Design strategies for lab-on-a-molecule probes and orthogonal sensing

Abstract

This review highlights the currently exploited working concepts of lab-on-a-molecule probes, with a particular focus on what is required for multianalyte detection and quantification in competitive assays. Both, chemosensor and chemodosimeter approaches are considered. The multifaceted design strategies and the orthogonal protocols are evaluated in order to identify and categorise the successful conceptions and to single out unknown territory and challenges for future work.

---

Kun Chen

Kun Chen received his BSc degree in chemical engineering from Tianjin University (China) in 2005. After obtaining his MSc in organic chemistry from Nankai University (China) in 2008, he received his PhD (2014) in Prof. Michael Schmittel's group at the University of Siegen (Germany), where his research focused on developing multianalyte chemosensors both in solution and at surfaces. Since 2014 he is a postdoctoral researcher at the MPI für Kohlenforschung in Mülheim, where he works on energy conversion and photocatalysis.

Qinghai Shu

Qinghai Shu obtained his doctoral degree from University Siegen, Germany in 2012 under the supervision of Prof. Dr Michael Schmittel. His PhD thesis focused on the design of Iridium(III)-based Lab-on-a-Molecules for the detection of anions. He joined the Beijing Institute of Technology as assistant professor in the Department of Material Science and Engineering in 2013, where he focuses on energetic materials, metal organic frameworks and multichannel chemosensing.

Michael Schmittel

Michael Schmittel studied chemistry and French in Freiburg and Paris. After habilitation at the University of Freiburg he joined the University of Würzburg as an associate professor in 1993 and since 1999 holds a Chair in Organic Chemistry at the University of Siegen. His broad interests, documented in ca. 250 publications and one textbook, cover research areas as diverse as non-statistical dynamics in cyclizations, orthogonal multicomponent assembly, self-assembled nanorotors, nanoswitches for catalysis and multichannel chemosensing.

---

1 Introduction

In the Computer Age, the rapid availability of precise and reliable information is of paramount significance to make correct decisions. Consequently, the systematic development and cost-effective fabrication of sensory devices have become vibrant fields of science and technology as analytical data on (bio)medical and environmental samples as well as their accountable utilisation in appropriate fields are increasingly required. Specifically, molecular sensing in the chemical and biochemical world will provide us with qualitative and quantitative information about the occurrence and amount of molecules and ions in all their facets. The array of analytes encompasses simple gases (hydrogen, oxygen and inert gases),¹ explosives,² and constituents in the environment,³ in food,⁴ in biological systems and medical applications.⁵

In recent years, the high-density spatial integration of analytical protocols has led to breathtaking developments in the field of lab-on-chip systems.⁶ Although most of the published work, if taken literally, does rather qualify for chip-in-a-lab systems (a small chip in a large laboratory) there is equally research elaborated being concerned with autonomous lab-on-chip systems. For example, a glass-substrate based lab-on-microchip (Fig. 1, left) was introduced by Schäfer et al.⁷ using electrochemiluminescence (ECL), while Wu, Baird and coworkers⁸ reported on lab-on-chip systems patterned at subcellular dimensions to target transmembrane processes.

Fig. 1 (left) Full and zoom-in view of a multipurpose test chip device erected on a glass substrate. The enlargement shows the test area for the electrochemiluminescence assay with its via (golden) to the outside world and the a-Si:H pin diodes (silver) on top of the cover plate.⁷ Adapted with permission from ref. 7. Copyright 2007 Springer. (right) Fluorescence micrograph of micrometer-size patterned lipid bilayers containing specific ligands used to cluster mast-cell receptors.^6a Adapted with permission from ref. 6a. Copyright 2006 Nature Publishing group.

An emerging and fascinating idea to increase spatial density of sensory information is to utilise probes that are designed to detect and quantify several analytes in competitive assays. They are denoted as lab-on-a-molecule⁹ probes. The underlying idea strongly contrasts the well-known conception of probes driven by a single input signal (the analyte), which have seen a roughly 150 year history starting off with simple colourimetric tests¹⁰ developing into the state-of-art environmental and medical probes as used today.

This review analyses the current design strategies for multianalyte lab-on-a-molecule probes with the aim to categorise all successful approaches as well as to identify challenges for future work.

2 Conceptual limitations of probes

According to the IUPAC, a chemical sensor is a device that transforms chemical information, ranging from the concentration of a specific sample component to total composition analysis, into an analytically useful signal. Chemical sensors contain two basic functional units: a receptor part and a transducer part.¹¹ Within the category of molecular chemical sensors, for which the word molecular probe is frequently used, most authors will nowadays differentiate between chemosensors and chemodosimeters. A chemosensor is a molecular probe, in which the interaction with the analyte is fully dynamic, reversible and thus labile. Restrictively, there is no clear definition on the acceptable rates in the equilibration used for sensing. In contrast, a chemodosimeter operates under kinetic control in an irreversible fashion.

The IUPAC definition remains quite vague with regard to what is an excellent and what is a mediocre chemical sensor. Accordingly, sensors are normally designed to operate under well defined conditions for specified analytes in certain sample types. Therefore, it is not always necessary that a sensor responds specifically to a certain analyte. Under carefully controlled operating conditions, the analyte signal may be independent of other sample components, thus allowing the determination of the analyte without any major preliminary treatment of the sample.¹¹

Given their multitude and diversity, even single-analyte chemosensors, the objective of countless research efforts, are difficult to assess in a substantiated and scientifically valid manner. One simple reason is that the quality of a chemosensors is not measured up against well defined standards. For example, selectivities of traditional single-analyte chemosensors with their single receptor are usually reported either with regard to (1) single-analyte solutions that are simply juxtaposed for comparison, or (2) restricted competition scenarios using a solution containing two analytes, or (3) in true competition with multianalyte mixtures. It is obvious that in scenario (1) we do not necessarily detect the most strongly binding analyte, but the one generating the largest signal change in the transducer unit. In two-analyte competitions, the signal changes depend on both complexation constants and signal differences. The most realistic setting, however, would require to report on the quantification of the target analyte in the presence of several other closely related analytes as would be typical for a real-world competitive assessment. One can readily imagine that with lab-on-a-molecule probes the situation becomes even much more complicated.

Considering atom economy and high-density integration, chemists have been developing methodologies to implement multianalyte assays in molecular-scale devices. As a result, sensor arrays were advanced in solution¹² and at surfaces.¹³ On the other hand, various concepts relying on a single type of probe have been fostered toward multianalyte analysis, such as lab-on-a-molecule,⁹ displacement assay,¹⁴ relay assay,¹⁵ keypad lock¹⁶ and so on. The outstanding ability of displacement assays, relay assays and keypad locks to evaluate multiple analytes, however, does not play a major role in real-world sensing because the responses usually depend on the addition sequence (queue assay). In contrast, lab-on-a-molecule probes respond to all analytes in one solution independent of the sequence (unrestricted assay).

When the lab-on-a-molecule approach was proposed first, it was devised by de Silva^9a as a molecular logic gate capitalising on Boolean logic¹⁷ with a perspective for medical diagnostics and therapy.^9b,18 Such examples, which require a semi-quantitative response to a multianalyte mixture and aim at diagnosing an illness on the basis of several analyte concentrations exceeding pre-chosen thresholds, have been reviewed at regular intervals.¹⁷ Obviously, orthogonal and quantitative detection of two or more analytes by a single probe is even more challenging, as shown in this review. For this purpose, we critically evaluate recent work focusing on lab-on-a-molecule probes that operate on multianalyte mixtures in a fully quantitative and orthogonal manner.

With regard to competition and quantification issues, unrestricted detection assays may be classified in three categories: firstly, true lab-on-a-molecule probes, secondly, pseudo-competitive multianalyte probes and thirdly, non-competitive multianalyte probes. In this context, a true lab-on-a-molecule probe is able to quantitatively sense two or more analytes in a multianalyte mixture. Such task may be accomplished by probes with several receptor sites or only a single one – in the latter case, the orthogonality of the analytical assay has to arise from using different conditions, in particular different interrogation channels, multi-signaling, kinetic differences (time or temperature), combinatorial chemistry, etc. In contrast, a pseudo-competitive multianalyte probe exhibits different responses to two or more analytes, but fails in a fully competitive quantitative assay. Rather, only one analyte is competitively analysed in the presence of all other constituents. Furthermore, some other pseudo-competitive multianalyte probes allow qualitative identification in a competitive assay but fail in quantification. Finally, according to our classification, a non-competitive multianalyte probe displays different responses to two or more analytes, but no competitive assay is provided (Scheme 1).

Scheme 1 Various categories of single-molecule multianalyte probes.

Since the concepts of both multiplexed and multimodal sensing have been widely used,¹⁹ sometimes though with an ambiguous connotation, we start out by providing definitions for their use in the present review. Thus, probes quantifying distinct analytes in a single interrogation channel are conceived on multiplexed sensing. In some model case below, different analytes were probed in two or more channels, but their quantification was realised in one of the channels only. This type of operation, we also acknowledged as multiplexed sensing. Quite differently, multimodal sensing involves the quantification of different analytes using two or more interrogation channels.

Some probes²⁰ display different selectivity in different solvents and at different pH. From an ideal point of view, such variations do not agree with our definition of a multianalyte assay in a single solution, so that they are not covered in this review.

3 True lab-on-a-molecule probes

The basic concept of a lab-on-a-molecule aims at implementing orthogonal sensing properties in a single probe. At first glance, the approach seems only to be a nice-to-have addendum to the traditional paradigm of “one probe for each analyte”, but it embraces far more than to simply connect two or more single-analyte receptors at a transducer scaffold. Quite evidently, the development of orthogonal sensing protocols demands to recognise the fact that different analytes may interfere both with each others' receptors and their transducer's signaling. Thus, the challenge pertains not only to set up orthogonal and highly selective binding protocols at the receptor sites, but equally to implement non-interfering signaling in the transducer unit.

There are different ways to avoid interference with signaling from the transducer unit. The most common ones involve signaling at different wavelengths in UV-Vis, photoluminescence (PL) or electrochemiluminescence (ECL). Alternatively, the clean recognition of different analytes may be achieved by interrogating different channels, e.g. one analyte is quantified in the UV-Vis channel, the other in the PL channel. Some seminal assays²¹ along these considerations have been performed as early as 1990, but quantitative detection and/or competitive assay and/or mechanistic explanations have remained obscure.

In this chapter, only probes will be endorsed that are able to selectively and orthogonally recognise two or more targets in competitive assays. Thus, the probe should be able to detect and quantify different analytes in a solution of all analytically detectable constituents. Up to now, close to two dozens of true lab-on-a-molecule probes have been reported that will be analysed and discussed. Aside of multianalyte studies using lab-on-molecules, recently a lab-on-graphene with triple-channel sensing²² and lab-on-nanoparticles²³ have also generated notable reverberations.

3.1 Lab-on-a-molecules operating on multiplexed sensing

3.1.1 Differentiation by using two or more binding sites. A true multianalyte probe has to display selectivity toward two or more analytes that are present in a mixture. A straightforward design is to use a reporter unit and to endow it with several distinct receptors for each analyte of interest. However, in practice, cross interference of targets may still arise from overlapping signals of the reporter unit. One way to distinguish analytes is to make use of mathematical algorithms, in particular principal component analysis,^12,24 artificial neural networks,²⁵ and partial least squares regression.²⁶

For example, probe 1, as reported by Suzuki and coworkers,²⁷ incorporated multiple chromogenic units linked together at a moderately selective (semiselective) binding site. The three heteroatoms (O, N, S) at the receptor site exhibit different hard-and-soft-acid–base (HSAB) characteristics suggesting distinct responses to individual transition metal ions (Scheme 2). Indeed, selective recognition of cations was possible using the different optical responses at the three donor sites (N-dye: 710 nm, O-dye: 460 nm, S-dye: 380 nm). For quantification, a library of 5 × 5 samples of Cu²⁺ and Fe³⁺ was analysed with a back propagation artificial neutral network (BP-ANN). Using the BP-ANN data, the simultaneous determination of Cu²⁺ and Fe³⁺ was possible in a mixed solution at a RMSEP (root-mean-square error of prediction) of 4.3 mM for copper and 5.0 μM for iron.

