Graphene Quantum Dots and Their Applications in Bioimaging, Biosensing, and Therapy

Abstract

Graphene quantum dots (GQDs) are carbon-based, nanoscale particles that exhibit excellent chemical, physical, and biological properties that allow them to excel in a wide range of applications in nanomedicine. The unique electronic structure of GQDs confers functional attributes onto these nanomaterials such as a strong and tunable photoluminescence for use in fluorescence bioimaging and biosensing, high loading capacity of aromatic compounds for small molecule drug delivery, and the ability to absorb incident radiation for use in the cancer-killing techniques of photothermal and photodynamic therapy. Here, recent advances in the development of GQDs as novel, multifunctional biomaterials are presented with a focus on their physicochemical, electronic, magnetic and biological properties, along with a discussion of technical progress in the synthesis of GQDs. Progress toward the application of GQDs in bioimaging, biosensing, and therapy is reviewed along with a discussion of the current limitations and future directions of this exciting material.

Graphical Abstract

Recent advancements in the development of graphene quantum dots (GQDs) in the field of nanomedicine are reviewed. Following an overview of the properties of GQDs, progress in top-down and bottom-up synthesis methods are presented. Application of GQDs in various modes of bioimaging, biosensing, and therapy are also discussed, with a subsequent discussion of future directions.

1. Introduction

Functional materials based on the crystalline allotropes of carbon have garnered significant interest from researchers in physics, chemistry, and materials science ever since the discovery of fullerenes in 1985. One carbon allotrope in particular, graphene, consists of just a single two-dimensional sheet of carbon atoms bonded to form a hexagonal lattice. This highly-ordered and closely packed monolayer of carbon atoms confers unique properties onto this material. One of these properties is the zero-energy bandgap which arises due to the idiosyncratic overlapping of graphene’s valence and conductance bands. However, zero-energy bandgaps only describe certain configurations of graphene. The zero-energy bandgap is attributed to defect-free graphene of infinite dimension; when the physical dimensions of graphene are restricted to finite values and defects are introduced within the lattice, the electronic structure is altered so that non-zero values of energy bandgaps are observed.^[1] Pertinently, graphene sheets of nanoscale dimensions are known to exhibit quantum confinement effects, where the bandgap of graphene may be tunably increased by decreasing the size of the graphene fragment. These graphene particles have been dubbed graphene quantum dots (GQDs), and this tunable electronic structure has been exploited for applications in electronics, energy harvesting, and medicine.^[2]

One of the most utilized properties of GQDs is their photoluminescence (PL) which describes the emission of light by graphene caused by the excitation of electrons within a GQD upon the absorption of incident photons and the subsequent emission of photons when those excited electrons relax back to a lower engery level. Through various studies, it has been shown that the PL excitation and emission wavelengths can be altered by controlling the size of GQDs, changing the surface properties, or through the introduction of dopants, such as nitrogen or boron, into the carbon lattice.^[3–5] The tunable PL of GQDs makes them an effective platform for application in bioimaging and biosensing systems since the optimal wavelengths of light dictated by the situation at hand may be harnessed by synthesizing GQDs of apporpriate size and/or dopant concentration to elecit the desired PL properties. In addition to dopants within the lattice, oxygen-rich functional groups such as hydroxyl and carboxyl groups exist on the edge of GQDs. These functional groups are a result of oxidation and self-passivation at the surface of GQDs during synthesis, and these polar constituents confer excellent aqueous solubility onto the GQDs, which is beneficial for biological applications.^[6] Furthermore, functional groups allow for the facile surface conjugation of useful moieties such as tumor targeting ligands or chemotherapeutics for use in anticancer drug delivery systems.^[7–10] In many other nanoparticle-based drug delivery systems, a solid core is stabilized by coating the core with polymers, which restricts the number of sites available for chemotherapeutic conjugation on the surface of the nanoparticle. Conversely, GQDs have a large surface area to volume ratio, owing to their planar structure; this allows for higher drug loading and more efficient delivery of chemotherapeutics. Furthermore, π-orbitals that are present throughout the sp²-hybridized GQD lattice can bind chemotherapeutics that have an aromatic ring in their structure through π-stacking, without covalent conjugation, further enhancing their ability for drug delivery.^[8,11,12]

The potential for GQDs to make a substantial impact in medicine has been demonstrated by a myriad of recent developments. Many studies have been aimed at developing novel GQD synthesis methods via simple ‘one-pot’ solutions and the utilization of more widely available precursors such as glucose and citric acid, as opposed to graphene and graphite.^[13–16] Advances have also been made in tuning the PL properties of GQDs. Emission in the near-infrared (NIR) region of the electromagnetic spectrum, as well as two-photon PL in GQDs reported in recent studies, have opened new windows for in vivo imaging given the transparency of biological tissue to NIR light.^[17,18] In addition, studies have shown that the PL of GQDs is altered in the presence of a chemical species, such as boron and nitrogen, allowing GQDs to be used in biosensing applications.^[3,19] While these studies have brought GQDs to the frontier of nanoparticle research with vast possibilities, there is still significant ground to cover ranging from consistent synthesis and purification methods, to a more thorough investigation of biological toxicity in vivo. Furthermore, a lack of a clear, universal definition regarding the structure and properties of GQDs hinders a more complete understanding of the system and its application in biological research.

In this review, an overview of the physicochemical, optical, electronic, magnetic, and biological properties of GQDs is first presented, followed by a discussion of recent progress in GQD research with a focus on advances in methods of GQD synthesis; these methods are divided into categories based on techniques relying on either top-down or bottom-up processes. In subsequent sections, the applications of GQDs are discussed including bioimaging via optical and magnetic modalities, in vitro and in vivo biosensing, and therapeutic approaches via drug delivery, gene delivery, and photodynamic therapy (PDT). The contents of this review are outlined in Figure 1. Table 1 summarizes the various properties and applications of GQDs gathered from the literature and discussed herein. Insight into the current limitations and future directions of GQD research are also provided.

Figure 1.

Advances in research of graphene quantum dots (GQD) have been made in their ease of synthesis, physical properties, and application in biological systems. Reproduced with permission.^{[8,34,41,51,113]} Copyright 2017, Multidisciplinary Digital Publishing Institute. Copyright 2016, John Wiley and Sons. Copyright 2018, John Wiley and Sons. Copyright 2019, Elsevier. Copyright 2017, Elsevier

Table 1.

Summary of recent advances in GQD synthesis, properties, and applications.

Starting material	Synthesis method	PL emission (nm)	Size (nm)	Application	References
Graphite rods	Electrochemical exfoliation	340-400	2.6±0.4	Facile synthesis of GQDs	[71]
Graphite rods	Electrochemical exfoliation	440-480	4	Tumor detection using pH-responsive change in PL	[131]
Graphite flakes	Microwave assisted liquid phase exfoliation	400-550	6	Synthesis without oxidation or reducing agents	[75]
Graphite	Graphite intercalated method	400-550	5	In vivo cell tracking	[122]
Fluorographene	Liquid exfoliation	N/A	10±2	MRI contrast agent	[52]
Carbon fiber	Chemical exfoliation	700-800	4.5±1.5	In vivo near-infrared imaging	[57]
Trisodium citrate	Pyrolysis	400-500	1.3±0.5	Fluorescence cell imaging	[16]
Citric acid, urea	Hydrothermal synthesis	410-470	12±2	Drug delivery for anticancer therapy	[11]
Vinylphenylboronic acid, boric acid	Solvothermal synthesis	370-530 420-530a)	5.8	MRI contrast agent	[51]
Nitropyrene	Hydrothermal synthesis	440-550	1-4	Cellular imaging via two-photon fluorescence microscopy	[113]
Pyrene	Hydrothermal synthesis	500-600	2.93±0.18	In vitro H2S sensor	[19]
Glucosamine, thiourea	Microwave assisted hydrothermal synthesis	448-539 800-890b)	3.9±2.0	One-step synthesis with microwave irradiation	[98]
Glucose	Hydrothermal synthesis	500-550	4.34	Detection of Cu2+ ions	[14]
Aminophenylboronic acid	Solvothermal synthesis	440-520 970-1020b)	4.7	In vivo imaging in the NIR-II window	[41]

^a)Two-photon excitation;

^b)NIR excitation

2. Graphene Quantum Dots and their Unique Properties

2.1. Physical Properties

The standard definition of a quantum dot (QD) is a “nanoparticle or region that exhibits quantum confinement in all three spatial directions”.^[20] Hence, a GQD can be described as a nanoparticle composed of graphene that exhibits the properties of a QD. Amongst the materials used to synthesize GQDs, graphene and graphene oxide (GO) both consist of a single hexagonal array of sp² hybridized carbon atoms, with the latter containing point defects and oxygen-containing functional groups throughout its lattice and on the edges.^[21] Amorphous carbon is another allotrope that consists of sp³ hybridized carbon in a non-crystalline arrangement.^[22] The structural differences between graphene, GO, GQDs, and amorphous carbon have important consequences for the observed properties of these carbon-based molecules (Figure 2). GQDs are reported as sub-100 nm nanoparticles consisting of either a single-layer or a few layers of graphene. The crystallinity of graphene is what distinguishes GQDs from the often-mislabeled carbon quantum dots (CQDs). The quantum confinement exhibited by GQDs results from the extensive sp² hybridized carbons in crystalline graphene, with doping, defects, and spatial dimensions altering their observed properties. On the other hand, CQDs are usually spherical nanoparticles consisting of amorphous carbon, exhibiting the PL property due to surface passivation.^[23] The PL behavior of CQDs has been shown to change after surface modifications.^[24,25] While graphitic domains exist throughout their structure, CQDs are reported to be much less crystalline than GQDs. Additionally, the structure of GQDs can also be altered by the presence of dopants within the crystal structure and chemical functional groups on the edges. The composition of GQDs is most commonly evaluated by X-ray photoelectron spectroscopy (XPS), and the amount of dopants present is analyzed through the XPS spectrum of the corresponding element of interest. By altering the chemical composition of GQDs, the desired properties can be achieved, making GQDs a versatile platform for biological applications. The methods for incorporating elements such as nitrogen and boron, as well as the effects of such doping will be further discussed in subsequent sections.

Figure 2.

Schematic illustration of structural differences between graphene, graphene oxide, graphene quantum dots, and carbon quantum dots.

2.2. Electronic Properties

The electronic characteristics of graphene are dictated by the arrangement of its electrons. In graphene, each carbon atom forms a trigonal planar orbital geometry with three sp² hybridized bonding orbitals, and each of these sp² orbitals forms a σ bond with another sp² orbital of a neighboring carbon atom. Thus, of the four valence electrons associated with each carbon atom composing the graphene lattice, three are localized in very stable single covalent bonds, whereas the fourth resides alone and unpaired either above or below the planar carbon lattice in a half-filled p orbital.^[26] This electronic structure is unique in that the resulting bandgap between the valence band formed by the electrons participating in bonding and the lone electrons of the p orbitals is zero. Thus, graphene is known as a zero-energy bandgap substance and actually conducts electricity like a metal due to the high mobility of the electrons in the conduction band.

