The Graphene Surface Chemistry and Adsorption Science

Abstract

Graphene (G) has attracted immense attention due to its exceptional physicochemical and electronic properties, and quite a large amount of literature has accumulated on this subject over the last few decades. The current work, based on a systematic review of the relevant literature, was designed to provide an overview of G surface chemistry with respect to its adsorption science. The aim was to improve knowledge of the graphene surface chemistry while informing new strategies for designing and implementing new G materials for emerging applications. The key G surface reactions include: molecular adsorption of gases, bandgap tuning, gas detection tests; alkali metal storage for battery technology; G vacancy engineering; environmental amelioration of pollutants; and sensors and biosensors technology. GO (graphene oxide) or G has been surface-modified using nonmetals, metals, metal oxides, or organics. In general, GO and related functionalized materials have high affinity and adsorption efficacy for ionic adsorbates, whereas pristine G, and reduced graphene oxide (rGO), exhibits enhanced hydrophobic surfaces with propensity to strong π-π interactions. The metals’ adsorption and doping can impart G magnetic and metallic character, whereas molecular intercalations tend to induce a G bandgap for nano-electronic and nanophotonic uses among other interactions.

1. Introduction

Graphene (G), a new layered carbon material, is one of the most studied materials over the last few decades. Investigations about G have centered on new fabrication methods [1]; improving understanding of its properties [2], designing and development of new derivatives [3], and exploring prospects for the new G materials [4]. In examining G properties, most studies focused on specific surface reactions, especially the adsorption of organic contaminants [5] and simple molecular gases [6], as well as heavy metal doping [7], and interaction with pharmaceutical agents [8]. The adsorption performance of G and its derivatives, primarily graphene oxide (GO), reduced graphene oxide (rGO), nonmetal- and transition metal-doped graphene (TM-Graphene), has been assessed extensively for their adsorption potential towards various adsorbates in terms of the reaction conditions; equilibrium, kinetic, and thermodynamic analyses; and the electronic characterization of adsorption energies, geometries, DOS, dipole moments, the reaction work functions, etc. The Langmuir [9], Freundlich [10], Temkin, Polanyi [11], and the Dubinin–Radushkevich [12] isotherms are widely used for evaluating the adsorption capacities, reaction mechanisms and thermodynamic of the adsorption processes. The adsorption kinetics, on the other hand, are often evaluated using the Lagergren first-order [13] and pseudo-second order [14] kinetics equations. Pristine graphene (G) surface presents basically three adsorption sites: the hollow site (H), at the center of a carbon hexagon; the bridge site (B), in the middle of a C∙C bond; and the top site (T), on the top of a carbon atom. Molecular orientations are used to determine the most stable adsorption configurations and the adsorption energy, E_ad, is defined according to Hess’ law as:

$$E_{ad}=E_{molecule}+E_{graphene}–E_{molecule@graphene}$$E1

where, E_molecule, E_graphene, and E_{molecule@graphene} are the energies of the isolated gas molecule, graphene, and molecule–graphene adsorption complex, respectively. The first-principles electronic calculations are used basing on the spin-polarized density-functional theory (DFT) via the Vienna ab initio simulation package (VASP) [15] with the exchange and correlation energy included through a generalized gradient approximation in PBE format [16]. Similarly projected augmented-wave (PAW) potentials [17] are also used to describe the ion-electron interactions.

However, comprehensive overview of the surface chemistry of G is lacking in the literature and the current work was, therefore, designed to provide an inclusive update of the surface adsorption of G and the related materials. The main goal was to improve the understanding of the graphene chemistry and, thus, contribute to information of emerging strategies for designing new G materials and their applications. GO is characterized by several reactive surface groups with strong acidity and high adsorption capacities for basic and cationic species. G on the other hand, exhibits a hydrophobic surface presenting high adsorption capacities for chemicals that show strong π-π interactions. Suitable modifications of GO or G with metal oxides and/or organic groups can produce nanocomposites with more enhanced tailored adsorption capacities and separation efficiencies for specific diverse applications, including fuel gas adsorption and storage, battery technologies, G vacancy electronic and photonic engineering, pollution ameliorating, and sensors and biosensors fabrications. The current work aims to consolidate relevant literature and provide new insights for understanding G chemistry and spur new strategies, and application of G and its derivatives.

2. Application of graphene in fuel gas storage

The search for sustainable clean energy remains one of the fundamental socioeconomic development goal of the 21st century. Actions toward realization of this goal include fuel gas storage techniques using efficient adsorbents and graphene (G) materials have attracted particularly intense attention due to their unique properties that lend them to diverse versatile modifications and applications. The large specific surface area of G has attracted researchers to explore it for diverse adsorptive applications. Ma et al. [18], for example, evaluated hydrogen adsorption using pristine single-layer G samples with a BET-specific surface area of 156 m²/g and found that only 0.4 wt.% hydrogen could adsorb on G at 77 K under 100 kPa pressure. Furthermore, even lower hydrogen adsorption rates of <0.2 wt.% were obtained at ambient temperature under elevated 6 MPa-pressure. The small gravimetric hydrogen uptake by G samples were attributed to low specific surface area of the materials and to the weak hydrogen binding into the G surfaces. Elsewhere, while using high pressures of about 40 bars, and Szczęśniak et al. [19], obtained optimum H₂ adsorption of 5.5–7 wt.% and 4.4–4.8 wt.% for activated rGO, and for transition metal (TM) nanoparticles doped-rGO, respectively.

Meanwhile, while investigating different hydrogen adsorption sites of G, Arellano et al. [20], observed that the most stable configuration was obtained when H₂ molecules physisorbed over the hollow (H) site, although the barriers for classical diffusion were, particularly, small. Nevertheless, in an ab initio study of atomic hydrogen adsorption on G, Ivanovskayaet al. [21], focused their investigations on the adsorption characteristics of the hydrogenation surface coverage of G. They observed that at high surface coverage, the resulting strain from substrate relaxation controlled H chemisorption leading to localized surface curvature. The chemisorption energy barrier was caused by the relaxation and by the adsorbent carbon atom rehybridization. This showed independence from the optimization technique and the system size. The H desorption barrier was, however, very sensitive to the correct structural relaxation and it was controlled by the degree of system hydrogenation.

According to Miura et al. [22], the carbon atoms reconstructions played an leading role in G interactions with H₂ molecules. The activation barrier for H₂ dissociation from an unrelaxed G was ~4.3 eV for a T–H–T geometry, and ~4.7 eV for a T–B–T geometry. The center of mass position of H₂ was, therefore, at the hollow site, and the two H atoms of the molecule were directed toward the top sites on the G structure. When the carbons relaxed, the activation barrier decreased to 3.3 eV for the T–H–T geometry and 3.9 eV for the T–B–T geometry, in which case, the two carbon atoms near the H atoms moved 0.33 Å toward the gas phase for the T–H–T geometry and 0.26 Å for the T–B–T geometry. Elsewhere, while investigating H atoms adsorption on G using first-principles while employing plane-wave based periodic density functional theory (DFT), Casolo et al. [23] selected a 5 × 5 surface unitcell to study H atoms’ single- and multiple-adsorptions on G surface. They computed the binding and barrier energies for sequential sticking of several H atoms’ configurations on top of G carbon. The authors recorded binding energies of 0.8–1.9 eV per atom with barriers to sticking of 0.0–0.15 eV. Magnetic structures formed in which spin density localized on a 3 × 3R30° sublattice and the binding (barrier) energies for sequential adsorption increased/decreased with the site-integrated magnetization. These results could be explained on the basis of the valence-bond resonance theory of planar π-conjugated systems. It suggested that preferential sticking due to barrierless adsorption was limited to the formation of hydrogen pairs.