Scheme 2 Structure of probe 1 and its different binding modes in the presence of cations.²⁷

Even if two or more receptor–analyte interactions are fully orthogonal, signal interference may occur when they share a single reporter unit and thus the competitive assay will simply show the sum of the individual responses. For example, probe 2 developed by Suzuki was able to selectively recognise Ca²⁺ at the O,O′-bis(2-aminophenyl)ethyleneglycol-N,N,N′,N′-tetraacetic acid site and Mg²⁺ at the β-dicarbonyl receptor under simulated biological conditions (pH 7.20, 50 mM HEPES, 130 mM KCl, 20 mM NaCl) using UV-Vis (λ_max for 2 only: 403 nm; +Ca²⁺: 358 nm; +Mg²⁺: 424 nm) and fluorescence (λ_em for 2 only: 480 nm; +Ca²⁺: 475 nm; +Mg²⁺: 485 nm).²⁸ However, when Ca²⁺ and Mg²⁺ were present in a single solution, UV-Vis and fluorescence showed the sum of all individual responses. To obtain quantitative results for each cation, mixtures of Ca²⁺ and Mg²⁺ were measured at different concentrations and a simple algorithm for analysis was applied assuming independent complexation of both analytes. Finally, it should be noted that 2 was successfully loaded into PC12 cells (PC12 is a cell line derived from a pheochromocytoma of the rat adrenal medulla)²⁹ for evaluating simultaneously the intracellular concentrations of Ca²⁺ and Mg²⁺, here, however, using their different kinetics in response to a stimulation with a mitochondria uncoupler (= carbonyl cyanide p-trifluoromethoxyphenylhydrazone).

Kaur, Singh and coworker³⁰ reported on the azacrown-ether appended thiadiazole 3 as a dual-analyte probe for Hg²⁺ and Pb²⁺. Among a variety of cations including lanthanides, only Hg²⁺ prompted a red shift in the UV-Vis channel from 491 to 535–574 nm, while Pb²⁺ caused a blue shift from 491 to 378 nm. The different response was rationalised by adverse effects on the HOMO–LUMO gap. While Hg²⁺ binds to the thiazolyl moiety, the LUMO residue of the probe, thus decreasing the HOMO–LUMO gap, Pb²⁺ associates with the crown ether, the HOMO unit, thus enlarging the HOMO–LUMO gap. UV-Vis titrations disclosed a 1 : 1 stoichiometry and comparable binding affinity for both cations, with log β = 5.50 for Hg²⁺ and 5.40 for Pb²⁺. Both analytes were analysed in a single solution by monitoring the absorbances at 574 nm for Hg²⁺ and 378 nm for Pb²⁺.

3.1.2 Differentiation by using excess of the lab-on-a-molecule. When exposed to different analytes, probes with a single receptor site usually give rise to more or less different signals due to different binding scenarios. In a dual assay, though, interference-free signaling will be disturbed by the competition for the same site. The interference may be remedied by using an excess of the lab-on-a-molecule compared to the total amount of strongly binding analytes.

For the simultaneous detection of Fe³⁺ and Cu²⁺, the Callan group³¹ developed the hybrid chemosensor 4 consisting of a CdSe/ZnS quantum dot with imine-linked phenol receptors mounted onto its surface. The system is characterised by two UV absorptions, one centred at 275 nm and the other at 355 nm. When exposed to various biologically and environmentally relevant cations in THF/HEPES (80 : 20, v/v, pH 7.0), only the addition of Fe³⁺ exhibited a substantial absorbance increase at both wavelengths, while addition of Cu²⁺ prompted a considerable bathochromic shift of both absorptions to 295 and 410 nm, respectively. In fact, these changes are so significant that they were readily detected by the naked eye: with Fe³⁺ a change from colourless to orange and with Cu²⁺ from colourless to green was witnessed. With both ions having similar binding constants, i.e. log β = 4.86 and 4.69 for Cu²⁺ and Fe³⁺, respectively, the selectivity was determined at 325 and 410 nm, because 325 nm represents an isosbestic point for all ions, except for Fe³⁺. Being a true competitive multianalyte probe, 4 was capable of detecting Cu²⁺ by a characteristic absorbance at 410 nm in the presence of Fe³⁺ and of monitoring Fe³⁺ at 325 nm in the presence of Cu²⁺ (up to a concentration of 20 μM for the individual cations).

Using Schiff bases as well, the Kumar group advanced two probes for multianalyte detection by monitoring solely the UV-Vis absorbance at different wavelengths (Fig. 2). The chromogenic chemosensor 5³² based on 1-aminoanthracene-9,10-dione showed characteristic colour changes only upon addition of Cu²⁺ (blue, λ_max at 600 nm) and Ni²⁺ (green, λ_max at 750 nm), whereas other metal ions did not affect its absorption spectrum (50 mM; pH 7.0; 10 mM HEPES; CH₃OH/H₂O, 4 : 1, v/v). Individual Cu²⁺ and Ni²⁺ concentrations were determined by plotting the absorbances at λ_max = 500 nm (due to 5), λ_max = 600 nm (due to the 5·Cu²⁺ complex) and λ_max = 750 nm (due to the 5·Ni²⁺ complex) vs. the conc. of Cu²⁺ and Ni²⁺. Competitive quantification was possible due to the similar binding constants log β_ML = 5.22 and 5.33 for Cu²⁺ and Ni²⁺, respectively. Below a total concentration of 25 μM, the UV-Vis assay was unperturbed when it was read out at well chosen wavelengths, i.e. 642 nm (for Cu²⁺) and 770 nm (for Ni²⁺).

Fig. 2 Structures of 5 and 6 (top) and the spectrophotometric response of receptor 5 to various ions (bottom).³²

In a related approach, the same authors replaced the 2-hydroxybenzaldehyde by a dipicolyl amino (DPA) group to prepare the chromogenic chemosensor 6.³³ At pH 4.0 (NaOAc–AcOH buffer), 6 experienced a characteristic colour change only upon addition of Co²⁺ (blue, λ_max at 620 nm), Ni²⁺ (yellowish pink, λ_max at 380, 460 and 510 nm) and Cu²⁺ (yellow, λ_max at 460 nm), while other cations did not affect the absorption spectrum. The differential response of 6 to Co²⁺, Ni²⁺ and Cu²⁺ depended on the pH, with only pH ≤ 4 being useful. From titration data, formation of 1 : 1 complexes with log β = 6.75, 5.86 and 7.10 (against Co²⁺, Ni²⁺ and Cu²⁺) was ascertained. As the binding constants differ by one order of magnitude, quantification in a single solution was limited to a total concentration below 30 μM. Given this limitation, absorbance measurements at 620 nm (Co²⁺) and 416 nm (ratio of Ni²⁺/Cu²⁺) allowed determination of Co²⁺ and Ni²⁺ or alternatively of Co²⁺ and Cu²⁺ within ± 5% error in two-analyte mixtures.

Similar to Kumar's approach, the Jang group explored two lab-on-a-molecule probes based on benzimidazole. With the tripodal receptor 7³⁴ they designed a dual-analyte chemosensor displaying rare metal-ion specific fluorescence enhancements at two different wavelengths. In MeCN, the free probe 7 exhibited a weak emission at 415 nm (λ_ex = 361 nm), but upon addition of Cu²⁺ the emission at 415 nm was almost 10-fold enhanced. In contrast, addition of Fe³⁺ generated an emission at 475 nm with the PL at 415 nm being fully quenched. The authors assumed that the emission enhancement at 415 nm in the presence of Cu²⁺ was owed to binding and rigidification of the ligand, which prohibited PET quenching from the lone-pair donors to the fluorophore. Likewise, emission increase at 475 nm was due to interference with PET quenching, while the wavelength shift was attributed to ICT. Simultaneous quantification of Cu²⁺ and Fe³⁺ was possible at 415 and 475 nm over a wide range of concentrations. The competitive assay worked because Cu²⁺ (K = 5.9 × 10³ M⁻¹) and Fe³⁺ (K = 1.3 × 10³ M⁻¹) have similar binding constants and both form 1 : 1 complexes with 7.

Based on a C₃-substituted benzene as platform for the benzimidazole receptors, Jang and coworkers³⁵ designed the dual-wavelength chemosensor 8 for the simultaneous analysis of Cu²⁺ and Fe³⁺ in MeCN/H₂O = 7 : 3, v/v (HEPES buffer, pH 7.1). Binding of Cu²⁺ at the receptor blocked quenching by PET leading to a ca. 1.6-fold emission enhancement at 418 nm, while binding of Fe³⁺ affected the CT behaviour triggering a new emission at 462 nm. With the help of UV-Vis and luminescence titrations, the 1 : 1 binding constants were determined as K = 1.1 × 10³ M⁻¹ and 1.0 × 10³ M⁻¹ for Fe³⁺ and Cu²⁺, respectively. Detection limits for Fe³⁺ and Cu²⁺ were determined to be as low as 1.19 × 10⁻⁵ M and 1.86 × 10⁻⁵ M. Importantly, quantification of Fe³⁺ and Cu²⁺ was not disturbed in the presence of equimolar amounts of other cations at 10 μM concentrations.

The pH-inert hetarylazo dye 9 with its N, S donor sites was developed by Kaur et al.³⁶ as a colourimetric sensor for Hg²⁺ and Fe³⁺ in MeCN while it was inert to other metal ions including lanthanides (Fig. 3). Due to the fact that the absorption of 9 at 480 nm was split into twin bands at 531 and 564 nm in the presence of Hg²⁺ or Fe³⁺, the ratio of Hg²⁺ : Fe³⁺ at a total concentration of 2 × 10⁻⁵ M was quantified in a mixture with the help of a calibration library (cf. c₍₉₎ = 3 × 10⁻⁵ M). Job plot analysis indicated a 1 : 2 (Hg²⁺ : 9) and 1 : 1 (Fe³⁺ : 9) binding model, with log β = 8.9 and 8.7. To illustrate its value for real-world applications, 9 was coated on ready-to-use paper strips. Dip-stick sensing of Hg²⁺ in aqueous solution by colour changes allowed to reach a detection limit as low as 5 × 10⁻³ M.

Fig. 3 Colours observed for 9 (i) alone and in the presence of (ii) 10⁻¹ M, (iii) 10⁻² M and (iv) 5 × 10⁻³ M Hg²⁺ in (a) solution and (b) on coated dip sticks.³⁶ Adapted with permission from ref. 36. Copyright 2011, Elsevier.

Using excess of the probe with regard to all strongly binding analytes works best if the analytes have comparable binding constants and generate nicely separated signals. This method is limited to samples with roughly known concentrations, though.

3.2 Lab-on-a-molecules operating on multimodal sensing

3.2.1 Differentiation by using excess of the lab-on-a-molecule. Using the same strategy as in Section 3.1.2 but now additionally collecting signals in two channels, Jang and coworkers prepared the Schiff base 10³⁷ as a probe for Fe³⁺ and HSO₄⁻ in aqueous solution (MeCN/Tris buffer, 8 : 2, v/v, pH = 7.6). Addition of Fe³⁺ generated complex 10·Fe³⁺ (K = 2.9 × 10⁵ M⁻¹) and triggered a 16-fold PL enhancement at 440 nm by suppressing PET and improving rigidity. In the UV-Vis channel, probe 10 selected HSO₄⁻ as a 1 : 1 adduct (K = 1.8 × 10² M⁻¹) over all other anions due to hydrogen bonding (–OH and –C=NH group) as monitored by a shift of the absorption band from 310 to 365 nm. Quantification of HSO₄⁻ and Fe³⁺ was achieved using the absorbance at 365 nm and emission at 440 nm, respectively. Notably, detection of Fe³⁺ in the PL channel was not disturbed in the presence of HSO₄⁻ and vice versa in a mixture of [Fe³⁺] = [HSO₄⁻] = 20 μM.

