Chiral Covalent Organic Frameworks for Enantioselective Sensing: Electrochemical, Electrochemiluminescence, and Photoelectrochemical Platforms

Abstract

Chiral covalent organic frameworks (CCOFs) merge the structural versatility of covalent organic frameworks with chirality, offering a powerful platform for enantioselective electrochemical sensing. This review comprehensively summarizes recent advances in CCOF-based sensors, focusing on signal transduction mechanisms. Three primary synthetic strategies (direct synthesis, post-synthetic modification, and chiral induction) and four electrode immobilization approaches (drop-casting, in situ growth, membrane fabrication, and nanopipette modification) are outlined and critically assessed with respect to their advantages and limitations. Based on photon involvement, the sensing applications are categorized into three operational modes: pure electrochemical, electrochemiluminescence (ECL), and photoelectrochemical (PEC) sensing. For each mode, the design principles, analytical performance, recognition mechanisms, and structure–property relationships are discussed. Current challenges and future directions are examined, spanning from analyte scope expansion and mechanistic elucidation to conductivity enhancement, aqueous stability, fabrication–performance correlations, multi-mode synergistic platforms, and real-sample validation. This review provides a coherent framework for understanding CCOF-based enantioselective electrochemical sensing and perspectives for future development.

1. Introduction

Chirality is a fundamental property that governs the structure and function of biomolecules [1,2,3,4,5]. In pharmaceutics, a drug’s stereochemistry often dictates its activity, metabolism, and toxicity [6,7,8]. One enantiomer may be therapeutic while the other is harmful, as tragically illustrated by thalidomide, whose (R)-enantiomer is a sedative while the (S)-enantiomer is a potent teratogen [9,10]. Developing rapid, sensitive, and reliable methods for chiral discrimination has therefore become a paramount objective in analytical chemistry, drug development, and clinical diagnostics [11,12,13,14,15,16,17].

Classical separation and detection techniques deliver high accuracy and resolution [18,19]. Typical examples include high-performance liquid chromatography, gas chromatography, capillary electrophoresis, and circular dichroism spectroscopy [20,21,22,23,24]. These methods, however, suffer from time-consuming sample preparation, expensive instrumentation, and the need for skilled operators [25,26,27]. They are consequently not well suited for rapid, on-site, or high-throughput analyses [28]. Electrochemical transduction methods have emerged as a compelling alternative [29,30,31]. These methods offer high sensitivity, fast response, operational simplicity, and intrinsic compatibility with miniaturized, portable platforms [32,33,34]. By directly converting chiral recognition at an engineered interface into a quantifiable signal, they enable real-time analysis with minimal sample consumption—attributes that make them particularly attractive for decentralized and point-of-care chiral analysis [35,36,37,38].

Sensor performance depends critically on the chiral recognition material at the electrode interface. Biomolecules such as proteins and DNA deliver exquisite enantioselectivity but suffer from poor stability and high cost [11,12]. Chiral small molecules and polymers are synthetically accessible, yet they lack the ordered porosity needed for efficient analyte enrichment [25,39]. Metal–organic frameworks (MOFs) provide high porosity and crystallinity [40,41], but their limited aqueous stability and difficult integration of electroactive functions restrict practical utility [42,43,44]. Chiral nanomaterials offer good conductivity and easy modification, but they generally lack uniform recognition sites and batch-to-batch reproducibility [29,45,46]. Historically, organized molecular and supramolecular sensing layers were the conceptual predecessors of COF-based sensing: molecular packing, aggregation, orientation, film morphology, and recognition layer–transducer coupling can profoundly affect selectivity and sensitivity. Bettini et al. showed that π–π-stacked perylene–bisimide aggregates act as sensitive recognition layers for aromatic biogenic amines, discriminating phenylethylamine and tyramine down to ~10⁻¹⁰ M even after layer-by-layer immobilization on an SPR transducer [47]. Neither a CCOF nor intrinsically enantioselective, this example illustrates the progression from molecular recognition to a functional sensing interface and the design principles—ordered packing, controlled film morphology, and stable coupling—that guide the construction of CCOF-based chiral interfaces discussed below.

Covalent organic frameworks (COFs) are crystalline porous polymers assembled from organic building blocks via robust covalent bonds [48,49,50,51,52]. Their unique attributes include predesignable topology, exceptional porosity, high surface area, tunable pore size, and excellent chemical and thermal stability [53]. Functional moieties can be integrated into the backbone with atomic precision, enabling rational materials design [54,55,56,57,58]. Among them, chiral COFs (CCOFs) represent a rapidly developing subclass [59]. They merge the inherent advantages of COFs with chirality, creating a versatile platform for enantioselective processes such as asymmetric catalysis, chiral separation, and sensing [60,61,62,63].

For electrochemical sensing, CCOFs provide several distinctive advantages [64,65,66,67,68]. First, their ordered pore structures define well-confined chiral microenvironments with a high density of recognition sites [69]. Second, modular synthesis allows simultaneous incorporation of chiral centers and electroactive reporters [70]. Third, robust covalent linkages ensure stability under diverse conditions [68,71]. Fourth, tunable pore size introduces a size-exclusion effect that amplifies enantioselectivity [72]. CCOFs thus combine structural designability, stability, and signal transduction efficiency in a single platform, although direct quantitative comparisons with other chiral materials under identical conditions remain limited and rigorous benchmarking studies are still needed to substantiate performance claims [59].

Substantial progress has been made in CCOF synthesis and electrochemical application. This review summarizes synthetic strategies and electrode modification approaches. We then introduce a rigorous photon-based taxonomy: pure electrochemical, electrochemiluminescence (ECL), and photoelectrochemical (PEC) sensing [73,74,75]. Pure electrochemical sensing includes Faradaic and non-Faradaic transduction and nanofluidic ion rectification; ECL involves electrical-to-optical conversion; PEC involves optical-to-electrical conversion. This framework accurately reflects the distinct operational principles. For each category, we discuss design principles, performance, and recognition mechanisms. We conclude with perspectives on future challenges and opportunities.

Before proceeding, we briefly clarify the contribution of this review relative to prior literature. Existing reviews cover chiral COFs from the perspectives of design, synthesis, and general properties (Han et al. [59]), electrochemical chiral sensing with supramolecular receptors such as cyclodextrins [33,35], and fluorescence-based CCOF sensing [14]. The present review differs in three respects: (i) it focuses on CCOF-based electrochemical sensing organized by a photon-based taxonomy (pure electrochemical, ECL, and PEC) that classifies transduction by fundamental operational principles rather than by material class or device architecture; (ii) it goes beyond cataloging examples to analyze structure–property relationships and recognition mechanisms, including unresolved mechanistic questions and anomalous observations that currently limit the field; and (iii) it provides quantitative cross-platform comparisons and an explicit account of the factors governing enantioselectivity (Section 3.4) to guide sensor selection and design. This combination of scope, taxonomy, and critical analysis is the distinctive contribution of the present work. Given the narrative scope of this work, a transparent literature search was performed (Web of Science, Scopus, and PubMed; keywords including chiral covalent organic framework, enantioselective sensing, electrochemical, electrochemiluminescence, and photoelectrochemical; up to August 2026), and representative studies were selected for critical discussion.

2. Construction and Electrode Immobilization of CCOFs

The successful application of CCOFs in enantioselective electrochemical sensing relies on two interconnected aspects: the chemical synthesis of CCOFs with well-defined chiral architectures, and the effective immobilization of these materials onto electrode surfaces to construct functional sensing interfaces [59]. This section summarizes the primary synthetic strategies for constructing CCOFs, and then discusses the electrode modification strategies that translate these materials into practical sensing platforms.

2.1. Synthetic Strategies for CCOFs

CCOFs are generally constructed via one of three routes: direct synthesis from enantiopure monomers, post-synthetic modification of achiral frameworks, or chiral induction during crystallization. The choice among them involves trade-offs in crystallinity, chiral density, and functional-group compatibility.