According to Kim et al. [24], however, presence of certain metals at G’s vacancy defects increased the binding energy and hydrogen adsorption of the metals tested; however, the Ca-vacancy complexes showed most favorable binding energy and overall hydrogen adsorption capacities. So, Ataca et al. [25], were able to demonstrate the capacity of Ca-adsorbed G has a recyclable hydrogen adsorption media. They found that Ca chemisorbed onto G by donating part of its 4 s charge to the empty π* G band imparting it with a positive charge and the metallic character, which increased its H₂ adsorption, in turn. The H₂ uptake capacity of Ca-modified G could be improved further by adsorbing Ca on both sides of the G nanosheets and utterly by saturating the Ca surface coverage as because Ca does not agglomerate on G surfaces at high surface concentration. This due to the high coulombic repulsions between the adsorbed Ca atoms.

Besides hydrogen, Gao et al. [26] reviewed G adsorbents for CH₄. They observed weak interactions and charge transfer from intrinsic G to CH₄. The authors noted that the affinity of methane for graphene surfaces could, however, be promoted by doping the Ni atom, setting a single vacancy defect, and/or adding oxygen-containing functional groups into the G surfaces.

3. Application of graphene in electronic technology: bandgap engineering

The unique electronic properties of G means that G could continue to attract attention in potential electronic and photonic applications. The charge carriers in G behave, such as massless Dirac fermions, and G shows ballistic charge transport ideal for circuit fabrication. However, the lack of appreciable bandgap around the Fermi level, which is the key concept for semiconductors critical for controlling electronic conductivity, restricts its applications in nano-electronic and nanophotonic devices. Recent interests in G were, therefore, inspired by its promise in fashioning two-dimensional semiconductor materials with tunable bandgap lending itself to diverse uses. For this, several bandgap engineering methods have been proposed and Berashevich & Chakraborty [27], for example, described adsorption of water and other gases on G and showed that the molecules behaved like defects on nanoscale G surface facilitating bandgap tunability and permitting magnetic ordering of localized states at the G edges. This showed that the molecules pushed the wave functions corresponding to α -spin (up) and β -spin (down) G states to the opposite (zigzag) edges breaking the sublattice and molecular point group symmetry of the material. The wave-function displacement was controlled by the adsorbed molecule lending itself to tunable bandgap opening.

Accordingly, Balog et al. [28] demonstrated a G bandgap tunability using atomic hydrogen adsorption on the Moiré superlattice positions of G supported on an Ir(111) substrate. Then Takahashi et al. [29], while performing high-resolution angle-resolved photoemission spectroscopy of oxygen-adsorbed monolayer G on 6H-SiC(0 0 0 1), found that the energy gap between the π and π-bands increased with oxygen adsorption. This led to a systematic shrinking of π-electron Fermi surface and highlighted the potential of monolayer G oxidization in its bandgap inducement and tunability. On their part, Yavari et al. [30] engineered a tunable G bandgap of ≈ 0.2 eV by reversible H₂O adsorption and showed that the energy gap decreased to ≈ 0.029 eV when humidity was reduced to zero.

According to Şahin & Ciraci [31], bonding of a single Cl atom in G occurs by ionic interaction through charge transfer from G to Cl. This results in a stable direct bandgap semiconducting structure with tunable bandgap controlled by applying uniform strain. Then, by examining structural and electronic properties of F₂ adsorption between the G bilayers, Shayeganfar [32] observed that charge transfer between F₂ and G, and the presence of sp² and sp³ orbitals of C∙C and C∙F bonds disrupted the G layers symmetry inducing an energy gap that depended on F orientation on the G surfaces. The adsorption of F₂ between the G bilayers also led to linear behavior between dipole moment and the energy gap and its electronic properties. This suggested existence of a tunable bandgap controlled by surface functionalization of G. Also, Tayyab et al. [33] predicted that Br-doping of G could induce sufficient bandgap width for optoelectronic applications.

On the other hand, Quhe et al. [34] demonstrated a tunable bandgap opening attributed to G metal adsorption. They illustrated a single-gated field effect transistor (FET) based on Cu-adsorbed ABC-stacked trilayer G. There was a clear transmission gap comparable to a bandgap showing the promise of metal-adsorbed G as a channel in single-gated FET device. Furthermore, while examining the metal dopants' effect on the structural, electronic, and charge transfer mechanism of G using DFT calculations, Tayyab et al. [33] observed that substituting Al atoms into G lattice induced a bandgap of 0.40 eV, and further predicted values of up to 0.82 eV for Al- and Be- doped G.

Other workers have assessed band gap tuning effects of organic agents intercalated into G surfaces. Chang et al. [35], for instance, while investigating the effect of borazine (B₃N₃H₆), triazine (C₃N₃H₃), and benzene (C₆H₆) on G’s electronic structure, found that the molecular adsorptions were accompanied by bandgap opening of up to 62.9 meV under local density approximation. The band opening scale was controlled by the adsorption site type, for C₃N₃H₃ and the heterocyclic molecules were more effective in inducing bandgap opening than the monocyclic ones. Also, most stable configurations led to the largest bandgap opening at the particular adsorption sites, and the charge redistribution patterns controlled the gap on-and-off bandgap switching, which opened whenever the charge redistributed to the bridge site position. This suggested that ionic ability of the dissimilar atoms in the heterocyclic molecules controlled the charge redistribution and the gap tuning efficacy in G. Elsewhere, Hildebrand et al. [36] proposed, a model of self-assembly of halogenated carbene layers on G. They predicted a tunable bandgap opening controlled by adsorbates’ self-assembly and surface coverage, which provided for a mechanism for modulated engineering of G electronic structure and its application to electronic technologies.

Thus, graphene is a gapless semiconductor that cannot be used in optoelectronic applications, such as solar cells in its pristine form. This necessitates for its bandgap tuning, which can be achieved by adsorptive doping with different species atoms including simple molecules, metal atoms or organic molecules etc.

4. Application of graphene in water adsorption

The laminar G structure confers unique physicochemical attributes usable in a wide range of applications to G, which are relating to material hygroscopy [37]. According to Liu et al. [38], the oxygen content of GO materials controls its H2O adsorption capacity, which diminishes with increasing levels of GO reduction. Therefore Lian et al. [37] have reported high water uptake capacity of 0.58 g/g for GO. Also, they found that the adsorption-desorption kinetics of H2O uptake by GO was 5-times higher due to high capillary pressure in the GO laminates and micro-sized tunnel-like wrinkles in the GO surfaces. As already reported elsewhere [38], the reduction of GO diminishes its oxygen content to form reduced graphene oxide, rGO [39]. Water adsorption onto rGO is, however, enhanced at high humidity due to water vapor multilayer formation at the rGO surface [38]. Thus, GO is a loosely defined G material containing variable oxygen content, which is dependent on conditions of its production, and it is generally characterized by high capacity to sorb and store H₂O molecules.

