) [66] A. Bendali, C. Agnes, S. Meffert, V. Forster, A. Bongrain, J.C. Arnault, et al., 18085e18092. Distinctive Glial and Neuronal Interfacing on Nanocrystalline Diamond, PLoS [93] A.G. Guell, K.E. Meadows, P.R. Unwin, J.V. Macpherson, Trace voltammetric One 9 (3) (2014) 8. detection of serotonin at carbon electrodes: comparison of glassy carbon, [67] Y.C. Chen, D.C. Lee, C.Y. Hsiao, Y.F. Chung, H.C. Chen, J.P. Thomas, et al., The boron doped diamond and carbon nanotube network electrodes, Phys. effect of ultra-nanocrystalline diamond films on the proliferation and dif- Chem. Chem. Phys. 12 (34) (2010) 10108e10114. ferentiation of neural stem cells, Biomaterials 30 (20) (2009) 3428e3435. [94] P. Actis, A. Denoyelle, R. Boukherroub, S. Szunerits, Influence of the surface [68] Y.C. Chen, D.C. Lee, T.Y. Tsai, C.Y. Hsiao, J.W. Liu, C.Y. Kao, et al., Induction and termination on the electrochemical properties of boron-doped diamond regulation of differentiation in neural stem cells on ultra-nanocrystalline (BDD) interfaces, Electrochem. Commun. 10 (3) (2008) 402e406. diamond films, Biomaterials 31 (21) (2010) 5575e5587. [95] T. Kondo, Y. Niwano, A. Tamura, J. Imai, K. Honda, Y. Einaga, et al., Enhanced [69] O. Babchenko, N. Romanyuk, P. Jendelova, A. Kromka, Tailoring morphologies electrochemical response in oxidative differential pulse voltammetry of of diamond thin films for neural stem cells culturing, Phys. Status Solidi B- dopamine in the presence of ascorbic acid at carboxyl-terminated boron- Basic Solid State Phys. 250 (12) (2013) 2717e2722. doped diamond electrodes, Electrochimica Acta 54 (8) (2009) 2312e2319. [70] D.J. Garrett, A.L. Saunders, C. McGowan, J. Specks, K. Ganesan, H. Meffin, et al., [96] A.N. Patel, S.Y. Tan, T.S. Miller, J.V. Macpherson, P.R. Unwin, Comparison and In vivo biocompatibility of boron doped and nitrogen included conductive- reappraisal of carbon electrodes for the voltammetric detection of dopamine, diamond for use in medical implants, J. Biomed. Mater. Res. Part B Appl. Anal. Chem. 85 (24) (2013) 11755e11764. Biomater. 104 (1) (2015) 19e26. [97] E. Popa, H. Notsu, T. Miwa, D.A. Tryk, A. Fujishima, Selective electrochemical [71] P.A. Nistor, P.W. May, F. Tamagnini, A.D. Randall, M.A. Caldwell, Long-term detection of dopamine in the presence of ascorbic acid at anodized diamond culture of pluripotent stem-cell-derived human neurons on diamond - A thin film electrodes, Electrochem Solid State Lett. 2 (1) (1999) 49e51. substrate for neurodegeneration research and therapy, Biomaterials 61 [98] D.A. Tryk, H. Tachibana, H. Inoue, A. Fujishima, Boron-doped diamond elec- (2015) 139e149. trodes: the role of surface termination in the oxidation of dopamine and [72] W. Tong, K. Fox, K. Ganesan, A.M. Turnley, O. Shimoni, P.A. Tran, et al., ascorbic acid, Diam. Relat. Mat. 16 (4e7) (2007) 881e887. Fabrication of planarised conductively patterned diamond for bio-applica- [99] J. Shalini, K.J. Sankaran, C.-L. Dong, C.-Y. Lee, N.