2.1.1. Direct Synthesis

The most straightforward route to CCOFs is the one-pot solvothermal condensation of enantiopure chiral monomers with achiral or complementary building blocks, integrating chiral centers directly into the framework backbone [76,77,78,79,80]. Recent representative examples include the work by Yuan et al., who synthesized an ionic chiral COF (triPhPy⁺-(S)-CHA) via a Zincke reaction between a trivalent Zincke salt and optically pure cyclohexane-1,2-diamine (Figure 1a) [65]. This one-step strategy directly incorporated both the chiral center and the electroactive pyridinium unit into the COF backbone, yielding a material with strong cathodic ECL response. Similarly, Liu and Kuang reported a triazine-based chiral COF (CC-MP CCTF) synthesized via a solvothermal method using (R)-2-methylpiperazine as the chiral source and cyanuric chloride as the triazine building block [81]. The resulting COF exhibited intrinsic ECL activity and enantioselective recognition toward phenylalanine. Ruan et al. developed a one-dimensional chiral ionic COF, (R)-PTCDA-RMP, through co-condensation of π-conjugated 3,4,9,10-perylenetetracarboxylic dianhydride (PTCDA) with chiral (R)-2-methylpiperazine [82]. This material demonstrated a unique bidirectional ECL response toward penicillamine enantiomers. Direct synthesis offers precise control over the distribution and density of chiral sites; its scope, however, is constrained by the limited availability of enantiopure monomers and the challenge of balancing crystallinity with chirality.

2.1.2. Post-Synthetic Modification

Post-synthetic modification (PSM) involves the covalent functionalization of pre-synthesized achiral COFs with chiral reagents [83,84,85,86,87]. This strategy is particularly attractive because it preserves the crystallinity and porosity of the parent framework while introducing chiral recognition sites. Tan et al. employed a PSM strategy to construct ECL-active chiral COFs (Ph-triPy⁺-(R)-Ru(II)) by coordinating a chiral Ru(II) complex to the bipyridyl units of an ionic COF skeleton [88]. The resulting material exhibited strong ECL signals and demonstrated enantioselective discrimination toward amino alcohols and amino acids. Yuan et al. adopted a stepwise coordination approach, first coordinating RuCl₃ to a bipyridine-containing COF and then introducing chiral cyclohexanediamine to afford Ru(bpy)(S-CHDA)₂²⁺ [89]. This flexible strategy allowed for the modulation of enantioselectivity by simply changing the chiral diamine configuration. Zhang et al. utilized post-synthetic modification to covalently functionalize an imidazole-containing COF with optically pure chiral epoxides, forming imidazolium-based ionic CCOFs with donor–acceptor (D-A) characteristics (Figure 1b) [90]. The imidazolium and pyrenyl units served as electron acceptor and donor components, respectively, enabling efficient intramolecular charge transfer (IRCT) and enhanced ECL emission. PSM offers high flexibility and utilizes existing COF libraries, but the degree of modification can be difficult to control, and harsh reaction conditions may compromise framework integrity.

2.1.3. Chiral Induction Synthesis

Perhaps the most elegant—yet most challenging—strategy is chiral induction, where chirality is imparted to an otherwise achiral framework through chiral catalysts, solvents, or templates during crystallization [91,92,93,94,95]. Zheng et al. reported a chiral induction strategy for the synthesis of self-standing chiral COF membranes via asymmetric catalytic interfacial polymerization [96]. Using (R)- or (S)-phenylethylamine (PEA) as the chiral inducer, they synthesized β-ketoenamine-linked COFs with controlled handedness. By adjusting the molar ratio of TP to PEA, they obtained two types of chiral COFs: cCOF-1 (chiral-induced only) and cCOF-2 (both chiral-induced and modified). Weng et al. developed a chirality-induced linkage exchange strategy to fabricate vinylene-linked chiral sp²C-COFs via an irreversible aldol reaction (Figure 1c) [91]. The approach involved pre-synthesis of a chiral Schiff-base precursor, followed by acid-catalyzed conversion from imine to vinylene linkages. This approach elegantly circumvents the need for pre-synthesized chiral monomers. The efficiency of chirality transfer and the underlying induction mechanisms, however, warrant further investigation before this strategy can be broadly applied.

Figure 1. General synthetic strategies for CCOFs: (a) direct synthesis of Ph-Py⁺-(S,S)-DPEA, (b) post-synthetic modification of MPIm⁺-(S)-MO, and (c) chiral induction synthesis of TT(∆)-COF. Adapted from ref. [91] with permission, copyright 2026, Royal Society of Chemistry.

Figure 1. General synthetic strategies for CCOFs: (a) direct synthesis of Ph-Py⁺-(S,S)-DPEA, (b) post-synthetic modification of MPIm⁺-(S)-MO, and (c) chiral induction synthesis of TT(∆)-COF. Adapted from ref. [91] with permission, copyright 2026, Royal Society of Chemistry.

Each synthetic strategy presents its own set of advantages and challenges. Direct synthesis offers precise chiral site control but is limited by monomer availability. Post-synthetic modification provides flexibility and preserves crystallinity but may result in heterogeneous functionalization. Chiral induction enables chirality transfer from non-chiral building blocks but faces challenges in controlling induction efficiency. The choice of synthetic strategy should be guided by the specific requirements of the target application, including the desired chiral environment, electrochemical activity, and sensing platform.

2.2. Electrode Modification Strategies for CCOFs

The rational design and fabrication of the sensing interface, particularly the immobilization of CCOFs onto the electrode surface, is a critical determinant of sensor performance, as it governs the loading and accessibility of chiral recognition sites, the kinetics of heterogeneous electron transfer, and the long-term stability of the device [97,98,99,100,101,102,103]. Equally important is the immobilization of CCOFs on the electrode surface. Four distinct modification strategies have been developed, from simple drop-casting to precision nanopipette integration, each offering specific trade-offs among fabrication simplicity, film quality, and sensing performance.

2.2.1. Drop-Casting

Drop-casting is the most widely used strategy due to its operational simplicity and universal applicability [104,105,106,107]. A pre-synthesized CCOF dispersion is deposited onto the electrode surface, and the solvent is evaporated to form a sensing film. This method is compatible with various electrode materials and CCOF compositions. In pure electrochemical sensing, Wang et al. drop-cast a suspension of Ph-Py⁺-(S,S)-DPEA·PF6⁻ onto a glassy carbon electrode (GCE), exploiting the intrinsic redox signals of the pyridinium units for chiral discrimination of multiple acids [64]. Similarly, Wang et al. deposited Fe₃O₄@COF@BSA composites onto 3D-printed nanocarbon electrodes for tryptophan enantiorecognition (Figure 2) [108]. For ECL sensing, Tan et al., Yuan et al., and Song et al. all employed drop-casting to fabricate CCOF-modified GCEs for enantioselective detection of amino alcohols, amino acids, and phenylalanine, respectively [65,88,109]. Despite its convenience, drop-casting suffers from uneven film thickness, batch-to-batch variability, and limited control over film morphology, which can compromise sensor reproducibility.

2.2.2. In Situ Growth

In contrast to post-synthesis deposition, in situ growth directly synthesizes CCOF films on the electrode surface by placing the electrode (or its modified layer) in the precursor solution under solvothermal or other reaction conditions [110,111,112,113,114,115]. This approach significantly enhances interfacial adhesion and yields more uniform and mechanically robust sensing layers. Wang et al. demonstrated this strategy by growing a COF layer directly on an anodic aluminum oxide (AAO) substrate, followed by post-synthetic chiral modification with L-penicillamine to construct a nanofluidic ion-rectification sensor for naproxen enantiomers (Figure 3) [116]. Weng et al. prepared an ultrathin chiral sp²C-COF film directly on an interdigitated capacitive (IDC) sensor via a chirality-induced imine-to-vinylene linkage exchange, enabling dielectric-based chiral sensing [91]. In situ growth, however, requires compatibility between the synthesis conditions and the electrode/device architecture, limiting its broader applicability.

