Biomolecule detection and quantification have become increasingly important in recent years, due to advancements in clinical diagnosis, which requires newer technologies for rapid and accurate detection of molecules at ultratrace concentrations. Since the historic development of enzymatic electrodes by Clark and Lyons in 1962 [1], there has been a quest among researchers for advanced sensing technologies and this has resulted in the development of more sophisticated and trustworthy sensors ^[2][3][4][2,3,4]. Though several techniques have emerged, Raman spectroscopy has sparked the most interest in biomolecule sensing due to its exceptional sensitivity rendered by the large signal amplification, chemical specificity, rapid recognition and non-destructive nature. Raman spectroscopy identifies the characteristic molecular vibrations and provides the fingerprints of the molecules with minimal to no sample preparation. However, the weak signal, due to low scattering probability (typically 10⁻⁴–10⁻⁶), was a bottleneck when deploying this versatile technique in the detection of ultratrace target molecules, until the discovery of Surface-enhanced Raman scattering (SERS) by Fleischmann et al., in 1974 [5]. The observation of an enhanced Raman signal of pyridine on roughed Ag electrodes eventually helped Raman spectroscopy to extend its applications up to the detection of a single molecule ^[6][7][6,7].

The electromagnetic (EM) and chemical (CM) mechanisms are the two important phenomena behind the Raman signal enhancement, proposed later by Van Duyne and Creighton groups, independently, in 1977 ^[8][9][8,9]. The EM enhancement originates from the excitation of surface plasmon on nanoscale plasmonic surfaces, mainly noble metal nanoparticles (Au and Ag), which contributes dominantly (10³ to 10⁸ times) to the SERS enhancement. It is mainly determined by the material morphology, dielectric constant of the medium and the localization of surface plasmon resonance (LSPR) and their coupling ^{[10][11][12][13]}[10,11,12,13].

The EM mechanism does not explain about the SERS enhancement with non-plasmonic substrates, e.g., oxides, nitrides, chalcogenides, etc. This can be well understood by the formation of charge-transfer complex, and thus new electronic states, of chemisorbed molecules with the substrates [14]. The CM enhancement is mainly determined by the Fermi level of the substrates and the molecules. The contribution from CM is relatively weaker (up to 10³ times) than that of the EM effect. However, CM has comprehensive advantages over EM, including cost-effectiveness, surface uniformity, signal reproducibility, muted photo-bleaching and blinking effects. Further information about the mechanism of SERS can be found in the excellent book by Eric and Pablo [15]. Considerable advancements in understanding charge-transfer complex formation and designing structurally, chemically engineered substrates have been made in the past two decades for the detection of multi-fold trace chemicals and biomolecules, which includes RNA analysis from plant tissues and multiplexed detection at a single-cell level ^[16][17][18][16,17,18].

The discovery of graphene by Novoselov and Geim in 2004 [19] opened a new era in the material sciences, which leads to the further development of various two-dimensional nanostructures (2DNS), including transition-metal dichalcogenides (TMDs), oxides, graphitic carbon nitride (g-C₃N₄), hexagonal boron nitride (h-BN), black phosphorus (BP) and 2D transition-metal carbide or nitride (MXenes) ^[20][21][22][20,21,22]. Recently, nanosheets of metal organic framework (MOF) and covalent organic framework (COF) have also joined the fascinating world of two-dimensional nanostructures. Apart from easy synthesis, these 2DNS and their nanocomposites have several advantages in SERS because of their unique physical and chemical properties, such as high uniformity with large specific surface areas, better chemical stability, excellent mechanical and optical properties with fluorescence quenching capability, π-π* interaction with biomolecules and good biocompatibility ^[23][24][25][23,24,25].

2. 2DNS as SERS Substrates

As mentioned earlier, the Raman signal enhancement by 2DNS is mainly through a charge-transfer mechanism. The electronic structure of the analyte–substrate interface, which is primarily accomplished by the transfer of an electron from the highest occupied molecular orbital (HOMO) to the lowest unoccupied molecular orbital (LUMO), determines the contribution of CM (charge-transfer) to Raman signal amplification. Moreover, the π-interaction facilitates the accumulation of analytes on their surface, which has a significant effect at lower concentrations. On the other hand, 2DNS can anchor the plasmonic nanostructures for better dispersion, i.e., prevention of agglomeration. Here, the SERS enhancement factor (EF), the degree of signal amplification [15], is improved as the essential nano-gaps are created by the well-separated plasmonic nanostructures. Therefore, 2DNS were widely deployed for the later purpose. Figure 1 2 illustrates the use of 2DNS as a SERS substrate and support for nanostructured plasmonic SERS substrates. Table 1 lists representative examples of various 2DNS employed as SERS substrates and support for plasmonic NPs.

