Detecting Cr⁶⁺ at ≈100 pM Concentration with Fluorescence Enhancement Signatures in a Novel Eco‐Fluorophore: Matching WHO's 96 pM Recommended Standard for Drinking Water

Abstract

Hexavalent chromium (Cr⁶⁺) ions in drinking water pose a significant risk to human health, being a leading cause for neurological disorders, organ damage, and infertility. This study introduces an ultrasensitive method for detecting trace Cr⁶⁺ over a wide concentration range (≈ 100 pM – 100 µM) through fluorescence enhancement signatures via integration of both covalent and non‐covalent interaction strategies on carbon quantum dots (CQD). The covalent functionalization is achieved from dual‐functionalized CQD (CQD‐(NH₂, COOH)) derived from coffee‐waste. Additionally, the covalent and non‐covalent approach integrates CQD‐(NH₂, COOH) with graphitic carbon nitride (g‐C₃N₄) to form a 2D/2D heterostructure. The synergy between CQD‐(NH₂, COOH) and g‐C₃N₄ introduces a mid‐gap band in their band structure, allowing multiple carrier excitation and recombination states, significantly enhancing the fluorescence quenching signal. This combination allows to achieve Cr⁶⁺ detection sensitivity down to ≈100 pM concentration—matching the World Health Organization's 96 pM permissible limit of total Cr in drinking water. Furthermore, a 70 pM detection limit is reported for Cr⁶⁺ in a mixture of twelve ions, including cations and anions, surpassing current state‐of‐the‐art detection limits. These results highlight the potential of dual covalent and non‐covalent modification strategy in nanomaterials to set new standards in ultrasensitive and wide‐range fluorescent sensing applications.

Through a dual‐functionalization strategy, carbon quantum dots (CQD) with exceptional fluorescence properties are engineered. These CQD are integrated with graphitic carbon nitride to form a 2D/2D heterostructure via both covalent and non‐covalent modification. This integration enables ultra‐trace Cr⁶⁺ detection with a limit of detection ≈70 pM—surpassing state‐of‐the‐art Cr⁶⁺ sensors.

1. Introduction

Fluorescence (FL) methods utilize the emission of light from a fluorophore, a molecule that absorbs light or other electromagnetic radiation, making electrons transition to a higher energy state with their subsequent return to ground state, emitting photons in the process. The characteristic FL signal of the emitted photons corresponds to the energy difference between the jump states.^{[ 1 ]} This spontaneous light emission phenomenon is widely used in characterizing biomolecular interactions through FL imaging, biomolecular conformational studies,^{[ 2 ]} environmental monitoring, and chemical sensing providing real‐time data. To push the detection limits of FL methods, engineering fluorophore materials with enhanced FL signals over a broad wavelength range are commonly pursued. Achieving higher sensitivity with an expanded dynamic range, improved signal‐to‐noise ratio, and lower limits of detection (LOD) has been approached through covalent strategies (such as surface modification and doping) and non‐covalent methods (including encapsulation, layer‐by‐layer assembly, and heterostructure formation^{[ 3 ]} to enhance FL behavior). Typically, the electronic transitions corresponding to FL are restricted to fast single band transitions giving sharp narrowband responses respective to the allowable band transition. Here, we integrate both covalent and non‐covalent interaction strategies in an engineered fluorophore to achieve higher sensitivity over a broad dynamic range. The combination of the covalent and non‐covalent modification strategies introduces a mid‐gap band (MGB) allowing multi‐level jump transitions enhancing the FL signature.^{[ 4 ]} The FL intensity increases with band broadening. We utilize this enhancement in a FL quenching measurement approach to detect trace amounts of hexavalent chromium (Cr⁶⁺) ions in complex mixtures of twelve ions including cations and anions.

Chromium (Cr), a versatile industrial heavy metal, is extensively used in industries such as steel production, metal coating, dyeing, leather tanning, and petroleum catalyst production. The discharge of various Cr compounds into the environment poses significant risks to biological and ecological systems^{[ 5 ]} worldwide. The trivalent chromium (Cr³⁺) and Cr⁶⁺ oxidation states are the most prevalent due to their stability. While Cr³⁺ is relatively harmless and plays an essential biological role with a daily human requirement of 50–200 µg, Cr⁶⁺ is highly toxic due to its strong oxidation potential. Cr⁶⁺ can easily penetrate cells causing damage^{[ 6} , ^{7 ]} and is considered an endocrine‐disrupting compound, leading to fertility issues etc.^{[ 8 ]} The World Health Organization (WHO) cites a provisional guideline of 0.096 nanomolar (nM) of total Cr in drinking water and the environmental protection agencies in different countries specifies 0.481 micromolar (µM) to 0.962 µM regulation for total Cr in drinking water.^{[ 6} , ^{7 ]} Therefore, it's crucial to distinguish Cr⁶⁺ from Cr³⁺ in our drinking water with high accuracy meeting the regulation standards. A variety of nanomaterials have been explored as potential fluorophore materials to achieve WHO's proposed concentration limit. But such fluorophores that include metals, metal organic frameworks, and metal composites like copper, gold‐doped copper nanoclusters,^{[ 9 ]} polyaniline/silver nanoparticles, and graphene oxide nanocomposites^{[ 10 ]}significantly fall short in achieved sensitivity. In addition, environmental toxicity,^{[ 11 ]} inherent instability,^{[ 12 ]} and high production costs associated with these fluorophore materials cause serious concerns. As a result, the pursuit of non‐metallic fluorophores that can achieve higher FL sensitivity meeting WHO's standards become paramount in advancing this field, offering a pathway for more sustainable and effective detection solutions.

Carbon quantum dots (CQD) are non‐metallic, carbon‐based nanomaterials smaller than 10 nm that have gained significant attention due to their unique properties, such as biocompatibility, eco‐friendliness, and tunable FL range. Notably, their nontoxicity, cost‐effective synthesis, and biological safety make them particularly appealing for Cr⁶⁺ sensing.^{[ 4 ]} In contrast to other toxic quantum dot semiconductors, CQD stand out as a promising alternative.^{[ 13 ]} However, unmodified CQD encounter problems including nonspecific interactions and insufficient quantum yield, which hinders FL sensitivity. We take a novel strategy to apply both covalent and non‐covalent interaction modalities to CQD. In the covalent strategy, we modify its surface by adding carboxyl and amino groups.^{[ 3 ]} We introduce a streamlined approach to synthesize dual‐functionalized CQD (CQD‐(NH₂, COOH)) directly from coffee waste through a single‐step ultrasonic synthesis process. Spent coffee grounds, which contain about 50% (hemi)‐cellulose and a substantial amount of carbon, provide consistent properties for CQD, regardless of the extraction or roasting methods.^{[ 14 ]} For the incorporation of covalent and non‐covalent strategies, we create a heterostructure combining two‐dimensional (2D) CQD‐(NH₂, COOH) and 2D graphitic carbon nitride (g‐C₃N₄) into 2D/2D CQD‐(NH₂, COOH)/g‐C₃N₄. The 2D g‐C₃N₄ is known for its optical properties in visible range, eco‐friendliness, high moisture resilience, and chemical stability. The CQD‐(NH₂, COOH) is decorated onto the g‐C₃N₄ structure using a single step ultrasound method. We compare the effects of covalency strategies and confirm that the synergic effect of covalent and non‐covalent modifications in the 2D/2D heterostructure furnish ultra‐sensitivity with a LOD of ≈70 picomolar (pM) over a broad sensing range from 0.1 nM to 100 µM. This allows us to demonstrate a highly selective portable sensor, selectively identifying our target Cr⁶⁺ ion in mixtures of Cr³⁺, Fe³⁺, Co²⁺, Al³⁺, Mo⁶⁺, Zn²⁺, and Cu²⁺ cations, SO₄ ²⁻, NO₃ ⁻, Cl⁻, and F⁻ anions that commonly end up in drinking water from industrial plants along with Cr⁶⁺.