So, Leenaerts et al. [40], while investigating H₂O adsorption on pristine G using first-principles calculations, indicated that there were four possible orientations of H₂O molecule on G surface. Water can adsorb, starting from the O atom, with the H∙O bonds pointing up (u), down (d), or parallel to the G surface (n) or based on another orientation (v) with one O∙H bond parallel to the surface while the other pointing to the surface. The energy of H₂O adsorption depends on its orientation in the G surfaces and on the site of adsorption. The energy difference between the various configurations is 5–6 meV between the orientations, and about 1–2 meV between the positions. It was noted that H₂O acts as an acceptor on G and the acceptor character (C, v) was the most energetically favored. So, Wehling et al. [41] while investigating the G electronic properties after water adsorption examined the effects of SiO₂ substrate and found that perfect suspended G was unresponsive to H₂O adsorption because doping required highly oriented H₂O clusters. However, adsorbed H₂O molecules shifted the SiO₂ substrate’s impurity bands and changed their hybridization with the G bands. So, the H₂O molecules led to doping of G supported on SiO₂ for lower concentrations than those for the free suspended G. This effect was, however, dependent on microscopic substrate properties.

Sanyal et al. [42] found that certain molecules interacted with divacancy in G layers, but, for H₂O molecules, the large DOS in the vicinity of the Fermi level was absent. Accordingly the authors observed a pseudo-gap at the Fermi level, which was comparable to that for pure G. The DOS of the H₂O-adsorbed system was intriguing and the peaks in the C and O atom projected DOS occurred around 4 eV above the Fermi level. Ma et al. [43] determined the H₂O adsorption energy on G using the quantum Monte Carlo and random-phase approximation. Elsewhere, Hamada [44] while using the van der Waals density functional (vdW-DF) to study water interaction with G highlighted the promise of G in water adsorption applications and indicated that the adsorption potency of the materials was controlled by their the intrinsic properties, mode of preparation, and surface modifications employed.

5. Application of graphene in atmospheric detection and amelioration

5.1 Adsorption of nitrogen species

Emmisions of nitrogen compounds, especially the oxides, NO_x, present one of the major environmental concerns today. NOx gases are responsible for various significant environmental problems and they contribute to the poor of human health by resulting in lung dysfunction and respiratory complications. Application of environmentally friendly, mild, and low-cost methods in mitigating presence of NOx in the environment is desirable both for a healthier and more eco-friendly environment [45]. The affinity of G surfaces for N-species has attracted the attention of a significant number of researchers. Sanyal et al. [42] studied G having divacancy defects on its surface, as an eco-friendly, mild, and inexpensive adsorbent for N₂ and found that N₂ molecule chemisorbed onto G in plane with the G surface with the two N atoms substituting in the usual positions of the G hexagonal network as substitutional impurity with chemisorption energies, E^ch, of −4.53 eV. The N₂ inclusion in the G plane resulted in a perfect honeycomb lattice with electrons doped into the G’s conduction.

Elsewhere, Leenaerts et al. [40] proposed two orientations for NH₃ adsorption on G. In one orientation, NH₃ adsorbs with the H atoms pointing away from the G surface (u) and in the other with the H atoms pointing to the surface (d). This indicated that both the adsorption site and the orientation influenced the adsorption energies of NH₃ but charge transfer was determined only by the surface orientation of the NH₃ molecule. The authors found that there was a small charge transfer of 0.03 eV from NH₃ to the G surface in the u orientation but almost no charge transfer in the d orientation. The u orientation, which was consistent with the donor character observed elsewhere [46], was therefore more energetically favored. So, Zhou et al. [47] then compared NH₃ adsorption on pristine G and on transition metal (TM) doped G (TM–graphene) and confirmed that the lowest-energy configurations for NH₃ adsorption on pristine and TM–graphene were through the N atom. However, for TM–graphene, NH₃ bonds via the N atom to the metal atom, by coordinating N lone-pair orbital with the metal valence bands.

It has been stated that adsorbates with a magnetic moment result in a larger doping [48]. So, Leenaerts et al. [40], while examining NO₂ adsorption on G and proposed that adsorbed NO₂ molecule can adopt three orientations on G. Thus, starting from the N atom, the NO₂ molecule can adsorb on G with the N∙O bonds pointing up (u), down (d), or parallel to the G surface (n). The authors found that the total G and NO₂ magnetic moment in the (B, d) orientation was 0.862 μ_B, and the corresponding charge transfer from G (M = 0 μ_B) to NO₂ (M = 1 μ_B) was 0.099 e. The orbital mixing, therefore, led to a charge transfer of ±0.039e from NO₂ to G but, when compared with NO, the charge transfers for the latter were an order of magnitude smaller. Accordingly, Zhou et al. [47] have stated that the NO₂ molecule can bind to G surfaces through the N atom (nitro configuration), through an O atom (nitrite configuration), or through both O atoms (cycloaddition configuration). Thus, for pristine G, NO₂ bonds by the cycloaddition configuration at the bridge site position with adsorption energy (E_ad) of ~0.18 eV and charge transfer of 0.2e to NO₂. However, for TM–graphene, the cycloaddition configuration is preferred with enhanced E_ad ∼ 2 eV, followed by the nitro configuration with E_ad ∼ 1.8 eV. This is by a large charge transfer of ∼0.6e from G to NO₂. However, Tang and Cao [49] found that the adsorption of NOx on GO was stronger than that on G leading to enhanced binding energies and charge transfers from NOx to GO through chemisorption interaction and with transition of the doping properties of NO₂ and NO from acceptor to donor character. Therefore, the interaction of NOx with GO forms H-bonds OHO (N) between -OH and the NOx besides other covalent bonds, including C‒N and C∙O. This is also accompanied by H abstraction to form nitrous acid- and nitric acid-like moieties. The spin-polarized DOS revealed hybridization of frontier orbitals of NO₂ and NO₃ with the electronic states around the Fermi level of GO resulting in strong acceptor doping and remarkable charge transfer characteristics from the molecules to GO.

Elsewhere, Dai et al. [50] evaluated NO and NO₂ adsorption on G modified with B, N, Al, and S dopants using DFT. They found that both B and N atoms retained their planar configuration on G surface but both Al and S atoms protruded out of the G plane. Nevertheless, both NO and NO₂ molecules were bonded to B-doped G but only NO₂ was adsorbed on S-doped G surfaces. The Al-doped G was, however, most reactive toward the gases although B- and S-doped G presented the most plausible prospects for practical NO and NO₂ gas adsorption and detection. So, the most relevant work to environment detection and remediation has been dedicated to NOx molecules mobilization on G, TM-graphene, and GO, but GO. Nonetheless, both TM-graphene and GO displayed enhanced NOx adsorption compared to pristine G adsorbents.

5.2 Adsorption of carbon oxides

Controlling toxic gas emissions is the foremost goal of the twenty-first century environmentalists. The adsorption technique offers a credible approach to dealing with these challenges and many carbon nanostructures, due to their unique surface morphologies and divergent potential for modification, have shown promise in fashioning plausible CO₂ sequestrators [19]. So, Mishra & Ramaprabhu, [51] reported high CO₂ adsorption capacity of 21.6 mmol/g for G at an 11-bar pressure and 25°C, based on a physisorption process. According to Szczęśniak et al. [19] polymers- and/or metal species-modified G nanomaterials possess high specific-surface areas and tailored surface attributes, which confers effective adsorbent properties for CO₂ and other gases.