-H. Tai, I.N. Lin, In situ tions, Mater Sci. Eng. C-Mater Biol. Appl. 43 (2014) 135e144. detection of dopamine using nitrogen incorporated diamond nanowire [73] G. Piret, C. Hebert, J.P. Mazellier, L. Rousseau, E. Scorsone, M. Cottance, et al., electrode, Nanoscale 5 (3) (2013) 1159e1167. 3D-nanostructured boron-doped diamond for microelectrode array neural [100] S. Raina, W.P. Kang, J.L. Davidson, Optimizing nitrogen incorporation in interfacing, Biomaterials 53 (2015) 173e183. nanodiamond film for bio-analyte sensing, Diam. Relat. Mat. 18 (5e8) (2009) [74] R.J. Edgington, A. Thalhammer, J.O. Welch, A. Bongrain, P. Bergonzo, 718e721. E. Scorsone, et al., Patterned neuronal networks using nanodiamonds and the [101] S. Raina, W.P. Kang, J.L. Davidson, Nitrogen incorporated nanodiamond film effect of varying nanodiamond properties on neuronal adhesion and with ‘ridge’ surface morphology for detection of bio-analyte, Diam. Relat. outgrowth, J. Neural Eng. 10 (5) (2013) 9. Mat. 18 (2e3) (2009) 574e577. [75] S.P. Khan, G.G. Auner, G.M. Newaz, Influence of nanoscale surface roughness [102] Y.L. Zhou, R.H. Tian, J.F. Zhi, Amperometric biosensor based on tyrosinase on neural cell attachment on silicon, Nanomedicine Nanotechnol. Biol. Med. immobilized on a boron-doped diamond electrode, Biosens. Bioelectron. 22 1 (2) (2005) 125e129. (6) (2007) 822e828. [76] A. Voss, H. Wei, C. Muller, C. Popov, W. Kulisch, G. Ceccone, et al., Influence of [103] M. Dankerl, M.V. Hauf, M. Stutzmann, J.A. Garrido, Diamond solution-gated the surface termination of ultrananocrystalline diamond/amorphous carbon field effect transistors: properties and bioelectronic applications, Phys. Sta- composite films on their interaction with neurons, Diam. Relat. Mat. 26 tus Solidi A-Appl Mat. 209 (9) (2012) 1631e1642. (2012) 60e65. [104] C. Lete, F. Teodorescu, M. Marina, Determination of dopamine and ascorbic [77] A. Thalhammer, R.J. Edgington, L.A. Cingolani, R. Schoepfer, R.B. Jackman, The acid using boron doped diamond microelectrode arrays, Rev. Chim. 64 (5) use of nanodiamond monolayer coatings to promote the formation of (2013) 540e544. functional neuronal networks, Biomaterials 31 (8) (2010) 2097e2104. [105] K. Kisler, B.N. Kim, X. Liu, K. Berberian, Q. Fang, C.J. Mathai, et al., Transparent [78] N.J. Yang, J. Hees, C.E. Nebel, Diamond ultramicro- and nano-electrode arrays, electrode materials for simultaneous amperometric detection of exocytosis in: N. Yang (Ed.), Novel Aspects of Diamond: from Growth to Applications, and fluorescence microscopy, J. Biomater. Nanobiotechnology 3 (2A) (2012) Springer-Verlag Berlin, Berlin, 2015, pp. 273e293. 243e253. [79] X.J. Huang, A.M. O'Mahony, R.G. Compton, Microelectrode arrays for elec- [106] J. Park, J.J. Galligan, G.D. Fink, G.M. Swain, In vitro continuous amperometry trochemistry: approaches to fabrication, Small 5 (7) (2009) 776e788. with a diamond microelectrode coupled with video microscopy for simul- [80] J. Park, V. Quaiserova-Mocko, K. Peckova, J.J. Galligan, G.D. Fink, G.M. Swain, taneously monitoring endogenous norepinephrine and its effect on the Fabrication, characterization, and application of a diamond