2. Results and Discussion

The covalent modification strategy adopts functionalization of the CQD with amino (NH₂) and carboxyl (COOH) groups in an in‐situ synthesis process, where the NH₂ from amino acids and COOH from chlorogenic acids come from the coffee and the coffe‐waste.^{[ 15 ]} The synthesis scheme of CQD‐(NH₂, COOH) is illustrated in Figure 1a(i). The covalent and non‐covalent modification strategies are implemented by decorating the 2D g‐C₃N₄ nanosheets with a tri‐s‐triazine structure (Figure 1a(ii)) in a hydrothermal process, resulting in a heterostructure with the CQD‐(NH₂, COOH) present in the solution forming 2D/2D CQD‐(NH₂, COOH)/g‐C₃N₄ as depicted in Figure 1a(iii). Hydrothermal treatment promotes the formation of covalent bonds between nitrogen (N) atoms in g‐C₃N₄ and the NH₂ on the CQD as highlighted in Figure 1a(iii), which we later confirm with material characterizations. Furthermore, the functional groups on the CQD, including the COOH and amino NH₂ groups, form hydrogen bonds (H‐bond), with their hydrogen atoms acting as donors and the electronegative atoms, such as oxygen or nitrogen, serving as acceptors. Simultaneously, the N atoms in g‐C₃N₄ can also act as H‐bond acceptors (Figure 1a(iii)). The introduction of the covalent and H‐bond interactions is facilitated by hydrothermal treatment, which further should be conducive to influencing the band structure of the synthesized material. Our goal is to induce a modified band structure that can significantly enhance carrier distribution between the two materials, ultimately improving the stability of the resulting 2D/2D heterostructure and boosting its overall performance in FL applications by optimizing transport properties.

Figure 1.

a) 2D chemical structure of i: CQD‐(NH₂, COOH), ii: g‐C₃N₄ and, iii: CQD‐(NH₂, COOH)/g‐C₃N₄, b) FL emission spectra of CQD‐(NH₂, COOH) (navy), CQD‐(NH₂, COOH)/g‐C₃N₄ (red) along with the excitation laser spectrum (violet, λ = 405 nm) at 298.15 K and pH = 7, c) Left: Schematic illustration of CQD‐(NH₂, COOH) decorated on g‐C₃N₄, Right: Charge transfer behavior before and after decoration.

In Figure 1b, the FL spectrum of CQD‐(NH₂, COOH) displays four distinct peaks at 574 nm, 595 nm, 618 nm, and 654 nm under 405 nm excitation, each corresponding to different emissive centers or states within the CQD‐(NH₂, COOH).^{[ 16 ]} The emissions at 574 nm and 595 nm, representing the yellow portion of spectrum, are likely attributed to ‐NH₂ surface functional groups or defects in the CQD structure.^{[ 17 ]} We may further surmise that the 618 nm peak, observed in the spectral orange‐red region, arises from quantum confinement effect,^{[ 18 ]} originating from our particle size distribution and its surface morphology (Figure S1a,b, Supporting Information, showing atomic force microscopy (AFM) characterization results) leading to a larger bandgap from restricted charge carrier movement in the modified band structure. In general, this peak may be associated with the size‐dependent properties or quantum confinement within the CQD‐(NH₂, COOH). The 654 nm peak, indicative of long‐wavelength rose‐red FL emission, further confirms the quantum confinement effect.^{[ 19 ]} Additionally, this spectral feature may be influenced by graphitization and the functionalization of surface carboxyl groups, which affect the electronic properties, including the bandgap.^{[ 16} , ^{20 ]} Surface functionalization, particularly with carboxyl groups, is expected to modify surface states, thereby altering the electronic and emissive properties of CQD‐(NH₂, COOH). The dominant role of surface functionalization is further supported by the Fourier‐transform infrared spectroscopy (FTIR) and X‐ray diffraction (XRD) data shown in Figure S2a,b (Supporting Information), confirming the presence of dual functionalization. Further evidence supporting the alteration of the electronic and emissive properties is discussed in detail later in the manuscript. The FL spectrum of the CQD‐(NH₂, COOH)/g‐C₃N₄ 2D/2D structure (Figure 1b) demonstrates a broader emission range, extending from 430 to 800 nm, with distinct peaks at 626 nm, 627 nm, 654 nm, and 716 nm. Additionally, a higher FL intensity (approximately 2 times) is observed in the CQD‐(NH₂, COOH)/g‐C₃N₄ heterostructure compared to that of CQD‐(NH₂, COOH) alone. The observed enhancement in FL intensity, along with the spectral broadening, can be attributed to improved carrier excitation and recombination processes enabled by the formation of the heterostructure. Heterostructuring promotes greater charge separation or excitation to MGB that emerges from these 2D/2D modifications. Specifically, the heterostructure modifies the donor‐to‐acceptor center ratio, thereby facilitating more efficient energy transfer.^{[ 21} , ^{22 ]} Additionally, the difference in bandgap and electronic structure between CQD‐(NH₂, COOH) and g‐C₃N₄ further enhances the carrier repartitioning in the modified band structure, ultimately influencing both the FL intensity and emission spectrum.^{[ 4} , ^{21 ]} The inherent dynamic nature of the carrier repartitioning in the modified energy bands gives the distinctive enhancement in FL intensity and broadening^{[ 22 ]} that we observe in our material. The enhanced charge separation, driven by energy state population–depopulation dynamics, plays a pivotal role in the energy transfer processes between donor and acceptor species, significantly enhancing the material's sensitivity.

Moreover, the observed peaks in the heterostructure exhibit a red‐shift compared to the CQD‐(NH₂, COOH) alone. This shift could be attributed to a reduction in quantum confinement^{[ 18 ]} within the CQD‐(NH₂, COOH) due to their interactions with g‐C₃N₄. When the effects of quantum confinement are diminished by the heterostructure formation (schematic in Figure 1c), the emissions shift to longer wavelengths^{[ 23 ]} from that of shorter wavelength emissions of CQD‐(NH₂, COOH) that are of higher energy.