Several scientists have focused their research on the mechanism of molecular adsorption onto G surfaces. While investigating adsorption of several molecules on G, Leenaerts et al. [40] showed that CO molecule assumed three different orientations on G, two of which involve the CO molecule in perpendicular position to the G surface, with the O atom toward the C atom (u) or away from the carbon (d), and one orientation with the CO molecule parallel to the G surface (n). In this case, CO acts as a donor molecule with charge transfer governed by the orientation assumed by the molecule on the G plane. Nevertheless, minor variations in charge transfer results due to differences in orbital overlap between the HOMO of the CO molecule and G. So, the C and O atoms remain bonded to each other, while CO is bonded to two pairs of G’s carbon atoms. However, the CO molecule does not align perfectly in the plane of G nanosheets so that the C atom of the CO molecule stays in the G plane, making the usual hexagonal planar structure while the O atom stays out of plane. So, Zhou et al. [47] found that the most stable adsorption sites for CO on pristine G involved physisorption interactions at the hollow site, and the CO molecule preferentially assumes a parallel orientation to the G surface at an adsorption distance of ~3.6 Å and adsorption energy of ~0.017 eV. For TM–graphene, however, CO bonds to the C atom bonding the metal atom with enhanced adsorption energy and charge transfer accompanied with shortened adsorption distance. Intriguingly, the magnetic moment of TM-graphene gets enhanced from 1 μB (without adsorption) to ~3 μB.

According to Lee and Kim [52], CO₂ chemisorption energies on graphene-C₄₀ at high pressure are 71.2–72.1 kcal/mol for the lactone systems are controlled by C∙O orientations at the UCAM-B3LYP level of theory. Nonetheless, the physisorption energies of CO₂ on G are only 2.1 and 3.3 kcal/mol at the single-point UMP2/6-31G** level of theory for the perpendicular and parallel orientations, respectively.

5.3 Dioxygen and ozone molecules

Ozone and other reactive oxygen species play an important role in atmospheric photochemistry of pollutants. Detection and removal of reactive oxygen species is, therefore, of essence. Dai et al. [50] analyzed the adsorption of several gases on G and showed that O₂ molecules adsorbed well onto B- and S-doped G samples. On the other hand, Sanyal et al. [42], have suggested that O₂ adsorbs on G by dissociating the two O atoms, which bond to two pairs of C atoms around divacancies in the G structure. The O atoms get oriented in opposite directions out of plane with the G surface. They, however, retain their bond length as in the isolated dimolecular state. In the absence of divacancy, however, O₂ adsorbs on pristine G via the hollow site with low adsorption energy of <0.1 eV and with the O∙O bond oriented perpendicularly to the G surface at an adsorption length 3.7 Å [47]. The length of the O∙O bond remains almost the same as in the free dioxygen molecule (1.24 Å). The Bader charge population analysis showed that the adsorption involved weak physisorption interaction. For metal-embedded G, however, O₂ bonds with the O∙O bond parallel to the G plane with increased adsorption energies (> 1 eV) and decreased adsorption lengths (~2 Å). This is accompanied by charge transfer for O₂ adsorbed on TM–graphene samples and a large expansion in the O∙O bond of 1.44 Å.

Lee et al. [53], who investigated O₃ adsorption on G using the ab initio DFT, found that O₃ molecule adsorbed on G basal plane with binding energies of 0.25 eV via physisorption interaction before it undergoes G-surface epoxidation with release of an O₂ molecule. The activation energy barrier for the physisorption to chemisorption reaction is 0.72 eV, and the binding energy of the chemisorbed state is about 0.33 eV. Ozone adsorption on G is, thus, a gentle reversible reaction relevant to covalent functionalization of the G basal plane.

5.4 Adsorption of sulfur species

Volatile sulphur compounds represent another class of species of relevance to environmental hygiene. Gao et al. [26] while using first principles based on DFT method , discussed in depth the stable configuration, adsorption energy, DOS, and charge transfer of H2S adsorption on intrinsic G, GO, Ni-doped G, and vacancy defect G samples. They observed weak adsorption and charge transfer for intrinsic G, which was enhanced by introducing oxygen-containing functional groups, doping with Ni atom, and setting its single vacancy defect. The single vacancy defect, in particular, promoted G interaction with H₂S molecule imparting excellent adsorption performance for H₂S molecule relevant to studying G-based sensors for the gas.

In another study, Shao et al. [54] investigated SO₂ adsorption on intrinsic G and on heteroatom-doped G samples containing B, N, Al, Si, Cr, Mn, Ag, Au, and Pt atoms using first principles based on DFT. They reported that the structural and electronic properties of the adsorption adducts of the molecules depended on the dopant type. SO₂ adsorbed weakly on intrinsic G, on the G samples doped with non-metals B-, and N. However rapid and strong SO₂ chemisorption occurred on metal and semimetal Al-, Si-, Cr-, Mn-, Ag-, Au-, and Pt-doped G. The analyses of the adsorption mechanisms showed the sensitivity of G-based SO₂ sensors was enhanced by introducing appropriate dopants with both Cr and Mn giving particularly promising results.

5.5 Halogens and halides

In Section 3 of this chapter, we discussed the role of Cl₂ and F₂ molecules in tuning the G bandgap. Şahin &and Ciraci [31] found that a single Cl atom bonding in G occurs by ionic interaction via charge transfer from G to Cl with local distortion in the underlying G layer. Meanwhile, a single Cl adatom migration on perfect G surfaces occurs unhindered but Cl adatom accumulation on the G surface to produce various conformations becomes unsustainable because strong Cl∙Cl interactions occur leading to desorption and formation of Cl₂ molecules. Fully chlorinated G with single Cl atoms bonded alternately to each carbon atom from opposite sides of the G sheets with sp³-type covalent bonds is buckled. So, Shayeganfar [32] discovered that F₂ adsorbs on G via mixed mechanisms involving both in-plane and out-of-plane molecular orientation on the G surface.

Consequently, Chen et al. [55] presented a detailed Br₂ adsorption and charge transfer study of G by combining in situ Raman spectroscopy and DFT techniques. They observed that when G is encapsulated in hexagonal boron nitride (h-BN) layers on either sides, the G surface is protected from Br₂ doping, but when G is supported on only one side by h-BN layer, it undergoes strong hole doping by adsorbing Br₂. The authors obtained molecular adsorption isotherm by plotting surface coverage versus pressure using a combination of Raman spectra and DFT calculations. They saw that the adsorption data fitted the Fowler-Guggenheim model with an adsorption equilibrium constant of ∼0.31 per Torr. The repulsive lateral interaction between adsorbed Br₂ molecules was ∼20 meV, whereas the binding energy for the Br₂ molecule was ∼0.35 eV. At the monolayer coverage, each Br₂ molecule accepts 0.09 e⁻ of charge from single-layer G, but when the adsorbent was supported on SiO₂ instead of h-BN, a threshold pressure was observed for which the diffusion of Br₂ along the SiO₂/G interface resulted in Br₂ adsorption on both sides of the SiO₂-doped G.

On the other hand, Sun et al. [56], while studying HF molecules’ adsorption on intrinsic and on Al-doped G by first-principles calculations, found that HF adsorption mechanisms were different for the adsorbents. The Al-doped G depicted greater adsorption energy and stronger interactions with HF molecule than the intrinsic G. The calculated net electron transfers, electronic density difference images, and DOS provided evidence that HF adsorption on Al-doped G was a chemisorption process, while the molecule’s uptake by intrinsic G was physisorption.