microelectrode contractile response of a rat mesenteric artery, Anal. Chem. 78 (19) (2006) for electrochemical measurement of norepinephrine release from the sym- 6756e6764. pathetic nervous system, Diam. Relat. Mat. 15 (4e8) (2006) 761e772. [107] B.A. Patel, J.J. Galligan, G.M. Swain, X. Bian, Electrochemical monitoring of [81] J.M. Halpern, M.J. Cullins, H.J. Chiel, H.B. Martin, Chronic in vivo nerve nitric oxide released by myenteric neurons of the guinea pig ileum, Neuro- electrical recordings of Aplysia californica using a boron-doped poly- gastroenterol. Motil. 20 (11) (2008) 1243e1250. crystalline diamond electrode, Diam. Relat. Mat. 19 (2e3) (2010) 178e181. [108] K. Yoshimi, Y. Naya, N. Mitani, T. Kato, M. Inoue, S. Natori, et al., Phasic [82] A.E. Hess, D.M. Sabens, H.B. Martin, C.A. Zorman, Diamond-on-polymer reward responses in the monkey striatum as detected by voltammetry with microelectrode arrays fabricated using a chemical release transfer process, diamond microelectrodes, Neurosci. Res. 71 (1) (2011) 49e62. J. Microelectromech Syst. 20 (4) (2011) 867e875. [109] R. Trouillon, M.I. Svensson, E.C. Berglund, A.S. Cans, A.G. Ewing, Highlights of [83] Najamuddin, S.A. Bazaz, H.Y. Chan, D.M. Aslam, Ieee. Impedance Character- selected recent electrochemical measurements in living systems, Electro- ization of MEMS Based All Diamond Neural Probe, Ieee, New York, 2008. chimica Acta 84 (2012) 84e95. [84] C. Hebert, E. Scorsone, A. Bendali, R. Kiran, M. Cottance, H.A. Girard, et al., [110] S. Natori, K. Yoshimi, T. Takahashi, M. Kagohashi, G. Oyama, Y. Shimo, et al., Boron doped diamond biotechnology: from sensors to neurointerfaces, Subsecond reward-related dopamine release in the mouse dorsal striatum, Faraday Discuss. 172 (2014) 47e59. Neurosci. Res. 63 (4) (2009) 267e272. [85] H.Y. Chan, D.M. Aslam, J.A. Wiler, B. Casey, A novel diamond microprobe for [111] G. Oyama, K. Yoshimi, S. Natori, Y. Chikaoka, Y.R. Ren, M. Funayama, et al., 454 D.J. Garrett et al. / Carbon 102 (2016) 437e454

Impaired in vivo dopamine release in parkin knockout mice, Brain Res. 1352 Diam. Relat. Mat. 19 (7e9) (2010) 772e777. (2010) 214e222. [122] Y.T. Kim, Y. Ito, K. Tadai, T. Mitani, U.S. Kim, H.S. Kim, et al., Drastic change of [112] N. Joye, A. Schmid, Y. Leblebici, Electrical modeling of the cell-electrode electric double layer capacitance by surface functionalization of carbon interface for recording neural activity from high-density microelectrode nanotubes, Appl. Phys. Lett. 87 (23) (2005) 3. arrays, Neurocomputing 73 (1e3) (2009) 250e259. [123] A.E. Hadjinicolaou, R.T. Leung, D.J. Garrett, K. Ganesan, K. Fox, [113] C.M. Hales, J.D. Rolston, S.M. Potter, How to culture, record and stimulate D.A.X. Nayagam, et al., Electrical stimulation of retinal ganglion cells with neuronal networks on micro-electrode arrays (MEAs). Journal of visualized diamond and the development of an all diamond retinal prosthesis, Bio- experiments, JoVE (39) (2010). materials 33 (24) (2012) 