Additionally, the formation of the heterostructure may lead to the creation of new emissive states (mid‐gap states) at the interface between CQD‐(NH₂, COOH) and g‐C₃N₄. This induces a characteristic change in electronic transport at the interface of these two materials resulting from a dynamic modulation of built‐in electronic field (Figure 1c) at the material junctions of the 2D structures. These new states, which likely require lower energy for FL emission than the original states in CQD‐(NH₂, COOH) and g‐C₃N₄, could further contribute to the observed red shift in the FL emission.^{[ 24 ]} The addition of the MGB facilitates higher probability of carrier transport via multi‐energy transitions close to the central excitation mode across the pseudo P‐N junction at the 2D/2D interface enhancing the overall FL intensity and bandwidth.^{[ 4 ]}

Figure 2a(i) shows CQD‐(NH₂, COOH) solution in a cuvette, and the corresponding transmission electron microscopy (TEM) image in Figure 2a(ii) clearly revealing the stacked edges of the particles. Their flat, thin structure provides a high surface‐to‐volume ratio, amplifying surface interaction effects with other materials.^{[ 25 ]} This stacked, multi‐layered architecture offers a platform for improved functionalization^{[ 26 ]} to achieve a highly sensitive FL material. Selected area electron diffraction (SAED) result (Figure S3a, Supporting Information) highlights the d‐spacing of 0.27 nm as shown in Figure 2a(ii), indicating the presence of the (100) plane, likely due to the graphitic core of the CQD.^{[ 27 ]} Figure 2b displays the layered structure of the 2D g‐C₃N₄ sheets, with a calculated d‐spacing of 0.318 nm (SAED in Figure S3b, Supporting Information), in agreement with literature reports for 2D g‐C₃N₄.^{[ 28 ]} Figure 2c(i) shows CQD‐(NH₂, COOH)/g‐C₃N₄ solution in a cuvette and Figure 2c(ii) presents the TEM image of the CQD‐(NH₂, COOH) decorated onto g‐C₃N₄ forming our prepared 2D/2D heterostructure. Notably, upon decorating CQD‐(NH₂, COOH) onto g‐C₃N₄, the d‐spacing at the interface increases to 0.375 nm (SAED in Figure S3c, Supporting Information), exceeding the d‐spacings of both CQD‐(NH₂, COOH) and g‐C₃N₄. Furthermore, the size of each g‐C₃N₄ sheet is estimated to be between 500 nm and 1 µm, as observed in both Figure 2b and the multiple flake structures in Figure 2c(ii).

Figure 2.

a) (i): Optical image of CQD‐(NH₂, COOH) solution in a cuvette, (ii): TEM image of CQD‐(NH₂, COOH), inset: zoom‐in TEM image of CQD‐(NH₂, COOH), b) TEM image of g‐C₃N₄, c) (i): Optical image of a CQD‐(NH₂, COOH)/g‐C₃N₄ solution in a cuvette, (ii): TEM image of CQD‐(NH₂, COOH)/g‐C₃N₄ inset: higher resolution image of the CQD‐(NH₂, COOH)/g‐C₃N₄, d) UV‐Vis absorption spectrum of CQD‐(NH₂, COOH) (navy) and CQD‐(NH₂, COOH)/g‐C₃N₄ (red), respectively. e) Tauc plot for direct bandgap calculation of CQD‐(NH₂, COOH) (navy) and CQD‐(NH₂, COOH)/g‐C₃N₄ (red), respectively.

This increase in d‐spacing can be attributed to several factors. One potential reason is interfacial interactions, which may arise from the formation of new covalent or H‐bonds (Figure 1a(iii)) or electrostatic interactions at the interface, leading to lattice distortion.^{[ 29 ]} Additionally, modifications to the electronic properties could influence the bandgap, the electronic distribution, and the overall reactivity of the material.^{[ 29 ]} In other words, by making a heterostructure, we effectively altered the d‐spacing. This means these quantum dots can exhibit pronounced quantum confinement effects within the introduced dimensional constraints.^{[ 30 ]} This could be verified from the Ultraviolet‐Visible (UV‐Vis) spectra and the Tauc plots in Figure 2d,e, respectively. UV‐Vis absorption spectrum of CQD‐(NH₂, COOH) shows two distinct peaks at 230 nm and 360 nm. The peak at 230 nm is typically associated to the π→π* transition of the C = C bond in the aromatic domains^{[ 31 ]} suggesting the presence of graphitic or sp² ‐hybridized carbon domains in the CQD. The presence of such a peak confirms the conjugated nature of the CQD, which is a characteristic feature of their electronic structure and is related to the extended π‐system of the carbon core. The peak at 360 nm is a more interesting feature. It is generally understood that it originates from n→π* transition associated with C = O or carboxyl groups.^{[ 32 ]} In CQD‐(NH₂, COOH)/g‐C₃N₄ structure, the peaks slightly shift to 225 nm and 330 nm, respectively, which we can generally attribute to the modified electronic interactions. The 225 nm peak is linked to π→π* transition of the C = C bond that is present in both the CQD‐(NH₂, COOH) and g‐C₃N₄.^{[ 31 ]} The 360 nm peak is associated with n‐π* transitions, which might be linked to either of the functional groups containing lone pairs, such as oxygen or nitrogen atoms on the surface of the CQD‐(NH₂, COOH)^{[ 32 ]} or the lone pair of nitrogen atoms in g‐C₃N₄.^{[ 33 ]} Moreover, the 2D/2D heterostructure adds an intriguing 438 nm peak like a signature that can be attributed to the surface states or defects^{[ 4 ]} in the structure originating from intraband energy levels, interfacial charge transfer states, or mid‐gap states^{[ 34 ]} facilitated by covalent and noncovalent strategy of our CQD modification. Therefore, instead of a simple broad tail (Figure 2d (navy)), the excitation shows additional gradient features increasing the selectivity signature (Figure 2d (red)). This suggests a diversification of excitation pathways, with the emergence of new routes replacing the continuation of the prior broad distribution.^{[ 35} , ^{36 ]}

Figure 2e shows the Tauc plots for both CQD‐(NH₂, COOH) and CQD‐(NH₂, COOH)/g‐C₃N₄. A direct bandgap (E_g) of 2.79 eV is estimated for CQD‐(NH₂, COOH) (refer to Equation S1, Supporting Information). This is significantly lower than the photon energy of 3.47 eV corresponding to the 360 nm peak (calculation shown in Equation S2, Supporting Information). Such an anomaly suggests that our observed 360 nm peak isn't purely a result of a direct band‐to‐band transition. Instead, it could be indicative of the role of the functional groups on the surface of the CQD or potentially due to quantum confinement effects inherent to the dimensional constraints we augment in our prepared quantum dot structures.^{[ 37 ]} Observation of this peak in our results confirms that the CQD have been either surface‐modified or possess intrinsic defects that bring about such transitions. Such spectral features can have implications on the FL behavior of the CQD‐(NH₂, COOH) with potential broad range of applications in photodetection^{[ 4 ]} and sensing. The Tauc plot of CQD‐(NH₂, COOH)/g‐C₃N₄ in Figure 2e (bottom panel) shows the presence of a dual direct bandgap. The first direct bandgap (E_g1) is 2.82 eV, and the second direct bandgap (E_g2) is 3.23 eV. E_g1 is close to the E_g of CQD‐(NH₂, COOH) (top panel). The slight increase could be due to the quantum confinement effect or interface interaction because of heterostructuring.^{[ 4 ]} The appearance of E_g2 might also be due to the heterostructuring that can result in modification of energy levels such as MGB formation, which has been reported before.^{[ 4} , ^{38 ]}

The alterations in the electronic and emissive properties and formation of a MGB in CQD‐(NH₂, COOH)/g‐C₃N₄ are supported by a combination of X‐ray photoelectron spectroscopy (XPS), structural, and optical analyses. In the XPS results, Auger shifts in the C KLL and O KLL regions (Figure 3a) confirm charge transfer and interfacial bonding effects.^{[ 39 ]} Additionally, binding energy shifts, peak broadening, and the appearance of new low‐energy components in the O 1s, N 1s, and C 1s spectra (Figure 3b–g) indicate significant charge redistribution and the formation of interfacial dipoles.^{[ 34 ]} Moreover, an approximation of the atomic percentages can be found in Table S1 (Supporting Information). In summary, structural characterization by TEM (Figure 2a(ii), 2b, and 2c(ii)) and SAED (Figure S3a–c, Supporting Information) reveals an increased lattice spacing (0.375 nm) and enhanced interfacial disorder, both of which are known to induce localized states. The XPS, TEM, and SAED results, together with the UV‐Vis spectra (Figure 2d) and Tauc analysis discussed above (Figure 2e), collectively indicate the emergence of MGB. This may be attributed to synergistic effects such as orbital hybridization, interfacial strain, defect generation, and charge imbalance, all contributing to the observed band structure modification. Moreover, covalent and non‐covalent modification using g‐C₃N₄ leads to a more uniform charge distribution, resulting in enhanced colloidal stability and a homogeneous zeta potential profile (Figure S4, Supporting Information). The detailed analysis of the XPS and zeta potential responses is provided in the Supporting Information.