6. Adsorption of alkali metals

Atomic-level investigations of metal adsorption and migration on G surface offer a plausible way of dealing with the challenges of developing novel alkali metal battery technologies [57]. Yang [58] pioneered work in which they reported that Li atoms bonded on a G layer interchangeably on both sides of the nanosheets by distorting the relative positions of the C atoms in the G honeycomb plane. They compared the results with graphane obtained by hydrogenation in which each C is pulled out of the plane by H and found that for Li-doped G, the carbon atom is pushed off by the attached Li instead. This resulted in a conducting counter-intuitive structure. Xue et al. [59], then, studied the adsorption of single Li and the formation of Li clusters on G using first-principles including van der Waals interactions. They found that while Li can exist on the surface of defect-free G under favorable conditions, the bonding was weaker and the surface concentration lower than in the metal adsorption on graphitic surfaces. At low concentrations, Li ions spread out on G surfaces due to coulombic repulsions between the adsorbing metal ions. However, as the Li content on the G surfaces increased, small Li clusters formed so that even though G has higher ultimate Li adsorption capacity than graphite, the nanoclusters nucleate Li dendrites, leading to failure. So, while studying Li adsorption using first-principles DFT calculations and surface diffusion on pristine and defect G structures, Fan et al. [60] observed that Li/C ratios lower than 1/6 for the single-layer G were energetically favorable. Further more, the existence of vacancy defects increased the Li/C ratio of the material. For double-vacancy and higher-order defects, however, Li ions diffused more freely in the direction perpendicular to the G sheets boosting the diffusion energetics in turn.

Indeed, Jin et al. [61] compared alkali metals; adsorption on single-layer G using first principles and observed a common trend in binding distance, charge transfer, and work function relating to increasing metal adsorption proportion ρ (adatom/C atom) on the G surface. There was a dip in the properties at ρ ≈ 0.04 for all metals except for Li, for which the dip occurred at ρ ≈ 0.08. This represented the transition of the adsorbed metals from individual atoms to two-dimensional metallic sheets exerting a depolarizing effect. Thus, G exhibited asymmetric function showing dependence on ρ with a dip on the adatom layer side and saturation on the G side, which was different from the case of bulk graphite.

Therefore, in order to appreciate the influence of G point defects on its Li adsorption, Zhou et al. [62] studied the uptake and diffusion of Li on G with divacancy and Stone−Wales defect using first-principles calculations. They found that in the presence of divacancy, Li adatom adsorbed on the hollow site above the center of an octagonal ring rather than on the top site of carbon atoms next to vacancy site. For the Stone−Wales defects, however, the Li atom adsorbed on the top site of the carbon atom in a pentagonal ring shared by two hexagonal rings. This resulted in buckling of the G sheet. In the case of both the divacancy and Stone−Wales defects, the interactions with Li adatom were attractive and the analysis of both the difference charge density and the Bader charge showed a significant charge transfer from Li adatom to the adjacent carbon atoms. Yang et al. [63], then, investigated sodium adsorption and intercalation into the G bilayers using DFT calculations. They systematically assessed the specific capacity, voltage, and migration energy barriers for Na storage in pristine and mono-vacancy defective bilayer G using DFT. The authors found that with an appropriate voltage (>0.5 V), the mono-vacancy defects improved the specific capacity from 123.97 to 382.54 mAh/g. Thus, Na⁺ ions transported from the defect-free regions (low energy barriers, 0.15–0.32 eV) to defect regions (large energy barriers, 0.56–0.59 eV) and decelerated. It demonstrated that a defect bilayer G was a promising material for making negative electrodes of the Na-ion batteries. Nonetheless, according to Olsson et al. [57], Li and K adsorption on pristine G was more favorable than Na adsorption. According to these workers, N- and O-containing defects dominated on the G surfaces, acting as metal trapping sites, and hindering the metal diffusion and migration on the G surfaces. This diminished the battery cycling performance.

Kim et al. [24] while studying the impact of intercalating G surfaces with alkali metal on its hydrogen uptake, however, found that G surface defects enhanced its metal binding energy and its dispersion and a considerable increase in binding energy was observed for alkaline earth metals. Furthermore, additional alkali metal adsorption of G was studied by Ataca et al. [25]. These studies could therefore give consequence to use of alkali-G material in metal storage devices.

7. The graphene-transition metal doping

Exploring new ways to diversify the basic properties of G continues to attract immense research interest. Transition metal-doping has also gained attention in the recent past, paving way to developing improved G materials with improved magnetism, and other fascinating properties with promise for diverse applications. Chan et al. [64], for instance, explored the adsorption of 12 metal adatoms on G using first-principles and computed the adsorption energy, geometry, DOS, dipole moment, and work function of each adatom-G system. They found that the adsorption groups I-III elements on G aligned with ionic bonding and it was characterized by large charge transfer and minimal change in the G electronic states. However, for the TMs, noble metals, and group IV metal, the adatoms interacted covalently with G surfaces leading to strong hybridization between adatom and G electronic states. Dipole moments across individual adatoms but the work-function shift correlated to the induced interfacial dipole of the G-adatom system and to the ionization potential of the isolated atom. According to Hu et al. [65], chemisorption of 3d and noble metals adatoms on G occurs at the bridge site and at the top site with hybridization between the adatom and the G electronic states, which resulted in distorted G layer. The half-filled 3d shell metal atoms as well as Zn, Ag, and Au atoms depicted low adsorption energies with a general decrease in the magnetic moment compared to the corresponding free adatoms. This was attributed to charge transfer and electron shift between the states of the adatom.

Elsewhere, Sevinçli et al. [66] while assessing the electronic and magnetic properties of 3d TM atoms’ adsorption on G and on G-nanoribbons, reported metal-dependent binding energies of 0.10–1.95 eV. The metals’ adsorption imparted magnetic and metallic properties of G with armchair edge shapes (AGNR’s) of adsorbed TM yielding minimum energy states. The resulting nonmagnetic semiconductor AGNR assumed metal/semiconductor and ferromagnetic/antiferromagnetic spin alignments. So, Cao et al. [67], while studying the geometries, electronic states, and magnetic properties of TM adatom and dimer adsorption on G, found that, except for Cu, chemisorption interactions yielded the most stable adsorption and dimerization on G surfaces controlled by exchange-correlation.

According to Valencia et al. [68], who modeled G and (8,0) single-walled carbon nanotubes (SWCNTs) functionalized with 3d TM atoms, the 4 s occupation with Pauli repulsion was responsible for the Cr, Mn, and Cu physisorption behavior of G. Using a new physical model involving coulomb interaction, 3dⁿ4s^x → 3d^n + x electronic promotion energy and occupation of the 1e₂(δ), 2e₁(π), and 2a₁(σ) metal orbitals, the authors found that Sc, Ti, Fe, and Co metals were present in the G surface as isolated individual atoms but all other 3d TM atoms diffused with clustering. In another work, Amft et al. [69] conducted density functional investigations on Cu, Ag, and Au adsorption on pristine G. While accounting for van der Waals (vdW) interactions using the vdW-DF and PBE + D2 methods, the authors analyzed the favorable adsorption sites, the adsorption-induced distortions in G sheets, and adatom diffusion paths and found from the vdW schemes that the three metal atoms adsorbed on the G sheets with buckling of the G layer. Only the results for Ag qualitatively differed from those obtained from generalized gradient approximation, which gives no binding for it. Otherwise, the results for the rest of the metals showed quantitative variations for the vdW-DF and PBE + D2 models..

Thus, by applying DFT calculations for elements of atomic number of 1–83, Nakada and Ishii [70] focused their work on the adsorbed adatoms' migration on G and showed that, adsorption favored the H6-site for the metals and the B-site for the nonmetals. Nonetheless, the migration energy was particularly high for the 3d TMs and for certain nonmetals. In the same way, Liu et al. [71], while applying first principles, noted that the H6 was the favored site for the rare earth adatoms of Nd, Gd, Eu, and Yb adsorption. The adsorption energies and the diffusion barriers of Nd and Gd were, however, larger than those of Eu and Yb, and all the adatoms induced significant electric dipole and magnetic moments in the adsorption complex. Eu formed flat islands on G attributed to its low diffusion barrier and large adsorption energy compared to the bulk cohesive energy. Nonetheless, the adsorption of Nd and Gd adatoms led to in-plane G lattice distortions.