2.2.3. Self-Standing and Composite Membranes

When CCOFs are fabricated as self-standing membranes or integrated with porous supports (e.g., poly(ethylene terephthalate), AAO), they serve as both the recognition element and the exclusive ion-transport pathway in nanofluidic sensing devices [117,118,119]. In this configuration, the analytical signal is the transmembrane ionic current, which is modulated by enantioselective binding events within the nanoconfined channels. The key advantage is the nanoconfinement effect: selective adsorption of one enantiomer alters the surface charge, wettability, or effective pore size of the CCOF channels, thereby amplifying subtle recognition events into large changes in ion flux. Zheng et al. prepared a self-standing chiral COF nanochannel membrane via interfacial polymerization and directly measured ion-current changes upon enantiomer binding (Figure 4) [96]. Wu et al. constructed a β-cyclodextrin-modified COF composite membrane on a porous poly(ethylene terephthalate) support, achieving a chiral selectivity factor of 19.2 for tryptophan enantiomers [66]. While this strategy offers label-free detection and ultrahigh sensitivity, it demands precise control over membrane thickness, pore alignment, and integration with the detection system.

2.2.4. Nanopipette Modification

At the ultimate limit of miniaturization, CCOF nanosheets can be assembled onto the tip of a single quartz or glass nanopipette to create a nano-confined sensing interface capable of single-molecule or single-entity sensitivity [120,121]. Yang et al. demonstrated this by depositing L-cysteine-modified COF nanosheets onto a nanopipette tip under an electric field (Figure 5) [122]. Scanning electron microscopy confirmed the successful coverage of the tip. Combined with a Fe³⁺ signal amplification strategy, this platform achieved an ultralow detection limit of 0.21 pM for L-DOPA, the highest sensitivity reported among CCOF-based nanofluidic sensors. The nanopipette configuration offers exceptional sensitivity and minimal sample consumption but requires specialized fabrication equipment and skilled operation, which currently limits its widespread adoption.

Collectively, the four strategies form a continuum from macroscopic planar electrodes to single-nanopore platforms, each addressing distinct sensing requirements (

Table 1. The choice of electrode modification strategy should be guided by the specific requirements of the target application.

Strategy	Advantages	Limitations	Typical Applications
Drop-Casting	Simple, versatile, widely applicable	Uneven film, batch-to-batch variation	Pure electrochemical, ECL sensing
In Situ Growth	Strong adhesion, uniform films	Requires compatible synthesis conditions	Nanofluidic ICR, capacitive sensors
Membrane/Nanochannel	Nanoconfinement amplification, label-free	Complex fabrication, difficult integration	Ion-transport sensing, chiral separation
Nanopipette	Ultrahigh sensitivity, minimal sample consumption	Specialized equipment, skilled operation	Single-channel nanofluidic sensing

). Drop-casting remains the first choice for proof-of-concept studies and material screening due to its operational simplicity, though its limited film uniformity and reproducibility constrain its use in quantitative or commercial applications. In situ growth provides superior film adhesion and homogeneity at the cost of stricter synthetic compatibility, making it suitable for integrated devices where long-term stability is critical. Membrane-based configurations exploit the nanoconfinement effect to achieve label-free detection with amplified signals, representing the platform of choice for nanofluidic ion-rectification and chiral transport studies. Nanopipette modification, while technically demanding, delivers the highest sensitivity and is uniquely suited for single-entity analysis and ultra-trace detection. A key challenge for the field is the lack of systematic comparative studies that correlate each fabrication method with quantitative metrics such as charge-transfer resistance, electroactive surface area, film stability, and batch reproducibility. Establishing such correlations would enable a transition from empirical selection to rational device engineering. Future efforts should also explore hybrid approaches, such as combining in situ growth with nanopipette assembly, to synergistically integrate the advantages of different strategies for next-generation CCOF-based chiral sensors.

We acknowledge that the advantages of CCOFs discussed above are relative rather than absolute, and that their performance superiority cannot be generalized without direct quantitative comparison. Conducting polymers, molecularly imprinted polymers (MIPs), molecular receptors, supramolecular architectures, and other nanomaterials can exhibit excellent enantioselectivity and detection limits in specific systems, and several of them have reached comparable or superior analytical figures of merit in particular applications [33,35]. The key differentiator of CCOFs lies in the combination of properties—ordered, designable chiral pores, high surface area, chemical robustness, and the ability to integrate recognition and transduction functions within one crystalline framework—rather than in any single parameter. To place the field on a quantitative footing, we compiled, wherever available, the linear range, detection limit, enantioselectivity factor, and stability data of representative platforms in Table 3 (Section 3.4). However, a rigorous head-to-head comparison between CCOF sensors and alternative chiral materials under identical conditions has not yet been reported, and we encourage the community to adopt standardized benchmarking protocols (same analyte, same transduction method, same measurement conditions) to unambiguously establish the practical advantages of CCOFs.

3. Enantioselective Electroanalysis of CCOFs

The conversion of chiral recognition events into detectable analytical signals is the core of CCOF-based enantioselective sensing [35,123]. At the highest hierarchical level, these signal transduction mechanisms are categorized according to the role of photons in the signal generation and readout process, yielding three mutually exclusive and collectively exhaustive classes [124,125]. Section 3.1 (Pure Electrochemical Sensing) encompasses all configurations where the signal is derived exclusively from electrical responses at the electrode/solution interface, without the involvement of external light sources or photon emission [126,127,128]. Section 3.2 (Electrochemiluminescence Sensing, ECL) involves an electro-to-optical conversion pathway, wherein electrical excitation generates an excited state that subsequently emits photons [129]. Section 3.3 (Photoelectrochemical Sensing, PEC) involves an optical-to-electrical conversion pathway, wherein illumination generates a photocurrent as the output signal [130]. This photon-based taxonomy eliminates the logical ambiguities inherent in device-morphology-based categorizations and establishes a physically rigorous framework that accurately reflects the distinct operational principles of each sensing platform. The three categories are compared side-by-side in

Table 2. Comparison of the three signal transduction modes based on photon participation in CCOF-based enantioselective sensing.

Feature	Pure Electrochemical	ECL	PEC
Photon role	None	Output (emission)	Input (excitation)
Core readout	Current/potential/impedance/ionic current	Luminescence intensity	Photocurrent
Excitation source	Electrode potential	Electrode potential	Light source + electrode potential
Instrumentation	Potentiostat/galvanostat (simplest)	Potentiostat + photomultiplier tube or CCD	Potentiostat + light source
Background signal	Non-Faradaic charging current (for Faradaic modes)	Near-zero optical background	Low (excitation and detection are separated)
Distinct advantage	Widely available instrumentation; rich readout strategies	Ultimate selectivity; zero-background; bidirectional response	Independent optimization of light absorption and charge transfer
Typical applications	Rapid screening; quantitative ee% analysis	Ultra-sensitive chiral discrimination; mechanistic studies	Multi-mode signal cross-validation; visual/colorimetric readout

.

3.1. Electrochemical Sensing

Electrochemical sensing generates signals solely from the electrical response of the CCOF/electrode interface, without light [131]. This modality offers distinct advantages including inherent simplicity of instrumentation, high sensitivity and selectivity, rapid response, cost-effectiveness, and intrinsic compatibility with miniaturized and portable platforms [132,133,134,135,136,137,138]. Based on the device architecture, this category is further divided into planar electrode sensing and nanofluidic ion-transport sensing.

3.1.1. Planar Electrode Sensing

Planar electrode sensing immobilizes CCOFs on flat conductive substrates such as glassy carbon electrodes or 3D-printed electrodes; the signal is generated perpendicular to the surface [108]. Depending on whether net electron transfer occurs across the interface, this category encompasses Faradaic transduction and non-Faradaic interfacial transduction [139,140].