Figure 3.

a) XPS survey spectra of (i): CQD‐(NH₂, COOH), (ii) CQD‐(NH₂, COOH)/g‐C₃N₄, High‐resolution XPS spectra for CQD‐(NH₂, COOH) b) O 1s, c) N 1s, d) C 1s, High‐resolution XPS spectra for CQD‐(NH₂, COOH)/g‐C₃N₄, e) O 1s, f) N 1s, g) C 1s.

The observed changes in bandgap—particularly the formation of MGB—along with variations in d‐spacing within the synthesized structures, significantly influence the electronic and emissive properties, including FL behavior.^{[ 40 ]} These modifications, evident in Figure 1b, are leveraged in our FL quenching‐based sensing methodology.

2.1. Fluorescence Quenching of Cr⁶⁺

To verify and demonstrate the efficacy of the enhanced FL material we synthesize, we test our materials with varying concentrations of Cr⁶⁺. For the CQD‐(NH₂, COOH) we observe a gradual decrease in the intensity of the prominent 595 nm peak for varying concentrations of Cr⁶⁺ ranging from 4.6 nM to 6.0 µM (Figure 4a). The observed quenching behavior could be due to the formation of non‐fluorescent complexes^{[ 41 ]} between Cr⁶⁺ and the CQD‐(NH₂, COOH) or be a result of dynamic collisions^{[ 42 ]} of Cr⁶⁺ with CQD‐(NH₂, COOH) that we will discuss as two primary mechanisms. Impressively, we determine the LOD for this material as low as 360 pM (Figure 4a), highlighting its high sensitivity towards Cr⁶⁺ even at extremely low concentrations. Details on the LOD estimation can be found in Equation S3 (Supporting Information). This superior FL quenching performance emphasizes the potential of CQD‐(NH₂, COOH) as an ultra‐sensitive sensing material, outperforming many existing CQD‐based systems for Cr⁶⁺ detection in environmental applications (Table 1 ). We must note here that the covalent modification strategy through surface functionalization significantly enhances the sensitivity not only in terms of LOD but also allows us to expand the range of detectable concentrations for Cr⁶⁺ as reported. A similar experiment is conducted with CQD‐(NH₂, COOH)/g‐C₃N₄, where an even broader range of Cr⁶⁺ concentrations (from 0.1 nM to 100 µM) is tested and the quenching of the main peak at 627 nm investigated (Figure 4b). For CQD‐(NH₂, COOH)/g‐C₃N₄, we estimate an even lower LOD of 70 pM. We thus demonstrate the ultra‐sensitive FL sensing capability of our 2D/2D heterostructure with a broad detection range and matching the WHO standard. The dual enhancements—functionalization of CQD and the formation of the CQD‐(NH₂, COOH)/g‐C₃N₄ heterostructure—enable us to achieve this sensitivity surpassing existing results in literature, summarized in Table 1. This represents a significant advancement over the state‐of‐the‐art FL sensing technologies, particularly in relation to the WHO's permissible limit of 96 pM^{[ 43 ]} for total Cr in drinking water. Our lowest measured concentration of 100 pM for Cr⁶⁺ aligns with the WHO limit, with potential to push sensitivity further to 70 pM, which we identified as our LOD.

Figure 4.

a) FL quenching_, inset: Stern‐Volmer plot as a result of increasing Cr⁶⁺ concentration in CQD‐(NH₂, COOH), b) FL quenching_, inset: Stern‐Volmer plot as a result of increasing Cr⁶⁺ concentration in CQD‐(NH₂, COOH)/g‐C₃N₄ for LOD measurement, c) cations, and d) anions at 1 µM concentration after 20 min of introducing in solution at 298.15 K and pH = 7.

Table 1.

A comparison of Cr⁶⁺ sensing in the present work with some of the previously published CQD data.

Sensing materials	Method	Linear range	LOD	Refs.
Cysm‐C. dots	Fluorometric	2.5–40 µM	1.57 µM	[51]
Au NDC@Ag NRs	Colorimetric	2.5–40 µM	1.69 µM	[52]
N‐CQD/MCD	Fluorometric	2.5–68 µM	4 µM	[53]
N‐CQD	Fluorometric	0–15 µM	0.3 µM	[54]
KYCD	Fluorometric	3.4–34 µM	0.11 µM	[55]
CQD‐(NH2, COOH)	Fluorometric	4.6 nM – 1.8 µM	0.36 nM	This work
CQD‐(NH2, COOH)/g‐C3N4	Fluorometric	0.1 nM – 100 µM	0.07 nM	This work

2.2. Determination of Selectivity and Stability of the CQD‐(NH₂, COOH) and CQD‐(NH₂, COOH)/g‐C₃N₄

To test and determine the levels of selectivity of our materials we measure the FL intensity with different ions in presence and absence of Cr⁶⁺. In our analysis of the results, represented as FL ratio – normalized F/F₀, F₀ represents the FL intensity of the fluorophore solution (CQD‐(NH₂, COOH) or CQD‐(NH₂, COOH)/g‐C₃N₄) in absence of ions while F denotes the intensity after quenching with a concentration of 1 µM for total ion content. Details regarding the normalization procedure are provided in the Supporting Information (Equation S4, Supporting Information). The normalized result for the tested Cr⁶⁺ ion represents the zero point on the 0 to 1 normalized scale, as shown in Figure 4c (top panel). This forms the benchmark of the most pronounced quenching effect on CQD‐(NH₂, COOH), demonstrating high selectivity even at trace concentrations as low as 4.6 nM of Cr⁶⁺. In contrast, Al³⁺ displayed the highest normalized F/F₀ value of 1, suggesting negligible quenching compared to Cr⁶⁺. Cr³⁺ showed noticeable but less dominant interactions with CQD‐(NH₂, COOH). These relative differences in quenching intensity highlight the potential of CQD‐(NH₂, COOH) as highly selective sensor material for FL‐based sensing. Upon introducing Cr⁶⁺ alongside other cations in a mixture, we observed a significant increase in quenching, further confirming the sensitivity of CQD‐(NH₂, COOH) to Cr⁶⁺. We note that CQD‐(NH₂, COOH) exhibits ≈13% greater sensitivity to Cr⁶⁺ compared to Cr³⁺. While this represents an improvement over previously reported results (Table 1), there remains scope for further enhancing the selectivity metric to more effectively distinguish Cr⁶⁺ from Cr³⁺ in mixed‐ion environments.^{[ 44 ]} We demonstrate this with our dual covalent and non‐covalent modified CQD‐(NH₂, COOH)/g‐C₃N₄ 2D/2D heterostructure. Figure 4c (bottom panel) highlights the enhanced selectivity towards Cr⁶⁺, as indicated by the marked difference in normalized F/F₀ intensity for Cr⁶⁺ compared to other cations including Cr³⁺ (90% more sensitive to Cr⁶⁺ compared to Cr³⁺). In addition, heterostructuring enables a significantly broader dynamic detection range—from 0.1 nM to 100 µM—whereas the covalently modified CQD‐(NH₂, COOH) alone exhibit a narrower measurable range of 4.6 nM to 6 µM. Likewise, Figure 4d shows the measurement results of FL quenching in presence of various anions on both CQD‐(NH₂, COOH) (top panel) and CQD‐(NH₂, COOH)/g‐C₃N₄ (bottom panel). In contrast to the results observed for cations, no significant quenching was detected for anions, even at concentrations as high as 1 µM. However, both the materials show significant FL quenching on the addition of Cr⁶⁺ alongside the anions. Together, Figure 4c,d showcase the remarkable selectivity achieved through our dual covalent and non‐covalent modification strategy to distinguish Cr⁶⁺ in the presence of dominant cations and anions expected in drinking water. Furthermore, CQD‐(NH₂, COOH)/g‐C₃N₄ shows significant difference in FL quenching with respect to Cr³⁺, highlighting its applicability to differentiate levels of Cr⁶⁺ in total Cr. We note here that we determined the LOD for total Cr in the order of 91.6 pM (Figure S5a, Supporting Information), which surpasses the WHO's recommended LOD (in the total Cr LOD test, Cr⁶⁺/Cr³⁺ = 0.2).