3.2.2 Differentiation by using different channels. Among the various kinds of chemosensors, the lab-on-a-molecule probe, which uses a single binding site but various detection channels, is not only valuable for sensing applications, but equally poses a particular design challenge. The first of its kind is the azacrown-ether modified ruthenium(II) complex 11 reported by Schmittel and Lin (Fig. 4b).³⁸ At the onset of the work, it was reasoned that the conformational space of the crown-ether receptor and its cation selectivity should be made responsive to the interrogation technique. Importantly, as a function of the interrogation method the electron demand of the signaling ruthenium unit is different (e.g. Ru²⁺vs. Ru³⁺) and as such the electron donating power of the azacrown-ether nitrogen, which depends on the coplanarity of its arrangement with the ruthenium phenanthroline, will be optimised. According to the design strategy, a steric abutment next to the nitrogen of the crown-ether receptor prevents a coplanar alignment of the macrocycle with the phenanthroline in the ground-state chemosensor. However, with increasing electron deficiency of the signaling unit, the system demands a better donation from the azacrown nitrogen. As a consequence, the crown ether is forced into the steric abutment (5, 6-CH unit of the phenanthroline) changing the conformational space of the crown ether and thus its metal ion selectivity.^38a

Fig. 4 (a) Explanation for the different selectivities of probe 11 in PL and ECL, (b) structure of 11, and (c) diagram of the relative UV-Vis, PL, ECL and redox potential changes of probe 11 in the presence of various cations.^38a

Various metal ions, such as Pb²⁺, Hg²⁺, Cd²⁺, Zn²⁺, Ag⁺, Cu²⁺, Ca²⁺, Mg²⁺, K⁺, and Na⁺ ions, were tested using UV-Vis/NIR, PL, CV and ECL as interrogation methods. In the UV-Vis channel, only Cu²⁺ ions caused a notable response in the MLCT intensity of 11 and a diagnostic metal-to-metal charge-transfer (MMCT) absorption in the near-IR (NIR) region at 1083 nm. Titration of 11 with Cu²⁺ ions at 429 nm revealed a one-step association sequence with log β_Cu²⁺ = 10.7 ± 0.2. In both CV and PL channels, 11 showed selectivity toward Pb²⁺, with PL being highly characteristic due to an emission enhancement with a parallel blue shift from 672 to 624 nm. In contrast, the ECL channel exhibited a high selectivity toward Hg²⁺ with minor interferences by Ag⁺ ions. Remarkably, a characteristic ECL enhancement (I/I₀ = 8.2) was found for Hg²⁺ ions, allowing to detect Hg²⁺ ions down to the ppm range in the presence of many other cations. Quantitative correlations enabled quantification of all three cations in a mixture without interference.

Another successful strategy for lab-on-a-molecule probes aims at utilising the different selectivities in ground and excited state. In 2009, the polymeric single-binding-site chemosensor 12 was investigated by the Callan group using two-channel interrogation.³⁹ The iminophenol receptor was readily prepared from commercially available 15 kDa poly(allylamine) (PAA) with 2-hydroxy-1-naphthaldehyde. Probe 12 is able to detect and quantify Ag⁺ and Mn²⁺ by absorbance and luminescence in THF : H₂O (9 : 1, v/v) at pH 7.0 (HEPES buffer). Upon addition of Ag⁺, notable absorption enhancements were observed at 310 and 418 nm leading to a yellow colour, which is not seen for other cations. A linear correlation between absorbance and concentration (at μM) was established. In the PL channel, Mn²⁺ showed a notable emission increase at 357 nm (along with a distinct small red shift) due to locking the imine configuration. Because the system exhibits an isoemissive point at 492 nm, a ratiometric test (I₃₈₅/I₄₉₂) was implemented to determine Mn²⁺. The different selectivity in both channels was explained by assuming that the preference for Ag⁺ in the UV-Vis channel arises from binding to the iminophenol tautomer, while in PL the ketoenamine tautomer produced by excited state intramolecular proton transfer (ESIPT) has a higher selectivity for Mn²⁺. While probe 12 is based on a powerful multianalyte strategy, unfortunately, its full potential for competitive multianalyte sensing was not demonstrated.

3.2.3 Differentiation by selective destruction of one analyte. A single-receptor probe may be upgraded to a lab-on-a-molecule by using different interrogation channels. In this context, ECL offers particular benefits, because strongly oxidative or reductive conditions are often not tolerated by all analytes. As a result, the combination of PL and oxidative ECL has received particular attention recently, because oxidative destruction under ECL conditions was used to eliminate the analyte that had dominated the response in the PL channel.

Recently, Schmittel⁴⁰ established the non-luminescent bis(ferrocenyl)phenanthroline iridium(III) complex 13 as a chemosensor for cyanide and fluoride in aqueous solution via PL and ECL. Cyanide triggered a 130-fold enhancement at 580 nm in the PL channel due to a displacement of the phenanthroline ligand by CN⁻. Under oxidative ECL conditions, the displacement was suppressed due to oxidation of CN⁻, while addition of F⁻ induced a 60-fold emission enhancement at 605 nm. The latter effect was due to preventing intramolecular electron transfer oxidation of the excited state through the ferrocenium unit by forming a strong F⁻⋯Fc⁺ ion pair (Scheme 3). A competitive assay corroborated the feasibility of quantitative detection of both anions without interference in the presence of not only each other but equally other anions (Cl⁻, Br⁻, I⁻, PF₆⁻, BF₄⁻, HSO₄⁻, AcO⁻ and H₂PO₄⁻).

Scheme 3 Sensing mechanism of probe 13 as a dual-channel lab-on-a-molecule for anion detection.⁴⁰

Taking iridium(III) complex [(pq)₂IrCl]₂ (14) as dual-channel lab-on-a-molecule, the Schmittel group has recently reported on a two-anion assay in aqueous solution at physiological pH.⁴¹ Probe 14 operates as a quantitative chemodosimeter for cyanide and acetate using PL and ECL interrogation (Scheme 4). In the PL channel, cyanide exhibited a selective response due to forming a strongly luminescent dicyano iridium complex emitting at 579 nm, while in the ECL channel only acetate displayed a strong luminescence enhancement (at 609 nm, >25-fold). The change of selectivity was revealed: in the ECL channel, cyanide did not survive under the oxidative conditions due to its low oxidation potential. In addition, the anodic potential triggered oxidation of acetate, followed by loss of carbon dioxide and formation of the methyl radical. The latter combined with the iridium(IV) complex to furnish the corresponding methyl iridium compound. In summary, the overall scenario corresponds to a Kolbe electrolysis and a follow-up radical–metal combination. In both channels, competition experiments were carried out in the presence of many anions without interference.

Scheme 4 Sensing mechanism of probe 14 as a dual-channel lab-on-a-molecule for anion detection.⁴¹

Dual-channel sensing was also useful in quantifying cysteine (Cys)–homocysteine (Hcy) and tryptophan (Trp) in an aqueous mixture of all proteinogenic amino acids at physiological pH (Scheme 5). The lab-on-a-molecule chemodosimeter 15⁴² with its reactive aldehyde group was designed by Chen and Schmittel to respond in the UV-Vis and PL channel toward Cys/Hcy through formation of the thiazolidine (or thiazinane) subunit enhancing PL at 606 nm to 210%. In contrast, Trp was selectively detected at the same wavelength in the ECL channel by fully quenching the emission. The unique quenching is based on the fact that under oxidative conditions Trp is selectively destroyed and that the slow reaction with Cys/Hcy does not interfere with the rapid ECL assay.

Scheme 5 Sensing mechanisms of probe 15 as a triple-channel lab-on-a-molecule for amino acids.⁴²

Along the same line, Sojic, Pinet and co-workers⁴³ reported on the 3,3′-functionalised guanidinium bipyridylruthenium(II) complex 16 as a differential chemosensor for L-glutamate and dihydrogenphosphate anions in acetonitrile (Scheme 6). Among a variety of anions, L-glutamate enhanced the PL intensity (up to four-fold) with a concomitant blue shift from 712 to 675 nm due to a selective binding between both guanidinium sites, while dihydrogenphosphate induced a similar response but a 3-fold enhancement. In contrast, ECL interrogation exhibited a high selectivity toward dihydrogenphosphate only, displaying a similar 4-fold enhancement at 712 nm with a blue shift to 675 nm. L-glutamate as a competitor for the same binding site was fully oxidised at the high anodic potential required for ECL generation and thus did not show up in ECL. A competitive ECL and PL assay of a mixture of anions confirmed 16 as a lab-on-a-molecule.

Scheme 6 Probe 16 as a dual-channel lab-on-a-molecule for glutamate and dihydrogenphosphate.⁴³

3.2.4 Differentiation by rendering one analyte invisible. A special trick to reduce interference in the assay is to make one analyte invisible in one of the interrogation channels, for example by using the PL quenching ability of many paramagnetic metal ions. This method additionally requires a limitation of the overall analyte concentrations and the analytes' ratio.

Besides iridium and ruthenium complexes, rhodamine derivatives have equally demonstrated their usefulness in multichannel lab-on-a-molecule chemosensing, preferentially due to intense spectral changes in UV-Vis and PL upon opening the spirolactam ring. In 2011, Nandhakumar, Tang and coworkers⁴⁴ described the rhodamine B Schiff base 17 as a dual chemosensor for both Cu²⁺ and Hg²⁺ ions in MeOH–HEPES buffered solution (pH = 7.4, 3 : 1, v/v). Cu²⁺ formed a 1 : 1 complex with 17 (K_a = 2.44 × 10⁵ M⁻¹) while triggering spirolactam opening along with a hyperchromic effect at 556 nm in the UV-Vis channel. Due to its paramagnetic nature, the strong Cu²⁺ complex did not show any significant emission in the PL channel. In contrast, 20 equiv. of Hg²⁺ induced a significant PL enhancement at 583 nm that did not interfere with other cations added in equimolar amounts to Hg²⁺. The PL titration disclosed a low binding affinity for 17·Hg²⁺ with K_a = 3.4 × 10² M⁻¹. The quantification of both cations in a mixed sample containing Hg²⁺ and Cu²⁺ = 1 : 1 was investigated. Up to 12 μM, quantification of Cu²⁺ was not disturbed by Hg²⁺ in the UV-Vis channel. On the other hand, quantification of Hg²⁺ in the presence of Cu²⁺ was feasible up to 1 mM.

Later, the same group⁴⁵ reported on a dual-analyte assay for Fe³⁺ and Cu²⁺ using the fluorochromic rhodamine B–coumarin conjugate 18 in HEPES buffered MeCN–H₂O aqueous solution (1 : 1, v/v, 50 mM HEPES, pH 7.0). With 18, Cu²⁺ formed a 1 : 1 complex (K_a = 2.44 × 10⁵ M⁻¹) that was readily recognised in the absorption mode due to the appearance of a new band centred at 563 nm indicative of spirolactam opening. The characteristic colour change from colourless to pink was reversed upon addition of EDTA. Other cations, even Fe³⁺ that formed a 1 : 1 complex (K_a = 1.7 × 10⁴ M⁻¹), did not interfere with the determination of Cu²⁺ in competitive assays, unless Fe³⁺ was added in huge excess (>25 equiv.). On the other hand, Fe³⁺ was highly selectively monitored by a significant PL enhancement at λ_em = 586 nm that was due to a reversible iron-induced ring-opening of the rhodamine's spirolactam. Although Cu²⁺ opens up the lactam even more effectively, the resultant complex is basically invisible in the assay, because its luminescence is rather low due to the paramagnetic nature of copper(II).

Using a related strategy, Miao, Zhao and coworkers⁴⁶ established probe 19 for Hg²⁺ and Cu²⁺ based on a 5-ferrocenyl-1,3,4-thiadiazole rhodamine B conjugate. In HEPES buffered MeCN–H₂O solution (8 : 2, v/v, pH = 7.2) Cu²⁺ formed a 1 : 1 complex with 19 (K = 1.73 × 10⁴ M⁻¹) leading to lactam-ring opening and a new absorption band centred at 554 nm. Due to the perceptible colour changes, 19 was a naked-eye probe for Cu²⁺. Although Fe³⁺, Cr³⁺ and Hg²⁺ reduced the absorbance toward Cu²⁺ in competitive assays, the absorption increase was still distinctive. Yet, the quantitative assay of Cu²⁺ may be somewhat hampered. One of the interfering cations, i.e. Hg²⁺, formed an equally strong 1 : 1 complex (K = 2.46 × 10⁴ M⁻¹) that was readily detectable in PL at 585 nm by a 27-fold enhancement. Importantly, the fluorescence enhancement was not disturbed by adding 2 equiv. of other cations. Even interference from Cu²⁺ was low, despite the similar binding affinities of Cu²⁺ and Hg²⁺ to 19 due to the low luminescence of the ring-opened paramagnetic complex 19·Cu²⁺.