For Faradaic transduction, three strategies have been developed based on the origin of the electroactive species. The earliest and conceptually most straightforward strategy exploits the inherent electroactivity of the target analyte itself. In this configuration, the CCOF serves exclusively as a chiral recognition interface that provides enhanced surface area and facilitated electron transfer, with the analytical signal originating directly from the analyte’s own redox process. Wang et al. demonstrated this approach using a Fe₃O₄@COF@BSA-modified 3D-printed electrode, where stronger hydrogen-bond interactions with L-Trp yielded a peak current ratio of 1.45 and a linear ee calibration (R² = 0.995) [108]. However, this strategy is inherently constrained to electroactive chiral molecules (e.g., tryptophan, tyrosine, cysteine) and offers no pathway for expanding the analyte scope to electrochemically silent species.

To overcome this limitation, Wu and Kong’s group developed two distinct strategies. The first embeds the electroactive reporter directly in the CCOF backbone: interfacial polymerization of a tripyridinium-based framework (Figure 6a) gives pyridinium units that simultaneously provide chirality and three stepwise redox signals, enabling discrimination of seven chiral acids with peak current ratios of 1.31–2.68 (Figure 6c) [64]. The second is modular, loading chiral ferrocene guests non-covalently into an ionic COF through π–π⁺ interactions (Figure 6b), which affords enantioselective sensing of amino alcohols and amino acids with current ratios of 1.46–1.72 (Figure 6d) [141]. DFT calculations rationalize the enantioselectivity of both systems. For Ph-Py⁺-(S,S)-DPEA, the L-Trp complex is more stable than the D-Trp complex, consistent with the higher peak current for L-Trp (Figure 6e). For the probe-mediated system, (S)-Fc binds D-Pro-OH 9.7 kcal mol⁻¹ more strongly than L-Pro-OH (Figure 6f), again consistent with the higher peak current observed for D-Pro-OH (I_p ratio 1.46) and with contact-angle measurements, in which the bound D-Pro-OH renders the film markedly more hydrophilic (46.2° vs. 55.4°). Because the two diastereomeric complexes share the same hydrogen-bond length (1.245 Å), this stability difference must arise from overall geometric complementarity rather than from a shorter hydrogen bond. The more informative observation is that the same (S)-Fc probe discriminates Pro-OH and penicillamine (Pen) in opposite directions. For Pen, acid–base pairing between the –COOH group and the C=N unit of (S)-Fc dominates, and the L-Pen complex is favoured by 15.7 kcal mol⁻¹. The recognition direction is therefore dictated by the dominant intermolecular force rather than by the absolute configuration of the selector; in terms of the three-point interaction model, changing the dominant force reassigns the contact points and reorganizes the whole diastereomeric complex. We thus tentatively propose that in probe-mediated systems the direction of enantioselectivity is an emergent property of the selector–analyte pair under given solution conditions (pH, solvent, ionic strength) rather than an intrinsic property of the probe. Measuring the recognition direction of (S)-Fc toward Pro-OH and Pen as a function of pH and ionic strength would test this directly. The two strategies also entail a trade-off. The intrinsic-skeleton approach couples recognition and transduction within one platform and covers the broadest analyte scope, but its rigid structure limits tunability and complicates mechanistic study; the probe-mediated approach decouples chirality from electroactivity and is synthetically flexible, yet leaves probe distribution, loading homogeneity, and host–guest dynamics poorly defined. Neither is yet predictive: peak current ratios vary widely without any established structure–selectivity correlation, and the near-absence of reported failed discriminations continues to relegate sensor development to empirical trial-and-error.

In contrast, non-Faradaic transduction detects changes in interfacial capacitance upon enantioselective adsorption, without net electron transfer [142,143]. Weng et al. demonstrated this by fabricating an ultrathin film of a chiral sp²-carbon-conjugated COF on an interdigitated capacitive sensor via chirality-induced imine-to-vinylene linkage ex-change (Figure 7a) [91]. The sensor operated on the principle of dielectric sensing: enantioselective incorporation of Trp into the chiral COF pores altered the local dielectric environment, producing measurable capacitance changes. A selectivity factor of 2.6 for L-/D-Trp discrimination and a linear capacitance response in the 10–40 mM range were achieved (Figure 7b,c). Notably, this capacitance-based readout requires no redox-active labels or added electrolytes, offering unique potential for miniaturized, low-power, and wearable chiral sensors. Yet the selectivity factors remain modest compared to Faradaic and nanofluidic approaches, and the underlying relationship between chiral pore chemistry and dielectric response has yet to be systematically explored.

3.1.2. Nanofluidic Ion-Transport Sensing

Unlike planar electrodes, which rely on electron transfer (Faradaic) or interfacial capacitance (non-Faradaic) at the electrode/solution interface, nanofluidic ion-transport sensing exploits ion flux within nanoconfined CCOF channels [144]. When CCOFs are fabricated as self-standing or composite membranes, the transmembrane ionic current becomes the analytical signal [145]. The nanoscale confinement amplifies subtle chiral recognition events—such as changes in surface charge, steric hindrance, local charge distribution, or redox state of confined species—into measurable modulations of ion permselectivity and rectification [122]. Based on the dominant physicochemical gating mechanism, four transduction strategies have been developed: electrostatic gating, exclusion gating, coordination-amplified gating, and redox gating [146].

Electrostatic Gating

Electrostatic gating relies on chiral recognition altering the fixed charge density (ζ-potential) on the channel walls, thereby modulating the electrical double-layer overlap and counterion permselectivity [147]. This mechanism has been demonstrated across three distinct nanofluidic platforms—employing β-cyclodextrin/poly(ethylene terephthalate), L-penicillamine/AAO, and self-standing COF membranes, respectively—sharing a common electrostatic gating principle while differing in chiral selectors, support substrates, and target analytes.

Wu et al. post-modified a COF membrane with β-cyclodextrin (β-CD) on a poly(ethylene terephthalate) support. L-tryptophan binding formed three hydrogen bonds with β-CD, shifting the ζ-potential from −10.6 to −17.1 mV and enhancing cation transport, yielding a chiral selectivity factor of 19.2 (Figure 8a) [66]. Wang et al. exploited asymmetric charge distribution in an L-penicillamine-modified c-COF/AAO heterostructured membrane. S-naproxen binding altered the surface charge via electrostatic adsorption, modulating the ion current rectification (ICR) effect and enabling detection down to 3.88 pM (Figure 8b) [116]. More recently, Zheng et al. reported an asymmetric catalysis strategy for interfacial polymerization to fabricate self-standing cCOF nanochannel membranes. Their cCOF-2-S-PEA membrane exhibited a ζ-potential shift from −5.43 to −3.42 mV upon (R)-limonene binding, with van der Waals forces driving the enantioselective interaction, yielding a selectivity factor of 11.6 (Figure 8c) [96].

A side-by-side comparison of these three systems reveals a critical trade-off. The β-CD-modified system achieves the highest selectivity (19.2) through multivalent hydrogen bonding, yet the poly(ethylene terephthalate) support introduces non-specific ion transport that complicates mechanistic analysis. The L-Pen-modified system leverages ICR for exceptional sensitivity (3.88 pM), but the AAO substrate’s large pores (∼50 nm) limit the COF layer’s contribution to ion transport, making the signal partially attributable to the support. The self-standing cCOF membrane eliminates support-induced artifacts, enabling direct interrogation of intrinsic transport properties, yet its lower surface charge density (ζ = −5.43 mV vs. −31.3 mV for the L-Pen-functionalized c-COF) results in weaker electrostatic gating and lower selectivity (11.6) [116]. Despite these differences, none of the systems provide predictive control over selectivity: selectivity factors span 11.6–19.2 without correlation to pore size, surface charge, or binding energy, and each platform uses different analytes, precluding direct benchmarking. Moreover, the literature exclusively reports successes, with failures conspicuously absent. This lack of structure–selectivity correlations and negative data prevents rational design, relegating sensor development to empirical trial-and-error.