To further validate the applicability of our FL sensor, real‐world water samples were collected from tap water and water fountains. These samples were analyzed using both inductively coupled plasma triple quadrupole mass spectrometry (ICP‐MS/MS) and ion chromatography for multi‐ion detection (Figure S5b,c, Supporting Information) and our fluorescence sensing platform (Figure S5d, Supporting Information). The total Cr concentrations detected by the FL sensor closely match those obtained via ICP‐MS/MS and are consistent with results from the synthetic Cr solution test (Figure S5a, Supporting Information), confirming our sensor's reliability and effectiveness in real‐water samples.

Furthermore, given the observed FL quenching trends across various ion concentrations in the absence and presence of Cr⁶⁺, a predictive model could be developed to estimate Cr⁶⁺ concentrations in complex ion mixtures based on the normalized F/F₀ value. By leveraging machine learning or regression‐based approaches^{[ 45 ]} along with the sensor's kinetic behavior study could facilitate precise artificial intelligence‐based quantification of Cr⁶⁺ in real‐world water samples, further enhancing the applicability of our sensor for environmental monitoring. However, the development and validation of such a predictive model are beyond the scope of this study.

2.3. Stability and Performance of the Materials with pH Change

We further conduct tests with varying pH in the range of 1 to 8.5 as shown in Figure 5a. Notably, CQD‐(NH₂, COOH) reveals significant changes in FL intensity at different pH levels, peaking at pH 4 and reaching a minimum at pH 8.5. This suggests that the covalently modified fluorophore (CQD‐(NH₂, COOH)) exhibits better stability in acidic conditions compared to alkaline ones. Interestingly, the dual covalent and non‐covalent modified CQD‐(NH₂, COOH)/g‐C₃N₄ shows stable FL response across the entire pH range from 1 to 8.5, indicating a broader pH tolerance and enhanced stability of the fluorophore. This is consistent with existing reports on the stability of g‐C₃N₄ ⁴ and highlights the fact that the g‐C₃N₄ adds overall enhancement in the stability of the 2D/2D heterostructure as suggested in Figure 1a(iii). The observed stability is particularly important, given that the pH of drinking water typically ranges from 5 to 8.5.^{[ 42 ]} As illustrated in Figure 5a, the CQD‐(NH₂, COOH)/g‐C₃N₄ heterostructure exhibits higher sensitivity compared to CQD‐(NH₂, COOH) under neutral pH condition (pH = 7), motivating further analysis to determine its LOD at the optimal pH. To evaluate performance, pH 5.7 and pH 8.5 were selected as representative lower and upper limit, respectively. At pH 5.7, where the highest FL intensity was observed (Figure 5a), the CQD‐(NH₂, COOH)/g‐C₃N₄ achieved a LOD of 69.7 pM (Figure S6a, Supporting Information). At pH 8.5, the LOD was determined as 86.8 pM (Figure S6b, Supporting Information). Notably, under neutral conditions (pH 7), which is most relevant for potable water, the heterostructure demonstrated the good sensitivity with an LOD of 70 pM (Figure 4b), thereby satisfying the WHO's permissible limit for Cr⁶⁺ in drinking water.

Figure 5.

pH‐dependent FL performance and photostability of the sensing materials. a) Radar plots of absolute FL intensity for CQD‐(NH₂, COOH) and CQD‐(NH₂, COOH)/g‐C₃N₄ across a pH range of 1.0 to 8.5, showing peak emission behavior. CQD‐(NH₂, COOH)/g‐C₃N₄ exhibits more consistent intensity across the range, suggesting improved pH tolerance and structural stability. b) Corresponding normalized FL response plots (F/F_i) for CQD‐(NH₂, COOH) and CQD‐(NH₂, COOH)/g‐C₃N₄. Images (right panels) show visual FL responses under 405 nm excitation after 0.5, 8, and 24 h, highlighting stable FL emissions.

2.4. Stability of the Materials Over Time

The stability of both materials in the presence of Cr⁶⁺ in water was evaluated over a 24‐hour period, with samples exposed to a 405 nm laser for 30 s at each interval. The corresponding results are presented in Figure 5b. Here, F represents the FL intensity in each experiment, while F_i indicates the FL intensity of the fluorophore in the presence of Cr⁶⁺ under the initial light exposure condition. Both diagrams in Figure 5b suggest high colloidal stability,^{[ 46 ]} particularly for CQD‐(NH₂, COOH)/g‐C₃N₄, which shows retention of FL sensitivity above 80% (F/F_i > 0.8) for an extended period. The high stability of both materials is a significant attribute, reflecting their resistance to photobleaching, aggregation, or degradation in aqueous environments.^{[ 47 ]} Such stable FL behavior is especially desirable for sensing applications, where long‐term consistency in emissions is critical. The maintenance of FL intensity indicates effective surface passivation of CQD‐(NH₂, COOH) and CQD‐(NH₂, COOH)/g‐C₃N₄, preventing oxidative damage or quenching interactions with water molecules or other species in solutions. Moreover, this stability under illumination suggests that CQD‐(NH₂, COOH) and CQD‐(NH₂, COOH)/g‐C₃N₄ possess robust core structures and are well‐dispersed, due to the careful implementation of both covalent and noncovalent modification strategies for enhancing CQD fluorescence. Such stability not only enhances their practicality for continuous, real‐time applications but also ensures minimal performance loss over time.^{[ 48 ]} The photobleaching (defined as the irreversible photochemical degradation of the fluorophores) of CQD‐(NH₂, COOH) and CQD‐(NH₂, COOH)/g‐C₃N₄ was tested over a four‐month period, with each illumination lasting one minute. The results are shown in Figure S7a,b (Supporting Information). As shown in Figure S7a (Supporting Information), the fluorescence intensity of CQD‐(NH₂, COOH) remains highly stable over time, retaining 98.79% of its initial emission, thereby demonstrating excellent photostability. This high photostability might be due to its intrinsic resistance to photobleaching.^{[ 49 ]} Figure S7b (Supporting Information) presents the same for CQD‐(NH₂, COOH)/g‐C₃N₄, clearly maintaining a fluorescence stability of 96.99% over the same period. Under illumination, the formation of a depletion layer at the heterojunction between CQD‐(NH₂, COOH) and g‐C₃N₄ is enhanced due to carrier repartitioning within the MGB. This redistribution increases the likelihood of covalent and non‐covalent bond cleavage within the heterostructure architecture, thereby facilitating a higher rate of photogenerated carrier activity, as observed experimentally. Consequently, the CQD‐(NH₂, COOH)/g‐C₃N₄ heterostructure exhibits slightly greater susceptibility to photodegradation compared to CQD‐(NH₂, COOH).^{[ 50 ]}