Vacancies or defects in G nanostructures present sites of altered chemical reactivity and open possibilities for tuning G properties by defect engineering. The understanding of the chemical reactivity of G defects is, thus, critical in implementing carbon materials in several advanced technologies. Pašti et al. [72] investigated atomic adsorption on G surfaces with single vacancy (SV) using DFT analyses of elements. They based their calculations on PBE, long-range dispersion interaction-corrected PBE (PBE + D2 and PBE + D3), and nonlocal vdW-DF2 functionals. They found that most elements, except groups 11 and 12 elements and the noble gases for which the contribution of dispersion interaction was most significant, bonded to the vacancy sites with interaction strengths that correlated to the cohesive energy of the elements in their stable phases. As most atoms could be trapped at the SV site, the calculated dissolution potentials showed that the adsorbed metals became more “noble” than they were in their respective stable phases. Then, Malola et al. [73] reported that Au adatoms adsorbed in-plane at G double-vacancies with diffusion barriers >4 eV and < 2 eV at the larger vacancies in line with the results earlier reported by Gan et al. [74].

8. Graphene adsorption of organic molecule

G materials present attractive pore volume, high conductivity, rich surface chemistry, and exceptionally large aspect ratio, among other promising properties, relevant to adsorptive catalytic applications in organic chemistry [5].

8.1 Pristine graphene

According to Lazar et al. [75] acetone, acetonitrile, dichloromethane, ethanol, ethyl acetate, hexane, and toluene adsorb on G surfaces with adsorption energies that range from 5.9 to 13.5 kcal/mol. The interaction strength of the organic molecules with G surface sites and the corresponding adsorption enthalpies were controlled by London dispersive forces, which dominated (∼60%) the interactions. Basing on DFT calculations, Chi & Zhao [76] investigated formaldehyde (H₂CO) adsorption on intrinsic and Al-doped G. They found that the process was characterized by high binding energies and short connecting distances controlled by chemisorption interactions between the adsorbing molecule and G surfaces. The DOS showed the presence of orbital hybridization between H₂CO and Al-doped G, but there was no evidence of hybridization between H₂CO molecule and intrinsic G surfaces. So, Cortés-Arriagada [77] analyzed 1,4-dioxane adsorption on Al-, Ti-, Mn-, and Fe-doped G and found that the metal-doped G displayed enhanced interaction with 1,4-dioxane molecules compared to intrinsic G. The adsorption energies (1.2–1.6 eV), in this case, were accompanied by changes in the electronic structure of the substrates, especially for the Mn and Fe dopants. The ab initio dynamics simulations showed stable adsorbent–adsorbate interactions of highest aqueous stability of interaction for Al- and Fe- followed by the Mn- and Ti-doped G, respectively.

While applying 50 mg/L initial adsorbate concentration at pH of 6.3 and 285 K reaction temperature, Li et al. [78], on their part, reported G adsorption capacity of 28.26 mg/g for phenol. The phenol adsorption equilibrium simultaneously fitted the Langmuir and the Freundlich isotherms while the adsorption kinetics could be described by pseudo-second order kinetics model. Elsewhere, Xu et al. [79], reported that aqueous bisphenol A (BPA) adsorption on G is also spontaneous and exothermic. According to the authors, the reaction was the Langmuir isotherm and the pseudo-second order kinetics with a high BPA adsorption capacity of 182 mg/g reported at 302.15 K. They suggested that both π-π interactions and H-bonding contributed to BPA immobilization into G. Notwithstanding, Pei et al. [80] investigated 1,2,4-trichlorobenzene (TCB), 2,4,6-trichlorophenol (TCP), 2-naphthol and naphthalene (NAPH) adsorption on both G and GO by applying batch equilibrium technique and micro-Fourier transform infrared spectroscopy. The results showed that the adsorption isotherms for the four aromatic compounds were nonlinear, which indicated that both general hydrophobic interactions and specific interactions played part and the four aromatic compounds depicted similar adsorption efficiencies at pH 5.0. In the case of G, however, greater adsorption efficacy of 2-naphthol displayed under alkaline conditions due to π-π interactions resulting from the higher π-electron density of anionic 2-naphthol than that of neutral 2-naphthol. For GO, however, the affinity for the four compounds was in the order of: NAPH < TCB < TCP < 2-naphthol, and the FTIR spectroscopy revealed dominant π-π interaction for TCB, TCP, and 2-naphthol adsorption on G.

8.2 Graphene oxide

According to Pei et al. [80], high GO adsorption for both TCP and 2-naphthol was attributed to existence of H-bonding between the adsorbates and O-containing groups in GO. However, Chen and Chen [81], while studying m-dinitrobenzene, nitrobenzene, and p-nitrotoluene adsorption on GO, rGO, and G-nanosheets using IR spectroscopy, observed that the hydrophilic GO displayed inferior adsorption capacity for the three nitroaromatic compounds (NACs). This was ascribed to the greater hydrophobic π-conjugation of the active sites in the adsorbents. Even so, greater adsorptions observed for rGO over those of pristine G were linked to the π-π electron donor-acceptor interactions between the NACs phenyls and the π-electron-enriched G lattice, and to charge electrostatics and the polar interactions between G defects and the NACs nitro groups groups. Accordingly, while investigating bisphenol A, nitrobenzene, phenol, benzoic acid, and salicylic acid adsorption on GO, Tang et al. [82], also opined that the aromatic compounds’ π-stacking ability controlled the adsorption processes of the compounds. So, Wang & Chen, [83] compared the adsorption and co-adsorption of naphthalene, 1-naphthol and Cd²⁺ on graphene oxide (GO), chemically-reduced graphene (CRG) and annealing-reduced graphene (ARG). They found that CRG had superior adsorption capacity the adsorbates than either ARG and GO. This, together with observation that the affinity of 1-naphthol for the adsorbents was greater than that of naphthalene was attributed to additional n-π electron-donor-acceptor (EDA) interactions between the 1-naphthol -OH groups and the electron-depleted G sites besides the π-π interactions. Furthermore, 1-naphthol uptake, by both the CRG and the ARG, had a maximum near the respective pKa values, which was consistent with the n-π EDA interaction mechanism. However, GO with greater surface functional groups density than both CRG and ARG, displayed greater affinity for Cd²⁺. This somehow enhanced naphthalene and 1-naphthol co-adsorption on GO and CRG through surface-bridging cation-π interactions. Nonetheless, somewhat suppressed naphthalene co-adsorption with Cd²⁺ on ARG was attributed to the sieving effect of hydrated Cd²⁺ on the micropore edges of ARG.

8.3 Reduced graphene oxide

Meanwhile, Yu et al. [84] synthesized rGO and utilized it in benzene and toluene adsorption under dynamic conditions. The synthesized rGO samples displayed maximum adsorption of 276.4 and 304.4 mg/g for benzene and toluene, respectively. In a comparative study, greater adsorption capacities and breakthrough times were observed for rGO than GO, and the spent rGO was readily regenerated by heating at 150°C. So, Wang et al. [85], while investigating effect of change in concentration of oxygen-containing groups in rGO observed that higher degrees of reduction of GO enhanced the interactions between the π system of G and the π unit of the phenolic molecules. Thus, the adsorption, which was an exothermic and spontaneous process, was promoted by enhancing GO reduction and introducing electron-donating/withdrawing functional groups on the rGO benzene ring.