The preceding scenario assumes a perfect two-dimensional carbon lattice free of defects. When discussing realistic forms of graphene, it turns out that the electronic properties of graphene, such as electron mobility and the bandgap, can be tuned by altering parameters such as the graphene’s lateral dimensions, concentration of defects, chemical functionalization (especially the presence of oxygen-containing groups), and physical conformation. Such alterations to the observed electronic properties of graphene and GQDs are largely owed to many-body effects, namely electron-electron and electron-hole interactions.

These behaviors have been investigated in one-dimensional materials such as carbon nanotubes (CNTs), nanoribbons, and boron nitride both experimentally and theoretically.^[27,28]

The electronic structure and energy bandgap of GQDs were commonly investigated using computational models in armchair (AM) and zigzag (ZZ) edge group conformations. The edge groups were saturated with hydrogens to observe the effects of lateral dimensions as well as edge group conformations. The results of this computational modeling show that as the lateral size of GQDs increases, the energy bandgap and the exciton binding energy decrease, and their values tend to be lower for ZZ conformation models (Figure 3).^[29]

Figure 3.

The effect of size and edge conformation on the energy bandgap of GQDs calculated through theoretical modeling. (a) Structures of GQDs of various sizes and edge conformation. The labels “AM” and “ZZ” refer to the edge type, while the following integer represents the number of aromatic rings. (b) Variation of energy bandgap with lateral size and edge type. The following letters indicate the modeling method used: Green’s function and screen Coulomb interaction (GW) and local density approximation (LDA). Reproduced with permission.^[29] Copyright 2015, American Chemical Society.

Electronic properties can also be altered by edge passivation with various functional groups. The effect of chemical functionalization on the energy bandgap was studied through theoretical modeling and calculations. The energy bandgap was shown to be more dependent on GQD conformation (3.269 eV in AM, 0.273 eV in ZZ), with small changes attributable to functionalization. However, when the hydrogen atoms were replaced with more electronegative atoms such as oxygen or fluorine, the bandgap was significantly decreased.^[30] While these theoretical studies provide help toward understanding the electronic structure of a single GQD with very specific dimensions and chemistry, many GQDs reported in the literature have a range of sizes and some variability in functional groups. Studies correlating these theoretical models to experimental observations would elevate the importance of computational modeling and perhaps facilitate modeling polydisperse, non-homogenous GQD systems. The significance of the bandgap of GQDs is demonstrated through the PL of GQDs. While the theoretical bandgaps are calculated using GQD models with specific structures, experimental data demonstrate that GQDs often exhibit excitation wavelength-dependent PL emission spectra.

2.3. Photoluminescence

PL describes the phenomenon of a molecule absorbing the energy of an incident photon through the excitation of an electron to a higher energy level and the subsequent emission of a photon when the excited electron relaxes. As such, the bandgap of GQDs determines the operative wavelengths that may participate in the PL spectra of GQDs. In fact, a perfect zero-energy bandgap form of graphene would not exhibit PL. Only by varying parameters such as adding surface groups, altering the concentration of dopants, or engineering the physical dimensions of GQDs can they be characterized with a non-zero bandgap and thereby exhibit PL. These parameters dictate the two main mechanisms by which PL in GQDs is controlled: (1) quantum confinement controlled by the size of the GQDs and (2) influencing the bandgap of a GQD due to the difference in energy levels of electrons associated with the sp² carbons of graphene and the electrons associated with surface groups, dopants, and edge states.^[31] By altering these parameters the emission windows of GQDs have been reported to range from deep ultraviolet (UV),^[6,32] blue,^[13,33–35] green,^[14,36] orange,^[37,38] red,^[39,40] and the NIR.^[17] PL is a property widely utilized in biological research, as it allows certain molecules to ‘glow’ under irradiation of a specific wavelength, providing a method to visually isolate a region of interest. The multitude of PL emission windows of GQDs allows them to be used in conjunction with other molecular dyes with a non-overlapping PL range that avoids any autofluorescence intrinsic to the cells and tissues to be imaged.^[41,42] These properties highlight the utility of GQDs in applications such as biosensing and bioimaging. The synthesis method, chemical composition, and surface passivation have all been found to contribute to variations in the PL properties of GQDs. Several mechanisms have been proposed to describe the interaction of GQDs with photons to explain the PL properties of GQDs.

Owing to their unique electronic structure, GQDs exhibit so-called tunable optical properties that may be altered by adjusting the physical and chemical characteristics of the GQDs under investigation. Studies have shown a trend toward redshift of the emission spectra as GQDs increase in diameter and are surface-passivated with functional groups.^[5,6] This redshift resulting from surface functionalization is due to the altered location of electrons and holes in GQDs. It was found by calculation of charge-transfer densities in several GQD models that the localization of holes near edge groups not only leads to the redshift but also an enhanced fluorescence due to the recombination of well-separated electron-hole pairs.^[43] The effect of doping in GQDs on the PL properties depends not only on the dopant atom, but also the doping method used, as well as the resulting structure of the GQDs. Doping the lattice of GQDs with nitrogen through irradiation in an NH₃ atmosphere has been shown to induce a blueshift by changing the charge-transfer densities within the lattice and by increasing the energy released upon recombination of the excited electron-hole pair (Figure 4).^[34,44] Increasing levels of nitrogen doping resulted in a blueshift due to the high electron affinity of nitrogen (Figure 4b). By analysis of the chemical structure of the nitrogen-doped GQDs, the nitrogen present in the structure was found to be largely amine-like, indicating that the nitrogen was present on either the edge or within the lattice with a vacancy defect as the nitrogen was bound to only one other carbon atom. Alternatively, when nitrogen is incorporated into the interior of the lattice as a part of the sp² hybridized network (aromatic nitrogen), the PL spectra is red-shifted and afforded an increase in intensity (Figure 4c). On the other hand, boron doping has been shown to increase the emission wavelengths of GQDs through the formation of trap states in the π-π* gap, lowering the energy of the emitted photons (Figure 4d,e).^[4] While many of the previously mentioned studies show emission spectra to be dependent on the excitation wavelength, excitation-independent PL has also been observed, particularly from GQDs synthesized from pristine graphite.^[37] In GQDs with large size distribution and various chemical structures, excitation-dependent PL is observed due to the response of GQDs with variabilities within the sample. That is, an individual GQD that has a different size or composition than another GQD in the same solution, albeit minor, will exhibit different PL spectra due to the differences in the energy bandgap associated with PL. Conversely, a greater degree of homogeneity in size and composition would result in a more excitation-independent PL emission. The lack of defects and the high uniformity of GQDs synthesized from materials of a higher degree of crystallinity allow them to exhibit excitation-independent PL. In addition to the wavelength of the PL, quantum yield (QY) is another key parameter that serves as a measure of the utility of GQDs as imaging probes. QY is the ratio of the number of photons emitted by a PL molecule to the number of photons absorbed by the molecule at the excitation wavelength and indicates the efficiency of the molecule as an imaging probe. The QY of a GQD is calculated by comparing its absorbance and emission to that of a molecular dye with a known QY, whose PL emission window is close to that of the GQD. Reported QY values have ranged from 2.9 to 83%.^[10,45]

Figure 4.

PL spectra of GQDs shift with the varying nature and degree of doping. (a) Photograph of GQDs and nitrogen-doped GQDs (NGQDs). The numbers following the NGQD labels indicate the irradiation time in min under NH₃ atmosphere. (b) PL spectra of NGQDs synthesized under NH₃ atmosphere showing a blue-shift as nitrogen atoms are incorporated into the interior of the GQDs. (c) PL spectra of aromatic amine functionalized GQDs. N_ar/N indicates the ratio of aromatic amines to total amine groups. A red-shift in the spectra with increasing aromatic amine-functionalization is apparent. UV-Vis (black) and PL spectra (colored) of (d) pristine GQDs and (e) boron-doped GQDs at various excitation wavelengths. A slight red-shift is seen as boron is incorporated into the structure. (a),(b) Reproduced with permission.^[34] Copyright 2017, Multidisciplinary Digital Publishing Institute. (c) Reproduced with permission.^[44] Copyright 2015, John Wiley and Sons. (d),(e) Reproduced with permission.^[4] Copyright 2014, Elsevier.

2.4. Magnetic Properties

Whereas many materials exhibit magnetic properties due to the presence of electrons in d and f shells, magnetic ordering in GQDs can be attributed to electronic states that are spin-polarized in ZZ conformations of GQDs; AM conformations exhibit a magnetic moment of zero.^[46] Alternatively, vacancies and dopants may also introduce magnetic moments at point defects within the lattice of a GQD. Early research on magnetism in GQDs was performed theoretically and led to different predictions. Several studies have reported that the geometry of GQDs plays an important role in their magnetic properties; defects at the edge were also reported to switch off the spin polarization.^[47,48] However, reports of robustness of the spin polarized edge states with respect to irregular geometries and edge roughness have also been presented.^[49] While discrepancies between theoretical models exist, magnetic properties of GQDs have also been explored experimentally. Paramagnetism has been observed in low-defect GQDs at very low temperatures; however, they show little intrinsic magnetism, due to the suppression of edge-state magnetism caused by the presence of defects or reconstruction of the edge structure. The observed magnetism is attributed to the passivation of the ZZ edge by hydroxyl groups.^[50] This notion of defect-induced magnetism is supported by other experimental studies where GQDs doped with fluorine or boron have shown significant magnetism.^[51,52] The paramagnetism exhibited by these doped GQDs is of interest in areas such as magnetic resonance (MR) imaging given that paramagnetic substances are often used to enhance image contrast in clinical MR imaging.

2.5. Biological Properties

While much of the current research is focused on improving optical properties of GQDs, biocompatibility remains an important aspect for in vivo applications of these materials. The scope of the biocompatibility of a material is subject to a specific application; however, the overarching definition of biocompatibility is the ability of the material to perform its intended function without inducing unintended biological reactions.^[53] Although toxicity due to aggregation has been reported in many carbon-based nanoparticles, GQDs display excellent solubility in water and provide a non-toxic alternative to metal-based nanoparticles by eliminating the toxicity concern due to metallic residues that are not cleared from the body since GQDs are entirely composed of carbon. One method of assessing biocompatibility is to evaluate cytotoxicity which describes an agent’s toxicity to cells in culture. Many studies have shown GQDs to be non-cytotoxic, with discrepancies in cytotoxicity resulting from surface modifications. GQDs modified with −NH₂, −COOH, and −CHO-N-(CH₃)₂ were shown to have low toxicity at concentrations up to 200 μg mL⁻¹, whereas −OH modified GQDs have shown some cytotoxicity at concentrations above 100 μg mL⁻¹ (Figure 5a).^[54,55] The mechanism of cytotoxicity induced by hydroxylated GQDs had previously been attributed to intracellular reactive oxygen species (ROS) generation and cellular senescence; however, a more thorough inspection revealed that the increase in ROS occurred only after cell viability had already decreased significantly, and the bioavailability of differentially functionalized GQDs, as measured by cellular uptake, did not correlate with their respective cytotoxicity.^[55,56] Cellular interactions with nanoparticles are complex, and specific studies on cytotoxicity would be more useful than an enveloping generalization in assessing the biocompatibility of GQDs.

Figure 5.