2.5. On the FL Quenching Mechanism

It is imperative to explain the underlying mechanism allowing us to achieve higher FL quenching sensitivity and selectively in presence of Cr⁶⁺. To begin with, as Cr⁶⁺ concentration increases, the chance of enhancements in fluorescence resonance energy transfer interactions between fluorophore and Cr⁶⁺ is expected.^{[ 56 ]} With more Cr⁶⁺ in the mix, there are more electron acceptors available for the excited carriers generated by the fluorophore, which in turn will reflect as a higher quenching gradient when the carrier relaxes. Additionally, greater concentrations of Cr⁶⁺ are expected to heighten the likelihoods of higher charge transfer rate between the fluorophore and Cr⁶⁺. Such higher transition rates are expected to promote non‐radiative decay with a diminished FL signal.^{[ 57 ]} The FL quenching behavior of CQD‐(NH₂, COOH) and CQD‐(NH₂, COOH)/g‐C₃N₄ in the presence of Cr⁶⁺ reveals distinctive behaviors for each system supporting the above‐mentioned mechanisms. It is easier to explain that the higher FL quenching gradients by analysing the responses in the frequency domain to gauge the transitional levels expected from the modified MGB introduced by the covalent and non‐covalent modifications in the 2D/2D heterostructure. The frequency rate of FL fluctuations is highlighted in Figure 6a obtained via Fast Fourier Transforms (FFT) of the FL spectra taken over time spaced by 6 mins over a period of 1 hr. The fluctuational rates show characteristic peak behavior with distinctive variations as a function of concentration (Figure 6a (right panel)).

Figure 6.

a) FFT analysis of FL quenching of CQD‐(NH₂, COOH) spaced by 6 mins over a period of 1 h for blank, 4.5 nM, 100 nM, and 6 µM, b) Van't hoff diagram, c) DFT energy change of CQD‐(NH₂, COOH), d) schematic of FL mechanism in CQD‐(NH₂, COOH), and CQD‐(NH₂, COOH)/g‐C₃N₄ in presence on different cations, e) FFT analysis of FL quenching of CQD‐(NH₂, COOH)/g‐C₃N₄ spaced by 6 mins over a period of 1 h for blank, 0.1 nM, 1 µM and 100 µM, f) Van't hoff diagram, g) DFT energy change for CQD‐(NH₂, COOH)/g‐C₃N₄.

The fluctuational characteristics as shown present the fluctuation spectra (via FFT analysis) of FL signal relaxation rates for CQD‐(NH₂, COOH), both in the absence (blank) and presence of Cr⁶⁺ at varying concentrations (4.6 nM, 100 nM, and 6 µM). In the blank sample, the relaxation spectrum is relatively smooth, indicating consistent and stable excited‐state dynamics with minimal external perturbation. However, upon the addition of Cr⁶⁺, a distinct, concentration‐dependent fluctuation behavior emerges. Notably, the spectrum corresponding to 100 nM of Cr⁶⁺ displays a pronounced peak, suggesting a maximum in fluctuation probability and thus enhanced sensitivity of carrier relaxation pathways at this intermediate concentration. This peak may arise due to optimal quenching interactions or transient complex formations, such as partial aggregation or dynamic binding states between Cr⁶⁺ions and functional groups on the surface.

Interestingly, at the highest tested concentration (6 µM), the FFT spectrum closely overlaps with the blank, indicating a suppression of fluctuation amplitude. This suggests that beyond a certain concentration threshold, the system may saturate or undergo self‐quenching or shielding effects that restore the relaxation dynamics to a more stabilized or averaged‐out state. Therefore, the observed trend does not follow a simple linear correlation with concentration. Instead, it reflects a non‐monotonic response, with 100 nM representing a critical concentration at which carrier relaxation dynamics are most disrupted, potentially offering insights into the optimal sensing window for Cr⁶⁺ detection using these CQD‐(NH₂, COOH).

The highlighted zoom‐in region (Figure 6a, right panel) identifies the transition rate range where these differences are most pronounced. A characteristic peak near 0.026 s⁻¹ suggests that the introduction of Cr⁶⁺ alters the emission stability of CQD‐(NH₂, COOH), particularly at intermediate concentrations. Interestingly, while quenching at rates above 0.1 s⁻¹ shows more pronounced fluctuational behavior, the overall fluctuation magnitude remains relatively low (Figure S8a, Supporting Information) with a broad distribution. This spectral behavior provides insight into the nature of the underlying FL quenching mechanism—whether it is predominantly static or dynamic in nature.

Static quenching occurs when non‐fluorescent complexes form between the fluorophore (CQD‐(NH₂, COOH)) and quencher (Cr⁶⁺) prior to excitation, typically leading to reduced FL intensity without significantly affecting the fluorophore's excited‐state lifetime. Under dominant static quenching, one would expect a decrease in fluctuational signal at lower fluctuation rates, as fewer fluorophores remain available for excitation and emission. However, our FFT spectra show increased fluctuational intensity at higher Cr⁶⁺ concentrations, which deviates from the expected trend for purely static quenching.

In contrast, dynamic quenching, mediated through collisional interactions between the fluorophore and quencher in the excited state,^{[ 1 ]} can lead to modifications in both intensity and lifetime. This is typically accompanied by increased fluctuation signatures. The enhanced fluctuational features observed in the FFT spectra, especially at intermediate concentrations, suggest changes in excited‐state dynamics indicative of dynamic quenching. This implies that the fluorophores undergo increased non‐radiative decay in the presence of Cr⁶⁺, likely through multiple jump transitions mediated by the MGB. The effective broadening of the FFT spectra at this intermediate 100 nM concentration supports this view, pointing toward the presence of additional energy dissipation pathways characteristic of dynamic mechanisms.

Furthermore, the FL response time, calculated from the FFT data (using Equation S5, Supporting Information) as described in Table S2 (Supporting Information), ranges from 7.62± 0.46 to 7.92± 0.34 s across Cr⁶⁺ concentrations from 4.6 nM to 6 µM. The concentration‐dependent variation in response time further reinforces the presence of a dynamic quenching process.

To further elucidate the quenching mechanism, Figure 6b presents the Van't Hoff analysis, where a small slope of –0.61 indicates weak temperature dependence—again suggestive of a dynamic process occurring via excited‐state collisions. This is also evident from the effective linewidth broadening seen in the FL responses with centration (Figure 4a). Complementing this, density functional theory (DFT) simulations (Figure 6c) reveal significant energetic shifts upon Cr⁶⁺ interaction, supporting the notion of strong excited‐state perturbations rather than ground‐state complexation alone. Taken together, these results indicate that although both static and dynamic quenching may coexist, the dominant pathway is dynamic quenching, as illustrated in the schematic representation in Figure 6d.

A similar concentration‐dependent behavior is observed for the CQD‐(NH₂, COOH)/g‐C₃N₄ heterostructure, as shown in Figure 6e. Like CQD‐(NH₂, COOH), the blank sample exhibits a smooth and stable decay profile, indicative of effective energy relaxation with minimal external perturbation. Upon the addition of Cr⁶⁺, however, the fluctuation spectrum becomes richer in transitional behavior, as highlighted in the zoomed region (Figure 6e, right panel). At the lowest concentration tested (0.1 nM of Cr⁶⁺), deviations from the blank remain minimal, suggesting that quenching at this level is negligible—likely dominated by initial static interactions. The characteristic peak for both the blank and Cr⁶⁺ treated samples appears at a lower fluctuational rate of ≈0.019 s⁻¹, supporting this observation.