9. DNA adsorption and bio-sensing tests

Interfacing DNA oligonucleotides with G materials has paved the way to the production of diverse new sensor devices. GO is an outstanding fluorescence quencher and fluorescently-labeled DNA molecules get quenched when they adsorb on GO. However, when a complementary DNA (cDNA) is introduced, it desorbs the probe DNA molecule from GO, and the fluorescence is restored. This permits for devising GO-based DNA optical sensors [86]. This also provides a fascinating topic for bio-interface science. DNA adsorb on GO via π–π stacking and H-bonding while overcoming electrostatic repulsion. The mechanism by which cDNA induces probe DNA desorption from G is, however, still a subject of intriguing discussion [87].

Whereas the analytical aspects of this phenomenon have been demonstrated extensively, the fundamental understanding of the binding mechanisms has lagged behind. Wu et al. [86] therefore studied 12-, 18-, 24-, and 36-mer single-stranded DNA adsorption on GO as a function of DNA length, pH, ionic strength, solvents, temperature, and cDNAs. They found that faster more robust DNA adsorption on G was experienced using shorter DNAs or cDNAs, at low pH, high ionic strength, appropriate solvents, and optimum temperatures. At the same time Lu et al. [39] compared DNA adsorption of GO and rGO using fluorescently labeled DNA and showed that, under similar conditions, DNA adsorbed with a 2.6-fold higher capacity on rGO than on GO. However, the corresponding GO systems showed a higher absolute rise in sensing and signaling kinetics.

10. Application of graphene materials in studies on remediation of water pollution

Environmental pollution is one of the leading global challenges of the twenty-first century and the removal of various pollutants from the environment is of essence [88]. Adsorption methods are effective in the remediation of water and air pollution, but their efficiencies depend on intrinsic porosity, surface characteristics of the adsorbent, and on the applied conditions. Graphene lends itself to tailored surface modification suited for wide range of application in water and atmospheric remediation and pollution prevention.

10.1 Heavy metals removal from water

10.1.1 Pristine graphene

Many studies on environmental water amelioration have focused on individual [89] and competitive [90] adsorption of hazardous metals. Hao et al. [91] obtained SiO₂/G composite for selective removal of Pb(II) ions from water with 113.6 mg/g adsorption efficiency Cortés-Arriagada & Toro-Labbé [92], on the other hand, while investigating adsorption capacity of Al- and Fe-doped G for trivalent and pentavalent methylated arsenic using quantum chemistry computations, observed that trivalent methyl arsenicals adsorption was reached under neutral pH conditions with adsorption energies of 1.5–1.7 eV whereas for the pentavalent methyl arsenicals’ the adsorption was achieved more optimally in acidic to neutral media with adsorption energies of 1.2–2.4 eV and 3.3–4.2 eV, respectively. The interacting σAs [sbnd] O bond weakening in the pollutant structure influenced the capacity and the stability of the adsorption complex.

10.1.2 Graphene oxide

Sitko et al. [93] while studying Cu(II), Zn(II), Cd(II) and Pb(II) adsorption on GO reported high adsorption capacities of 294, 345, 530, and 1119 mg/g for the metals, respectively. The metal affinities for GO surfaces, which was based on chemisorption reactions, was in the order: Pb(II) > Cu(II) ⪢ Cd(II) > Zn(II). Meanwhile, Wang & Chen [83] reported strong affinity of GO for Cd²⁺ ions, while studying Zn(II) adsorption on GO, Wang et al. [94] reported a high Zn(II) adsorption capacity of 246 mg/g, which was achieved in 20 min under adsorption conditions of neutral pH of 7.0, 2 mg adsorbent dosage, and 20°C temperature. The process had a Langmuir equilibrium constant was 5.7 L/g based on pseudo second-order kinetics [14]. Elsewhere, Lingamdinne et al. [95] elucidated Co(II) adsorption properties of GO and obtained a maximum adsorption of 21.28 mg/g while using medium pH of 5.0–8.0 and GO adsorbent dosage of 1.0 g/L. It was found that the adsorption process followed pseudo-second order [14] kinetics based on mixed reaction mechanism involving the π∙π bonds electrons interaction Langmuir, which was described by Langmur, Freundlich, and Temkin isotherms.

10.1.3 Functionalized graphene oxide

While investigating EDTA-functionalized GO (EDTA-GO) for aqueous Pb(II) removal, Madadrang et al. [96], reported high adsorption capacity of 479 ± 46 mg/g achieved at pH 6 in 20 min. They showed that experimental data fitted the Langmuir adsorption model and the EDTA-GO adsorbent was regenerated by washing in HCl. Yet, in another study, White [97] while evaluating aqueous Cu(II) adsorption on GO and on carboxylated graphene oxide (GO–COOH) under conditions of ambient temperature, pH of 6 and 60 min equilibration, Cu²⁺ removal reached 97–99.4%, with 277.77 and 357.14 mg/g maximum adsorption capacities for GO and GO‒COOH, respectively. The process, which was spontaneous and exothermic, was also described by the Langmuir-type adsorption equilibrium. Kumar & Jiang [98], at the time proposed a chitosan-functionalized GO adsorbent for aqueous arsenic adsorption and reported that arsenic oxyanions adsorption onto the novel adsorbent was facilitated via cation-π interactions, based on RNH₃⁺-aromatic π moieties; electrostatic interactions involving H₂AsO₄⁻, HAsO₄²⁻–⁺NH₃R; intra- and intermolecular H–bonding; and on anion-π interaction based on the R-COO^--aromatic π moieties with the oxygen-containing groups on GO surfaces.

10.2 Adsorption of nutrient ions from water

On their part, Vasudevan & Lakshmi [99] evaluated aqueous phosphate adsorption on G at varying pH, ionic strength, and temperatures conditions and determined the G adsorption capacity for phosphate to be 89.37 mg/g using 100 mg/L initial phosphate concentration and 303 K temperatures. The time-dependent phosphate adsorption fitted the second order kinetics [14], while the adsorption equilibrium was a spontaneous, endothermic process governed by Langmuir adsorption mechanism.

10.3 Dyes removal from aqueous solutions

While investigating aqueous methylene blue (MB) adsorption on G, Liu et al. [100] found that the adsorption data followed the Langmuir isotherm with maximum adsorption capacity of 153.85 mg/g at 293 K. The corresponding kinetics fitted the pseudo-second order kinetics model [14], and the process was both spontaneous and endothermic. At the time, Yao et al. [101], testing the potential of magnetic Fe₃O₄@Graphene composite for aqueous MB and Congo red (CR) removal, reported respective adsorption capacities of 45.27 and 33.66 mg/g for the dyes.

Wu et al. [102], then, utilized rGO and also evaluated MB adsorption alongside that of acrylonitrile (AN), p-toluenesulfonic acid (p-TA), and 1-naphthalenesulfonic acid (1-NA) and found that the larger molecules with greater number of benzene rings showed greater propensity to adsorb on GO than the smaller compounds. Accordingly, MB uptake by G was a π-π stacking process, which allowed for up to five adsorbent reuse cycles. The recorded MB, p-TA, and 1-NA adsorption capacities were ∼1.52 ∼ 1.43, and∼1.46 g/g, respectively. At the same time, Sun et al. [103] studied the enhanced GO for aqueous adsorption of acridine orange based on in situ GO reduction by sodium hydrosulfite and observed an improved GO adsorption capacity from 1.4 g/g to 3.3 g/g, which was based on sodium hydrosulfite conversion of carbonyl groups on GO surfaces into hydroxyl groups forming the principal dye adsorption sites in rGO.