Exclusion Gating

Exclusion gating leverages the physical size mismatch between chiral selectors and the nanochannel aperture [148]. In this design, the chiral selector is deliberately chosen to be larger than the COF pore size, preventing its entry into the channels while allowing it to gate ion transport from the membrane surface. Zheng et al. exploited this principle using L-cysteine-modified gold nanoparticles (L-Cys-AuNPs, ~3.5 nm) in conjunction with a COF membrane featuring 1.7 nm pores (Figure 9a) [149]. L-Cys-AuNPs exhibited preferential binding to S-limonene, triggering selective nanoparticle aggregation that effectively covered the pore mouths. This steric blockade translated into a dramatic 50.4% current reduction for S-limonene—contrasting sharply with the mere 15.3% decrease observed for R-limonene—and yielded a detection limit of 36 ng L⁻¹ (Figure 9b). The elegance of this mechanism lies in its physical amplification strategy: a molecular recognition event occurring outside the nanochannel is converted into a macroscopic barrier for ion passage, thereby decoupling the recognition step from the transport readout. Yet this exterior gating leaves the intrinsic chiral environment of the COF channels unused, and the reliance on nanoparticle aggregation introduces batch-to-batch irreproducibility that complicates quantitative calibration.

Coordination-Amplified Gating

Coordination-amplified gating harnesses the powerful electrostatic effects of metal-ion coordination to amplify subtle recognition events into robust signal changes, all without altering the metal’s oxidation state [150]. Yang et al. deposited L-cysteine-modified COF nanosheets onto the tip of a quartz nanopipette [122]. L-DOPA bound to the L-Cys sites via homochiral interactions and subsequently chelated Fe³⁺ ions. The highly charged Fe³⁺ complex effectively shielded the negative surface charges of the nanochannel and partially blocked ion transport, while D-DOPA showed negligible binding and chelation (Figure 10). This strategy achieved an ultralow detection limit of 0.21 pM for L-DOPA, the highest sensitivity reported among CCOF-based nanofluidic sensors. Nevertheless, the requirement for exogenous Fe³⁺ addition complicates the workflow, and the generalizability of this amplification strategy across diverse analyte families remains untested.

Redox Gating

Redox gating introduces a distinct chemical dimension to ion-transport modulation by employing an internal redox reaction as a switch that alters membrane surface properties [151]. Critically, the analytical signal remains the ionic current—not the electron-transfer current—making this a chemically amplified non-Faradaic sensing strategy. Zheng et al. constructed a chiral copper-coordinated COF (COF-Cu(II)) nanochannel membrane (Figure 11) [152]. Enantioselective binding of reducing amino acids, exemplified by L-cysteine, triggered the reduction in framework-confined Cu(II) to Cu(I). This seemingly simple electron-transfer event induced profound physicochemical changes: the water contact angle plummeted from 58.9° to 43.0°, while the ζ-potential shifted dramatically from −4.02 to −17.83 mV. These transformations significantly modulated the ion current rectification ratio, achieving detection limits of 13–21 pg L⁻¹ for cysteine enantiomers. This design elegantly demonstrates how an internal Faradaic process can be harnessed as a chemical amplification strategy within a purely ion-transport sensing framework. However, the consumption of redox-active sites during each sensing event raises concerns about long-term stability, and the mechanism is intrinsically restricted to reducing analytes.

Within the realm of pure electrochemical sensing, Faradaic transduction relies on the redox current generated at the electrode interface, non-Faradaic sensing detects changes in interfacial capacitance upon enantioselective adsorption, and nanofluidic sensing amplifies chiral recognition events into modulations of transmembrane ion current through electrostatic, exclusion, coordination, and redox gating. Despite this diversity, the quantitative relationship between chiral binding and signal output remains elusive, and nearly all reported platforms have been validated exclusively in pristine buffers rather than complex matrices. Future efforts should combine operando spectroscopy with theoretical calculations to resolve selectivity mechanisms and broaden the analyte scope to complex multifunctional chiral molecules.

3.2. Electrochemiluminescence (ECL) Sensing

ECL combines the advantages of electrochemistry (spatiotemporal control) and chemiluminescence (high sensitivity, low background) [153,154]. In CCOF-based ECL sensing, the chiral COF undergoes redox reactions on the electrode surface to generate an excited state, which emits light upon relaxation [81,155,156]. Although the underlying excitation step involves Faradaic electron transfer, the final measurable output is luminescence intensity, which fundamentally distinguishes this mode from pure electrochemical measurements. Enantioselective binding alters the electron transfer efficiency or energy level structure, modulating the luminescence intensity. Based on the origin and integration mode of the ECL-active species, three subcategories are identified: intrinsic skeleton emitters, covalent/coordination modification, and physical loading [157].

3.2.1. Intrinsic Skeleton Emitters

Here, ECL-active units and chiral recognition sites are integrated within the CCOF backbone itself [65,90]. Enantioselective binding modulates ECL emission through two principal pathways, distinguished by the stage at which the chiral recognition event intervenes in the luminescence process.

The first pathway involves modulation of excited-state formation or decay dynamics. In this regime, the ECL emitter’s intrinsic electronic structure remains largely unchanged; instead, enantioselective binding alters the efficiency of electron transfer pathways that lead to or depopulate the excited state. Yuan et al. synthesized triPhPy⁺-(S)-CHA, where pyridinium units served simultaneously as recognition sites and ECL emitters [65]. L-arginine quenched cathodic ECL via photoinduced electron transfer (PET) from the excited pyridinium to the analyte, achieving an I_D/I_L ratio of 33.0 (Figure 12a,b); DFT calculations further showed that the L-Arg complex exhibits a lower binding energy (8.53 kcal mol⁻¹) than the D-Arg complex, and this energy difference is the direct cause of its greater stability (Figure 12c). Contact angle measurements further corroborated the stronger affinity between (S)-iCCOF and L-arginine, confirming the preferential electrostatic attraction between the chiral pyridinium cations and L-arginine (Figure 12d). Zhang et al. engineered ionic D–A CCOFs via post-synthetic modification, where imidazolium (acceptor) and pyrenyl (donor) units enabled intramolecular charge transfer (IRCT)-based ECL [90]. Here, enantioselective binding does not change the material’s band structure; rather, the formation of a more stable homochiral complex blocks the IRCT process, suppressing excited-state generation and yielding a consistent “quench-like” pattern with selectivity ratios up to 13.1 (Figure 12e–g). Liu and Kuang’s triazine-based CC-MP CCTF similarly operates through excited-state formation efficiency: the chiral pocket of (R)-2-methylpiperazine preferentially accommodates D-phenylalanine, enabling more efficient generation of the excited state via the coreactant pathway and yielding a 1.96-fold higher ECL response toward the D-enantiomer [81].

The second pathway involves direct modulation of the CCOF’s electronic structure. Here, enantioselective binding alters the HOMO–LUMO gap of the framework itself, changing the thermodynamic driving force for excited-state formation. Ruan et al. demonstrated this with (R)-PTCDA-RMP, which exhibited bidirectional ECL—D-penicillamine binding narrowed the HOMO–LUMO gap, facilitating electron transfer and enhancing ECL emission, while L-penicillamine binding widened the gap and suppressed it (detection limits: 9.74 μM for L-PA and 49.6 μM for D-PA) [82].

Overall, intrinsic skeleton emitters seamlessly couple chiral recognition with signal transduction within a single material. However, the quantitative relationship between chiral pocket geometry, binding thermodynamics, and transduction efficiency remains elusive, and current designs rely on empirical optimization. Future work should employ isothermal titration calorimetry with operando ECL spectroscopy and energy decomposition analysis to establish predictive design rules, and expand the library of ECL-active backbones beyond pyridinium, triazine, and perylene to include aggregation-induced emission or thermally activated delayed fluorescence units.