At higher Cr⁶⁺ concentrations, however, the fluctuational behavior shifts. Specifically, at rates above 0.1 s⁻¹, the range of fluctuations narrows relative to the blank (Figure S8b, Supporting Information), implying a transition in the quenching regime. This suggests that while static quenching through complex formation may dominate at lower concentrations, dynamic quenching becomes increasingly significant at elevated Cr⁶⁺ levels due to enhanced collisional interactions. Moreover, the increased number of jump transitions observed at lower fluctuational rates points to the creation of additional allowable energy states, likely introduced through complex formation, through which excited carriers can relax (Figure 1c).

The reduced fluctuation broadening at higher rates further supports a hybrid quenching mechanism for CQD‐(NH₂, COOH)/g‐C₃N₄, where both static and dynamic processes coexist, with dynamic quenching emerging as the dominant pathway at higher Cr⁶⁺ concentrations.

Consistent with these observations, the response time of CQD‐(NH₂, COOH)/g‐C₃N₄ demonstrates faster kinetics compared to CQD‐(NH₂, COOH). As presented in Table S3 (Supporting Information), response times range from 4.63± 0.44 to 5.83± 0.31 s across Cr⁶⁺ concentrations from 0.1 nM to 100 µM. This time‐dependent behavior, mirroring that of CQD‐(NH₂, COOH), provides further evidence of dynamic quenching mechanisms playing a key role in the FL response of the heterostructure.

Additionally, the dual‐modified CQD‐(NH₂, COOH)/g‐C₃N₄ system exhibits a significantly steeper Van't Hoff slope of 727 (Figure 6f), indicating a pronounced temperature dependence consistent with a static quenching mechanism. This suggests the formation of a greater number of non‐fluorescent ground‐state complexes, a characteristic behavior of static quenching.^{[ 58 ]} DFT calculations presented in Figure 6g reinforce this interpretation, revealing a substantial energy shift upon Cr⁶⁺ addition, from −177.8 Ry to −771.9 Ry, indicating strong ground‐state interactions aligned with static quenching behavior. Moreover, the LOD for Cr⁶⁺ determined at 343.15K (Figure S9, Supporting Information) improves to 40.8 pM. This enhancement in sensitivity at elevated temperatures results from the formation of ground‐state complexes under thermal excitation, which again supports the static quenching behavior.^{[ 58]}

Notably, the DFT energy changes observed for Cr⁶⁺ are significantly more pronounced than those for other cations, correlating well with the experimental trends shown in Figure 4c. This alignment between theoretical and experimental data further substantiates the unique interaction between Cr⁶⁺ and the dual‐functionalized system. Interestingly, however, UV‐Vis absorption spectra (Figure S10, Supporting Information) show no discernible change following Cr⁶⁺ addition, supporting the coexistence of dynamic quenching processes. This apparent contradiction underscores the complex quenching behavior in the system, where both static and dynamic mechanisms are at play. The relative dominance of each likely depends on Cr⁶⁺ concentration and local interaction dynamics within the heterostructure. In summary, the fluctuational behavior highlights the jump transitions via the multi energy band states as introduced by the covalent as well as non‐covalent modifications.

Furthermore, to assess the recyclability and post‐use stability of the material as a FL sensor, the CQD‐(NH₂, COOH)/g‐C₃N₄ heterostructure was subjected to a recovery process. The sensing solution was treated with activated carbon^{[ 59 ]} at 60 °C for 2 h to facilitate Cr⁶⁺ adsorption, followed by filtration to remove residual ions. The recovered solution was then analyzed via FTIR spectroscopy and compared with the pristine CQD‐(NH₂, COOH)/g‐C₃N₄ material. As shown in Figure S11 (Supporting Information), the structural integrity of the heterostructure was retained, with only minor differences in concentration, likely attributed to partial loss of material due to adsorption onto activated carbon during the recovery process.

Figure 7 summarizes the proposed FL quenching mechanism based on comprehensive material characterization, emphasizing the critical role of both covalent and non‐covalent interactions in facilitating efficient charge transfer. The valence band minima of CQD‐(NH₂, COOH) and CQD‐(NH₂, COOH)/g‐C₃N₄ obtained from XPS (Figure S12, Supporting Information), along with bandgap values derived from Tauc plots (Figure 2e), were used to estimate Fermi level energies. Upon heterojunction formation, a downward shift of –1.17 eV was observed, suggesting strong interfacial electronic coupling and band bending due to Fermi level pinning. This shift promotes the accumulation of excited carriers within the MGB, effectively reducing the probability of recombination and enabling more stable photoexcited electrons for interaction with Cr⁶⁺ ions. The emergence of the MGB at the interface thus acts as a key pathway for enhanced charge transfer during the FL quenching process. Given that Cr⁶⁺ is a strong oxidizing agent and a known FL quencher via electron capture, the lower conduction band minimum of the CQD‐(NH₂, COOH)/g‐C₃N₄ heterostructure (Figure 7) further supports the thermodynamic favorability^{[ 60 ]} of electron transfer. This electronic configuration ultimately enhances the sensitivity of FL‐based Cr⁶⁺ detection.

Figure 7.

Schematic illustration of the ultrasensitive and selective FL detection of Cr⁶⁺ with the CQD‐(NH₂, COOH)/g‐C₃N₄. Bottom panel highlights the effective change in Fermi level of the materials obtained from Tauc plot and XPS data analysis.

3. Conclusion

This study presents a sustainable and effective strategy for developing ultrasensitive FL sensors for Cr⁶⁺ detection by integrating covalent and non‐covalent modifications into CQD. Using a one‐step, in situ sonochemical approach and coffee waste as a carbon source, we synthesized dual‐functionalized CQD‐(NH₂, COOH), which exhibited ultra sensitivity (LOD: 360 pM) and a dynamic detection range of 4.6 nM to 6.0 µM. Upon heterostructuring with g‐C₃N₄, the resulting 2D/2D CQD‐(NH₂, COOH)/g‐C₃N₄ system achieved an exceptional ultra‐low LOD of 70 pM for Cr⁶⁺, with a broadened detection range spanning 100 pM to 100 µM—meeting the WHO's permissible limit for drinking water. Notably, the dual‐modified material also demonstrated a LOD of 91.6 pM for total Cr, highlighting its selectivity. The enhanced performance is attributed to band structure modulation via MGB formation, improved charge separation, and interfacial electronic coupling, as confirmed by FFT, DFT, and XPS analyses. These results underscore the potential of dual covalent–non‐covalent strategies in engineering advanced fluorophores for real‐time, low‐concentration detection of hazardous ions in complex environmental settings.