5812e5820. [114] V. Maybeck, R. Edgington, A. Bongrain, J.O. Welch, E. Scorsone, P. Bergonzo, [124] K. Fox, H. Meffin, O. Burns, C.J. Abbott, P.J. Allen, N.L. Opie, et al., Develop- et al., Boron- doped nanocrystalline diamond microelectrode arrays monitor ment of a magnetic attachment method for bionic eye applications, Artif. cardiac action potentials, Adv. Healthc. Mater 3 (2) (2014) 283e289. Organs (2015). [115] P. Ariano, A. Lo Giudice, A. Marcantoni, E. Vittone, E. Carbone, D. Lovisolo, [125] M. Cottance, S. Nazeer, L. Rousseau, G. Lissorgues, A. Bongrain, R. Kiran, et al., A diamond-based biosensor for the recording of neuronal activity, Biosens. Diamond Micro-Electrode Arrays (MEAs): a New Route for In-vitro Appli- Bioelectron. 24 (7) (2009) 2046e2050. cations, Ieee, New York, 2013. [116] M. Dankerl, S. Eick, B. Hofmann, M. Hauf, S. Ingebrandt, A. Offenhausser, et [126] L.T. Hall, G.C.G. Beart, E.A. Thomas, D.A. Simpson, L.P. McGuinness, J.H. Cole, al., Diamond transistor array for extracellular recording from electrogenic et al., High spatial and temporal resolution wide-field imaging of neuron cells, Adv. Funct. Mater 19 (18) (2009) 2915e2923. activity using quantum NV-diamond, Sci. Rep. 2 (2012) 9. [117] J.M. Halpern, S.T. Xie, G.P. Sutton, B.T. Higashikubo, C.A. Chestek, H. Lu, et al., [127] M.W. Doherty, N.B. Manson, P. Delaney, F. Jelezko, J. Wrachtrup, Diamond electrodes for neurodynamic studies in Aplysia californica, Diam. L.C.L. Hollenberg, The nitrogen-vacancy colour centre in diamond, Phys. Rep- Relat. Mat. 15 (2e3) (2006) 183e187. Rev Sec Phys. Lett. 528 (1) (2013) 1e45. [118] E.W. Keefer, B.R. Botterman, M.I. Romero, A.F. Rossi, G.W. Gross, Carbon [128] S. Steinert, F. Dolde, P. Neumann, A. Aird, B. Naydenov, G. Balasubramanian, nanotube coating improves neuronal recordings, Nat. Nanotechnol. 3 (7) et al., High sensitivity magnetic imaging using an array of spins in diamond, (2008) 434e439. Rev. Sci. Instrum. 81 (4) (2010) 5. [119] Z. Yang, E.W. Keefer, W. Liu, Noise characterization, modeling, and reduction [129] D. Le Sage, K. Arai, D.R. Glenn, S.J. DeVience, L.M. Pham, L. Rahn-Lee, et al., for in vivo neural recording, in: Y. Bengio, D. Schuurmans, J.D. Lafferty, Optical magnetic imaging of living cells, Nature 496 (7446) (2013) 486e489. C.K.I. Williams, A. Culotta (Eds.), Advances in Neural Information Processing [130] D.A. Simpson, J. Tetienne, K. Ganesan, L.T. Hall, S. Petrou, R.E. Scholten, et al., Systems 22, Curran Associates Incorporated, Vancouver, 2009, pp. Magneto-optical imaging of thin magnetic films using spins in diamond, 2160e2168. arxiv (2015). arXiv:1508.02135 (cond-mat.mes-hall). [120] H.X. Ji, X. Zhao, Z.H. Qiao, J. Jung, Y.W. Zhu, Y.L. Lu, et al., Capacitance of [131] D.R. Glenn, K. Lee, H. Park, R. Weissleder, A. Yacoby, M.D. Lukin, et al., Single- carbon-based electrical double-layer capacitors, Nat. Commun. 5 (2014) 7. cell magnetic imaging using a quantum diamond microscope, Nat. Methods [121] T. Watanabe, T.K. Shimizu, Y. Tateyama, Y. Kim, M. Kawai, Y. Einaga, Giant 12 (8) (2015) 736e738. electric double-layer capacitance of heavily boron-doped diamond electrode,