Biocompatibility of GQDs assessed through cell metabolic assay, histology, and cell viability assay. (a) Evaluation of cell metabolic activity of A549 cells after exposure to three kinds of GQDs at different concentrations. (b) Histological evaluation of the vital organs of rats 4 weeks after intravenous injection of GQDs. (c) Cell viability at various concentrations of GQDs and GQDs modified by reaction with the following: phenylhydrazine (PH), benzoic anhydride (BA), and 2-bromo-1-phenylethanone (BrPE), in the dark, or (d) under irradiation. (a) Reproduced with permission.^[54] Copyright 2014, Springer. (b) Reproduced with permission.^[58] Copyright 2018, Future Science Group. (c),(d) Reproduced with permission.^[64] Copyright 2017, Royal Society of Chemistry.

Investigation of in vivo toxicity of GQDs is still in its early stages; however, recent studies have extensively shown the effect of specific GQDs in mouse models. Biodistribution and cytotoxicity of carboxylated GQDs were investigated after intravenous injection in mouse models; in vivo imaging of the mice and ex vivo analysis of the organs showed that the GQDs had begun to be cleared from the body after 12 h. The results of complete blood count and histological analysis of GQD-treated and saline-treated mice also showed no significant differences between the experimental and control groups;^[57] similar results were observed in another study.^[58] In both studies, minor changes were found in the histological evaluation of the liver and kidney at high dosages of GQDs after 30 d (Figure 5b). The lack of acute toxicity of GQDs shown in these in vivo studies is a promising indicator of the biocompatibility of GQDs. In addition to toxicity studies, the immune response to GQDs, as well as degradation mechanisms, have been studied to further characterize the biocompatibility of GQDs.^[59–61] However, as demonstrated in in vitro studies, alterations of the chemical functionalities of GQDs can yield significant differences in their interaction with living organisms; hence the validity of these in vivo studies as comprehensive indicators of biocompatibility of GQDs is still dubious.

One of the reported prominent effects of GQDs in cellular environments is their generation of photoinduced ROS and resulting cytotoxicity.^[62,63] While this ability of GQDs to potentiate ROS concentration upon exposure to incident radiation can be used advantageously in areas such as PDT, it is important to address the extent of ROS generation by GQDs upon photo-irradiation for practical applications.^[58] The effect of functional groups, specifically the oxygen containing ketone, hydroxyl, and carboxylic groups, on the ability of GQDs to generate ROS, was investigated. The results indicate that all the oxygen containing functional groups increase ROS generation, with the ketonic group inducing the highest ROS levels. The increased levels of ROS generation were closely related with decreased cell viability under continued irradiation, while cell viability after storage in dark conditions remained unchanged (Figure 5c,d).^[64] Other studies have also confirmed that without irradiation, no cytotoxicity was observed in cells treated with GQDs.^[65] According to these studies, significant cytotoxicity is caused by the ROS generated via irradiation of GQDs. Although much of the application of GQDs in biological research utilizes the PL properties of GQDs, and hence requires irradiation from a light source, no cytotoxicity has been reported in these studies. As recent studies have focused on GQDs utilizing longer wavelengths for their PL properties, further studies on the exact mechanism of ROS production by GQDs, as well as the effects of structure and functional groups are required to better understand the observed phenomenon and its effect on practical applications of GQDs.

3. Synthesis Methods

The synthetic routes of making GQDs with high levels of monodispersity, uniformity, and crystallinity can be divided into two major categories: top-down and bottom-up. Properties of GQDs have been shown to vary depending on their synthesis method as well as the precursor material used. Many of the synthesis methods for GQDs are relatively simple and inexpensive, with the reactants ranging from carbon black, collected from the combustion of paraffin, to sugar molecules such as glucose.^[66,67] The processes often include elevated temperature and reported methods have utilized conventional microwave ovens and a ‘one-pot’ syntheses through use of autoclave reactors.^[33,68] Nevertheless, more sophisticated synthesis methods exist for applications that require specific GQD properties such as carbon crystal orientation.^[69,70] The ease of synthesis and availability of reactants have promoted recent advancements in carbon nanomaterials research. The various methods for GQD synthesis are outlined in Figure 6.

Figure 6.

Schematic illustration of recent advances in GQD synthesis via top-down and bottom-up methods. The specific reaction conditions and precursors are representative of the studies covered in this report.

3.1. Top-down Methods

Top-down approaches start with bulk precursors, such as large graphene sheets or CNTs, which are then cleaved into appropriately sized GQDs via chemical or physical means. The precursor materials are separated via intercalation or oxidation, which also introduces point-defects within the carbon lattice. Through chemical reduction or ultrasonication, the defects can serve as sites where the carbon bonds are cleaved, and GQDs are separated from the precursors. By using various synthetic methods and tuning reaction conditions, the properties of the synthesized GQDs can be manipulated. Hydro- and solvothermal synthesis,^[68,71,72] electrochemical exfoliation^[73], and liquid phase exfoliation^[74,75] are some of the techniques employed as top-down methods.

3.1.1. Hydro/solvothermal Synthesis

One of the most common routes for synthesis of GQDs is hydrothermal/solvothermal synthesis, where carbon precursors, commonly graphene sheets, are cut into GQDs through oxidation followed by treatment at high temperatures.^[76] In a study, large graphene sheets were oxidized in concentrated H₂SO₄ and HNO₃ to form GO, introducing oxygen-containing functional groups in the edge and basal plane of the graphene sheets. The defect-rich GO was then dispersed in an aqueous NaOH solution and treated in an autoclave reactor at 200°C. The solution was dialyzed through a 0.22 μm membrane to separate the GQDs from the remaining bulk graphene precursor, and the resulting GQDs had a 9.6 nm average diameter and displayed strong blue fluorescence.^[71] Similar results have been achieved by dispersing oxidized graphene in organic solvents, instead of aqueous solutions, as Zhang et al. demonstrated using GO and dimethylformamide.^[77] The proposed mechanism of cutting oxidized graphene into GQDs is similar to that of unzipping a CNT. Briefly, the oxygen containing epoxy and carbonyl groups can exist in a linear chain within the lattice, which is a thermodynamically unstable structure. In order to form a more favorable state, the lattice is cut along this chain, inducing the cleavage of the underlying C-C bonds in the graphene lattice.^[71,78] The reaction conditions of hydro/solvothermal syntheses provide the dispersion media and energy through high temperatures to facilitate this reaction. The effect of oxidation of the precursor and the properties of the resultant GQDs were investigated. Here, GO was further oxidized by refluxing in HNO₃ before being treated hydrothermally. The resulting GQDs (GQD1) were collected through dialysis to collect particles smaller than 3500 Da; a sediment collected inside the dialysis bag was refluxed again in HNO₃, and the resulting solution was dialyzed to obtain a second batch of GQDs (GQD2). GQD2, which had undergone a second refluxing step, was measured to be smaller than GQD1, and exhibited PL at longer wavelengths than GQD1 did. XPS analysis revealed the presence of nitrogen-containing functional groups within the structure of GQD1 and GQD2, and the difference in PL properties was attributed to a greater extent of nitrogen doping in GQD2.^[79] In a more thorough investigation of the effects of precursor and temperature, bulk carbon precursors, including CNTs, GO, and carbon black, were treated hydrothermally in a mixture of ethanol and H₂O₂; the resulting GQDs were collected as a pellet through centrifugation. While the measured sizes of the three GQD batches were not significantly different, their defect structures, as measured by Fourier transform infrared spectroscopy and Raman spectroscopy, exhibited noticeable differences. GQDs derived from CNTs had the greatest density of defects. The difference was also apparent in PL measurements of the three types of GQDs, with CNT-based GQDs exhibiting the strongest PL intensity, further highlighting that the PL of GQDs is closely tied to their defect structures.^[80]

Recent efforts in hydrothermal synthesis have been aimed at improving the ease of synthesis, as well as minimizing detriment to the environment by avoiding the use of strong oxidizing agents. In a recent investigation, the hydrothermal approach to GQD synthesis was combined with a flow synthesis method for a rapid, green production of GQDs. In this method, an aqueous solution of GO and calix arene tetrasulfonic acid was mixed with a flow of KOH, which was then mixed in a rapid flow with supercritical water into a reactor, from which, GQDs were collected in a cooling chamber. The resulting GQDs were highly soluble in aqueous solutions and exhibited tunable PL properties by varying the concentrations of the reagents, and therefore controlling the degree of defects. Furthermore, this method allows for a rapid synthesis, compared to lengthy hydrothermal treatments required of using autoclave reactors.^[81] The fluid-based nature of hydro/solvothermal methods allows facile modifications to the process, as demonstrated by this study. Furthermore, synthesis from bulk materials allows large scale production of GQDs, while imparting high water solubility due to oxygen-containing functional groups present on the edge. However, these methods require special equipment such as autoclave reactors and harsh conditions that cause an environmental safety concern. Hydro/solvothermal synthesis methods do not allow precise control over size, morphology, or other properties, as these processes involve cleavage of large molecules at random sites throughout their lattice. Recent efforts have been aimed at overcoming these obstacles to allow straightforward and controlled synthesis of GQDs through variation of reaction conditions in one-step processes.

3.1.2. Electrochemical Exfoliation

In electrochemical exfoliation, layered bulk materials, such as graphite rods, are exfoliated with the aid of an applied electric field. Initially used as a method to obtain pristine graphene sheets,^[82,83] electrochemical exfoliation has more recently been used to synthesize GQDs with precise control over size and shape. In electrochemical exfoliation, materials with sp² hybridized carbon, usually graphite or multiwalled CNTs, undergo a two-step process: (1) the material is placed in an electrolyte solution, and an anodic potential is applied versus a reference electrode, initiating the cleavage of sp² bonds and intercalation of the electrolytes or solvent molecules between the layers of the bulk material; (2) a cathodic potential is applied for electrochemical reduction of the product.^[84] The control over the applied electric potential allows more selective oxidation in the bulk material, in contrast to hydro/solvothermal synthesis methods, in which oxidation is not controlled. Furthermore, electrochemical exfoliation does not require toxic oxidizing or reducing agents and can be performed in ambient conditions. The potential for electrochemical exfoliation in GQD synthesis was developed, where multi-walled CNTs (MWCNTs) were initially exfoliated into multilayered graphene nanoribbons, by first oxidizing the MWCNTs at a controlled potential, then followed by a reduction process. The presence of an electric field provided more control over the orientation of the MWCNTs, allowing synthesis of uniform size and defects in the product.^[84] In their subsequent work, GQDs exhibiting green luminescence with diameters ranging from 3 to 8 nm were synthesized from MWCNTs through electrochemical exfoliation at 90°C. In addition, the study also showed the temperature dependence of the resulting GQDs, with the average diameter increasing to 23 nm at 30°C, hinting at possibilities of tuning optical properties through altering the size of the GQDs.^[85] While the previous study used an aprotic solvent with LiClO₄ as the electrolytic solution, aqueous solutions have also been used. Tan et al. synthesized 3 nm wide GQDs exhibiting red PL through electrochemical exfoliation of graphite in an aqueous solution of K₂S₂O₈, and these GQDs were used in cell imaging. The very active SO₄^−● radicals evolved in the solution acted as electrochemical “scissors” to cleave the bulk material into pristine, intact sp² structures.^[86] The effect of multiple electrolytic species and their relative concentration were also investigated. Two graphite rods were submerged in a solution of NaOH and citric acid. Varying the concentration of NaOH led to changes in defect density of the resulting GQDs due to the availability of OH⁻ ions to intercalate between the layers of graphite. Analysis by XPS and Raman spectroscopy revealed that the abundance of the OH⁻ ions in the synthesis setup correlated with the amount of oxygen-containing functional groups observed in the GQDs. This was further supported by the red-shift of the PL spectra with increasing defect concentration.^[73]