Very recently, Yan and coworkers⁴⁷ adopted the same strategy to advance the reversible chemosensor 20 for Cu²⁺ and Hg²⁺ in EtOH–HEPES buffered solution (20.0 mM HEPES, 30 : 70, v/v, pH 7.0). Cu²⁺ triggered a ring-opening at the spirolactam with a concomitant absorption at 558 nm in the UV-Vis channel but without significant PL enhancement due to its paramagnetic nature. UV-Vis titration of 20 with Cu²⁺ indicated formation of a strong 1 : 1 complex (K_a = 1.19 × 10⁴ M⁻¹). In contrast, Hg²⁺ induced ring-opening of the spirolactam without significant change in the UV-Vis channel but distinct enhancement in the PL channel. A PL titration of 20 with Hg²⁺ equally disclosed a 1 : 1 stoichiometry for 20·Hg²⁺ (K_a = 3.4 × 10⁶ M⁻¹). Both responses were reversed by adding EDTA to remove both metal ions. Due to its higher binding affinity, the fluorogenic response to Hg²⁺ (2 equiv.) was not disturbed in the presence of 10 equiv. of other cations. Similarly, the colourimetric response to Cu²⁺ (2 equiv.) was not compromised by 10 equiv. of other cations due to the higher extinction coefficient of 20·Cu²⁺.

By virtue of the dramatic spectroscopic changes that go along with the spirolactam-rhodamine ring opening and the severe quenching effect of Cu²⁺, probes 17–20 were able to quantify Cu²⁺ aside of other metal ions. In most cases, limitations of the overall analyte concentrations and the analyte ratio were crucial to achieve orthogonal quantification.

Aside of rhodamines, other emissive molecules may also be adopted for lab-on-a-molecule probes. Inspired by the chromophore in the green fluorescent protein, Long and James⁴⁸ designed the highly conjugated dual-channel probe 21 to assess Hg²⁺ and Zn²⁺. The probe formed a 1 : 1 complex (2.3 × 10⁴ M⁻¹) with Hg²⁺ showing an intense hypochromic effect at 400 nm and a little hyperchromic effect at 550 nm in the UV-Vis channel. This response was not disturbed by other cations. On the other hand, Zn²⁺ induced a strong emission at 474 nm by forming a 1 : 1 complex with 21 at K_a = 2.1 × 10³ M⁻¹. In competition experiments, the PL enhancement was quenched by Hg²⁺ and Cu²⁺ due to spin–orbit coupling and paramagnetic effects, respectively. Using thiourea as a masking agent, the disturbance from Hg²⁺ and Cu²⁺ was eliminated and quantification of Zn²⁺ became feasible.

Very recently, Chen and Schmittel⁴⁹ developed the first triple-anion probe based on the iridium(III)–imidazolium conjugate 22. F⁻ deprotonated the phenanthroline methylene group of 22 – as indicated by H/D exchange and a 1 : 2 ratio in 22·(F⁻)₂ (log β = 9.71 ± 0.30) – leading to a new charge transfer band at 457 nm in the UV-Vis. This response was disturbed by the acidic anions H₂PO₄⁻ and HSO₄⁻. In the PL channel, H₂PO₄⁻ triggered a 7-fold enhancement along with a blue shift from 660 to 607 nm, which is readily explained by replacing the original luminescence from a low-lying triplet state (of the auxiliary ligand) with an MLCT emission. Among a series of anions, the disturbing F⁻ was masked by adding Me₃SiCl. In the ECL channel, 22 exhibited a remarkable selectivity toward AcO⁻ over all other anions by displaying an enhancement at 605 nm with a slight blue shift. With the help of single-anion calibration curves, probe 22 (10 μM) quantified F⁻ as 10.9 μM, H₂PO₄⁻ as 8.2 μM and AcO⁻ as 0.18 mM from a mixture of three anions ([F⁻] = [H₂PO₄⁻] = 10 μM, [AcO⁻] = 0.20 mM).

4 Pseudo-competitive multianalyte probes

Often authors claim to detect and quantify more than one analyte with their probe, but do not perform an analytic assay in a solution containing all target analytes. As we have emphasised above, lab-on-a-molecule probes use orthogonal protocols to detect and quantify two or more analytes without (severe) interference. In this chapter, we will acknowledge multianalyte probes that do not rigorously match the requirements of true lab-on-a-molecules, i.e. they lack a report on the quantification or even fail in a fully competitive assay. We denote these systems as pseudo-competitive multianalyte probes.

It should be noted that pseudo-competitive multianalyte probes have to exhibit clearly distinguishable spectroscopic output signals for each target analyte to fit the category. As such, their evaluation tells us a lot about currently used strategies and tricks for analyte differentiation.

4.1 Multianalyte probes operating on multiplexed sensing

4.1.1 Differentiation by using two or more (kinds of) interaction/binding sites.
4.1.1.1 Profiting from cross-interference among interaction/binding sites. In comparison with a single-site multianalyte probe, the combination of two or more receptors on a scaffold is rather straightforward, if each receptor has its own signaling unit. Nonetheless, integrating two or more receptors usually is not the simple sum of all individual receptors because cross-interference is widespread. Fedorova and coworkers⁵⁰ reported on an intelligent multitasking probe bearing two independent ditopic receptors for multianalyte analysis. The device 23 consisted of two specific functional groups, one being a benzo-15-crown-5 ether displaying selective binding toward alkaline earth metal ions such as Mg²⁺ and Ba²⁺, while the other was a phenylazathia-15-crown-5 ether showing strong affinity to the soft and heavy metal ions Hg²⁺ and Ag⁺. NMR measurements indicated that formation of the bimetallic complex was affected by negative heterotopic cooperativity due to Coulomb repulsion. Electrochemical results revealed the HOMO to be located on the crown ether motifs and the LUMO on the pyridinium fragment. Steady-state and time-resolved fluorescence disclosed that ICT processes in 23 from both crown ether residues to the N-methylpyridinium did not interfere. Binding of different cations, however, altered the ICT processes and relaxation mechanisms generating a PL at 396 nm for Mg²⁺ and Ba²⁺ and one at 490 nm for Ag⁺, Hg²⁺ and H⁺. The lifetimes were used to qualitatively differentiate the ions, thus providing a new strategy toward multianalyte probes.

Baitalik and coworkers⁵¹ amalgamated both cation (terpyridine) and anion (imidazole) receptors in probe 24. Among a series of cations, only Fe²⁺ induced an intense hyperchromic effect at 575 nm by interacting with 2 equiv. of 24. Although other cations equally triggered some changes at other wavelengths, they did not disturb the response towards Fe²⁺ (24 : Fe²⁺: other cation = 1 : 0.5 : 0.5). Among a series of anions, only F⁻ and OH⁻ generated a shift from 341 to 420 nm due to deprotonation at the imidazole. The quantification of F⁻ in the presence of Fe²⁺ and reciprocally of Fe²⁺ in the presence of F⁻ was not realised.

With cross-interference being common, one searches for ways to benefit from interferences for better differentiation of analytes. For instance, placing two or more receptors in close proximity may give rise to major interferences, by e.g. conformational hindrance, resonance interaction etc., providing additional strategic options, e.g. FRET, ESIPT etc., to read out multianalyte mixtures.

Along these considerations, a fluorescent probe with BODIPY and rhodamine units has been designed by Li and coworkers⁵² for the selective detection of Hg²⁺ and Ba²⁺ ions. Upon adding Hg²⁺ to 25 (Scheme 7), the appearance of an absorption at 558 nm attested ring opening of the rhodamine chromophore. In contrast, addition of Ba²⁺ induced a PL enhancement due to suppressing PET quenching from the calixarene to BODIPY. No competitive assay was demonstrated. Owed to the occurrence of FRET from BODIPY to rhodamine in the presence of both Hg²⁺ and Ba²⁺, this system may also be used as an INHIBIT and YES logic gate.

Scheme 7 Proposed complexation scenario of the reaction of probe 25 with different ions.⁵²

A similar system, i.e.26, was conceived for the detection of reactive oxygen species (ROS) and reactive nitrogen species (RNS), important targets due to their role in physiological processes,⁵ by Lin and coworkers.⁵³ H₂O₂ cleaved the borate ester of 26 thus reducing PET quenching along with fluorescence recovery of the coumarin residue at 460 nm. Quite differently, NO triggered the ring-opening of the spirolactam accompanied by rhodamine emission at 580 nm. A mixture of H₂O₂ and NO, lastly, promoted a FRET from coumarin to rhodamine. In sum, 26 was able to sense H₂O₂, NO, and H₂O₂/NO via characteristic fluorescence signal patterns (at three λ_em using two λ_ex) of blue–black–black, black–black–red and black–red–red, respectively (Scheme 8). Macrophage experiments demonstrated that probe 26 was able to respond to either or both of H₂O₂ and NO in a cell environment.

Scheme 8 Sensing mechanism of probe 26 for the detection of H₂O₂ and/or NO.⁵³

4.1.1.2 Profiting from mathematical algorithms. Combining lab-on-a-molecule and sensor array concepts, Margulies and coworkers⁵⁴ developed probe 27 by mounting four dye molecules with different emissive wavelengths onto a cis-amino-L-proline scaffold. The coexistence of PET, ICT and FRET made it possible to set up unique optical signatures. By monitoring the PL intensity at different wavelengths, probe 27 was able to differentiate 13 drug molecules via a 2D recognition pattern using principal component analysis (PCA). For instance, it distinguished different ratios of D-xylose and rifampicin in human urine and unknown urine–drug samples with a 97% success rate, pointing the way toward lab-on-a-molecule diagnostics. However, quantification was not fully addressed.

4.1.2 Differentiation at a single interaction/binding site.
4.1.2.1 Profiting from prototropic effects at the binding site. Because the wavelength of absorption or PL is quite sensitive to electron density, an analyte-induced (de)protonation next to the reporter unit is an effective strategy to generate different signals with different analytes.

While metal cations of similar size or from the same group in the periodic table often lead to comparable spectral changes, they may entail vastly different responses if they affect deprotonation at the chemosensor. For example, Qian and coworkers⁵⁵ successfully combined two ICT (internal charge transfer) processes to discriminate Cd²⁺ and Zn²⁺. By mounting both a DPA and a 2-picolyl binding motif onto a 1,8-naphthalimide reporter unit, the resultant probe 28 (Scheme 9) was able to distinguish Cd²⁺ by a hypochromic effect at 460 nm in the UV-Vis channel (with an isosbestic point at 420 nm) and an emission enhancement along with a blue shift from 531 to 487 nm in the PL channel. Binding of Cd²⁺ at both DPA and 2-picolyl group is able to suppress electron donation by the NH moiety. On the other hand, addition of Zn²⁺ prompted a shift from 460 to 492 nm in the UV-Vis channel and emission quenching along with a shift from 531 to 558 nm. Both the red shift and quenching of emission in 28·Zn²⁺ were attributed to deprotonation at one of the secondary amines. PL titrations disclosed 1 : 1 complexes for both 28·Zn²⁺ (1.65 × 10⁵ M⁻¹) and 28·Cd²⁺ (5.75 × 10⁵ M⁻¹). Due to the stronger binding of Cd²⁺, its quantification was not disturbed by Zn²⁺, whereas the heavy quenchers Co²⁺ and Cu²⁺ erased any emission enhancement of 28·Cd²⁺.

