An overview of recent applications of reduced graphene oxide as a basis of electroanalytical sensing platforms

The academic literature using graphene within the field of electrochemistry is substantial. Graphene can be fabricated via a plethora of routes with each having its own unique merits (e.g. cost, fabrication time, quality and scale) and reduced graphene oxide (rGO) is more often the material of choice for electrochemical sensors and associated applications due to its ease of fabrication and ability to be mass produced on the kilogram scale. This review overviews pertinent developments in the use of rGO as the


Introduction
Graphene is now extensively researched since the advent of its reported range of physical properties back in 2004 and 2005 [1,2] and it is acknowledged that graphene has been theoretically explored since the 1940s and known to exist since the 1960s [3]. There are a substantial number of papers reporting the use of graphene in the field of electrochemistry making it now impossible to summarise the field in a single review. Consequently in this paper, we focus and only overview the recent field of electroanalytical sensors from 2016 to 2017.
The fabrication of graphene is achieved by a plethora of ways, as identified within Figs. 1 and 2 each of which having their own merits; there are many papers and reviews that focus extensively upon the variety of fabrication methodologies and as such, will not be covered here. Cleary pristine graphene can be fabricated allowing the fundamental understanding of graphene to be deduced but is not readily used as the basis of electrochemical sensors due to its high cost and lack of manufacturing scalability. Other forms of graphene are thus explored, due to these very reasons. As such, reduced graphene oxide (rGO) is widely utilised, which is usually fabricated from the oxidation/exfoliation of graphite to graphene oxide (GO) and then its reduction to graphene via a variety of routes such as chemical, thermal or electrochemical; this is summarised within Fig. 2. There are clearly a wide range of ways to produce graphene, ranging from producing pristine graphene through to what is termed rGO.
Initially, it is pertinent to define exactly what graphene oxide (GO) and rGO is. Dreyer et al. [4] have usefully provided a summary: GO is chemically modified graphene, which is prepared by oxidation and exfoliation that is accompanied by extensive modification of the basal plane. GO is a monolayer with a high oxygen content, typically characterised by a C/O ratio of less than 3:1 and  The most common graphene production methods. Each method has been evaluated in terms of graphene quality (G), cost aspect (C; a low value corresponds to high cost of production), scalability (S), purity (P) and yield (Y) of the overall production process.
Reprinted with permission from Nature [41]. typically closer to 2:1, respectively. Note that GO is not a new material given that it has been known to exist since the 1840s [4][5][6]. It has been largely overlooked in this field of research, being considered predominately as a precursor for graphene synthesis. Brodie [6] reported that graphite oxide is defined as: bulk solid made by the oxidation of graphite through processes that functionalise the basal planes and increase the interlayer spacing. Graphite oxide can be exfoliated in solution to form (monolayer) graphene oxide or partially exfoliated to form few-layer graphene oxide. Thus, rGO is defined (as above) that has been reductively processed by chemical, thermal, microwave, photo-chemical, photo-thermal or microbial/bacterial methods to reduce its oxygen content.
Interestingly, GO is reported to have limited electrical and thermal transport due to the oxygen functionalisation on the basal plane surface [7], and it has been largely overlooked in this field of research, being considered predominantly as a precursor for the synthesis of graphene. GO has been reported to be beneficial in a very few electrochemistry areas [8,9], such as in energy storage devices [10], and nucleic acid monitoring [11]. Recent work by Brownson et al. [12] has demonstrated that GO, used as is, gives rise to unique electrochemical responses where the oxygenated species encompassing GO strongly influence and dominate the observed voltammetry, which is crucially coverage dependant. rGO is an interesting member of the graphene family since it, along with GO, are the only variants where fabrication can be scale-up and be manufactured on the kilogram scale [7]. The fundamental understanding of graphene is generally explored using graphene that is made via mechanical exfoliation or Chemical Vapour Deposition (CVD) [13][14][15][16], i.e. pristine graphene, which allows stricter control over the number of graphene layers, but fundamental work on rGO is yet to be fully explored, although general trends can be readily transferred such as the density of edge plane likesites/defects/number of graphene layers dominating the electrochemical response [17,18]. Note the C/O ratio is not zero, i.e. like pristine graphene and the C/O amount and type of C O groups may influence its electrochemical response.