Electrochemical synthesis is also used to impart distinct chemical functionalities onto GQDs to alter their properties. A reported method for synthesis of amine-modified GQDs (NGQDs) was through in situ functionalization of graphene with amine groups via electrochemical exfoliation. Carefully-cut, highly-oriented pyrolytic graphite was placed in a solution of (NH₄)₂HPO₃; then, an anodic potential was maintained across the sheets for varying amounts of time. The graphene nanosheets (GN) were separated by size using polymeric filters, and NGQDs were collected through centrifugation. The resulting NGQDs displayed greater amination levels when synthesized with lower applied potentials, as the rate of GN exfoliation was slower, allowing sufficient in situ functionalization with the ammonium ions in the solution.^[87] Zhang et al. demonstrated the synthesis of boron-functionalized GQDs by doping graphite rods; XPS spectra verified the presence of boron functional groups on the synthesized GQDs.^[88]

Electrochemical synthesis provides a pathway for more precise synthesis of GQDs compared to hydro/solvothermal methods, by selectively oxidizing the precursor material according to the applied electric potential. Further studies demonstrated that electrochemically exfoliated GQDs can be further functionalized by varying the electrolyte solution. The relatively simple setup, as well as the absence of strong oxidizing agents, demonstrated electrochemical exfoliation to be an efficient synthesis route for production of GQDs.

3.1.3. Liquid-phase Exfoliation

Another method for exfoliation of larger carbonaceous materials into GQDs is to use ultrasonication in an organic solvent, eliminating the need for oxidizing reagents and providing a pathway for scalability. Sonication provides the energy necessary to cleave the larger molecules into nanoscale particles. Such processes, dubbed liquid-phase exfoliation (LPE), have been used for treatment of many two-dimensional materials. LPE is especially useful in preserving the integrity of the original carbon source. Acetylene black and nano-graphite were dissolved in N-methyl-2-pyrrolidone (NMP) to serve as sources for high-defect GQDs (HD-GQDs) and low-defect GQDs (LD-GQDs), respectively, and were subsequently ultrasonicated for 1h, after which the residual precipitates were removed via centrifugation (Figure 7a). The ratio between the disorder band and the graphitic band in the Raman spectrum of the HD-GQDs was 0.966 and 0.413 for LD-GQDs, indicating a lower edge roughness and good crystallinity in LD-GQDs. The effect of the difference in crystallinity is consequently demonstrated in the PL properties of HD-GQDs and LD-GQDs. The PL spectra of HD-GQDs showed a strong excitation-dependent behavior, whereas the PL spectra of LD-GQDs was largely independent of excitation wavelength (Figure 7b,c). Interestingly, the size distribution of GQDs was much greater in LD-GQDs (2 to 9 nm) than in HD-GQDs (2 to 6 nm). This can be attributed to the greater amount of defects present throughout acetylene black compared to nano-graphite serving as cleavage sites to form GQDs, resulting in more cleavages overall. This is supported by the differences in morphology of HD-GQD and LD-GQDs, as LD-GQDs were shown to have more linear edges than the curved rough edges in the HD-GQDs. The difference in properties between HD-GQDs and LD-GQDs showed that the crystallinity of the starting material is an important factor in determining the properties of the resultant GQDs.^[74]

Figure 7.

Characterization of GQDs synthesized through liquid phase exfoliation. (a) Schematic presentation of liquid phase exfoliation and the mechanism of GQD synthesis. (b) PL spectra of HD-GQDs and (c) LD-GQDs at various excitation wavelengths. Reproduced with permission.^[74] Copyright 2016, Elsevier.

The mechanism of exfoliation in LPE is the vibration of solvent molecules between the graphite sheets acting as an intercalation lever, leading to structural defects and further disintegration of graphite sheets into smaller fragments. A convenient, easily accessible device capable of generating the vibrations necessary in LPE is a microwave oven. Microwave irradiation has shown to lead to expansion of the graphene sheets to facilitate the intercalation prior to the sonication step. In a microwave-assisted LPE method, graphite flakes with size < 300 μm were expanded by irradiating them for 1 min in a household microwave oven. Then, these expanded graphite flakes were dispersed in NMP and ultrasonicated for 9 h, after which GQDs were obtained by removing sediments via centrifugation.^[75] In contrast to electrochemical exfoliation, LPE does not provide precise control over the morphology of the resulting GQDs, because the cleavage of the bulk material occurs at the intercalation sites that are randomly dispersed within the lattice. However, the use of sonication instead of chemical oxidation and reduction as a means of synthesis demonstrates a safer and simpler alternative to other top-down methods. LPE presents a low-cost, simple, and environmentally friendly method for scalable, high-yield synthesis of GQDs.

3.2. Bottom-up Methods

In contrast to top-down methods, bottom-up methods involve the assembly of GQDs from small precursor molecules. The bottom-up methods of GQD synthesis yield GQDs with controlled morphology, as well as monodispersed particles. The diversity of carbon precursors used in bottom-up synthesis is reflected in the various synthesis methods reported. While top-down synthesis methods adhere to an overall scheme of oxidation followed by reduction of the precursor, the mechanisms for bottom-up synthesis range from organic chemistry^[5] to pyrolysis.^[6,16] Among these methods, recent progress in stepwise organic synthesis, pyrolysis of precursors, and microwave-assisted synthesis are reported here.

3.2.1. Stepwise Organic Synthesis

Organic synthesis or solution chemistry methods involve building GQDs from small carbon precursors through organic reactions and allow precise control of the structure of the final product. In a stepwise synthesis, small organic molecules with benzene moieties, such as 3-iodo-4-bromoaniline, were used as building blocks to synthesize GQDs consisting of exactly 168, 132, and 170 conjugated carbon atoms. The resulting GQDs were remarkably uniform, and this technique provides for complete control over the desired physical and optical properties of the synthesized GQDs.^[5] Subsequent work has shown the capability for doping with nitrogen, as well as the addition of functional groups to tune the bandgap of resulting GQDs.^[89,90] Another synthesis method was reported that was based on the dehydration polymerization of nitriloacetic acid to form nitrogen-doped GQDs in a high temperature and high pressure environment. The proposed mechanism is the formation of the double bonds between carbon atoms as a result of nucleophilic addition after the dehydration step. The resulting GQDs had an average diameter of 29 nm, which is quite large compared to most reported diameters of GQDs (<10 nm). However, they exhibited blue fluorescence and high QY. The reason for the blue fluorescence exhibited by this GQD, as opposed to red emission which is expected for larger GQD systems, may be attributed to the extensive nitrogen doping throughout the structure, increasing the energy bandgap of the GQDs.^[91] Even though this work is presented as a stepwise synthesis, the control of the reaction product was not observed, as shown through the large size distribution of the GQDs. The GQDs synthesized through these methods have been reported to be soluble only in organic solvents, which is a barrier in the transition of these GQDs to biological applications. Furthermore, these methods require very specific chemicals, as well as conditions for the organic reactions to occur. Owing to the specific reactions involved and little potential for scalability, stepwise organic synthesis seems to have lost its viability as a novel synthesis method for biological applications.

3.2.2. Pyrolysis of Precursors

Pyrolysis involves the decomposition of organic molecules at a high temperature in order to irreversibly alter their chemical composition and structure.^[92] Molecules such as citric acid, glutathione, and ethanolamine are treated at high temperatures, after which, the products are purified.^[93,94] These reactions are labelled as ‘one-pot’ syntheses. Utilizing only a few reactants in a reaction vessel, this one-step synthesis presents great potential for scalability. Pyrolysis of molecules such as citric acid was shown to yield uniform, monodisperse GQDs. Trisodium citrate was heated for 4 min at 200°C, dissolved in deionized water, then purified using centrifugal filtration devices to obtain GQDs between 3 and 10 kDa. The resulting GQDs displayed high crystallinity, as well as high QY of 3.6% with quinine sulfate as the reference dye.^[16] The relative ease and economic benefits position pyrolysis as a favorable method for GQD synthesis. However, these processes are often accompanied by lengthy purification steps such as passive dialysis. Due to the nature of pyrolysis, the product consists of various carbonaceous materials such as mesoporous carbon^[95,96] and carbon nanotubes,^[97] resulting from decomposition of the precursors. Identification of reaction conditions that favor synthesis of GQDs over other carbon-based materials, as well as effective purification methods would further propel pyrolysis to a more relevant synthesis method for GQDs.

3.2.3. Microwave Assisted Hydrothermal Synthesis

The pyrolysis methods described above utilize heating elements such as convection from an oven or conduction through a heating mantle. Such methods of applying heat inevitably result in a thermal gradient within the reaction vessel, resulting in discrepancies within the synthesized product. Microwave assisted hydrothermal synthesis has been reported as an alternative to traditional hydrothermal synthesis. By irradiating the reactants with microwaves in an aqueous solution, the heat can be distributed more evenly throughout the reaction vessel, resulting in highly uniform GQDs. Furthermore, microwave irradiation can dramatically shorten reaction times and lead to the synthesis of functionalized GQDs by utilizing precursors containing various chemical functional groups. In an early approach to microwave assisted GQD synthesis, an aqueous solution of glucose was irradiated using a conventional microwave oven in the presence of ammonia. Variation of the reaction time and irradiation power led to changes in the fluorescence properties exhibited by the resulting GQD solutions.^[6] The proposed mechanism of synthesis for such GQDs is the breakdown of the carbon bonds in the precursor molecules leading to polymerization and carbonization of the molecules, which then serve as nucleation sites for GQDs to grow from.^[32] Recent efforts have been aimed at using various precursors to further functionalize these microwave-synthesized GQDs to alter their fluorescent properties. A mixture of glucosamine and thiourea was used in order to impart nitrogen (N-GQDs) and sulfur (NS-GQDs) into the GQDs (Figure 8). The resulting N-GQDs and NS-GQDs displayed different fluorescence properties in both the visible light and NIR range, as well as different physical characteristics.^[98] Microwave assisted hydrothermal synthesis presents a simple and easy method for GQD synthesis, with a modification to the traditional heating methods. The uniform heating provided by microwave irradiation leads to a more consistent and homogenous product, and studies have demonstrated synthesis of GQDs using a conventional kitchen microwave, eliminating the need for intricate reaction vessels. However, many studies that have used similar methods have since labelled the resulting particles as CQDs, rather than GQDs, owing to their lack of crystallinity, and more so due to the ambiguity between the definitions of the two particle systems.^{[3,33,99,100]} Regardless, the microwave-assisted methods allow a rapid, simple, and environmentally friendly synthesis of QDs that display many of the same properties as GQDs.

Figure 8.

Microwave synthesis of GQDs. Schematic drawing of microwave synthesis of (a) N-GQD and (b) NS-GQD showing the functionalization of GQDs with amine- and sulfo-functional groups. Reproduced with permission.^[98] Copyright 2018, John Wiley and Sons.