Scheme 9 Structure of 28 and suggested binding mode upon addition of Zn²⁺ and Cd²⁺.⁵⁵

Later, Yoon, Shin, Xu and coworkers⁵⁶ removed the 2-picolyl group to set up the ratiometric fluorescent probe 29 for Zn²⁺ and Cd²⁺ in aqueous solution (MeCN/0.5 M HEPES, pH = 7.4, 1 : 1, v/v). In alkaline medium (pH = 6.3–12.8), the emission of 29 is effectively suppressed by PET. Among a series of cations, Cd²⁺ binds to the DPA unit and the amide oxygen leading to a 21-fold PL enhancement with a concomitant shift from 483 to 514 nm. In contrast, addition of Zn²⁺ prompted the formation of the imidic acid tautomer resulting in a 22-fold enhancement with a blue shift to 446 nm. The tautomerisation of the amide group in the presence of Zn²⁺ was supported by NMR data, see Scheme 10. The binding affinity of 29 to Zn²⁺ (K_d = 1/K_a = 5.7 nM) is stronger than that to Cd²⁺ (K_d = 48.5 nM), so that quantification of Zn²⁺ was not disturbed by Cd²⁺. In contrast, detection of Cd²⁺ was only possible in the absence of Zn²⁺.

Scheme 10 Structure of 29 and suggested binding modes upon addition of Zn²⁺ and Cd²⁺.⁵⁶

In 2009, Jiang and coworkers⁵⁷ designed probe 30 as a variant of the well-known trispyridylamine chelate motif for distinguishing Cd²⁺ from Zn²⁺. Addition of Cd²⁺ generated a 40-fold PL enhancement (along with a red shift to 422 nm) by blocking PET from the amine to the quinoline fluorophore. In contrast, addition of Zn²⁺ caused a large 87 nm red shift to 476 nm along with only a slight emission enhancement (Φ_Zn/Φ₀ = 1.7) as a result of an ICT process upon metal complexation. NMR investigations indicated that the different emission responses were a result of different binding patterns (Scheme 11). Cd²⁺ inhibited PET by coordination at both nitrogen and oxygen atoms of the amide, while Zn²⁺ prompted a deprotonation at the amide. The loose binding to the DPA nitrogens was not able to fully shut down PET. Addition of Cd²⁺ to 30·Zn²⁺ induced a blue shift and enhanced PL intensity by forming 30·Cd²⁺ (K_d = 0.25 ± 0.03 pM), indicating that quantification of Cd²⁺ was not disturbed by Zn²⁺. Indeed, the PL intensity of 30·Cd²⁺ was only quenched to 10% in the presence of 1 equiv. of Zn²⁺.

Scheme 11 Structure of 30 and suggested binding modes upon addition of Zn²⁺ and Cd²⁺.⁵⁷

Only little later, the same group designed the related Zn²⁺ chemosensor 31⁵⁸ and a strategy for the detection of Cd²⁺ by employing a ratiometric metal ion displacement. Addition of Zn²⁺ induced a large PL shift from 405 to 458 nm along with a 20-fold enhancement due to chelation enhanced fluorescence (CHEF). When 31·Zn²⁺ was exposed to Cd²⁺, the emission shifted from 458 to 425 nm exhibiting an isoemissive point at 445 nm. This feature allowed a ratiometric fluorescent calibration for the quantification of the stronger binding Cd²⁺.

Liu and coworkers⁵⁹ designed the probe 32 with 8-aminoquinoline and 8-hydroxyquinoline binding sites. Addition of Zn²⁺ prompted deprotonation at the amide group triggering emission enhancement at 486 nm and a decrease at 410 nm with an isoemissive point at 450 nm (Scheme 12). On the other hand, Cd²⁺ preferred to bind at the 8-hydroxyquinoline residue and the amide's oxygen thus inhibiting PET, which resulted in a 14-fold PL enhancement at 410 nm. The latter enhancement was quenched by AcO⁻, though. Among all cations, Cd²⁺ displayed the strongest affinity (9.8 × 10⁵ M⁻¹) and even displaced Zn²⁺ from 32·Zn²⁺ (1.8 × 10⁴ M⁻¹), so that the response toward Cd²⁺ was not disturbed by Zn²⁺. The crystal structure of 32·Zn²⁺ and 32·Cd²⁺ provided convincing support for the different binding modes.

Scheme 12 Structure of 32 and its binding modes upon addition of Zn²⁺ and Cd²⁺.⁵⁹

Aside of cations, anions can also be differentiated by this strategy. For instance, Kumar et al.⁶⁰ reported on the thiacalix[4]arene-based pyrene probes 33 and 34 for sensing of F⁻ and CN⁻. Both anions induced a similar change in the UV-Vis channel by interacting with the thiourea unit. However, their PL response was different. Cyanide (0–10 equiv.) quenched the pyrene monomer luminescence while enhancing the excimer emission. On the other hand, low amounts of F⁻ (0–5 equiv.) induced a similar response as CN⁻ but higher quantities (F⁻ now used in 6–80 fold amount) enhanced both monomer and excimer emission due to a deprotonation of the thiourea group. With such a response, the authors were able to differentiate F⁻ (I − I₀) and CN⁻ (I_E/I_M) using differential monitoring. While detection of F⁻ was not disturbed by CN⁻, the addition of F⁻ interfered with the response of 33 to CN⁻ (I_E/I_M), preventing detection of CN⁻ in the presence of F⁻. Probe 34 exhibited a similar detection behaviour as 33.

In the present subchapter, strong shifts of 54–76 nm (probes 28–30 and 32) were documented when the deprotonated groups was additionally involved into complexation with the metal ion. A somewhat smaller effect was recorded, when deprotonation and complexation were decoupled (shift = 33 nm, see probe 31). In such situation, even when the binding constants for the different analytes differ by one or two orders of magnitude, full quantification is possible by limiting the overall concentration of analytes and keeping the probe in excess.

4.1.2.2 Profiting from different binding modes at the binding site. Different binding modes at a single site may lead to vastly different signals, if the site is equipped with two different reporter units operating at different wavelengths. For instance, Kim, Vicens and coworkers⁶¹ developed the calixarene-rhodamine-based FRET (fluorescence resonance energy transfer) probe 35 (Scheme 13). On one hand, Hg²⁺ and Pb²⁺ opened the lactam ring of 35 triggering a strong PL enhancement due to FRET from the excited pyrene(s) to the rhodamine. On the other hand, by coordinating to the tren-diamide unit, Al³⁺ triggered a strong emission of the pyrene excimer but hardly any rhodamine emission. Addition of Al³⁺ to 35·Hg²⁺ (39 070 M⁻¹) gradually revived the pyrene emission band and weakened the emission of rhodamine. In a reverse way, addition of Hg²⁺ to a solution of 35·Al³⁺ did not induce significant PL changes. Quantification of Al³⁺ is not disturbed by other cations while that of Hg²⁺ is only feasible in the absence of Al³⁺.

Scheme 13 Proposed mechanism of probe 35 for sensing Hg²⁺/Pb²⁺ and Al³⁺.⁶¹

By linking a rhodamine to a naphthalimide unit, 36 was shown by Kim and coworkers⁶² to operate as a light-up probe for Cu²⁺ and Zn²⁺. Addition of Cu²⁺ to 36 led to a 2 : 2 complex (K_a = 0.52 × 10⁴ M⁻³) showing the characteristic absorption and emission patterns of ring-opened rhodamines. In contrast, Zn²⁺ formed a 1 : 2 complex with 36 (K_a = 0.34 × 10³ M⁻²) accompanied by appearance of a PL band at 595 nm indicating a promoted ICT process (Scheme 14). Other cations did exert little or no effect. Due to the lower K_a of the zinc complex, addition of Cu²⁺ to a solution of 36₂·Zn²⁺ induced a gradual blue shift in PL to 550 nm, indicating that Cu²⁺ is able to displace Zn²⁺ with concomitant formation of 36₂·(Cu²⁺)₂. As a result, detection of Cu²⁺ was not disturbed by Zn²⁺ while that of Zn²⁺ was only feasible in the absence of Cu²⁺.

Scheme 14 Structure of 36 and its possible binding modes upon addition of Zn²⁺ and Cu²⁺.⁶²

Xiao and coworkers⁶³ prepared the rhodamine–naphthalene conjugate 37 to sense Hg²⁺ and Mg²⁺ by monitoring the PL intensity at different wavelengths. Addition of Hg²⁺ prompted an opening of the spirolactam ring of the rhodamine dye along with a new absorption at 558 nm and a new emission at 589 nm. In contrast, Mg²⁺ induced a different emission centred at 523 nm due to coordinating to the naphthol residue. The presence of Mg²⁺ did not affect the emission of 37·Hg²⁺ (1 × 10⁵ M⁻¹). As a result, quantification of Mg²⁺ was only feasible in the absence of Hg²⁺.

Even in the presence of a single reporter unit, different binding modes may also lead to distinct signals with different analytes. Pyrene is famous for its excimer vs. monomer emission properties. Along those lines, Lee and coworkers⁶⁴ developed probe 38 by merging the tryptophan unit (Trp) with a pyrene reporter unit. As a result, the PL response allowed detection of Pb²⁺ and Hg²⁺ at different wavelengths. Notably, the 1 : 1 complex of Pb²⁺ and 38 enhanced the PL of the monomeric pyrene by inhibiting PET between pyrene and Trp. In contrast, Hg²⁺ interacted with 2 equiv. of 38 displaying a typical pyrene excimer emission. Because 38₂·Hg²⁺ (K_a = 1.12 × 10¹³ M⁻²) was much stronger bound than 38·Pb²⁺ (3.11 × 10² M⁻¹), Hg²⁺ could easily be detected in the presence of Pb²⁺.

Not only pyrene dimer formation but also its separation may be used for detecting different analytes. Mounting pyrene via a click reaction onto a binaphthol scaffold, as explored by Zhu, Cheng and coworkers,⁶⁵ furnished a new probe for Ag⁺ and Hg²⁺. Ag⁺ formed a 1 : 1 complex (K_a = 3.4 × 10⁵ M⁻¹) with 39 (Scheme 15) by coordinating to both triazole units which entailed some spatial separation of both pyrene residues. As a result, the PL of the pyrene excimer was reduced and that of the pyrene monomer enhanced allowing a ratiometric quantification of Ag⁺ in the presence of many metal ions. This assay was disturbed by Hg²⁺, though, because it formed a 1 : 2 complex with 39 at a higher affinity constant (4.3 × 10⁸ M⁻²), which led to a quenching of both monomer and excimer emission. Thus, the assay for Ag⁺ only worked in the absence of Hg²⁺.

Scheme 15 Structure of 39 and its possible binding mode upon addition of Hg²⁺ and Ag⁺.⁶⁵

Similarly as in Section 4.1.2.1, the different binding modes at a single binding site are subjected to competition between the different analytes. To nonetheless implement quantification schemes for more than one analyte, the assay needs to be restricted to analytes with similar binding affinity. Otherwise excess of the probe is required. As a way out, the analyte exhibiting the strongest affinity may be masked when quantifying the other analyte.

4.1.3 Differentiation by forming vastly different products. Vastly different products give rise to highly distinct signals, a protocol generally exploited in chemodosimeters. The latter recognise analytes based on irreversible product formation and, along these lines, chemodosimeter approaches have been applied not only for one but also for several analytes in multianalyte detection. To ensure high sensitivity and dynamic range, they are often built using luminophores with excellent quantum yield and reactive sites for facile conversion. Rhodamine and BODIPY are regularly used due to their excellent optical and chemical properties.