In this review, we overview recent developments utilising rGO as the basis of electroanalytical sensors. Table 1 provides an overview summarising the last two years contribution to this field; we limit this timeframe due to the sheer volume of publications. a Value not stated within the paper but estimated from presented data, -; information not stated within the paper, 3D; three dimensional, mp; macroporous, AFP; ␣-fetoprotein, AuNPs-PEDOT/PB-rGO; gold nanoparticlespoly(3,4-ethylenedioxythiophene)/Prussian blue-reduced graphene oxide nanocomposite, BFE; bismuth coated glassy carbon, CRP; C-reactive protein, Cu/AMT; opper-2-amino-5-mercapto-1,3,4-thiadiazole, DNA; deoxyribose nucleic acid, DOPA; 3,4-dihydroxy-l-phenylalanine, EDS; energy-dispersive X-ray spectroscopy, EELS; electron energy loss spectroscopy, ERGO; electrochemically reduced graphene oxide, FESEM; field emission scanning electron microscope, FS; Fluorescense spectra, FT-IR; Fourier transform infrared spectroscopy, GC; glassy carbon electrode, GO; graphene oxide, HPLC; high performance liquid chromography, MWCNTs@rGONRs; multiwalled carbon nanotubes@reduced graphene oxide nanoribbons/chitosan, NMR; nuclear magnetic resonance, NP; nanoparticle, P(l-Cys); poly(l-Cysteine), PGE; pencil graphite electrode, POM; polyoxometalate, rGO; reduced graphene oxide, SEM; scanning electron microscope, SPE, screen-printed electrode, SPR; Surface Plasmon Resonance, STM; scanning tunnel microscope, TEM, transmission electron microscope, TGA; thermogravimetric analysis, TNM; tert-nonyl mercaptan, UV-vis; ultraviolet-visible spectroscopy, VSM; vibrating sample magnetometer, XPS; X-ray photoelectron spectroscopy, XRD; X-ray diffraction analysis.  Table 1 is grouped into analytical targets, the supporting electrode substrate, fabrication method, analytical outputs and characterisation techniques and number of graphene layers reported for each sensor/composite. We provide a critical summary and suggest future directions for this field. Table 1 lists selected literature over the period 2016-2017 describing electroanalytical sensors reported in the academic literature which utilise rGO in one form or another. As noted in Table 1, many of the graphene production methods rely upon solvent-based synthesis of reduced graphene oxide (rGO), which can be created by one of a number of production methods and are sometimes not reported in sufficient detail [19][20][21]. Each method yields rGO of different qualities, and therefore producing differing electrochemical responses that have to be accounted for. To quote one example, one recent report highlights a voltammetric determination method for the anti-inflammatory drug furazolidone [22], but the methods employ commercially available GO with an unknown production method, adding a layer of ambiguity to the work that requires removal using a range of characterization techniques, discussed previously. This is demonstrated succinctly in the work by Poh et al. that shows rGO synthesized from Staudenmaier, Hoffman, and Hummers methods yield significantly different electrochemical responses [23]. This is due to the different reagents used for the oxidation of graphite, which not only introduce different levels of oxygenated species, but also introduce impurities, especially in the case of the Hummers method, which is purported to introduce N heteroatoms to the graphene backbone. Note other reports have contradicted this [24]. It is interesting to note that the O/C ratio of Hummers-type GO has been demonstrated to vary insignificantly with respect to reagent concentrations and conditions, however the crudely-termed "oxygen-related bonds" may be more informative as to how reaction conditions in the Hummers method affects the GO structures, and hence the electrochemistry [24]. It is found from such work that controlling the residence time of the chemical reactions at 35 • C in the Hummers method has a significant effect upon the total number of oxygen bonds in the GO product. Despite this, Hummers methods are the most common due to the relative safety of the chemical reaction, as it does not use explosive chemicals. Aside from the variable oxygenated species, there is also a question of the impurities in Hummers-type rGO, often in the form of manganese (IV) structures as reported elsewhere [25]. While such materials can be beneficial, one must be aware of the level of impurity and take this into account when designing rGO-based sensors. It is clear from inspection of Table 1 that characterisation methods are not consistently applied nor are the number of layers of the material which limits comparison of the electroanalytical sensors and their fundamental understanding.