3.2.2. Coordination Modification

In this mode, highly efficient ECL emitters—primarily chiral Ru(II) complexes—are coordinatively attached to the COF skeleton, combining the quantum efficiency of metal complexes with the framework’s chiral discrimination capability. Wu and Kong’s group developed two such systems. Ph-triPy⁺-(R)-Ru(II) (Figure 13a) exhibits cathodic ECL at −1.6 V with Na₂S₂O₈ as the coreactant [88]. Its enantioselectivity reverses between analyte classes: for amino alcohols (e.g., leucinol, valinol), D-enantiomers give higher ECL intensity than L-enantiomers (D > L), whereas for amino acids (e.g., serine, aspartic acid, glutamic acid, leucine, valine), the order is L > D, with intensity ratios of 1.25–1.94 (Figure 13b,c). Mechanistically, photoinduced electron transfer (PET) involving the electron-withdrawing pyridinium cation of the skeleton and the Ru luminophore intrinsically suppresses emission; enantioselective binding increases steric hindrance around the luminophore, suppresses PET, and thereby restores ECL (Figure 13d). The second system, Ru(bpy)(S-CHDA)₂²⁺, discriminates tryptophan, leucine, methionine, threonine, and histidine with intensity ratios up to 1.75 [89]. Both systems show an unexplained reversal of enantioselectivity—between analyte classes for Ph-triPy⁺-(R)-Ru(II), and within one analyte class for Ru(bpy)(S-CHDA)₂²⁺ (L > D for Trp/Leu/Met but D > L for Thr/His)—and systematic DFT studies correlating binding energetics with selectivity direction are still lacking. Three pieces of evidence, however, allow a testable hypothesis. First, contact-angle measurements show that the preferentially bound D-leucinol renders the film markedly more hydrophilic (52.6° vs. 74.4° for L-leucinol), confirming a genuine difference in interfacial binding rather than an artifact of transduction. Second, the selectivity is strongly pH-dependent: the ratio for Leu-OH falls from 1.38 at pH 7.0 to 1.1 at pH 5.5, an effect the original authors attributed to protonation of the analyte –NH₂ group. This implicates the analyte protonation state, and hence its electrostatic contribution, in setting the recognition direction—consistent with zwitterionic amino acids (–COO^-) and essentially neutral amino alcohols (–CH₂OH) engaging the Ru(II) pocket through different dominant forces. Third, the concentration dependence is asymmetric: the ECL signal scales linearly with L-leucinol but remains essentially unchanged even at 20 mM D-leucinol, implying different binding stoichiometry or dynamics rather than simple competition for a single site. Taken together, these observations suggest that the analyte family dictates which of two opposing contributions governs PET efficiency: greater steric bulk around the luminophore suppresses PET and enhances ECL, whereas greater local polarity, conveyed by an oriented polar side chain, facilitates PET and quenches ECL. Because amino acids present an ionizable zwitterionic head group, amino alcohols, and a neutral hydroxyl terminus, the most tightly packed diastereomeric complex may simultaneously present the most—or the least—polar microenvironment to the Ru(II) centre, thereby inverting the net response. The same framework accounts for the within-class inversion reported for Ru(bpy)(S-CHDA)₂²⁺, where the recognition direction correlates with side-chain polarity (L > D for non-polar Leu and Met; D > L for polar Thr and His). We stress that this competition between steric suppression and polarity facilitation of PET is a working hypothesis rather than an established mechanism, and that the two contributions have not yet been deconvoluted. It could be tested by measuring ECL intensity ratios as a function of pH and temperature, by determining the binding thermodynamics independently (isothermal titration calorimetry or contact-angle titrations), and by DFT or energy-decomposition analysis that explicitly includes solvation and the correct protonation state of each analyte.

3.2.3. Physical Loading

The simplest and most flexible approach loads chiral selectors or ECL probes into COF pores via non-covalent interactions such as physical adsorption or host–guest inclusion, decoupling the structural requirements for chirality and electroactivity [158]. Song et al. demonstrated this concept by post-synthetically modifying β-CD onto an aminal-linked COF, with rhodamine B (RhB) loaded into the β-CD cavities as a quencher of the ECL emission through electrochemiluminescence resonance energy transfer (ERET) (Figure 14) [109]. The aminal-linked COF exhibited intrinsically strong ECL due to the non-polarizable aminal linkages, which suppressed non-radiative relaxation and gave a 4.5-fold signal enhancement over the amorphous analogue, corresponding to a relative ECL efficiency of 35.07% versus Ru(bpy)₃²⁺. In this design, L-phenylalanine, with higher β-CD binding affinity, displaced RhB from the cavities and restored the ECL signal, while D-phenylalanine caused negligible displacement, achieving a detection limit of 0.045 μM and serum recoveries of 96.9–107.2%. Zhao et al. physically absorbed an optically pure Co(III) complex ((R,R)-CHA-Co-OTs) into an ionic COF (TriPhPy⁺-BiPy) [159]. The (R,R)-Co complex gave ECL enhancement with (S)-amino alcohols and quenching with (R)-amino alcohols, achieving a record I_S/I_R ratio of 47.7 for alaninol, along with 10.2 for valinol, 8.7 for leucinol, and 2.8 for 2-amino-1-butanol. The differential coordination of enantiomers with the Co(III) center determined the ECL response direction, and good ee calibration enabled analysis with errors below 2.5%. Notably, while the physical loading strategy offers exceptional flexibility and record-high selectivity, non-covalent loading risks leaching of the chiral selector or reporter and poorly defined spatial relationships between the recognition site and the ECL emitter—issues that affect reproducibility. Future efforts should explore covalent tethering of chiral selectors for improved stability and more predictable signal transduction.

Collectively, ECL-based chiral sensing exploits a combination of mechanisms, including photoinduced electron transfer (PET), electrochemiluminescence resonance energy transfer (ERET), modulation of the framework’s intrinsic electronic structure, and competitive binding dynamics, to achieve exceptional selectivity and ultra-low background through electrical-to-optical signal conversion. Nevertheless, mechanistic interpretation remains largely phenomenological: the sign and magnitude of ECL changes are rationalized by HOMO–LUMO alignment and PET/ERET efficiency, but the quantitative contributions of individual intermolecular forces to enantioselective binding are rarely dissected. Future work should employ energy decomposition analysis (EDA) or independent gradient model (IGM) approaches, combined with operando spectroelectrochemistry, to establish a physically transparent basis for the rational design of ECL-active CCOFs with predictable enantioselectivity. Moreover, the unexplained reversal of enantioselectivity between amino alcohols and amino acids in some Ru-modified systems, as well as the limited analyte scope across all ECL platforms, warrant systematic structural investigation and expansion toward complex chiral drugs.

3.3. Photoelectrochemical (PEC) Sensing

In contrast to ECL, where electrical excitation generates light as the output, PEC sensing reverses the optical–electrical pathway [160]: light excitation of a CCOF-modified photoelectrode produces electron–hole pairs, and the resulting photocurrent serves as the analytical signal [91,97,98,99]. Enantioselective binding at the chiral interface modulates charge separation or electron-transfer efficiency, thereby changing the photocurrent [161]. This separation of excitation (light) and detection (current) sources affords low background and high sensitivity, while also enabling unique opportunities for multi-mode signal cross-validation [162].

Luo et al. developed a D-Cys-functionalized COF (D-COF) via thiol–ene click chemistry for PEC chiral sensing (Figure 15) [162]. Driven by homochiral interactions, the D-Cys moieties anchored within the pores imparted a markedly stronger adsorption affinity toward D-Cys than toward L-Cys. The adsorbed D-Cys then acted as an electron donor, significantly enhancing the photocurrent, whereas the adsorption of L-Cys was negligible and produced little change, yielding a photocurrent response ratio ΔI_D/ΔI_L of 12.0 and a detection limit of 0.13 nM. Building on this single-mode PEC sensor, the same D-COF platform was integrated with smartphone-based RGB colorimetry. The protonation of imine bonds by adsorbed D-Cys induced a distinct yellow-to-red color shift, quantifiable by smartphone RGB analysis (detection limit 69.0 nM), whereas L-Cys produced no visible change. This dual-mode design allowed independent signal cross-validation, effectively eliminating false positives or negatives, and demonstrated practical applicability in commercial L-Cys capsules.