4. Bioimaging

Bioimaging is an important methodology that is used in both research and clinical settings, allowing for observation of biological processes such as targeted delivery, cellular uptake, and biodistribution of therapeutics in an isolated, detailed manner using various portions of the electromagnetic spectrum.^[10,57,101] In cancer diagnostics, the role of imaging is especially important as sensitive imaging allows early detection of tumors, as well as identification of metastasis and the resurgence of cancer. The intrinsic PL of GQDs allows them to be used as optical probes in fluorescence imaging, with no further conjugation of fluorescent dyes as is needed with other nanoparticle platforms. Recent advances have extended the scope of the usefulness of GQDs in bioimaging, as they have shown capabilities for both NIR fluorescence imaging and MR imaging. Compounded with their excellent biocompatibility, GQDs are ideal imaging probes with application in various modalities of bioimaging.

4.1. Fluorescence Imaging

Fluorescence imaging is an important tool in biomedical applications that uses visible light and NIR spectra to analyze the distribution of molecules of interest in cells, tissues, and whole animals.^[102] Utilizing specific molecules that display PL as probes, fluorescence imaging is widely used in both laboratory settings and clinical practice for applications such as cell tracking,^[103] monitoring of therapeutics subcellular imaging,^[104,105] and disease diagnostics.^[106] Organic dyes, or fluorophores are commonly used fluorescent probes in vitro, in vivo, as well as ex vivo. An important factor in determining whether a material can be used in the body is solubility. Many fluorophores suffer from poor water solubility and often require an extra bioconjugation step to render it soluble in bodily fluids.^[107] If water-insoluble substances are introduced into the body, they will likely aggregate and be marked for removal by the body’s immune system, or, if introduced in a large enough amount, they could cause catastrophic damage such as blockage of blood flow upon introduction to the vascular system. Furthermore, organic fluorophores have been reported to be toxic at high doses.^[108] Interest in semiconductor QDs as fluorescent probes has been on the rise due to their optical properties. These QDs, such as CdSe, require surface coating with polymers or ligands to become soluble in water.^[107] Biocompatibility of semiconductor QDs have long been a concern as their constituents have proven to be toxic, and their incomplete clearance from the body can lead to chronic damage in tissue.^[109,110] In the case of GQDs, their intrinsic water solubility and low toxicity have been widely reported,^[6,12,58,86] positioning GQDs as a viable platform for fluorescence imaging.

The bioimaging potential of GQDs was explored by using GQDs synthesized from carbon black as a fluorescent probe in confocal microscopy. The product showed high biocompatibility and displayed green fluorescence once internalized in MCF-7 human breast cancer cells.^[111] Due to their amphiphilicity, arising from the hydrophilic edge functional groups and hydrophilic basal plane, and small size, GQDs are often able to enter cells without a targeting ligand. While non-specific cellular uptake is advantageous in some applications, the ability of an imaging system to have high selectivity for identifying and targeting specific cells or subcellular components is often sought after in both diagnostic and drug delivery systems. To demonstrate the targeting ability of a GQD-based imaging system, insulin was conjugated onto GQDs for labelling and tracking of insulin receptors in 3T3-L1 adipocytes. Small clusters of insulin-GQDs were observed in adipocytes, and by using GQD fluorescence, the authors were able to track the lateral movement of GQDs on the receptors located within the cell membrane.^[112] The stability of the GQD-insulin conjugate in different intracellular environments was effectively demonstrated in this study.

Many methods of modifying GQDs also allow them to be used as probes for imaging cell nuclei. Amphiphilic GQDs functionalized with nitrogen and chlorine were synthesized via a one-step solvothermal method. These GQDs displayed good solubility in aqueous environments, a large positive zeta potential (28.5 mV), which was favorable in crossing the cell membrane, as well as the ability to bind to histones in cell nuclei. Confocal fluorescent images of HeLa cells stained with both GQDs and 4’,6-diamidino-2-phenylindole showed nuclear targeting of GQDs without the aid of any conjugated ligands.^[113] GQDs have been demonstrated to be stable fluorescent probes for facile in vitro imaging, while causing minimal cytotoxicity. Their high water solubility and nanoscale size allow them to be internalized in cells without the aid of targeting molecules. The tunable PL of GQDs also allows them to be used in conjunction with other fluorescent probes to simultaneously observe various processes within the cell. Coupled with the ability to deliver therapeutic material, GQDs can serve as an effective multifunctional probe in various biological applications.

Two-photon fluorescence has also been demonstrated in many GQD studies for bioimaging, as well as biosensing applications. In two-photon microscopy, the emitted photon has a shorter wavelength than the excitation wavelength as the electrons are excited by the absorption of two photons of longer wavelength, instead of a single photon with shorter wavelength. This moves the excitation wavelengths away from the region where there is background interference, allowing for a more efficient fluorescence imaging process. Hydrothermally synthesized GQDs were amine- and sulfo-cofunctionalized to be used as a two-photon microscopy probe by Wang et al. 1,3,6-trinitropyrene was dissolved in an aqueous solution of ammonium sulfide, which was then placed in an autoclave for 12 h at 200°C. The resulting GQDs displayed blue fluorescence with both one-photon and two-photon fluorescence spectroscopy. HeLa cells were treated with the GQDs and irradiated under an 800 nm laser for two-photon confocal imaging. The images of the treated cells show uptake of GQDs by the cells, as well as strong fluorescence from the GQDs, showing that these GQDs are suitable for long-term two-photon fluorescence microscopy and observation.^[114]

The extent of enhanced tissue penetration of two-photon fluorescence imaging using GQDs was also investigated. Nitrogen-doped GQDs (N-GQDs) were synthesized from GO and dimethylformaldehyde as two-photon fluorescence probes. The N-GQDs emitted blue-green fluorescence and were shown to be in the cytoplasm but not inside the nuclei of Hela cells (Figure 9a–c). Various thicknesses of Intralipid were used as tissue phantoms to replicate the scattering and absorbance of tissue (Figure 9d). The tissue phantoms were placed between N-GQDs and an objective lens of a microscope, and tissue penetration was measured by observing the intensity of the fluorescence image of N-GQDs at various thicknesses of tissue phantoms. Using two-photon fluorescence, a high signal-to-noise ratio was observed from 0 to 1300 μm, and even at 1800 μm, the fluorescence of N-GQDs were observed through the tissue phantoms, while utilizing normal fluorescence demonstrated tissue penetration of less than 400 μm (Figure 9e).^[115] A similar experiment performed with rat liver tissue and NIR emission showed that the tissue penetration was up to 320 μm, which is lower than that demonstrated by the previous study. However, the NIR-emitting GQDs were compared to GQDs with green fluorescence, which demonstrated a tissue penetration depth of only 180 μm.^[116] The large imaging depths demonstrated by the two-photon and NIR fluorescence of GQDs highlight the potential of utilizing GQDs for non-invasive detection of biological activities and diagnostics for diseases.

Figure 9.

Nitrogen doped GQDs for cell and tissue imaging. (a) PL emission spectra of one-photon fluorescence (OPF) and two-photon fluorescence (TPF). (b) Bright field and (c) fluorescence images under 800 nm excitation of Hela cells (scale bar = 10 μm). (d) Schematic of setup used for two-photon fluorescence imaging (TPFI) using tissue phantoms with varying thickness. (e) Penetration depth of N-GQDs for TPFI and one-photon fluorescence imaging (OPFI) in tissue phantoms (scale bar = 100 μm). Adapted with permission.^[115] Copyright 2013, American Chemical Society.

Recent efforts in engineering the PL of GQDs for bioimaging have focused on the use of a portion of the electromagnetic spectrum known as the second near-infrared window (NIR-II) which is typically defined as the range of wavelengths from 1000 to 1700 nm. The NIR-II imaging window may be viewed as an optimal fluorescence imaging region given tissue’s relative transparency to this sector of the electromagnetic spectrum and the maintenance of excellent spatial resolution. The amount of light scattered when photons encounter matter is inversely proportional to the fourth power of the wavelength of the incident light. As such, it may prove desirable to establish a bioimaging system that operates at wavelengths longer than the visible light or traditional NIR windows (750-900 nm) in order to further increase tissue penetration depth and increase signal strength. However, the spatial resolution of images deteriorates as longer wavelengths of electromagnetic radiation are employed. Thus, the ability of GQDs to exhibit PL in the NIR-II region brings them to the frontier of bioimaging techniques. In a study that utilized GQDs for NIR-II imaging in a mouse model, GQDs were injected into the tail veins of mice at a dose of 2.5 mg kg⁻¹, and the whole-body fluorescent signal throughout the mice was monitored for a period of 16 h. The biodistribution of GQDs in different organs were analyzed ex vivo. After 30 min post-injection, fluorescent signal was observed throughout the mice, and the intensity remained for 8 h. Recently, Wang et al. developed GQDs that were simultaneously doped with both nitrogen and boron (N-B-GQDs) for NIR-II bioimaging.^[41] These N-B-GQDs were synthesized via a one-pot pyrolysis of 3-aminophenylboronic acid monohydrate in a solution of acetone and hydrogen peroxide at 230°C, which yielded highly monodisperse 5 nm diameter particles with two or three layers of graphene per particle (Figure 10a). When irradiated with an 808 nm excitation source, the N-B-GQDs exhibited NIR-II PL emission in the range of 950 to 1100 nm (Figure 10b). The authors theorized that the extensive nitrogen and boron doping created a considerable number of vacancy defects in the graphene lattices of the N-B-GQDs which resulted in a red-shifting of the PL emission peak into the NIR-II window. After showing N-B-GQDs to be non-cytotoxic in three different cell lines (Figure 10c), the authors investigated the in vivo PL capabilities of their NIR-II fluorophores by acquiring a series of PL images beginning one minute after intravenous tail-vein injection of N-B-GQDs in nude athymic mice over a duration lasting 120 min (Figure 10d). During this experiment, PL was measured in the wavelength range of 1000 nm to 1700 nm with an 808 nm excitation source at an illumination density of 3 W m⁻². Note that minimal autofluorescence was observed in tissue not containing N-B-GQDs, whereas tissue with N-B-GQDs such as the kidneys and blood vessels are clearly visible immediately upon injection.

Figure 10.

Nitrogen and boron dual-doped GQDs for NIR-II bioimaging. (a) Transmission electron microscopy image showing the monodisperse, 5 nm diameter particles. The inset at the top right reveals the lattice fringe indicative of graphene, whereas the inset at the bottom left is a photograph of the GQDs in solution demonstrating their excellent solubility. (b) PL spectrum of N-B-GQDs exhibiting NIR-II emission when excited with an 808 nm laser source. The insets display an optical image and a PL image of N-B-GQDs in aqueous solution. (c) In vitro cytotoxicity study of N-B-GQDs performed by assessing the viability of SF763, 4T1, and B16F10 cells 72 h after incubation with N-B-GQDs. Statistical analysis was performed using the Student’s two-tailed t-test (**p < 0.01, ***p < 0.001) (d) In vivo NIR-II imaging of live mice. Panel in the top left is a photograph of a nude mouse and subsequent panels in the top row depict PL images of mice that were injected with phosphate buffered saline (PBS), but not any contrast agent. The second and third rows of images depict a time-course of PL images after intravenous injection with N-B-GQDs. The bottom row of images highlights PL in the blood vessels. Adapted with permission.^[41] Copyright 2019 Elsevier.