GO utilised in electrochemical sensing applications
Interesting work utilising rGO have reported a gold nanoparticles-poly(3,4-ethylenedioxythiophene)/Prussian bluereduced graphene oxide (AuNPs-PEDOT/PB-rGO) nanocomposite which has been utilised as a stable and sensitive label-free electrochemical immunosensor for detecting ␣-fetoprotein (AFP) [26]; see Fig. 3 for an overview of the electroanalytical platform. The use of the rGO was reported to increase the sensors performance through an increased surface area and being a useful supporting material for the Prussian blue compound. This rGO based immunosensor showed a sensitive response to AFP in a linear range from 0.01 to 50 ng mL −1 with a low detection limit of 3.3 pg mL −1 and was shown to allow the determination of AFP within human serum. Another approach for rGO fabrication is chemically reduced rGO. One recent success of this technique was reported by Kiranş an et al., who utilized HI to chemically reduce wet-synthesized graphene oxide using the method reported by Kovtyukhova et al. [20,27]. Their work describes graphene paper, fabricated using a sonicated suspension rGO, Ag nanoparticles, and pyronin Y. The mixture was transferred to a polycarbonate membrane and vacuum filtered, leaving a paper-like structure behind consisting of a rGO/AgNP/pyronin Y composite. After treatment with HI and electropolymerisation of pyronin Y using voltammetry, the paper electrode was ready to use. The electrode was employed towards dopamine detection, finding a similar performance to several other graphene composites studied previously. However, the novelty of graphene paper allows this sensor to stand out, especially since it would be feasible to scale this up into larger sheets that could be easily shaped and sized by hand. Graphene GO can also be incorporated into epoxy resin electrodes, carrying the advantage of long term electrode stability, durability, and repeatability over proof-of-concept approaches such as dropcasting. Muñoz et al. performed amperometric experiments using rGO made through a Hummers method with ascorbic acid as the reductant [28]. The electrode utilised gold nanoparticles and beta-cyclodextrin (see Fig. 4), which is a molecule often used in host-guest chemistry as an acceptor site for some simple biomolecules, including thyroxine. The epoxy resin approach was demonstrated to detect nanomolar concentrations of thyroxine, which are required for adequate diagnosis of thyroid activity in adults. Host-guest interactions are also exploited for the detection of BrO 3 − in water samples by Palanisamy et al. [29]. Their work exploits the host-guest properties of beta-cyclodextrin to immobilise haemoglobin upon rGO. The haemoglobin in this instance is the signal reporting aspect of the electrode due to its specificity towards the BrO 3 ion, which reduces to Br(V) to Br − , resulting from haemoglobin's affinity for dioxygen. Microwave-assisted synthesis of graphene is the preferred approach in the work by Li et al. [30] In what is essentially a Hummers-type GO, the microwave assistance speeds up the fabrication process for the rGO. The rGO is then subjected to further microwave treatment in the presence of Ag(NO 3 ) 2 and Zn(CH 3 CH 2 COO) 2 in the presence of ethanol and NaOH to create a platform for glucose oxidase to attach efficiently. The approach reports lower detection limits than similar works, though the linear ranges are notably smaller. Electrochemically reduced graphene oxide (ERGO) has been used as the basis of an ethanol biosensor where nicotinamide adenine dinucleotide (NADH) is used as a mediator with gold-silver bimetallic nanoparticles (Au-AgNPs), poly(l-Cysteine) (P(l-Cys)) immobilised upon the ERGO [31]. Fig. 5 provides a summary of the various steps required to produce the sensor. This composite electrode is reported to exhibit an excellent electrocatalytic response towards NADH at a low oxidation potential (+0.35 V) and minimization of surface contamination due to the synergistic effects of the Au-AgNPs, polymer and ERGO [31]. This ERGO based sensor was also used as a sensitive ethanol biosensor, which was prepared with alcohol dehydrogenase (ADH) via glutaraldehyde, bovin serum albumin and nafion (Naf). There was a linear response for ethanol in the concentration range from 0.017 to 1.845 mM with a low detection limit of 5.0 M reported. The GCE/Au-AgNPs/P(l-Cys)-ERGO/ADH/Naf based sensor was utilised as the basis of an ethanol sensor in different commercial beverages (Fig. 6).
Other notable works include the detection of dopamine by Kim et al., who utilised chemically cleaned rGO on a glassy carbon electrode as the working electrode in their work [32]. In many respects, their work demonstrates that simplicity can be a good thing in electroanalytical sensing devices, because they only use the two elements in their electrode design. Such electrodes may not exhibit the selectivity characteristics that an enzymatic system may offer, but for the case of dopamine detection, this is not an issue provided the electrochemical oxidation peaks of dopamine, ascorbic acid, and uric acid can be independently discerned. Their work observes dopamine the presence of 1 mM ascorbic acid with a sensitivity of 0.588 A M −1 cm −2 . The limit of detection was estimated to be 2.64 M, which is around the average level of dopamine passed in urine by an adult in the western world.