Very recently, the chromo-fluorescent probe 40 for the detection of Zn²⁺ and Fe³⁺ was designed by Luxami et al. (Scheme 16).⁶⁶ In aqueous solution (20 μM HEPES/MeOH, 1 : 1, pH = 7.2), addition of Zn²⁺ and Fe³⁺ triggered the appearance of UV-Vis absorptions at 380 nm and 335–500 nm, respectively, while other metal ions did not produce any significant colour changes. In the PL channel, Zn²⁺ and Fe³⁺ caused emissions at 440 and 585 nm, respectively. The difference arises, because Zn²⁺ binds to 40 at the hydroxy, imine and amide oxygen atoms without eliciting ring opening of the spirolactam. On the other hand, Fe³⁺ induced a ring-opening that was followed by an irreversible hydrolysis (Scheme 16). Individual Zn²⁺ and Fe³⁺ concentrations were determined by monitoring the PL intensity at 440 nm and 585 nm, respectively. Addition of Fe³⁺ to a solution of 40·Zn²⁺ resulted in an instantaneous PL quenching at 440 nm and enhancement at 585 nm suggesting to use 40·Zn²⁺ as a good displacement ratiometric probe for Fe³⁺. In contrast, the optical change produced from 40 and Fe³⁺ was not reversed by Zn²⁺. Thus, quantification of Zn²⁺ is only feasible in the absence of Fe³⁺.

Scheme 16 Sensing mechanism of probe 40 as a dual-wavelength probe for cation detection.⁶⁶

Very recently, Emrullahoğlu and coworkers⁶⁷ connected a rhodamine and BODIPY unit via a hydrazone group. Due to the isomerisation about the C=N linkage and the spirolactam form of the rhodamine dye, probe 41 is non-emissive. Hg²⁺ prompted a ring-opening of the spirolactam entailing a PL enhancement at 585 nm, whereas Au³⁺ induced not only ring-opening but also C=N bond hydrolysis leading to emission recovery of the BODIPY as carbaldehyde. As a result, probe 41 displayed emissions at both 585 and 506 nm in the presence of Au³⁺, which were not disturbed by any other cation including Hg²⁺. Actually, complex 41·Hg²⁺ was not stable in the presence of Au³⁺ as indicated by PL enhancement at 506 nm suggesting C=N bond cleavage. Thus, quantification of Hg²⁺ was disturbed by Au³⁺. In fact, an excess of probe may allow to set up 41 as a true lab-on-a-molecule, but such experiment was not reported.

Besides rhodamine and BODIPY, other luminophores have also been adopted. Qian, Xu and coworkers⁶⁸ developed the real-time chemodosimeter 42 with large Stokes shifts to discriminate ClO⁻ and OONO⁻ from other reactive oxygen species (ROS). In buffered solution at pH 7.4 (0.1 M sodium phosphate, 1% DMF as cosolvent), ClO⁻ triggered PL enhancements at 460 (71-fold) and 570 nm (63-fold) that were assigned to the formation of an intermediate and the final product, respectively. On the other side, OONO⁻ only generated PL increase at 460 nm (68-fold) (Scheme 17). A quantitative assay of ClO⁻ was performed using the intensity ratio at 460 nm and 570 nm, but those of OONO⁻ and both analytes were not realised.

Scheme 17 Mechanism of 42 as a dual-wavelength probe for ClO⁻ and OONO⁻.⁶⁸

It is well-known that the PL of iridium complexes originates from LC and/or MLCT transitions leading to characteristic emission spectra. Inspired by this wealth of information, a chemodosimeter based on iridium(III) complex 43⁶⁹ was reported by Chen and Schmittel that is able to detect Hg²⁺ and ClO⁻ simultaneously at pH = 6 (aqueous solution) in a single sample. Oxime residues were purposely selected as reactive sites and attached at the cyclometalated ligands of the iridium luminophore. The design was chosen because (1) PL in 43 should be low due to PET and oxime isomerisation, (2) the oxime is known to undergo facile conversion to other functional groups, and (3) the resultant functionality should influence the PL spectra in terms of shape and intensity. At pH 5–7, the oxime groups of 43 experienced dehydration selectively in the presence of Hg²⁺, while at pH > 5 and in the presence of ClO⁻ they were increasingly oxidised to the carboxylate, resulting in an enhancement at two distinct wavelengths (505 nm for Hg²⁺ and 589 nm for ClO⁻) (Scheme 18). From the PL intensity ratio I_577nm/I_504nm at pH 6, the ratio of both analytes was obtained and their individual concentrations determined after assessing the concentration of Hg²⁺ in DMF–phthalate buffer at pH 5. At present, 43 is a selective two-analyte chemodosimeter for the largest library of analytes (33 species) ever reported.

Scheme 18 Sensing mechanism of probe 43 as a dual-wavelength probe for Hg²⁺ and ClO⁻.⁶⁹

In summary, use of chemodosimeters provides a powerful strategy toward distinguishable signals for each analyte with high selectivity and quantification of each analyte. Reactions with large favourable free energy change (ΔG) and products with high quantum yield ensure highly sensitive detection. Nevertheless, the simultaneous quantification of two or three analytes in a single solution remains a main challenge. If orthogonal reactions are available then a fully competitive assay can be readily achieved by using an excess of the probe. However, if one product reacts further on with another analyte (see 42), then quantification of all analytes becomes difficult and sometimes even impossible.

4.1.4 Differentiation by different rates (chemodosimeter). Kinetic differences may serve as another venue to differentiate analytes in multianalyte sensing, but so far none of the approaches represents a true lab-on-a-molecule, either due to the lack of competition or quantification. Nonetheless, the examples below convey some important insight how to use kinetic differences for differentiation in sensing.

Strongin et al.⁷⁰ reported on probe 44 to discriminate glutathione (GSH), cysteine (Cys) and homocysteine (Hcy) from other amino acids under physiological condition. All three targets, GSH, Cys and Hcy, underwent a Michael addition to trigger a PL enhancement (λ = 377 nm) by suppressing the alkene-related PET. However, only the adducts generated from Cys and Hcy are able to undergo intramolecular cyclisation at the phenol ester site that are distinct in their kinetic rates (Scheme 19). Because the generated 2-(2′-hydroxy-3′-methoxyphenyl)benzothiazole (HMBT) experiences tautomerisation upon ESIPT with its phenol (λ = 377 nm) and keto (λ = 487 nm) tautomers displaying emissions at distinct wavelengths, the concentration of Cys, GSH and Hcy may be determined by monitoring the PL intensity over time. With cetyltrimethylammonium bromide (CTAB) to enhance the reaction rate in a micellar medium, the concentration of Cys, GSH and Hcy was read out at 487 nm after 9 min, 377 nm after 9 min and at both 487 and 377 nm after 9 min, respectively. A competitive assay was not realised.

Scheme 19 Sensing mechanism of 44 as a multianalyte probe for Cys–Hcy detection.⁷⁰

More recently, the same group⁷¹ used the rhodol thioester 45 in combination with S–N acyl transfer and ESIPT to differentiate GSH and Hcy/Cys (Scheme 20). The probe displayed a weak emission at 587 nm due to PET quenching from the nitrothiophenoxy unit. GSH induced a transthioesterification that removed the nitrothiophenol thus recovering the PL at 587 nm. Hcy/Cys prompted an analogous reaction followed by an S, N-acyl shift to trigger ESIPT at the HMBT along with a PL enhancement at 454 nm. Other amino acids did not induce any significant PL enhancement. Neither did addition of Cys to a mixture of 45 and GSH cause any change indicating that 45 can differentiate GSH from Hcy/Cys. The addition of a mixture of GSH and Cys to 45 led to dramatic enhancements at both 454 and 587 nm. This result as well as experiments on human breast cancer cells indicated that the probe can detect GSH and determine the overall concentration of Hcy/Cys in a single solution.

Scheme 20 Sensing mechanism of 45 as probe for GSH and Cys/Hcy.⁷¹

Guo and coworkers designed the coumarin–hemicyanine probe 46⁷² with its three reaction sites as a fluorescent chemodosimeter for Cys and GSH. Cysteine induced a S_NAr-based halogen–thiol exchange at site 1 followed by an thiol–amine rearrangement with the liberated thiol quickly attacking site 2. In the UV-Vis channel, the whole process led to a hypochromic effect at 535 nm and a hyperchromic effect at 360 nm. The concentration of cysteine was monitored by a new PL band at 420 nm. Hcy intrigued a similar substitution and rearrangement but a slower attack at site 2. As a result, it quenched the emission of probe 46. GSH initiated a S_NAr at site 1 and a slow attack at site 3 accompanied by a hypochromic effect at 535 nm and a hyperchromic effect at 450 nm in the UV-Vis channel. In the PL channel, the new emission at 512 nm was useful to monitor the concentration of GSH. In the presence of high concentrations of Cys/Hcy/GSH (100 μM/100 μM/1 mM), 46 responded first to Cys due to its higher reactivity than that of GSH. In the presence of low concentrations of Cys/Hcy/GSH (1/1/10 μM), 46 elicited a PL enhancement at 512 nm. Unfortunately, quantitative detection in competition experiments was not addressed.

Inspired by the large Stokes shifts and substituent-dependent emissions of pyronin dyes, the same group⁷³ designed the chemodosimeter 47. In the UV-Vis channel, GSH triggered a slight red shift to 593 nm in line with thiophenol/GSH exchange at the thiopyronin. On the other hand, Cys/Hcy (15 equiv.) caused a hypochromic effect at 588 nm along with the emergence of a band centred at 460 nm indicating the generation of aminopyronin dyes. Probe 47 was inert to other thiol-free amino acids evocative of a reaction pathway involving first S_NAr and then S,N-rearrangement for Cys/Hcy. Individual quantification was realised by adopting different excitation wavelength. In the PL channel, after excitation at 580 nm, GSH (15 equiv.) elicited a 63-fold enhancement at 622 nm, while Cys/Hcy did not generate any increase. When excited at 455 nm, though, Cys/Hcy (15 equiv.) triggered a 90-fold enhancement at 546 nm but GSH only induced a poor emission at 622 nm. Cell imaging experiments indicated the feasibility of a simultaneous detection of both GSH and Cys/Hcy, however, the corresponding quantification scheme was not provided.

The examples demonstrate that kinetic differentiation is a powerful strategy to distinguish Cys/Hcy and GSH. By virtue of different rates (44, 45 and 47) or different sites of attack (46), analytes are differentiated. Probe 45 can simultaneously detect both GSH and Hcy/Cys in a dilute human serum sample. Moreover, excess of 46 might be able to quantify both GSH and Cys in a single solution.

4.1.5 Differentiation by inducing superstructures. James and Jiang⁷⁴ developed the boronic acid–pyrene conjugate 48 as an amphiphilic probe to differentiate glucose and fructose. Fructose interacted with the boronic acid forming a 1 : 1 complex (353 ± 21 M⁻¹), which entailed a PL enhancement of the pyrene monomer at 377 nm. On the other hand, glucose interacted with two equiv. of 48 forming a 1 : 2 complex (1.9 × 10⁶ M⁻²) that assembled into an aggregate (diameter: 2000 nm) at pH = 10 by virtue of pyrene–pyrene π–π stacking entailing a characteristic excimer enhancement at 510 nm. With phenylboronic acid (10 mM) as a “knock-out”, the interference of fructose (5 mM) could be reduced to 4%. Unfortunately, quantification of fructose in the presence of glucose was not addressed.

Inducing a superstructure is a smart strategy to generate vastly different signals for different analytes. Adversely, such approach requires high concentration of the probe. In the case of 48, 0.1 mM of probe was used. For quantification of two analytes, the analyte that is unable to induce a superstructure should exhibit the higher affinity with the probe and be quantified first. Quantification of the second analyte can then be realised by masking the first one.

4.2 Multianalyte probes operating on multimodal sensing

4.2.1 Differentiation by using two or more (kinds of) interaction/binding sites. Sometimes pairs of analytes display “synergetic interactions” leading to quite a different response from that of the individual compounds. For instance, Ihmels and coworkers⁷⁵ reported on benzo[b]quinolizinium 49 and its utility to detect Hg²⁺ and/or calf thymus (ct) DNA by inducing a 7 nm blue shift and 15 nm red shift in the UV-Vis channel, respectively. In the PL channel, probe 49 proved to be almost non-luminescent due to PET quenching (path a) and rotational deactivation about the C–N bond (path b). Notably, the very low emission intensity of 49 in aqueous buffer solution (HEPES, 25 mM, pH = 7.0) was hardly influenced by complexation of Hg²⁺ (K = 1.0 × 10⁵ M⁻¹) or by intercalation into DNA (K = 1.6 × 10⁵ M⁻¹). When both Hg²⁺ and ct-DNA were present, though, the PL intensity was 8-fold enhanced. It seems that only in the presence of both inputs the two relevant deactivation pathways are suppressed, so that this signature can be used as a 2-input logic gate. Quantification of DNA in the presence of Hg²⁺ was not addressed.