4.2. In vivo Optical Imaging

While in vitro imaging is used as the primary demonstration of the capability of a fluorescent probe in optical imaging methods, in vivo imaging presents a challenge in comparison, as more factors need to be considered, such as the in vivo biocompatibility, biotoxicity, and the dynamic metabolism of the organism under study. Perhaps an indicator of the recent progress in GQD research is the number of studies that utilize GQDs for in vivo applications such as cell tracking and nuclear targeting.

Annexin V (A5)-modified GQDs were used to demonstrate the possibility of tracking apoptotic cells in zebrafish. Fluorescent GQDs synthesized from neem leaf extract were conjugated with A5, which has been widely used to detect apoptotic cells with other fluorescent probes. A5-fused-enhanced modified green fluorescent protein (EGFP) (A5-EGFP) transgenic zebrafish and wildtype zebrafish were treated with A5-GQDs. The EGFP and A5-GQDs illuminated a similar region in the zebrafish, indicating that the A5-GQDs were able to target the apoptotic cells.^[121] Though there are many more studies to be done before the results obtained in a zebrafish model can be translated and replicated in a mammalian model, this study implicates the feasibility of tracking apoptotic cells in vivo using a GQD-based probe. In another study, GQDs synthesized from graphite were used to track human adipose-derived stem cells (hADSCs) in a mouse model using two-photon fluorescence microscopy.^[122] The two-photon PL spectra showed emission wavelengths between 400 and 520 nm with 680 to 860 nm excitation and no significant cytotoxicity or inhibition of cell function. The hADSCs were treated with GQDs, encapsulated in methacrylated hyaluronic acid hydrogel, and were injected into the dorsal region of athymic mice. Under 670 nm irradiation, significant fluorescence signal was observed even after 24 h. In both of these studies, the GQDs maintained their PL property without significant photobleaching and demonstrated no significant toxicity or insolubility in aqueous environments in vivo. These studies indicate that GQD-based systems could be used as fluorescence probes in cell tracking as opposed to conventional organic dyes. In addition to optical imaging in the visible range, GQDs have also been utilized in imaging in the NIR window, as well as in MR imaging, demonstrating the remarkable versatility of GQDs as a platform in bioimaging. By improving the intensity of the signal from the GQDs, applications such as metastatic cell tracking or immune cell labelling using GQDs could well be within reach.

4.3. MR Imaging

MR imaging utilizes radio frequency signals to alter the spin of protons present throughout the body to obtain anatomical images and to observe physiological processes. MR imaging is a preferred method of clinical imaging due to its noninvasive nature, high spatial resolution, and virtually unlimited tissue penetration depth. Contrast agents (CAs) can be used to highlight the biological features under study by making them brighter (T1 CAs) or darker (T2 CAs). While advances have been made in T2 CAs using safe superparamagnetic iron oxide nanoparticles, T1 CAs are typically limited to transition metal ion chelates, especially those based on gadolinium (Gd). However, transition metals are well known to be toxic to the body, and Gd-based CAs have recently been shown to cause nephrogenic systemic fibrosis and accumulate in tissue after chronic use.^[117–119] To address these toxicity issues, various forms of doped graphene nanomaterials have been studied as safe, metal-free alternatives to transition metal-based CAs. Doping of graphene nanoparticles with various atoms leads to development or enhancement of magnetic properties by inducing local paramagnetic moments around the dopants.^[50] Fluorographene quantum dots (FGQDs) were synthesized by simple LPE of fluorinated GO. The resulting FGQDs had a high-defect density and displayed paramagnetic characteristics due to the presence of carbon-fluorine bonds acting as paramagnetic centers. The transverse MR relaxation rate r₂ is a measure of the spin-spin relaxation time, which indicates the time taken for the magnetization vector of the protons in an MR experiment to decay towards equilibrium. The presence of T2 CAs in a sample to be imaged shortens this time. The r₂ of the FGQDs were calculated to be 39 s g⁻¹ ml⁻¹, which is indicative of a good T2 CA. The presence of fluorine atoms also enables the FGQDs to be used in ¹⁹F MR imaging, allowing for dual mode imaging.^[52] Compared to using a single mode to obtain images from tissues, dual modal imaging provides an enhanced visualization of the tissue, and offers better reliability to the data, greatly increasing the accuracy and efficiency of MR imaging. In another study of doping GQDs to enhance their magnetic properties for use in MR imaging ^[120], boron doped GQDs (B-GQDs) were synthesized via a hydrothermal route. Here, the boron-dopants served as paramagnetic centers and the B-GQDs were evaluated as T1 CAs in vitro and in vivo. The in vitro longitudinal MR relaxation time of B-GQDs in aqueous solution was comparable to that of clinical Gd-based CAs. Furthermore, the MR imaging capability of B-GQDs was evaluated by administering B-GQDs through intravenous injection in mice, and high T1 signal intensities were observed in the heart, stomach, and kidney of the mice compared to the pre-injection signal (Figure 11).^[51] These advances have shown the development of graphene-based, metal-free MR imaging CAs. The inherent fluorescence of GQDs make these CAs multimodal imaging probes, further broadening the scope of their application in bioimaging. Modification of their surface functional groups with targeting ligands and other biologics remains to be explored.

Figure 11.

In vivo T1-weighted MR images of abdominal cross-sections of mice treated with boron-doped GQDs with dynamic time-resolved MR imaging acquired at various time points after intravenous injection. The arrows indicate various organs: heart (H), liver (L), kidneys (K), spleen (Sp), and stomach (St). The heart and stomach show the greatest contrast 68 min after administration. Reproduced with permission.^[51] Copyright 2016, John Wiley and Sons.

5. Biosensing

In addition to bioimaging, the optical properties of GQDs can be used in biosensing. While both bioimaging and biosensing applications utilize the PL of GQDs and require detection of emitted photons, the use of GQDs in bioimaging has allowed isolated visualization of specific cells and tissues of interest and enhanced contrast in MR images. On the other hand, the role of GQDs in biosensing systems is to detect and indicate the presence of biomolecules. As previously noted, altering the electron structure of the edge groups can alter the optical properties of GQDs. GQD-based biosensors utilize the affinity between specific functional groups within GQDs and the analyte biomolecule. When a functional group that is conjugated onto the GQD binds to the analyte, the association between the pair can provide different electronic states. By altering the electronic structure of the GQD, the detection of an analyte can then be measured as a change in PL intensity. Biosensor systems based on GQDs have shown to be capable of detecting ions, DNA, and various other metabolites.^[99,123] Biosensing requires high selectivity, sensitivity, and simplicity – the photostability of GQDs and rapid response of PL-based systems situate GQDs to be a promising platform for biosensing.

5.1. Ion and Small Molecule Detection

Whether they are essential to biological mechanisms or present acute toxicity even at small doses, ions need to be regulated and transported at the cellular level. Hence, it is important that in vitro ion sensors are selective, as well as sensitive.^[124,125] Biosensors utilizing the affinity of certain functional groups for specific ions have previously been developed using the PL properties of GQDs. There have been remarkable leaps in improving the selectivity and sensitivity of these sensors recently. A Ni²⁺ sensor based on GQDs modified with ethylenediamine (E-GQDs) was reported to have a QY of 83% and a detection limit of 3 × 10⁻⁸ M. The E-GQDs displayed a strong yellow PL, which was significantly quenched with the addition of Ni²⁺. This was further demonstrated in vitro by treating rat adipocyte-derived stem cells with E-GQDs, and the quenching of the PL upon introduction of Ni²⁺ to the cells.^[45]

In addition to ions, GQD-based biosensors can also be used to detect the levels of other chemicals in vitro. A turn-on sensor consisting of GQDs functionalized with (2,4-dinitro-phenoxy)tyrosine (DNPTYR) was developed to indicate a H₂S attack. Abnormal levels of H₂S in cells are related to illness including Alzheimer’s disease and cancer. This novel design utilizes the photoinduced electron transfer between GQDs and DNPTYR, an electron-withdrawing group. The PL of the GQDs was quenched upon covalent conjugation of DNPTYR; in the presence of H₂S, however, PL was recovered as H₂S cleaved the dinitrophenoxyl group. The design of this biosensor is illustrated in Figure 12a. The biosensor was able to indicate the presence of H₂S dynamically in vitro. MCF-7 cells were treated with GQD-DNPTYR and incubated until the particles were internalized, as indicated by small specs of green in confocal images (Figure 12b). The PL intensity of GQD-DNPTYR was continuously enhanced with the addition of H₂S, until phorbol myristate acetate (PMA) was added to decrease the H₂S levels in the cells, upon which the PL of GQD-DNPTYR stopped increasing. The detection limit was reported to be as low as 2 nM, highlighting the significance of this system.^[19] The excellent photostability, biocompatibility, and water solubility of GQDs present advantages over other systems such as semiconductor QDs and organic dyes in biosensing.

Figure 12.

PL turn-on nanoprobe for H₂S sensing based on DNPTYR-GQD. (a) Schematic illustrations of the synthesis and PL quench mechanisms of GQD-DNPTYR. (b) Confocal images of MCF-7 cells without incubation of GQD-DNPTYR (left most), 1 hour after treatment with GQD-DNPTYR (second from left), in the presence of H₂S for 25 min (third from left), and in presence of H₂S and PMA (right most). Green fluorescence is emitted by GQD-DNPTYR and blue fluorescence is emitted by NucBlue. Scale bar = 10 μm. Reproduced with permission.^[19] Copyright 2013, Royal Society of Chemistry.