CCOF-based PEC sensing has demonstrated the advantages of independent light excitation and multi-mode integration for reliable chiral analysis. However, the field is still in its infancy, with only a handful of reported examples and an analyte scope limited primarily to cysteine. The mechanistic interplay between chiral recognition, charge separation dynamics, and photocurrent generation—especially in the presence of colorimetric or other integrated readouts—has not been systematically investigated. Future work should broaden the library of photoactive CCOFs, extend the analyte scope to more complex chiral pharmaceuticals, and explore other hyphenated strategies to achieve orthogonal signal validation with molecular-level structural insight. Furthermore, quantitative studies on how multi-mode integration may alter the intrinsic photoelectrochemical properties of the CCOF are essential to guide rational sensor design.

3.4. Section Summary: From Physical Principles to Practical Selection

From the physical perspective of photon participation, the three signal transduction modes occupy fundamentally distinct domains: pure electrochemical (no light), ECL (electrical-to-optical), and PEC (optical-to-electrical). A quantitative comparison of the analytical performance of representative platforms—including linear range, detection limit, enantioselectivity factor, and reproducibility/durability where reported—is compiled in

Table 3. Comparative summary of representative CCOF-based enantioselective sensing platforms.

CCOF Platform	Transduction	Analyte	Enantioselectivity Factor (Definition, Concentration in mol L−1)	Linear Range (mol L−1)	Detection Limit (mol L−1)	Reproducibility/Durability	Ref.
Ph-Py+-(S,S)-DPEA	Electrochemical	7 chiral acids (Trp, Asp, Glu, Ser, Tyr, MA, MDA)	IL/ID = 1.31–2.68 (peak-current ratio, 5.0 × 10−4)	5.0 × 10−6–5.0 × 10−4 (Trp)	n.r.	RSD < 2.3% (n = 5)	[64]
triPy+Ph-(S)-Fc		Amino alcohols (Pro-OH, Val-OH, Leu-OH, Ala-OH); amino acids (Met, Ser, Pen)	IP ratio = 1.46–1.72 (peak-current ratio at 0.46 V, 1.0 × 10−3)	1.0 × 10−6–1.0 × 10−3 (Pro-OH)	n.r.	RSD < 2.6% (n = 3)	[141]
Fe3O4@COF@BSA/3DE		L-/D-Trp	IL/ID = 1.45 ± 0.008 (peak-current ratio, 1.0 × 10−4)	1.0 × 10−4	n.r.	repeatability (triplicate)	[108]
TT(D)-/TT(L)-COF		L-/D-Trp	selectivity factor = 2.6 (capacitance-response ratio, 1.0 × 10−2)	1.0 × 10−2–4.0 × 10−2	n.r.	stable over 5 sensing–regeneration cycles	[91]
CD-COF membrane		L-/D-Trp	η = RL/RD = 19.2 (transmembrane ionic-current ratio, 1.0 × 10−3)	n.r. (tested 1.0 × 10−9–1.0 × 10−2)	2.8 × 10−10	regenerable (30 min water rinse)	[66]
c-COF/AAO		S-/R-naproxen (Npx)	ΔI at +1.0 V (ion-current change, qualitative, 1.0 × 10−4)	1.0 × 10−11–1.0 × 10−4	3.88 × 10−12	stable in KCl for 15 days	[116]
cCOF-2-S-PEA		(R)/(S)-limonene	ΔIR/ΔIS = 11.6 (current-change-rate ratio, 7.3 × 10−6)	7.3 × 10−10–3.7 × 10−6	1.4 × 10−10	intra-/inter-assay RSD 3.9–4.6%/3.7–4.4% (n = 6)	[96]
L-Cys-AuNP/COF membrane		(S/R)-limonene and other flavor enantiomers	ΔIS/ΔIR = 50.5/14.9 (gating ratio, 7.3 × 10−6)	7.3 × 10−10–7.3 × 10−6	2.6 × 10−10	intra-/inter-assay RSD < 5%	[149]
CCNC nanopipette		L-/D-DOPA	(I0 − I)/I0 at −0.8 V (current-change ratio, 1.0 × 10−6)	1.0 × 10−12–1.0 × 10−5	2.1 × 10−13	reversible (2 h water rinse)	[122]
(Λ)-COF-Cu(II)		L-/D-Cys and other reducing amino acids	(I − I0)/I0 = +20.5% (L-Cys)/−11.7% (D-Cys) (current-change rate, 8.3 × 10−6)	8.3 × 10−11–8.3 × 10−5	1.1 × 10−13 (L-Cys)/1.7 × 10−13 (D-Cys)	RSD < 2.3% (intra/inter, n = 6); stable over 5 days	[152]
triPyPh+-(S)-CHA	ECL	L-/D-Arg (also Leu, Lys, Ala)	ID/IL = 33.0 (Arg) and 2.4–2.6 (Leu/Lys/Ala) (ECL intensity ratio, 1.0 × 10−3)	1.0 × 10−7–1.0 × 10−3	n.r.	RSD 0.79% (10 scans in 180 s)	[65]
MPIm+-(S)-MO		(R)/(S)-MDA and other chiral acids/amino acids/alcohols	IR/IS = 1.6–13.1 (ECL intensity ratio, 1.0 × 10−3)	1.0 × 10−7–1.0 × 10−3	n.r.	RSD < 0.94% (10 scans); inter-/intra-batch < 2.7%; stable 2 months	[90]
CC-MP CCTF		D-/L-Phe	ID/IL = 1.96 (ECL response ratio, 2.0 × 10−4–1.0 × 10−3)	2.0 × 10−4–1.0 × 10−3	1.0 × 10−4	n.r.	[81]
(R)-PTCDA-RMP		D-/L-penicillamine (PA)	sign of ECL–C regression slope: D-PA +1780/L-PA −2731 a.u. mM−1 (5.0 × 10−5–1.0 × 10−3)	5.0 × 10−5–1.0 × 10−3	9.74 × 10−6 (L-PA)/4.96 × 10−5 (D-PA)	n.r.	[82]
Ph-triPy+-(R)-Ru(II)		Amino acids and amino alcohols	IL/ID = 1.30–1.94 (amino acids)/ID/IL = 1.25–1.38 (amino alcohols) (ECL intensity ratio, 1.0 × 10−3)	1.0 × 10−4–2.0 × 10−2 (L-Leu-OH)	n.r.	RSD 2.06% (600 s continuous scans)	[88]
Ru(bpy)(S-CHDA)22+		Amino acids (Trp, Leu, Met, Thr, His)	ECL intensity ratio up to 1.75 (5.0 × 10−4)	5.0 × 10−4 (ee calibration)	n.r.	RSD 1.92% (600 s); 2.2% (3 electrodes); 3.4% (5 scans)	[89]
β-CD/aminal-COF + RhB		L-/D-Phe	IL/ID = 8972/1022 (ECL intensity ratio, 5.0 × 10−6)	5.0 × 10−8–1.0 × 10−4	4.5 × 10−8	RSD 3.2% (n = 6); stable 2 weeks	[109]
TriPyPh+-BiPy + Co(III)		Amino alcohols (Ala-OH, Val-OH, Leu-OH, Pro-OH, 2-amino-1-butanol)	IS/IR = 47.7 (Ala-OH) and 8.7 (Leu-OH) (ECL intensity ratio, 1.0 × 10−3)	1.0 × 10−5–1.0 × 10−3 ((R)-Leu-OH)	n.r.	RSD 1.29% (10 consecutive scans)	[159]
D-COF	PEC	D-/L-Cys	ΔID/ΔIL = 12.0 (photocurrent ratio, 1.0 × 10−4)	1.0 × 10−9–1.0 × 10−3	1.3 × 10−10	RSD 1.65% (PEC, 5 platforms)	[162]

Note: n.r. = not reported in the original study; ECL = electrochemiluminescence; PEC = photoelectrochemical; ee = enantiomeric excess. All concentrations are given in mol L⁻¹; values originally reported as mass concentrations were converted using M(limonene) = 136.2 g mol⁻¹ and M(Cys) = 121.2 g mol⁻¹. For each platform, the enantioselectivity factor is given with its definition and the analyte concentration at which it was measured.