2.1. Graphene SERS (GERS) Substrates

Graphene is a single sheet of sp²-bonded carbon atoms in a hexagonal honeycomb lattice. It is the well-known and most explored two-dimensional allotrope of carbon with unusual electronic, optical properties, and high theoretical surface area ^[61][62][67,68]. The free π-electron, rich in graphene, can make π-interaction with other systems and accumulate on its surface. Consequently, the charge-transfer between the graphene substrate and the adsorbed molecules is enhanced to observe the SERS signal augmentation ^[63][69]. This phenomenon has been exploited in graphene-enhanced Raman scattering (GERS) for a wide range of applications, including materials development ^[64][70], energy ^[65][71] and biomedicine ^[66][67][72,73].

2.2. Nitrides SERS Substrates

The lone-pair electrons in nitrides have an advantage while using them as SERS substrates. A hexagonal lattice made up of boron and nitrogen atoms makes up the equivalent of graphene, known as hexagonal boron nitride (h-BN). Boron nitride possesses a dipole-coupled Raman amplification mechanism, according to a recent investigation ^[68][74]. Highly sensitive, label-free, and non-destructive biomolecule detection is achieved using h-BN nanostructures ^[69][75]. However, their wider band gap (~6 eV) requires high excitation energy for a conventional CM signal enhancement, which is not suitable for biological molecules ^[70][76]. Carbon nitrides are other important 2DNS for Raman signal amplification. Redemann et al. discovered in 1940 that graphitic carbon nitride (g-C₃N₄) possesses a graphite-like van der Waals layered structure ^[71][77]. Despite having good physicochemical stability, the poor signal enhancement from pristine g-C₃N₄ has hindered its use as an independent SERS substrate for sensing applications. However, chemical and structural (e.g., induced disorders to the heptazine chain) modifications may help improve the enhancement factor. Few compound nitride thin films have also been reported as SERS substrates due to their resonant plasmonic characteristics. For example, Shaoli et.al. have prepared titanium nitride (TiN), aluminium nitride (AlN) and titanium-aluminium nitride (TiAlN) thin film SERS substrates with 95% higher signal strength compared to bare glass substrate ^[72][78]. A highly stable niobium nitride thin film with good uniformity has been prepared by reduction nitridation that enhances the Raman signal of Rh6G by 4 × 10³ factor ^[73][79].

2.3. Black Phosphorous (BP) SERS Substrates

Bulk BP was first synthesized in 1914, however, an atomically thin BP 2DNS is realized just recently ^[74][80]. Compared to red and white phosphorus, BP is the most stable form of elemental phosphorus ^[75][81]. The 2D zig-zag structure of BP sheets consists of phosphorus atoms with three covalently bonded nearest neighbours, while the sheets are bound together by weak van der Waals forces. These layers can be easily exfoliated into 2D BP nanosheets, since the multi-level quantum chemical calculations indicate an exfoliation energy of −151 meV per atom ^[76][82]. These wrinkly sheets of honeycomb lattice have armchair and zigzag forms, as in graphene. The layer-dependent band gap, from 0.3 (bulk) to 2.0 eV, of BP allows the use of a wide range of excitation light in the UV, visible and NIR ranges for SERS analysis ^[77][83]. Interestingly, Lin et al. reported an anisotropic SERS substrate using few-layered BP and ReS₂, which exhibited polarization-dependent signal enhancement ^[78][84]. Therefore, BP nanosheets have recently received great attention for a wide range of applications, particularly in biomedicine, photothermal therapy, photodynamic therapy, drug administration, 3D printing, bio-imaging, and theranostics ^[79][80][81][85,86,87].