Many of the GQD-based biosensors are designed so that the analyte interacts with a molecule attached to the GQD, resulting in PL recovery through an irreversible process. In other GQD-based detection probes, the analyte interacts directly with GQDs, often modified with various functional groups. In such designs, the affinity between the analyte molecule and the functional groups are of utmost importance. In a nitrogen-doped GQD-based biosensor for detection of 2,4,6-trinitrophenol (TNP), tris(hydroxymethyl)aminomethane was used as a surface passivating agent to impart amine functionalities on the surface of the N-GQDs. To assess the specificity of the N-GQDs for TNP, the extent of PL quenching in the presence of TNP was compared to that in the presence of various metal ions, as well as other aromatic compounds with structures similar to TNP. The results showed that significant photoquenching occurred only in the presence of TNP. This was attributed to the overlap of the absorbance spectrum of TNP and the emission spectrum of the N-GQDs, resulting in efficient fluorescence resonance electron transfer.^[126] However, the selectivity of the N-GQDs in binding with TNP was not tested, and the affinity between the N-GQD and the TNP could not be analyzed. In another study, N-GQDs were used as catalysts in the reduction of hydrogen peroxide, as well as a colorimetric indicator of glucose and hydrogen peroxide. The kinetics of N-GQDs were compared with that of horseradish peroxidase (HRP), with 3,3’,5,5’-tetramethylbenzidine (TMB) and H₂O₂ as substrates. The Michaelis-Menten constant (K_m) of N-GQDs with H₂O₂ as the substrate (0.10) was much greater than that of HRP (2.39), indicating that higher concentrations of H₂O₂ were needed to achieve maximum activity. On the other hand, the K_m value when using TMB as the substrate (11.19) was much greater for N-GQDs than it was for TMB (0.18), showing the affinity of N-GQDs for aromatic compounds.^[127]

While these in vitro studies demonstrate that GQDs can be used to detect analytes in a cellular environment, in vivo biosensing through GQDs could bring them closer to clinical relevance. Advances in GQD research, such as utilizing the NIR-II window for PL emission for deeper tissue penetration, advocate that GQDs could be used as in vivo biosensors. Although these studies suggest that the biocompatibility and water solubility of GQDs allow them to be used in in vivo experiments, only a few studies of in vivo GQD-based biosensors have been reported. Nanoparticles synthesized in a one-pot hydrothermal synthesis method from citric acid and neutral red, an organic dye, were used in in vivo detection of the noble metal ions Pt²⁺, Au³⁺, and Pd²⁺ in zebrafish. While the authors have labelled these nanoparticles as carbon dots (CDs), the physico-chemical and optical characterization of these CDs show results similar to many GQD systems, especially those synthesized using hydrothermal methods, further highlighting the need for a unified distinction between CQDs and GQDs. The resulting CDs demonstrated red PL emission and had an average size of 3.4 nm. To assess the in vivo detection capability of the CDs, zebrafish were fed with different concentrations of Pt²⁺, and placed in a CD solution for 4 h. In the subsequent fluorescence imaging of the zebrafish, the intensity of the PL emission correlated with the concentration of Pt²⁺ fed to the zebrafish.^[128] While barriers to translation to mammalian models exist, such as the immunogenicity of GQDs and greater signal absorbance by the tissue, this study suggests that GQDs can be used as a biosensor in living systems.

5.2. GQDs as Diagnostic Platforms

In addition to detecting small molecules in cells, GQD-based biosensors are also able to serve as diagnostic tools for diseases such as cancer. The nutritional and metabolic environment in tumors is distinct from those of surrounding healthy tissues. Notably, the pH in tumors is consistently lower than that of healthy tissues due to the production of lactic acid and hydrolysis of ATP in the anaerobic and energy-deficient conditions found in the tumor.^[129] This discrepancy in cellular environments has been exploited for improved and efficient cancer diagnostics.^[130] Several studies have shown a change in PL of GQDs correlating to the pH of the aqueous medium in which they reside. A pH responsive sulfur-nitrogen-doped GQD system (pRF-GQD) was reported to exhibit blue PL in pH above 6.8 and transition into green PL in pH below 6.8. The pH at which the optical property would change was tuned so that it would be responsive to the pH of the tumor environment. An intravenous injection of pRF-GQD was administered to mice bearing HeLa tumors, and the PL signal was observed for 24 h post-injection. Strong green PL emission was observed from the tumor site, and upon imaging various tissues after euthanizing the mice, it was found that the tumor tissue had a high concentration of pRF-GQD due to the enhanced permeation and retention effect. The utility of pRF-GQD in detecting different tumors was demonstrated by fluorescence imaging of mice bearing PANC-1, HepG2, A549, U87MG, and HeLa tumors.^[131] The spatial resolution of the tumor is far better from what is necessary to detect early stage cancer or metastatic clusters in clinical settings. However, the novel pH-responsive fluorescence of this system demonstrates the possibility of cancer diagnosis with passive targeting. Furthermore, the pRF-GQDs exhibit upconversion PL, allowing further tissue penetration and greater sensitivity as the effect of the autofluorescence of biological bodies is minimized.

While commercial applications of GQDs in clinical diagnostics has yet to be realized, GQDs have been utilized as chemical sensors in other applications. One example is a paper strip-based chemical sensor for environmental and food samples using Forster resonance energy transfer (FRET) in GQDs. In this study, GQDs were synthesized from citric acid, and were embedded into small areas confined by a wax lining on a nitrocellulose matrix. A mobile phone was used as both a source of UV excitation of GQDs and a detector for the fluorescence of the GQDs (Figure 13). Various compounds with aromatic rings were tested for PL quenching of the GQD strips and the capability of this sensing system to detect different levels of analyte were demonstrated.^[132] While the current study pertained more to food and environmental samples and detection of certain aromatic compounds, it presents the potential for a non-invasive and inexpensive diagnostic system that could be viable in clinical settings.

Figure 13.

Schematic representation of the paper embedded GQD-based sensor via FRET phenomenon. (a) Examples of phenolic compounds that are able (yellow) and unable (red) to quench the PL of GQDs. (b) Schematic representation of the device used to illuminate the paper sample. (c) Electric circuit of the UV LED connected to a USB port. (d) Schematic of nitrocellulose paper strip with embedded GQDs in circular areas separated by printed wax. (e) Image of the sensor system in use. The screen shows the readout of the sample via the fluorescent spot detected by the camera. (f) Schematic of a typical yes/no (ON/OFF) result. Reproduced with permission.^[132] Copyright 2017, Springer Nature.

6. Therapeutic Applications

The physicochemical and biological properties of GQDs have led to their use in therapeutic applications. The variety of functional groups available on the edge of GQDs can be conjugated with targeting ligands and therapeutics. In addition to the edge functional groups, the presence of sp² hybridized carbon in the GQD lattice allows therapeutic molecules, which also often contain aromatic rings, to be bound onto the basal planes of the GQDs, increasing the drug loading capacity and allowing for more efficient drug release. In addition to chemotherapeutic drugs, GQDs have been reported to deliver DNA for gene therapy. Furthermore, GQDs can enhance the generation of ROS upon incident irradiation, allowing them to be used in PDT. Figure 14 illustrates the role of GQDs in therapeutic applications.

Figure 14.

Therapeutic applications of GQDs. Chemotherapeutic drugs can be loaded onto the basal plane of GQDs through π-orbital stacking. Doxorubicin (DOX) is used as a representative small-molecule chemotherapeutic drug here (top left). Conjugation of edge groups with cationic peptides confers a positive charge onto GQDs, which allows them to form complexes with negatively charged DNA or RNA for gene therapy (top right). GQDs can act as photosensitizers, generating reactive oxygen species upon irradiation at a specific wavelength (bottom). GQDs have been shown to be a potent photosensitizer in photodynamic therapy.

6.1. Drug Delivery

GQD-based drug delivery systems have garnered much attention not only due to their minimal toxicity and the presence of functional groups, but also their high drug loading capacity and ability for simultaneous tracking due to their PL. This has led to many studies using GQDs as chemotherapeutic delivery systems for disease treatment. Anti-cancer therapeutics such as doxorubicin (DOX) and methotrexate (MTX) are potent but cannot be delivered efficiently in their free forms to desired tissues in the body due to poor solubility. The molecular structure of these therapeutics allows them to be bound onto the basal plane of GQDs via π-stacking and to be delivered in a hydrophilic carrier. The large surface area-to-volume ratio of GQDs also allows for high drug loading compared to traditional drug delivery systems. The capability of GQDs as a DOX delivery system was shown through biotin-conjugated GQDs (GQD-BTN). The drug loading process consisted of simply mixing GQD-BTN and DOX, during which the π-orbitals would stack to form GQD-BTN-DOX complexes. By observing the fluorescence signal from DOX, the cellular uptake of DOX was 27% higher from the administration of GQD-BTN-DOX than from the administration of free DOX and GQD-DOX.^[8] Nitrogen-doped GQDs (N-GQDs) were also studied as an MTX delivery system. MTX was loaded onto the N-GQDs through mixing a dilute solution of MTX and N-GQDs (Figure 15a). The complex was observed to release 60% of the drugs in the first 9 h, followed by a slow liberation for a total release over 48 h. The cytotoxicity of the MTX-N-GQDs in MCF-7 cells were analyzed; compared to free MTX, the MTX-N-GQDs initially showed similar cell viability, but at later times, they exhibited greater cytotoxicity, which the authors attributed to the slower diffusion of the MTX-N-GQD complex (Figure 15b). On the other hand, no changes in cell viability were observed after incubation with N-GQD without MTX after 48 h.^[11] In these studies, GQDs have been shown to be capable of enhancing the therapeutic effects of anti-cancer drugs, demonstrating the possibilities of a GQD-based drug delivery system.

Figure 15.

GQDs for MTX delivery. (a) A schematic representation of N-GQD synthesis via the hydrothermal method, formation of the MTX-(N-GQD) complex, and proposed intracellular release. (b) In vitro cytotoxicity of MTX-(N-GQDs) assesd 48 h after incubation. MCF-7 cells were exposed to different concentrations of free MTX (dark bars) and MTX-(N-GQDs) (light bars). Statistical significance of MTX viability compared to MTX-(N-GQDs) is denoted by * and ** at the level of p < 0.05 and p < 0.01, respectively. (c) Evaluation of cytotoxicity of N-GQDs without any MTX. Reproduced with permission.^[11] Copyright 2017, Elsevier.

It should be noted, however, that the anticipated concurrent treatment and tracking utilizing the PL of GQDs has not been demonstrated in the aforementioned studies. Due to the π-π interaction of the drugs and GQDs, the PL intensity of GQDs have actually been shown to be quenched. Furthermore, reports of the mechanism of drug release from the GQD-drug complexes are at times conflicting – while the homogenous fluorescence intensity is attributed to the detachment of DOX from the GQD in the low-pH environment of the endosomal compartment, the greater cytotoxicity of the MTX-GQD complex compared to free MTX is attributed to its slower diffusion at longer times.^[8,11] The mechanism of cellular internalization of GQDs could provide an insight into the drug release mechanism. In studies in which targeting ligands are conjugated onto GQDs, the GQD-drug complex is reported to be internalized via receptor-mediated endocytosis. This was demonstrated by observing the fluorescence intensities of GQD and DOX in a GQD-DOX-folic acid system in HeLa cells. The fluorescence of GQD was observed in the cytoplasm 0.5 h after treatment, while significant fluorescence of DOX was observed 8 h after treatment. These results indicate a rapid receptor-mediated endocytosis, and a slow dissociation between the GQD and DOX.^[12] A more thorough investigation of cellular internalization of GQDs showed that GQDs were present in the endoplasmic reticulum, as well as the cytoplasm, through TEM images of MGC-803 cells, indicating the possibility of GQD internalization through caveolae-mediated endocytosis. Fluorescence imaging of MCF-7 cells showed that some of the GQD can enter the nucleus, as fluorescence was observed in the nucleus 3 d after treatment.^[133] In studies of drug release from GQDs, however, the drug is shown to be present in the nuclei, while the GQDs are not. The timeframe of when these images are taken, is often in scale of hours, instead of days, which indicates that the release of the drugs occurs sooner than the proposed nuclear internalization of GQDs.

In addition to chemotherapeutic agents, GQDs can also be utilized to deliver other therapeutic agents that suffer from poor water solubility. A GQD-folic acid (FA) system was utilized to deliver IR780 iodide (GQD-FA-IR780), a photothermal therapy (PTT) agent, and FA was used as a targeting ligand to specifically target tumor cells. PTT agents convert irradiation to heat and eradicate tumors by raising the local temperatures; however, IR780 suffers from poor water solubility, presenting a barrier to delivery of IR780 at effective dosages while minimizing toxicity. The basal plane of GQDs presents an ideal carrier for IR780, due to the π-π interaction between the sp² hybridized carbon in the GQDs and IR780. Furthermore, due to the nature of PTT, efficient targeted delivery is important in reducing side effects. The results showed that the GQD-FA-IR780 was able to completely eradicate tumors in mouse models through irradiation with an 808 nm laser.^[134]

GQDs provide an effective delivery system for chemotherapeutic drugs, improving the delivery dosage with high drug loading; however, a more in-depth study of the release of chemotherapeutics from the basal plane of GQDs may render GQD-based drug delivery systems to be more viable in the future.