4. Experimental Section

Materials

Dual‐functionalized CQD were synthesized from coffee waste (Figure S13a, Supporting Information). Melamine (99%, ACS reagent, Sigma Aldrich) was used as a precursor for g‐C₃N₄. Deionized (DI) water served as the solvent for all the synthesis. Also in order to perform the sensing test, chemical reagents Cr (VI) oxide (CrO₃) from VWR (99% purity), iron (III) chloride hexahydrate (FeCl₃·6H₂O, ACS reagent, Sigma Aldrich), cobalt (II) nitrate hexahydrate (Co(NO₃)₂ ·6H₂O, reagent grade, 98%, Sigma Aldrich), aluminum chloride hexahydrate (AlCl₃·6 H₂O, reagent plus, 99%, Sigma Aldrich), ammonium molybdate tetrahydrate ((NH₄)₆Mo₇O₂₄·4H₂O, Bio Ultr >99.0%, Sigma Aldrich), zinc nitrate hexahydrate (Zn(NO₃)₂·6H₂O, reagent grade, 98%, Sigma Aldrich), copper (II) acetate hydrate (Cu(CH₃COO)₂·H₂O, 98%, Sigma Aldrich), sulfuric acid (H₂SO₄, 93–98%, ARISTAR ULTRA, Ultrapure for trace metal analysis, VWR Chemicals), sodium hydroxide (NaOH, reagent grade, ≥98%, pellets, anhydrous, Sigma Aldrich), hydrazine solution (35 wt. % in H₂O, Sigma Aldrich) are used. In addition, Cr⁶⁺ was subjected to chemical reduction with a stochiometric amount from hydrazine solution to generate Cr³⁺ for the sensing test.

Preparation of Precursor Solutions and Synthesis of CQD‐(NH₂, COOH): Here, we employed coffee waste as the precursor. To synthesize the dual‐functionalized CQD (CQD‐(NH₂, COOH)), an in‐situ functionalization process is carried out, utilizing NH₂ from amino acids and COOH from chlorogenic acids present in the coffee.^{[ 15 ]} Using high‐intensity sound waves helps to in‐situ synthesis of CQD‐(NH₂, COOH) possible through cavitation phenomena. 10 g of coffee waste is subjected to an ultrasonication treatment using a probe sonicator (Q500, QSonica) with a power of 400 W and frequency of 20 kHz, alternating 1‐minute pulse on and off for a total duration of 10 min. Following this, the mixture is subjected to filtration in two stages: the first stage using an 11 µm Whatman filter paper and subsequently a 0.65 µm syringe filter for finer particle removal (and Figure S13b, Supporting Information).

Preparation of Precursor Solutions and Synthesis of g‐C₃N₄: Preparation of 2D g‐C₃N₄ was carried out using a bottom‐up approach, with melamine as the precursor. 3 g of melamine underwent thermal polymerization process to synthesis the bulk g g‐C₃N₄. The melamine was placed in a furnace and heated to 520 °C at a rate of 5 °C per minute,^{[ 29 ]} where it was maintained for 4 h. Following the heating process, the obtained bulk g‐C₃N₄ was subjected to a top‐down exfoliation method to obtain 2D g‐C₃N₄. The bulk g‐C₃N₄ was dispersed in 60 mL of DI water and sonicated for 5 min at a power of 400 W and a frequency of 20 kHz. The resulting suspension was then centrifuged for 10 min at 5000 rpm, and the supernatants, containing 2D g‐ C₃N₄ were collected and filtered through a 0.22 µm PTFE hydrophobic syringe filter to remove bigger particles.

Decoration of CQD‐(NH₂, COOH) on g‐C₃N₄: 2 mL of the CQD‐(NH₂, COOH) solution was mixed with 20 mL of the g‐C₃N₄ solution prepared earlier. The resulting mixture was then subjected to a hydrothermal process, heated at a rate of 7 °C per minute until reaching 180 °C, and maintained at that temperature for 1 h.

Materials Characterization Techniques: Characterizing the synthesized materials was done using techniques such as TEM, AFM, XRD, XPS, UV‐Vis, and FTIR spectroscopy, to determine particle size and morphology, crystalline structure, elemental composition, chemical states, and electronic structure, optical properties, and surface functional groups, respectively. Furthermore, FL behavior of the synthesized materials was measured by monitoring emission intensity upon excitation at 405 nm using a FL measurement setup.

FL Quenching Tests

To prepare the Cr samples, 1 mg of CrO₃ was dissolved in 1L of DI water and then 927 mL of as‐prepared solution was diluted with DI water to 1L solution. In each step, the solution was vortexed for 1 min at 600 rpm. Subsequently, a specified volume of the sample was dispensed using a micropipette into 2 mL of as prepared CQD‐(NH₂, COOH) and CQD‐(NH₂, COOH)/g‐C₃N₄.

Initially, a predefined concentration of metal elements is introduced into the sample. Subsequently, the mixture is homogenized using the Vortex‐Genie 2 Shaker at a constant speed of 1000 rpm for a duration of 30 s.

Following the mixing phase, a 2 mL of the resultant sample was transferred into a quartz cuvette. The sample within the cuvette was then subjected to excitation using a 405 nm violet laser. The FL of the sample because of this excitation was measured. The data acquisition was performed using the Ocean Optics Flame Miniature Spectrometer, which was interfaced with OceanView software version 1.4.1, enabling the detailed analysis and visualization of the FL properties of the sample. The schematic of the FL sensor structure, setup, and overall mechanism could be found in Figure 7.

DFT Calculation Method

To explore the energy changes resulting from the introduction of different ions into the CQD‐(NH₂, COOH) and CQD‐(NH₂, COOH)/g‐C₃N₄ system, DFT calculations were conducted using the Quantum ESPRESSO code. The initial structure was obtained from materialsproject.com and subsequently modified using Vesta software. The calculations utilized the Perdew‐Burke‐Ernzerhof variant of the Generalized Gradient Approximation, which includes van der Waals interactions. Since standard PBE functionals do not adequately account for weak interactions, a semi‐empirical dispersion‐corrected DFT approach developed by Grimme was employed. This model is well‐established for providing a more accurate description of weak vdW interactions. Grimme's method incorporates the Rappe‐Rabe‐Kaxiras‐Joannopoulos model of ultra‐soft pseudopotentials to address these interactions effectively.^{[ 4 ]} CQD‐(NH₂, COOH)/cation: 100/1, g‐C₃N₄/cation: 100/1. Cation charges were applied using RESP charges in GAMESS software package.

For each material, a mesh of 5×5×1 k‐points was utilized, with further calculations based on the total energy plotted against the number of k‐points. A vacuum distance of over 20 Å was maintained to minimize interactions between adjacent layers. All structures were optimized using Avogadro software, applying the Marzari‐Vanderbilt cold smearing method with a small broadening width of 0.00146 Ry. Additionally, the diagonalization method was set to “devid” with a mixing factor of 0.4 for self‐consistency. For energy calculations, 250 cycles were selected, and average values were selected as the energy presented in Figure 6c,g. The supercell comprised two (3×3) CQD‐(NH₂, COOH) and three (4×4) g‐C₃N₄. Detailed information on the DFT result could be found in the supporting information (Figure S14a–c and Table S4, Supporting Information).

Conflict of Interest

The authors declare no conflict of interest.

Author Contributions

The manuscript was written through the contributions of all authors. All authors have approved the final version of the manuscript. P.Z. generated the idea, designed and conducted the experiments, and analyzed the data under the supervision of S.K., P.Z. and A.P. prepared AFM samples and conducted the tests. P.Z.wrote the manuscript. A.P. and S.K. contributed to idea generation, experimental design, data analysis, and revising the manuscript.

Supporting information

Supporting Information

Zandi P., Phani A., Kim S., Detecting Cr⁶⁺ at ≈100 pM Concentration with Fluorescence Enhancement Signatures in a Novel Eco‐Fluorophore: Matching WHO's 96 pM Recommended Standard for Drinking Water. Adv. Mater. 2025, 37, 2504142. 10.1002/adma.202504142

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.