4.2.2 Differentiation by selective blanking of one competitor. As already discussed in chapter 3, the multichannel approach is an effective strategy to quantify different analytes in a competitive manner. In principle, probes have a potential for working as a true lab-on-a-molecule when different analytes exhibit a pronounced response in one channel but are silent in the other channels. If they fail in one of the competition experiments, we denote the probe as a pseudo-competitive one. For example, Chang and coworkers⁷⁶ used the cyclam 50 bearing two different reporter units for the selective recognition of Hg²⁺ and Cu²⁺ in both UV-Vis and PL channels. Hg²⁺ induced a slight blue shift (−11 nm) in the NBD (7-nitrobenzo-2-oxa-1,3-diazolyl) region with some splitting of the pyrene absorptions in the UV-Vis channel. In contrast, Cu²⁺ generated a hyperchromic effect in the pyrene region and a hypochromic effect with a blue shift (−11 nm) in the NBD region. In the PL channel, Hg²⁺ induced a 10-fold emission enhancement at 538 nm by blocking PET and PCT of the NBD residue,⁷⁷ while a marked quenching at 538 nm was observed (90%) with Cu²⁺ due to its paramagnetic nature. However, no true lab-on-a-molecule was developed. Addition of Cu²⁺ into a solution of 50·Hg²⁺ quenched the emission at 538 nm suggesting that Cu²⁺ formed a more stable complex with 50 than Hg²⁺. As a result, detection of Hg²⁺ was disturbed by Cu²⁺.

By attaching three pyridyl residues on a phenyl scaffold, Wei, Yu and others⁷⁸ generated a new probe for the dual detection of Zn²⁺ and Cu²⁺ (Scheme 21). In the UV-Vis channel, Zn²⁺ induced a slight absorbance increase (ca. 20% at 244 nm) while Cu²⁺ caused an intensive hyperchromic effect (2.9-fold) at the same wavelength. Because Zn²⁺ coordinated to the two ortho-pyridine rings of 51, it induced a 6.4-fold PL enhancement due to hindering the rotation about the C–C bond. In contrast, Cu²⁺ quenched the emission by a factor of 4.3 due to its paramagnetic nature. Among a series of cations, small interferences were reported: Cd²⁺ enhanced while Hg²⁺ slightly attenuated the PL intensity of 51. Addition of Cu²⁺ to a solution of 51·Zn²⁺ resulted in a gradual emission decrease at 375 nm due to formation of 51·Cu²⁺. As a result, the UV-Vis channel may be used to quantify Cu²⁺ and the PL channel for detecting Zn²⁺ in the absence of Cu²⁺.

Scheme 21 Structure of 51 and its possible binding modes upon addition of Zn²⁺ and Cu²⁺.⁷⁸

Using the distinct metal-ion binding properties of porphyrin and DPA, 52 and 53 were designed by Jiang to selectively detect Cu²⁺ and Pb²⁺.⁷⁹ In MeOH/CH₂Cl₂ (1 : 1, v/v), Pb²⁺ formed a 3 : 2 complex (2.1 × 10⁴ M⁻⁴) with 52 that showed a new absorption at 466 nm and an absorbance decrease at 417 nm as well as a new emission band at 604 nm and PL quenching at 650 nm. In contrast, Cu²⁺ formed a 2 : 1 complex (7.4 × 10⁵ M⁻²) with 52, which did not prompt any significant changes in the UV-Vis channel but PL quenching at 650 nm. Cu²⁺ exhibited a higher binding affinity than lead as it displaced Pb²⁺ from 52·Pb²⁺. The concentration of Cu²⁺ was determined by PL while Pb²⁺ may be quantified in both PL and UV-Vis channels in the absence of Cu²⁺. Probe 53 exhibited a similar detection behaviour.

Li and coworkers⁸⁰ reported on the amino pyridine and benzimidazole conjugate 54 as a multianalyte sensor. Among a series of cations, only Al³⁺ hindered the ESIPT in 54 and enhanced the emission along with shifts from 395 (54_N*) and 528 (54_T*) nm to 430 and 504 nm, respectively (Scheme 22). On the other hand, Cu²⁺ quenched the PL of 54 due to its paramagnetic properties. In the UV-Vis channel, Al³⁺ induced a red shift from 345 to 355 nm without any change in absorbance, while addition of Cu²⁺ prompted a dramatic hypochromic effect at 345 nm and a hyperchromic effect at 390 nm after a 45 nm red shift. Job plots indicated a 1 : 1 stoichiometry for both 54·Al³⁺ (1.35 × 10⁶ M⁻¹) and 54·Cu²⁺ (6.53 × 10⁵ M⁻¹). Quantification of Al³⁺ was not disturbed by other cations in the PL channel. Due to the higher affinity of 54·Al³⁺, however, Al³⁺ disturbed the detection of Cu²⁺ in the UV-Vis channel. Upon addition of F⁻ and EDTA to 54·Al³⁺ or 54·Cu²⁺, fluorescence and absorption spectra were identical to those of 54 demonstrating the reversibility of binding. Complex 54·Al³⁺ exhibited a high selectivity for F⁻ over Cl⁻, Br⁻ and I⁻.

Scheme 22 ESIPT mechanism of probe 54.⁸⁰

The key to successful multianalyte sensing is to implement a differential selectivity. As one of the strategies, a strongly binding analyte is made “silent” in one channel, e.g. Cu²⁺ as a strong quencher usually is unable to generate PL enhancement, while it is the dominant analyte in another channel. Other cases of “blanking an analyte”, but less in numbers, are known for UV-Vis channel detection. For example, the binding of Cu²⁺ to 51 resulted in a marked hyperchromic effect in the absorption spectrum, while Zn²⁺ as the competitor only induced a negligible increase.

4.2.3 Differentiation by different rates (chemodosimeter). Obviously, as the temperature affects both reaction rate and product partitioning, the product ratio will depend on the temperature. Using such strategy, Jiang and coworkers⁸¹ recently reported on the dual-channel chemodosimeter 55 for anion detection (Scheme 23). Among the various anions, only the addition of F⁻ in DMSO, due to a deprotonation of 55, entailed distinct colour changes, as expressed by a change from pale yellow to pink. With CN⁻, 55 underwent cyclisation at room temperature to a PL-active benzoxazol with hardly any interference from other anions, except for F⁻. Thus, in competitive assays containing F⁻ and CN⁻, the protocol suggested to determine first F⁻ by absorption spectroscopy, then for evaluation of CN⁻ the system was heated to 70 °C for 160 min to effect formation of the cyclised product. No correlation for quantification of cyanide was provided.

Scheme 23 Proposed sensing mechanism of probe 55 for detecting fluoride and cyanide.⁸¹

4.2.4 Differentiation by inducing superstructures. Formation of superstructures through controlled aggregation after analyte binding provides another highly sensitive way for analysis. Kobayashi, Shen and coworkers⁸² designed the BODIPY probe 56 for sensing Cu²⁺ and Hg²⁺. Due to the coordination preferences of Cu²⁺, the square-planar complex [Cu²⁺(56)₂] underwent J-aggregation as visible from a new absorption at 569 nm and a new emission at 570 nm. The aggregation resulted in a high sensitivity for Cu²⁺ (1 μM) and a notable selectivity over other cations including Hg²⁺ (well visible in the UV-Vis channel). Importantly, Hg²⁺ also formed a 1 : 2 complex with 56 that was silent in the UV-Vis channel but showed a 7-fold PL enhancement at 510 nm. This response was not disturbed by any of the usual cellular ions (Mg²⁺, Na²⁺, Ca²⁺ and Zn²⁺). Quantification of Hg²⁺ in the PL channel, however, was only feasible in absence of Cu²⁺.

5 Non-competitive multianalyte chemosensors

Chemosensors in this category are often denoted as two-analyte chemosensors in their original papers but they do not exhibit/prove selectivity in competitive assays. Thus, they are incompatible with single-solution multianalyte quantification. Herein, a competitive assay would mean either the competition among different target analytes or among analyte and another unrelated species. If either of the competition was missing, the example was assigned to chapter 5. For a synopsis of the operating principles, probes 57–86 are listed in Table 1 along with their target analytes, the detection method(s) and spectroscopic responses (Scheme 24).

Table 1 Multianalyte detection using probes 57–86

Probe	Analytes	Detection method	Response
57 ^83	Hg^2+	UV-Vis	Red shift from 421 to 515 nm
		PL	Red shift from 475 to 508 nm with an enhancement factor of 30
	Cu^2+	UV-Vis	Red shift from 421 to 520 nm
		PL	Red shift from 475 to 590 nm with an enhancement factor of 8
58 ^84	Na^+	UV-Vis (20 °C)	Hyperchromic effect at 566 nm
	K^+	UV-Vis (50 °C)	Hyperchromic effect at 566 nm
59 ^85	H^+	PL	Quenching at 420 nm and enhancement at 448 nm (Ratiometric response)
	Cu^2+		Quenching at 420 nm
60 ^86	Cu^2+	UV-Vis	Hypochromic effect at 489 and hyperchromic effect at 383 nm (Isosbestic point at 410 nm)
	Hg^2+		Hypochromic effect at 489 and hyperchromic effect at 413 nm
61 ^87	Zn^2+	PL	Quenching at 411 nm and enhancement at 458 nm (Ratiometric response)
	Cu^2+		Quenching at 411 nm
62 ^88	Zn^2+	PL	Enhancement (16–18 fold)
	Ni^2+		Quenching
63 ^89	Hg^2+	PL	Enhancement at 421 nm (8.3-fold)
	Cu^2+		Enhancement at 405 nm (2.8-fold)
	Zn^2+		Enhancement at 393 nm (3.6-fold)
64 ^90,91	Hg^2+	UV-Vis	Hypochromic effect at 575 nm (5-fold reduction, blue-shift to 535 nm) and a new band appearing at 405 nm
	Fe^3+		Hypochromic effect at 550 nm
	Cu^2+		Hypochromic effect at 575 nm
	Fe^3+	CV	Anodic shift (28 mV)
	Cu^2+		Anodic shift (12 mV)
	Hg^2+		Anodic shift (31 mV)
65 ^91	Cu^2+	UV-Vis	Hypochromic effect at 585 nm and blue-shift (from 585 to 560 nm)
	Fe^3+		Hypochromic effect at 555 nm and blue-shift (from 585 to 555 nm)
	CN^−		Hypochromic effect at 585 nm
	Cu^2+	CV	Anodic shift (11 mV)
	Fe^3+		Anodic shift (18 mV)
	Hg^2+		Anodic shift (24 mV)
66 ^91	Fe^3+	UV-Vis	Hypochromic effect (10-fold reduction) and blue-shift (from 584 to 554 nm)
	Cu^2+		Hypochromic effect (1.5-fold reduction) and blue-shift (from 584 to 574 nm)
	CN^−		Hypochromic effect at 584 nm
67 ^92	Cu^2+	UV-Vis	Hyperchromic effect at 557 nm (>30 fold)
	VO^2+	PL	Enhancement at 583 nm (>40 fold)
68 ^93	Cu^2+	PL	Quenching with shift from 492 to 471 nm
		UV-Vis	Blue shift from 420 to 373 nm
	Ni^2+	PL	Quenching
		UV-Vis	Red shift from 420 to 478 nm
69 ^94	Hg^2+	UV-Vis	Blue shift from 518 to 499 nm
		PL	Blue shift from 548 to 532 nm (7-fold enhancement)
	Ag^+		Enhancement at 548 nm (1-fold)
70 ^95	Cu^2+	PL	Quenching at 517 nm
	F^−		Quenching at 517 nm (0–5 μM) Enhancement at 478 nm (5–350 μM)
		UV-Vis	Hypochromic effect at 265 and 342 nm Hyperchromic effect at 447 nm (Isosbestic point at 379 nm)
71 ^96	Ag^+	PL	Enhancement (56%) at 458 nm
	Cu^2+		Quenching (89%) at 458 nm
		UV-Vis	Blue-shift from 390 to 357 nm (Isosbestic point at 281 nm)
72 ^97	Fe^2+	PL	Enhancement (F516/F650 from 16.4 to 26.67)
	Hg^2+		Quenching (F516/F650 from 16.4 to 1.24)
73 ^98	Cu^2+	UV-Vis	Hypochromic effect (441 nm)
	F^−		Hypochromic effect (441 nm) Hyperchromic effect (572 nm) (Isosbestic point at 480 nm)
	Cu^2+	PL	Quenching
	F^−		Quenching
74 ^99	Cu^2+	UV-Vis	Hyperchromic effect at 527 nm
	Fe^3+	PL	Enhancement at 552 nm
75 ^100	K^+	PL	Quenching at 430 nm with 20 blue-shift
	F^−		Quenching
76 ^101	Hg^2+	UV-Vis	Hyperchromic effect at 248 nm (4.9-fold) and 395 nm (9.7-fold)
	Zn^2+	PL	Enhancement at 390 nm (10.5-fold)
77 ^102	F^−	UV-Vis	Hypochromic effect (465 nm) Hyperchromic effect (565 and 600 nm)
	CN^−		Hypochromic effect (465 nm) Hyperchromic effect (520 nm)
78 ^103	Zn^2+	PL	Enhancement at 480 nm (50-fold)
	Cu^2+		Quenching at 390 nm
79 ^103	Zn^2+	PL	Blue shift from 585 to 480 nm with 41-fold enhancement
	Cu^2+		Quenching (85%) at 585 nm
	F^−		Quenching at 585 nm
80 ^104	Zn^2+	PL	Enhancement (85-fold) Red-shift (from 445 to 520 nm)
	Cd^2+		Quenching Red-shift (from 445 to 580 nm)
81 ^105	Zn^2+	PL	Enhancement with red shift from 475 to 492 nm
	F^−		Quenching at 475 nm
82 ^106	Sr^2+	PL	Enhancement at 365 nm
	Cd^2+		Quenching at 365 nm
83 ^107	Cu^2+	PL	Enhancement at 476 nm
	F^−		Quenching at 476 nm
84 ^108	FMN	PL	Quenching at 525 nm
	FAD		Enhancement at 525 nm
85 ^109	NAD^+	PL	Monomer quenching
	NADP^+		Excimer formation at 526 nm
	FMN		Mutual quenching
	FAD		Time-dependent enhancement
86 ^110	F^−/CN^−	PL	Quenching at 608 nm
	H2PO4^−		Enhancement (3.2 fold) with red shift from 608 nm to 615 nm
	Fe^2+	UV-Vis	Red shift from 318 to 323 nm and a new peak at 575 nm