Other methods use rGO as a growth substrate for heterostructures of metal oxides. One such example is the work by Zhao et al., who grew ZnO nanorods on rGO that are used as molecular wires [33]. Their work uses a rGO thin film prepared by filtering rGO suspensions through a membrane, collecting the rGO upon the membrane. The rGO was dried and transferred to a SiO 2 substrate, then ZnO was hydrothermally grown at 90 • C for 1 h. The approach is fairly simple in its nature and provides a more robust Reproduced from Ref. [28]. method for a repeatable graphene-based electrode. The electrode was used to anchor glucose oxidase, providing a centre for glucose oxidation that could be efficiently transmitted to the potentiostat through the synergistic effect of the rGO network and the ZnO molecular wires transferring the signal from glucose oxidase. The sensor was demonstrated to have a sensitivity to glucose of 17.64 A mM −1 , which is comparable to similar methods in the current literature (see Table 1). Electrochemically reduced graphene oxide (ERGO) films have been utilised to modify screen-printed electrodes (SPE) for the detection of Zn/Cd and Pd [34], and pencil graphite electrodes for the detection of Pd in milk Samples [35]. In both cases the detection mechanism was via stripping voltammetry. Improvements in sensor performance was observed using rGO, adsorbed onto large surface area electrodes, giving rise to favourable electron transfer sites.
The summary of the use of rGO as the basis of the sensor is that it is mainly used to increase the assessable electrochemically active area and also provide stability for the immobilisation of electrocatalytic compounds/moieties. In the case that the electrochemical mechanism is diffusionally based, the peak current is governed by the Randles-Ševćik equation [36], and the increase in the electrochemically active area, where the electroanalytical sensor is based upon rGO (rather than say a compound modified upon its surface), will give rise to larger voltammetric currents and hence a larger signal output which should improve the sensitivity and limit of detection. However, control experiments using just graphite (or similar) are not routinely performed and needs to be undertaken in cases where rGO is reported to give rise to significant outputs. One area that benefits greatly from the use of rGO is where the underlying electrochemical mechanism is adsorption based, as is the case via stripping voltammetry, where the increased surface area gives rise to a greater proportion of adsorption sites and hence improvements in the sensors performance will be realised. This has been reported, for example by Göde et al. [37] modified a glassy carbon electrode (GCE) with calixarene functionalized reduced graphene oxide (CA/RGO) and applied it to the sensing of, Fe(II), Cd(II) and Pb(II), in aqueous solutions. The CA/RGO was found to significantly improve the sensitivity of electrochemical responses and separation of the target heavy metals. The electrochemical oxidation of Fe(III), Cd(II), and Pb(II) was performed by square wave voltammetry (SWV) with analytically useful detection limits possible with the sensor applied into pharmaceutical formulations [37]. Clearly such approaches, have merit capitalising upon the underlying electrochemical mechanism, i.e. adsorptive over diffusional which really shows the benefit of using rGO.

Concluding remarks
This review has explored the recent field of using rGO as the basis of electroanalytical sensors, where mainly it is found that rGO is the basis of sensor and is usually modified with nanoparticles or electrocatalytic moieties. In this approach, rGO is reported to provide a larger surface area and stable surface onto which nanoparticles and electrocatalytic compounds are reproducibility immobilised. However, in other cases there are no justifications of why rGO is beneficial or why it is being used when graphite could equally be usefully utilised. The large surface area of rGO is beneficially utilised when the electrochemical mechanism is adsorptive in nature, this is exemplified when used in stripping voltammetry. Our analysis of the field suggests the following is addressed in future publications, which are currently seldom considered: • Consider the amount of rGO used in sensors and consider coverage experiments to determine the optimal amount of rGO required. • The number of layers of rGO should be reported to allow direct comparison and reproducibility in the field. • Since rGO is not fully reduced, i.e. to pristine graphene, the C/O ratio, amount and C O groups/types needs to be deduced as these might contribute to the electrochemical response. • Controls are made with graphite for the proposition of rGO as in most instances, based on the design of the sensor, the former is more than adequate. • The methodology of the rGO fabrication is made, i.e. following the work of Pumera et al. [38] which overviews the terminology that should be incorporated, i.e. chemically rGO should be notated as CrGO or similar to allow future understanding and differentiation. • The underlying supporting electrode material is not usually explored and this will have an effect upon the electrochemical sensor due to the supporting electrodes roughness and rGOelectrode substrate interactions. Generally however, if a sensor is really going to be implemented into the field, the research should be undertaken upon screen-printed electrodes which allow translation of research from the laboratory in-to-the field due their scales of economy and high reproducibility [39,40].