Figure 1 . (A) Classification of CDs as GQDs, CQDs, or CPDs. (B) Schematic diagram of GQDs synthesis. (C) Schematic diagram of CQDs synthesis. (D) Schematic diagram of CPDs synthesis.
GQDs were first prepared by Ponomarenko et al.[35] in 2008 based on the previous report on CDs by Xu et al. [13] in 2004. GQDs can be regarded as a subset of graphene and/or graphene oxide with sp2structure, exhibiting similar chemical and physical properties to it. However, unlike two-dimensional graphene sheets with zero band gap, the GQDs is typically anisotropic characterized by single or less layers atomic graphite planes with lateral size typically <10 nm. The bandgap of GQDs is opened due to the quantum confinement effects caused by the small size, and is closely related to the size (mainly refers to the conjugated sp2 domain), which is the major feature of GQDs.[36] Density functional theory (DFT) calculations have shown that the bandgaps of GQDs decrease from around 7 eV for benzene to 2 eV for GQDs consisting of 20 fused aromatic rings.[37] Moreover, the photoluminescence (PL) of GQDs can be precisely regulated by control of the attached chemical functionalities, heteroatom doping, defects, edge configuration, and shape.[38, 39] Many regard GQDs as giant polyaromatic molecules.[40] Compared to simple polyaromatic molecules, GQDs are more complex in that they are larger, carry abundant functional groups, and have narrower and more complex band gap structures. The intense PL emissions of GQDs can be roughly divided into two primary types: PL originating from created or induced energy bandgaps in single sheet GQDs and PL that is associated with defects in single- and/or multiple-layer graphene. There are many different opinions about the number of layers in GQDs, with 1-3 layers,[41] <5 layers,[42] or <10 layers being advocated in different literature. In the discussion below, we mainly focus on GQDs with less than 5 layers (GQDs with more layers are classified as CQDs in the current work).
There are two types of synthetic routes towards GQDs: top-down and bottom-up methods (Figure 1B). The top-down method involves controlled fragmentation of bulk carbonaceous materials (e.g., graphite, graphene sheets, carbon nanofibers, and carbon nanotubes) into small pieces. The most commonly used methods include oxidative/reductive cutting,[39, 43] physical grinding,[44] or combination of cutting and grinding.[45] Top-down methods have the advantages of abundant raw materials, simple operation, and large scale production, but do not allow accurate control over the size and morphology.[46] The formation of GQDs by top-down methods does not involve the aggregation of precursors. Accordingly, the C:H ratio in GQDs prepared by top-down methods is typically near-unity, with C mainly existing in sp2 form. However, top-down preparation methods result in partial oxidation of some carbons, whilst also introducing certain functional groups (e.g. hydroxyl, carbonyl, carboxylic acid, and epoxy/ether groups) and/or doping of heteroatoms into the GQDs. These surface modifications give GQDs good dispersibility, with the introduced defects often acting as fluorescent centers into the GQDs.[47] π−π stacking interactions between adjacent layers in GQDs can cause interlayer quenching, similar to that seen for large conjugated aromatic hydrocarbons or conjugated fluorescent polymers. The effect of such “aggregation” on the optical properties of GQDs is explored in section 4.
Compared with the top-down methods, synthesis of GQDs by bottom-up methods is more chemically challenging, but allows the preparation of GQDs with more uniform morphologies and controlled size. Progressive reaction in solution is the dominant approach for the bottom-up synthesis of GQDs. In such bottom-up methods, aggregation is critical to the synthesis of GQDs, with the aggregation of precursors being a critical initial step. Such methods have more stringent precursor requirements, typically utilizing molecules with conjugated structures or molecules that combine to form six-membered rings, such as hexabromobenzene [48], hexa-peri-hexabenzocoronene,[49]1,3,6-trinitropyrene,[50] citric acid (CA),[51] glucose,[52] and so on. The reaction precursors react and aggregate into larger graphene-like sheets, with aggregation controlling the lateral size and thickness (number of layers). Most GQDs prepared by such routes are anisotropic. Liu et al. [49] used hexa-peri-hexabenzocoronene as precursor to prepare GQDs through a stepwise process of aggregation, carbonization, oxidization, surface functionalization, and reduction (Figure 2A). Wide-angle X-ray scattering show that the unit cell parameter of the GQDs was consistent with the length of the hexa-peri-hexabenzocoronene molecule, with a strong diffraction peak observed attributed to the π−π stacking distance between two overlapped molecules of hexa-peri-hexabenzocoronene. Moreover, by performing control experiments using hexa-peri-hexabenzocoronene derivatives with different functional groups, it was demonstrated that the condensed stacking of hexa-peri-hexabenzocoronene in the GQDs could be attributed to the lack of steric hindrance from the substituents, leading to the generation of graphitic framework with few defects during the carbonization step. Lee et al. [52] prepared GQDs through catalytic solution chemistry (Figure 2B). The D-glucose carbon source aggregated horizontally through spontaneous dehydration, resulting in the formation of graphene. Dong et al. [51] similarly developed the synthesis of monolayer GQDs using a CA precursor (Figure 2C). Yan et al. [53, 54] reported the preparation of GQDs with a tunable size through solution chemistry (Figure 2D) involving surface modification with 2, 4, 6-triakyl phenyl. This synthesis route allowed accurate control of the number of conjugated carbon atoms. Crowdedness at the edges of the graphene cores caused twisting of the phenyl groups of 2, 4, 6-triakyl phenyl away from the plane of the core, leading to alkyl chains projecting in all dimensions. This reduces the face-to-face interaction between the developing graphene sheets, allowing large GQDs to be synthesized with good solubility.
For all of the aforementioned GQDs, the band gap is closely related to the size (especially the size of the sp2 domain) and number of layers resulting from aggregation. Yeh et al.[55] showed that the band gap in GQDs decreased with increasing GQDs lateral size (Figure 2E). The PL color varied from red-orange to blue as the GQDs size was reduced (from 8 to 1 nm). Results are consistent with quantum confinement with decreasing size in the sp2 conjugate domains. Dong et al.[56] prepared GQDs by chemically oxidizing CX-72 carbon black (Figure 2F). This synthesis route produced monolayer and multi-layer GQDs simultaneously, with sizes of 15 nm and 18 nm, respectively. Monolayer and double-layer GQDs have strong green and yellow PL emissions, respectively. Yan et al. [57]proposed two different strategies to narrow down the electronic band gap of GQDs (Figure 2G). Since the band gap of pristine GQDs is negatively correlated with the size of the sp2 domain, bandgap narrowing was realized via conjugating GQDs with polyaromatic molecules to enlarge the sp2 carbon network. The sensitivity of GQDs to this type of modification is due to the anisotropy and small size of GQDs. This strategy may not work well for other types of CDs whose bandgaps are less sensitive to quantum confinement effects.[40]
2.2. CQDs
Fluorescent carbon nanoparticles were first named CQDs by Sun et al. in 2006 [58]. Like GQDs, CQDs also have quantum confinement and crystalline core structures. The main different between GQDs and CQDs is their shape and core composition. GQDs are anisotropic with crystalline sp2 cores, whereas CQDs are quasi-spherical carbon nanoparticles with less crystallinity and containing more defects. CQDs typically possess as crystalline core containing a mixture of sp2 and sp3carbons. Usually, CQDs are terminated by oxygenic/nitrogenous functional groups, with heteroatom contents ranging from 5-50 wt.%. The differences between GQDs and CQDs mainly originate from the different precursor aggregation pathways in their respective syntheses.