. To enable direct cross-platform comparison, all concentrations in Table 3 are expressed in mol L⁻¹, and for each platform, the definition of the enantioselectivity factor together with the analyte concentration at which it was measured is specified. In practical terms, the field has converged on four characteristic paradigms—pure electrochemical (planar-electrode) sensing, nanofluidic ion-rectification, ECL, and PEC sensing—each with complementary strengths. Pure electrochemical sensing offers the simplest instrumentation and richest readout strategies, making it ideal for rapid screening and disposable sensor applications. Nanofluidic ion rectification achieves the highest sensitivity through nanoconfinement amplification and chemical signal enhancement strategies. ECL provides superior selectivity with near-zero optical background, while PEC uniquely enables multi-mode signal cross-validation through the independent optimization of light absorption and charge transfer. The selection of an appropriate sensing mode should therefore be guided by the required sensitivity, selectivity, instrumentation availability, and the nature of the target analyte, rather than by a single universal metric. Relative contributions of intermolecular forces and the distinction between binding and signal-transduction selectivity.

The enantioselectivity of CCOF-based sensors originates from a complex interplay of non-covalent interactions, including electrostatic interactions, hydrogen bonding, host–guest interactions, π–π stacking, van der Waals forces, and steric hindrance. Their relative contributions vary markedly among the systems reviewed here. For example, in the triPhPy⁺-(S)-CHA system, electrostatic attraction between the cationic pyridinium framework and the anionic carboxylate of arginine was identified as the dominant driving force, supported by DFT binding energies and contact-angle measurements [65]. In the β-CD-modified nanochannel system, multivalent hydrogen bonding between L-tryptophan and the β-CD cavity governs the enantioselective response [66], whereas in the (S)-Fc probe-mediated system, π–π⁺ interactions primarily control the incorporation of the ferrocene probe into the ionic COF, while the analyte discrimination itself relies on the hydrogen-bonding pattern between the probe and the analyte [141]. In nanofluidic platforms, steric exclusion and van der Waals interactions dominate in the L-Cys-AuNP gating design [149], whereas electrostatic and redox effects prevail in the coordination- and redox-gating configurations [150,152]. This diversity underscores that no single interaction universally dictates CCOF enantioselectivity; instead, the dominant force is dictated by the chemical identity of the chiral selector, the analyte, and the transduction geometry.

Equally important is the conceptual distinction between binding selectivity and signal-transduction selectivity. Binding selectivity refers to the thermodynamic preference of the chiral recognition site for one enantiomer (i.e., a lower binding energy of the diastereomeric complex), whereas signal-transduction selectivity denotes the difference in the measurable analytical signal (current, ECL intensity, or photocurrent) generated upon interaction with the two enantiomers. A larger analytical signal does not necessarily correspond to stronger molecular binding, as directly illustrated by the (S)-Fc system in Section 3.1.1, where the same (S)-Fc probe discriminates Pro-OH and Pen enantiomers in opposite directions, reflecting a shift from hydrogen-bonding to electrostatic acid–base pairing as the dominant interaction, and by the Ru(II)-functionalized ECL systems in Section 3.2.2, where the direction of the response (L > D vs. D > L) varies with analyte family without a simple correlation to binding affinity [88,89]. The observed analytical response is thus a convolution of binding thermodynamics, interfacial transport, electron-transfer kinetics, and transduction efficiency. Recognizing this distinction is essential for the rational design of CCOF sensors: optimizing the binding selectivity alone does not guarantee an enhanced analytical signal, and mechanistic studies should therefore evaluate both binding and transduction contributions independently, for example by combining isothermal titration calorimetry or DFT binding energies with kinetic and transport measurements.

4. Conclusions and Future Perspectives

CCOF-based electrochemical chiral sensors have matured into a mechanism- and design-oriented field, built upon three fundamental physical transduction modes—pure electrochemical, ECL, and PEC—with the pure electrochemical domain further branching into direct planar-electrode and nanofluidic ion-rectification sensing. These four complementary paradigms offer a diverse toolbox, achieving ultrahigh sensitivity, exceptional selectivity, operational simplicity for rapid screening, and built-in multi-modal cross-validation. The field’s rapid progress has, in turn, sharply defined its core challenges, each of which points toward a concrete research direction.

Electrical Conductivity. Most CCOFs are wide-bandgap insulators, restricting Faradaic and ECL efficiency [163]. Innovations in ionic backbone engineering and donor–acceptor architectures have lowered charge-transfer resistance, yet achieving high conductivity in concert with high crystallinity and well-defined chirality remains a non-trivial synthetic challenge. Future efforts should target intrinsically conductive CCOFs through mixed-valence frameworks or controlled incorporation of conductive fillers, while preserving the structural order essential for chiral recognition [164,165].

Aqueous Stability and Framework Robustness. Imine linkages, commonly used in COF synthesis, are susceptible to slow hydrolysis during prolonged potential cycling or at extreme pH [166,167,168]. Systematic long-term stability data under realistic operational sensing conditions are scarce. Priority must be given to hydrolytically robust linkages such as vinylene, dioxin, and phenazine, along with operando spectroscopic techniques for real-time monitoring of framework integrity during sensor operation [169,170,171,172].

Rigorous Validation and Analyte Scope. The field must move beyond simply reporting differential signal intensities. Standardized validation protocols—including tests against achiral structural analogs, competitive binding assays, and linear enantiomeric excess calibration curves—are needed to establish sensor reliability. Equally importantly, the current analyte scope is overly narrow, heavily focused on simple amino acids and alcohols. Extension to complex, pharmaceutically relevant molecules with multiple chiral centers remains largely uncharted, and systematic studies of analytes that fail to be discriminated are essential to delineate the true boundaries of sensor performance [173,174,175].

Electrode Modification–Performance Correlation. While multiple electrode modification strategies have been proposed—drop-casting, in situ growth, membrane fabrication, and nanopipette integration—a quantitative understanding of how the fabrication method impacts charge-transfer resistance, effective surface area, film stability, and batch reproducibility is conspicuously absent [176,177,178,179]. Comparative studies that directly correlate film morphology and fabrication conditions with key sensor metrics are essential for moving from artisanal fabrication to rational device engineering.

Real-Sample Performance and Practical Deployment. The vast majority of sensors have been validated only in pristine buffer solutions. Pioneering work demonstrating recovery in serum and urine confirms the feasibility of operation in complex biological matrices but also reveals the gap to clinical deployment. Closing this gap demands antifouling coatings, integrated on-chip sample pretreatment, and matrix-matched calibration, ideally implemented on robust, portable platforms such as 3D-printed electrodes and paper-based microfluidic chips [180,181,182,183,184,185,186,187]. Furthermore, the discrimination of pure D- and L-enantiomer solutions represents only the first step toward practical chiral analysis. Of equal importance is the quantitative determination of enantiomeric composition in mixtures, which requires the ability to distinguish chiral discrimination, selective detection of each enantiomer, and verification of enantiomeric excess (ee). Only a few of the reviewed studies have demonstrated linear ee calibration, and systematic validation in racemic or partially resolved mixtures remains largely unexplored. Future work should therefore prioritize matrix-matched calibration, spike-recovery experiments in biological fluids, and cross-validation with reference methods (e.g., chiral HPLC) to firmly establish the practical utility of CCOF-based sensors.

Scalability, Reproducibility, and Multi-Mode Integration. Solvothermal batch syntheses are inherently sensitive to subtle variations in conditions, leading to inconsistencies in crystallinity and sensor response. Scalable membrane-based fabrication routes offer a path forward, but their inter-batch reproducibility requires systematic optimization [188,189,190]. Furthermore, while multi-mode integration has demonstrated the value of signal cross-validation, the potential for mutual interference between readout mechanisms and the challenging co-design of hybrid PEC–Raman platforms represent exciting frontiers for future exploration [191,192,193,194].

In summary, CCOFs have emerged as a sophisticated and versatile platform for enantioselective electrochemical sensing. Concerted advances in synthetic methodology, mechanistic understanding, and rational device engineering will further enhance the capabilities of CCOF-based sensors. These platforms are poised to meet the growing demand for rapid, sensitive, and reliable enantioselective analysis across the pharmaceutical, clinical, and environmental sectors.