6.2. Gene Delivery

Gene therapy describes the technique of delivering nucleic acids to cells in order to treat diseases. Instead of delivering chemotherapeutics to inhibit cellular pathways and induce cell death of tumor tissue, gene therapy provides a new method of treatment by delivering DNA or RNA that can replace a mutated gene, knockout an improperly functioning gene, or express new gene products to combat disease.^[135–137] Similar to the requirements for drug delivery systems, effective carriers in gene therapy are non-toxic and possess the ability to bypass biological barriers to ensure delivery of the nucleic acid payloads into cell cytosols and nuclei. While viral vectors are effective in gene delivery due to their natural ability to invade and deliver their genetic material, the barrier of immunogenicity induced by viral vectors has kept them far from safe for clinical use.^[138] As the traits of an effective gene carrier have been observed in GQDs, efforts have been aimed at developing GQDs as non-viral vectors in gene therapy. GQD-based gene nano-carriers were developed by forming a complex of GQDs, chimeric peptides, and plasmid DNA (pDNA). The chimeric peptide (MPG-2H1) contained a positively charged motif for complexation with pDNA, a ‘fusion peptide’ motif for disruption of the endosomal membrane and accelerated endosomal escape, and a nuclear localizing signal motif. The complex was formed through a non-covalent interaction. At an optimized ratio of the components of the complex, the exhibited transfection efficiency was eightfold compared to that of just the peptide-pDNA complex. In addition to enhancing the transfection efficiency, the cellular uptake of the GQD-peptide-pDNA complex could be observed through confocal microscopy due to the PL of the GQDs.^[139] This study suggests that GQDs are a promising transfection vector that could lead to more in vitro and in vivo applications of GQD-based gene therapy. Compared to the use of GQDs as drug delivery agents, however, the utility of using GQDs over other non-viral vectors seems less apparent. Due to their sp² hybridized structure and π-stacking, the GQD systems allow higher drug loading than that observed in other nanoparticle-based drug delivery systems. On the other hand, the sites for conjugation of signaling molecules and a cationic polymer for DNA loading seem to be restricted to the edge functional groups of GQDs, as these molecules need to be covalently bound onto the edge groups. The basal planes that were highly utilized in drug delivery systems do not provide any obvious advantage in gene therapy. While the ability to monitor gene delivery using the PL of GQDs is useful, innovative research in utilizing the basal plane in gene delivery would propel GQDs as a viable non-viral vector for gene therapy.

6.3. Photodynamic Therapy

In PDT, tumor cells are killed by ROS produced by GQDs under irradiation and in the presence of oxygen. The ROS production of GQDs under irradiation may be a cause for concern in practical applications; on the other hand, it presents an opportunity to use GQDs as PDT agents by enhancing its ability to kill tumor cells. Compared to current treatment methods, such as chemo- and radio-therapy, PDT is noninvasive and has fewer side effects. GQDs present a new class of photosensitizers for PDT that provides high water dispersibility and photostability compared to current PDT agents such as porphyrins, and low toxicity and biocompatibility compared to semiconductor QD alternatives.

GQDs can be used as imaging probes, as well as PDT agents. The ¹O₂ QY produced by GQDs under irradiation was found to be 1.3, the highest efficiency reported for any PDT agent. The GQDs also demonstrated superior photostability compared to Ppix, a clinically approved photosensitizer, and CdTe, a semiconducting QD. (Figure 16a) GQDs showed similar levels of cytotoxicity as Ppix under irradiation, but were also shown to be non-toxic without irradiation, while Ppix demonstrated some toxicity. To test the in vivo imaging capabilities of the GQDs, an aqueous solution of GQDs were injected into the back of a mouse. The fluorescence images clearly show a greater PL signal around the injection site compared to the rest of the body (Figure15 b,c). In an in vivo PDT study, mice with subcutaneous breast cancer xenografts were split into three groups which were treated with PDT, GQDs only, and irradiation only. The tumors in the PDT group were destroyed in 17 days, whereas the tumors in the other two groups continued to grow, suggesting the efficiency of GQDs as photosensitizers in PDT (Figure 16d,e).^[140]

Figure 16.

GQDs for PDT. (a) Photostability of GQDs, Ppix, and CdTe measured as a ratio of absorbance at 470 nm over time after irradiation using a 500 W xenon lamp. (b) Bright-field and (c) red-fluorescence images of a mouse after subcutaneous injection of GQDs. (d) Tumor volumes over time in the three treatment groups (n = 5 per group). P < 0.05 for each group. PDT: GQD + irradiation, C1: GQD only, C2: light irradiation only) (e) Photographs of mice after various treatments (the integers following the treatment labels indicate days after first treatment). Reproduced with permission.^[140] Copyright 2014, Springer Nature.

More recently, multi-functional GQD systems for combined photothermal therapy and PDT were reported. Using organic matter derived from mechanical shearing of plant leaves, GQDs were synthesized via a hydrothermal reaction. The resulting GQDs were reported to have photothermal capabilities, raising temperatures to 43°C at 100 μg mL⁻¹ after 5 min of irradiation with an 808 nm laser. ROS generation was also reported to occur under 808 nm irradiation in MDA-MB-231 cells. The viability of the cells was lower in the cells treated with GQDs and the 808 nm irradiation compared to the cells treated with GQDs alone. The effects of PDT and PTT could not be isolated due to the processes occurring under the same irradiation, as reported.^[141]

GQDs have been demonstrated as potential PDT agents in in vitro and in vivo experiments. Under irradiation, GQDs were effective in reducing tumor volumes, as well as inducing cell death, while without irradiation, they exhibited lower cytotoxicity compared to conventional PDT agents. Furthermore, traditional PDT agents require irradiation in the UV/Vis range, which limits tissue penetration, and consequently, the extent of PDT; GQDs can be irradiated at longer wavelengths to address this limitation and are shown to be more efficient PDT agents. The photoinduced cytotoxicity of GQDs from ROS production opens a new avenue for cancer therapy.

7. Conclusion and Future Perspectives

Recent advances in GQD research have shown the potential of GQDs as novel platforms in biotechnology and nanomedicine. Methods for doping and functionalizing GQDs have been investigated through novel synthesis methods, and the consequent effect on optical, electronic, magnetic, and biological properties have been leveraged to apply GQDs in medical applications. Novel experimental methods have been developed to optimize the physicochemical characteristics of GQDs so that they express the necessary properties for a specific application. Furthermore, various studies have focused on developing safe and straightforward ‘one-pot’ synthesis methods, utilizing a few simple precursor molecules and readily available equipment such as a household microwave. These accomplishments present GQDs as accessible and suitable systems for applications in nanomedicine.

While initial research of GQDs as bioimaging probes has focused on fluorescence imaging, current efforts have been aimed at developing alternative modes of imaging, such as MR imaging in combination with fluorescence imaging and tuning the PL window of GQDs to be used in NIR imaging. These new developments have allowed GQDs to be used in in vivo experiments due to their non-invasive nature and high tissue depth penetration. In addition to bioimaging, the electronic structure of GQDs can be modified through chemical alteration which allows them to be used in biosensing applications. In the presence of specific analytes found in cellular environments, the PL of GQDs can be either quenched or recovered, depending on the design of the system, and thereby act as an ‘on-off’ switch biosensor. GQDs have also been studied as a platform for therapeutic applications, mostly in cancer research. The presence of functional groups as well as π-orbitals allow high drug loading and conjugation of cationic polymers for DNA complexation in gene therapy applications. Finally, ROS production by GQDs under irradiation allows them to serve as good photosensitizers for PDT. GQDs are particularly useful in this application, as their inherent PL properties allows the in vivo tracking of therapeutic payloads. Furthermore, the different modes of therapy facilitated by GQDs could be used in conjunction with one another, leading to a more effective concurrent treatment.

The future of GQD-based biological research seems boundless, as it is a relatively new class of material. The amount of reports involving GQDs has increased acceleratingly every year. However, there are some identifiable barriers to the ascent of GQDs to the forefront of nanotechnology in medicine. The lack of accurate characterization methods for GQDs has led to discrepancies in physicochemical and optical properties reported in literature and in proposed mechanisms of altering PL properties amongst various reports. Strict codification of the different classes of carbon-based nanomaterials would help to eliminate some of the existing imprecision in vocabulary used in discussing GQD research. More clearly delineated definitions of the categories of carbonaceous QDs will help facilitate more accurate theoretical models and more easily applicable empirical results described in the literature. Reconciliation of calculations of electronic properties generated from theoretical models with experimental data would also further our understanding of the properties of GQDs. In addition, conflicting claims with respect to several mechanisms involving GQDs have been reported, such as the mechanism of ROS production and release of chemotherapeutic drugs bound to the basal plane. A more in-depth approach to understanding these mechanisms will allow these unique properties of GQDs to be further harnessed and lead to more efficient GQD-based therapy.

The two-dimensional structure of GQDs is unique compared to other nanoparticle systems which are three-dimensional, and often spherical. As mentioned, the capability to complex with therapeutic agents within the basal plane greatly enhances drug loading compared to spherical nanoparticles. In addition to their superior drug loading, the innate optical and electronic properties of GQDs allow for concurrent monitoring of therapeutic delivery, as well as controlled release through the interaction between GQDs and incident light. The possibilities of using GQDs as a platform for biological imaging, sensing, and therapy is vast. While short term in vitro toxicity tests are far from enough to justify GQDs as safe in biological systems, the initial reports of low in vivo toxicity and low residual accumulation in mouse models are indeed encouraging. More extensive and long-term toxicity studies will shine more light on the unique properties of GQDs in biological research.

Biographies

Seokhwan Chung earned his B.S. in materials and science engineering from the University of Illinois at Urbana-Champaign. He is currently pursuing his Ph.D. degree in materials science and engineering at the University of Washington. His current research interest is in synthesis and application of iron oxide and carbon-based nanomaterials for cancer therapy and diagnostics.

Richard A. Revia earned his B.S. in biomedical engineering from the University of Texas at Austin in 2008 and his M.S. in electrical engineering from Montana State University in 2013. He is currently pursuing his Ph.D. degree in materials science and engineering at the University of Washington. His research interests include the development of nanoparticle platforms as contrast agents in magnetic resonance imaging and designing devices for use in tissue engineering.

Miqin Zhang is Kyocera Chair Professor of Materials Science & Engineering and Neurological Surgery, and adjunct professor in the Departments of Bioengineering, Radiology, and Orthopaedics & Sports Medicine at the University of Washington (UW). She joined the faculty of the UW in 1999 after receiving her Ph.D. from University of California at Berkley. Her research focuses on nanomedicine for cancer diagnosis and treatment, biomaterials for tissue engineering, and biosensors for detection of chemical and biological agents.