Yan et al. [104], then, assessed aqueous MB adsorption on oxidized G samples with varying oxidation oxygen levels and found that GO samples displayed a fast and efficient pH independent adsorption process, which increased with the increasing oxidation of G surface. The MB adsorption behavior was transformed from Freundlich-type at low G oxidation levels to Langmuir-type at high oxidation levels. The binding features of MB-adsorbed GO also changed from parallel stacking of MB molecules on the graphitic plane through hydrophobic π-π interaction to a vertical standing interaction via groups forming MB surface adsorption. The adsorption efficiency of the regenerated GO displayed little loss up in efficacy to four cycles of reuse. Later, Molla et al. [105] when studying MB, rhodamine B, and methyl orange adsorption on GO samples, observed that, unlike negative dyes, which did not adsorb, the positive dyes (methylene blue and rhodamine B) adsorbed on GO surfaces through electrostatic interactions between their positive dipoles and the GO negative dipole of the oxygen surface groups. This process was fast and reached 97% and 88% in just 15 min, for MB and for rhodamine B, respectively. Accordingly, the ab initio molecular dynamics showed that favorable adsorption configuration was at 2298 fs for methyl orange and 2290 fs for MB but MB was more strongly (−2.25 eV/molecule) adsorbed than methyl orange (−1.45 eV/molecule).

At the time, Nguyen-Phan et al. [106] studied the role of GO in photocatalytic removal of MB by titanium dioxide/GO composites reporting enhanced adsorption-photocatalysis for TiO₂/GO composites compared to those of pure TiO₂. Both the removal efficacy and the corresponding MB photodegradation increased with the GO proportion in the composite up to 10 wt%. The enhancement of these properties was attributed to the components’ synergy in increasing the composite specific surface area, π-π conjugation between the dye and the aromatic rings, and in the ionic interactions between MB and oxygen-containing functional groups in GO.

10.4 Pesticides and other organic pollutants

Cationic surfactants present a considerable water pollution problems. They are difficult to degrade, leading to persistence in water sources. Chen et al. [107] evaluated aqueous adsorption of a cationic dodecyl amine hydrochloride (DACl) surfactant on GO by analyzing its zeta potential and applying FTIR and X-ray photoelectron spectroscopy (XPS). They found that the adsorption equilibrium was consistent with the Freundlich isotherm while the adsorption kinetics followed pseudo-second order model [14]. Accordingly, they reported an adsorption process that was endothermic and consistent with electrostatic interaction and H-bonding between DACl and GO.

Then, Wuest & Rochefort [108] while studying amino triazines adsorption on G using DFT calculations reported strong adsorbate affinity for G, which they attributed to specific attractive interactions of -NR₂ groups within the underlying surfaces as the driving force for the process. Afterwards, the authors [109] reported unprecedented aqueous GO and rGO adsorption of capacities of ∼1200, 1100, and 800 mg/g for chlorpyrifos (CP), endosulfan (ES), and malathion (ML) respectively. They noted that the process was mediated through water because direct interactions between G and the pesticides were weak and unlikely. However, the adsorption of the compounds was both pH- and background ions-independent.

10.5 Pharmaceuticals and other emerging pollutants

Pharmaceutical agents have been become common persistent pollutants in in many water bodies around the world [7]. The occurrence of such emerging contaminants in water bodies poses unique threats to living things. Conventional techniques of pharmaceuticals removal from water are complex, expensive and generate secondary hazardous residues. Due to inefficient wastewater management systems, new treatment technologies are needed to deal with new products for the safety of the environment [110]. The adsorption technique has gained popularity but the engineering of new adsorbents for emerging trace pollutants is desired. G materials have lent themselves to application in various fields due to their robust physicochemical attributes and several researchers are already exploring them in environmental remediation. Gao et al., for instance, investigated removal of three tetracycline antibiotics, including tetracycline, oxytetracycline, and doxycycline [111] from water using GO. They reported that the tetracycline antibiotic deposited on GO surfaces via the π-π and cation-π interactions. The adsorption data could be fitted the Langmuir and Temkin isotherms with Langmuir adsorption capacity of 313, 212, and 398 mg/g for three antibiotics, respectively. The kinetics adsorption data fitted the pseudo-second order kinetics model [14] with a sorption rate constant (k) of 0.065 g/mg/h. Nonetheless, the adsorption capacities of GO for tetracycline decreased with increasing solution pH and increasing Na⁺ concentration.

In another study, Nam et al. [112], while studying the adsorption behaviors of diclofenac (DCF) and sulfamethoxazole (SMX), evaluated the effect of GO dosage, contact time, pH, and sonication. They predicted binding energies between the drugs and GO surface groups and found that GO adsorption of the drugs was controlled by the oxygen-containing functional groups in GO, which exhibit negative surface charge over a wide range of pH values of 3–11. DCF showed favorable binding energy of −18.8 kcal/mol compared to that of SMX, which was −15.9 kcal/mol. The removal efficiencies for the two agents reached 35 and 12% within the initial 6 h respectively. however they were strongly enhanced to 75 and 30% under sonication, respectively. This is because sonication facilitates dispersion of exfoliated GO particles diminishing their surface density of oxygen-containing functional groups. It was found that the equilibrium data for both adsorbates fitted the Freundlich model.

Zhu et al. [110], on the other hand, while testing GO for metformin removal from water, observed an initial rapid and efficient metformin uptake, which was also strongly temperature-, pH-, ionic strength-, and the background electrolyte-controlled. The GO optimum adsorption of metformin was achieved at pH 6.0 and 288 K. The authors then found that the adsorption process was both spontaneous and exothermic and suggested that both π–π interactions and H-bonds played the leading role. The adsorption kinetics showed that 80% metformin removal was achieved within 20 min with high-rate constants of k₁ (0.232 per min) and k₂ (0.007 g/mg/min) for Lagergren first order [13] and pseudo-second order kinetic [14], respectively.

In the same way, Pavagadhi et al. [113] employed GO the removal of microcystin-LR (MC-LR) and microcystin-RR (MC-RR) –two common algal toxins in natural water. The adsorption experiments employed typical aqueous matrix containing environmental water anions and cations. The authors reported 1700 and 1878 μg/g GO adsorption capacities for MC-LR and MC-RR, respectively. The adsorption kinetics for the process were achieved within 5 min and the GO samples could be recycled up to ten cycles without significant loss in their adsorption potency.

11. Conclusions

G exhibits hydrophobic surfaces presenting high adsorption to chemicals that exhibit strong π-π interactions. GO, on the other hand, has a number of reactive functional groups with strong acidity suited for high adsorption for basic and cationic adsorbates. We find that the key application for G materials in the last decades have included hydrogen and other fuel gas adsorption and storage; alkali metal storage and battery technology; design and development of nano-electronic and nanophotonic devices; vacancy engineering for modifying the G surface reactivity for various application; G environmental application in water and atmospheric amelioration of both organic and inorganic pollutants; and in fashioning sensors and biosensors for most diverse applications. Suitable modifications of GO or G with nonmetals, metals, metal oxides, or organics can produce nanocomposites with enhanced adsorption capacities and separation efficiencies toward various groups of adsorbate species. The metals’ adsorption, for example, imparts magnetic and metallic character to G. A number of molecular intercalations have also been found useful in inducing sufficient bandgap opening into G, which is critical for controlling electronic conductivity and opening up the materials to application in nano-electronic and nanophotonic devices among other applications.