2.4. MXenes SERS Substrates

Transition metal carbides, nitrides or carbonitrides make a new class of 2D material, known as MXenes. They typically have a layered structure with (n + 1) layers of M connected by n layers of X in the pattern [MX]_n-M, where M is an early transition metal (such as Sc, Ti, Zr, Hf, V, Nb, Ta, Cr or Mo), and X is either carbon or nitrogen. A general formula for these compounds is M_n+1X_n (n = 1–3) ^[82][83][88,89]. Since its first discovery (Ti₃C₂) in 2011, MXenes have attracted immense attention in a variety of applications, including energy, environmental and healthcare sectors. The high electrical conductivity of highly metallic MXenes, having unique electronic and optical properties and intense LSPR effect in the visible or near-infrared range, makes them a promising SERS substrate ^[52][58]. Here, both EM and CM contribute to boosting the Raman signal ^[84][90]. Their flexibility and hydrophilic nature make functionalization or tagging with Raman reporters, easy.

2.5. Transition Metal Dichalcogenide (TMD) SERS Substrates

Compounds with the generalised formula MX₂, where M is a transition metal and X is a chalcogen, such as S, Se or Te, make up the family of layered materials known as “transition metal dichalcogenides”. Strong intralayer bonding and weak interlayer binding enable the exfoliation of these van der Waals solids into 2D nanosheets ^[85][91]. A layer of transition metal sandwiched between two saturated chalcogen layers makes these less reactive 2D TMD layers. The confinement of charge carriers in two dimensions in TMDs dramatically alters their characteristics for a variety of applications ^[86][87][30,92]. These atomically flat sheets enable effective charge transfer between the probe molecules and substrates through weak contacts, such as π–π* interactions, and make them suitable for chemical Raman signal enhancement ^[88][89][93,94]. These TMDs are particularly interesting since they facilitate attachment of probed molecules to induce the CM effect ^[90][95].

2.6. Metal Oxide SERS Substrates

Most semiconductors exhibit weak SERS signals due to their large band gaps and lack of surface plasmon resonance. Oxygen incorporation in semiconductors increases the Raman enhancement factor as good as 10⁵ times, probably due to the enhanced charge-transfer from the semiconductor band edges to the adsorbed molecules ^[91][96]. Metal oxide semiconductors, such as titanium oxide (TiO₂), tungsten oxide (WoOx) and molybdenum oxide (MoOx), were recently tested as SERS substrates ^[37][43]. The surface polarisation effect due to the oxygen defect states boosts the Raman signals in these substrates ^[92][97]. For instance, few-layered MoO₃ nanosheets act as a sensitive SERS substrate, which enhances the Raman signal up to 2.28 × 10⁴ times and makes it capable of detecting 2 × 10⁻⁸ M of an Rh6G molecule ^[93][98]. Similarly, ultrathin, chemical vapour-deposited MoO₂ nanosheets show enhancement of the Raman signal up to 2.1 × 10⁵ and possess excellent reusability and uniformity ^[94][99]. In both cases, it has been found that the EF further increased by decreasing the thickness of the MoO_x nanosheets.

2.7. 2D MOFs/COFs SERS Substrates

Metal-Organic Frameworks (MOFs) are crystalline porous materials consisting of metal ions or cluster nodes linked by organic ligands such as carboxylate ligands and other negatively charged ligands ^[95][96][100,101]. MOFs show excellent SERS performance that is generally attributed to the charge transfer enhancement mechanism ^[97][102]. Several studies have been carried out to deploy MOFs as SERS substrates. For the first time, Yu et al. reported the Raman signal enhancement of Methyl Orange adsorbed on Matériaux Institut Lavoisier (MIL)-type MOFs ^[98][103]. Later, several other MOFs, including ZIF-67, Co-TCPP MOFs and Co-MOF-74 were employed directly as SERS substrates, which shows an EF of about 10⁶ for an Rh6G molecule ^[99][104]. Covalent Organic Frameworks (COFs) are ordered structures built up from organic building blocks via covalent bonds ^[100][105]. The use of COFs as SERS substrate is still in its infancy, while MOFs gained more popularity because of the plasmonic hybrids. Two-dimensional allotropes of these MOFs and COFs are attracting increasing research attention due to their ultrathin morphology, which offers a high surface-to-volume atom ratio ^[95][100]. Their high surface area with molecular structure facilitating a π–π* interaction is a critical advantage for their application in SERS substrates.