---

Scheme 24 Structures of probe 57–86.

6 Conclusions

From our enumeration, it has become clear that orthogonal multianalyte detection by a single probe has attracted growing attention in recent years. The field will receive even more interest, when the first applications have emerged, in which small volume segments of different composition (e.g. see the pioneering work with probe 2 in cells)²⁸ are screened for several analytes at high spatial resolution, because probe partitioning will not matter for the analysis. Moreover, replacing the single-analyte probes in sensor arrays with lab-on-a-molecules will vastly improve the power of analytic assays.

In the last three years, numerous novel systems have emerged and new protocols have surfaced. Multianalyte chemosensors are successful when they operate on selective sensing and signaling modes that are based on different detection channels, orthogonal binding, orthogonal reaction mechanisms and/or differing kinetic features. Up to now, however, most multianalyte chemosensors are only able to quantify two analytes. Only in few probes, i.e. with 11 and 22, quantification of three species has been realized. As a result, additional work needs to be invested to improve the number of detectable analytes in large-sized libraries by using a single lab-on-a-molecule assay.

A wide range of analytes, selectivity in real-world chemical and biological samples, orthogonal detection of closely related analytes, reusability and practicability are desired in combination with surface-immobilized systems for handy applications. Furthermore, the sensitivity to different analytes should be in the relevant detection range for real world assays (e.g. the concentrations of GSH and Cys in the cell are 1–10 mM and 30–200 μM, respectively).⁷¹ Toward the end, low-cost, real time detection, and dynamics of response will become important in real-world applications. With an increasing amount of these requirements fulfilled, multianalyte chemosensors will find an increasing utility in environmental, industrial and medical analysis.

Acknowledgements

We thank the DFG for financial support in the DFG/ANR framework (Schm 647/17-1) and are indebted to Dr S. Khatua (North-Eastern Hill University, India) for some initial discussions on lab-on-a-molecule probes.

Notes and references

1. (a) L. E. Kreno, K. Leong, O. K. Farha, M. Allendorf, R. P. Van Duyne and J. T. Hupp, Chem. Rev., 2012, 112, 1105; (b) T. Wagner, S. Haffer, C. Weinberger, D. Klaus and M. Tiemann, Chem. Soc. Rev., 2013, 42, 4036.
2. (a) M. E. Germain and M. J. Knapp, Chem. Soc. Rev., 2009, 38, 2543; (b) Y. Salinas, R. Martínez-Máñez, M. D. Marcos, F. Sancenón, A. M. Costero, M. Parra and S. Gil, Chem. Soc. Rev., 2012, 41, 1261.
3. M. T. Albelda, J. C. Frías, E. García-España and H.-J. Schneider, Chem. Soc. Rev., 2012, 41, 3859.
4. A. Mills, Chem. Soc. Rev., 2005, 34, 1003.
5. (a) A. R. Lippert, G. C. van de Bittner and C. J. Chang, Acc. Chem. Res., 2011, 44, 793; (b) Y. Zhou, Z. Xu and J. Yoon, Chem. Soc. Rev., 2011, 40, 2222; (c) X. Chen, X. Tian, I. Shin and J. Yoon, Chem. Soc. Rev., 2011, 40, 4783; (d) J. Chan, S. C. Dodani and C. J. Chang, Nat. Chem., 2012, 4, 973; (e) A. V. Jentzsch, A. Hennig, J. Mareda and S. Matile, Acc. Chem. Res., 2013, 46, 2791.
6. (a) H. Craighead, Nature, 2006, 442, 387; (b) T. F. Didar and M. Tabrizian, Lab Chip, 2010, 10, 3043; (c) A. M. Foudeh, T. F. Didar, T. Veres and M. Tabrizian, Lab Chip, 2012, 12, 3249; (d) Y. Du, B. Li and E. Wang, Acc. Chem. Res., 2013, 46, 203.
7. H. Schäfer, H. Lin, M. Schmittel and M. Böhm, Microsyst. Technol., 2008, 14, 589.
8. M. Wu, D. Holowka, H. G. Craighead and B. Baird, Proc. Natl. Acad. Sci. U. S. A., 2004, 101, 13798.
9. (a) D. C. Magri, G. J. Brown, G. D. McClean and A. P. de Silva, J. Am. Chem. Soc., 2006, 128, 4950; (b) D. C. Magri, M. C. Fava and C. J. Mallia, Chem. Commun., 2014, 50, 1009.
10. L. Rosenfeld, Four centuries of clinical chemistry, Taylor & Francis, New York, London, 1999. See comment on page 255.
11. A. Hulanicki, S. Glab and F. Ingman, Pure Appl. Chem., 1991, 63, 1247.
12. (a) A. T. Wright and E. V. Anslyn, Chem. Soc. Rev., 2006, 35, 14; (b) P. Anzenbacher Jr., P. Lubal, P. Buček, M. A. Palacios and M. E. Kozelkova, Chem. Soc. Rev., 2010, 39, 3954; (c) K. L. Diehl and E. V. Anslyn, Chem. Soc. Rev., 2013, 42, 8596; (d) S. Stewart, M. A. Ivy and E. V. Anslyn, Chem. Soc. Rev., 2014, 43, 70.
13. (a) C. N. LaFratta and D. R. Walt, Chem. Rev., 2008, 108, 614; (b) D. R. Walt, Chem. Soc. Rev., 2010, 39, 38; (c) D. R. Walt, Anal. Chem., 2013, 85, 1258.
14. (a) L. E. Santos-Figueroa, M. E. Moragues, E. Climent, A. Agostini, R. Martínez-Máñez and F. Sancenón, Chem. Soc. Rev., 2013, 42, 3489; (b) X. Li, X. Gao, W. Shi and H. Ma, Chem. Rev., 2014, 114, 590.
15. (a) M. Dong, Y. Peng, Y.-M. Dong, N. Tang and Y.-W. Wang, Org. Lett., 2012, 14, 130; (b) Y. Peng, Y.-M. Dong, M. Dong and Y.-W. Wang, J. Org. Chem., 2012, 77, 9072.
16. (a) D. Margulies, C. E. Felder, G. Melman and A. Shanzer, J. Am. Chem. Soc., 2007, 129, 347; (b) J. Andréasson and U. Pischel, Chem. Soc. Rev., 2010, 39, 174; (c) U. Pischel, J. Andréasson, D. Gust and V. F. Pais, ChemPhysChem, 2013, 14, 28; (d) B. Rout, L. Motiei and D. Margulies, Synlett, 2014, 25, 1050.
17. (a) A. P. de Silva and S. Uchiyama, Nat. Nanotechnol., 2007, 2, 399; (b) A. P. de Silva, Chem. – Asian J., 2011, 6, 750; (c) U. Pischel, Angew. Chem., Int. Ed., 2007, 46, 4026; (d) U. Pischel, Aust. J. Chem., 2010, 63, 148.
18. S. Ozlem and E. U. Akkaya, J. Am. Chem. Soc., 2009, 131, 48.
19. (a) H. Kobayashi, M. R. Longmire, M. Ogawa and P. L. Choyke, Chem. Soc. Rev., 2011, 40, 4626; (b) S. Carregal-Romero, E. Caballero-Díaz, L. Beqa, A. M. Abdelmonem, M. Ochs, D. Hühn, B. S. Suau, M. Valcarcel and W. J. Parak, Annu. Rev. Anal. Chem., 2013, 6, 53.
20. See examples: (a) M. Dong, Y.-W. Wang and Y. Peng, Org. Lett., 2010, 12, 5310; (b) H. J. Jung, N. Singh, D. Y. Lee and D. O. Jang, Tetrahedron Lett., 2010, 51, 3962; (c) Z.-Q. Hu, C.-S. Lin, X.-M. Wang, L. Ding, C.-L. Cui, S.-F. Liu and H. Y. Lu, Chem. Commun., 2010, 46, 3765; (d) Y. Liu, Y. Sun, J. Du, X. Lv, Y. Zhao, M. Chen, P. Wang and W. Guo, Org. Biomol. Chem., 2011, 9, 432; (e) L. Wang, W. Qin, X. Tang, W. Dou and W. Liu, J. Phys. Chem. A, 2011, 115, 1609; (f) N. Kumari, S. Jha and S. Bhattacharya, J. Org. Chem., 2011, 76, 8215; (g) L. Xu, Y. Xu, W. Zhu, B. Zeng, C. Yang, B. Wu and X. Qian, Org. Biomol. Chem., 2011, 9, 8284; (h) L. Wang, J. Yan, W. Qin, W. Liu and R. Wang, Dyes Pigm., 2012, 92, 1083; (i) X. Sun, Y.-W. Wang and Y. Peng, Org. Lett., 2012, 14, 3420; (j) P. N. Basa and A. G. Sykes, J. Org. Chem., 2012, 77, 8428; (k) M. S. Moorthy, H.-J. Cho, E.-J. Yu, Y.-S. Jung and C.-S. Ha, Chem. Commun., 2013, 49, 8758; (l) Y. Li, J. Wu, X. Jin, J. Wang, S. Han, W. Wu, J. Xu, W. Liu, X. Yao and Y. Tang, Dalton Trans., 2014, 43, 1881.
21. (a) M. E. Huston, C. Engleman and A. W. Czarnik, J. Am. Chem. Soc., 1990, 112, 7054; (b) J. M. Castagnetto and J. W. Canary, Chem. Commun., 1998, 203.
22. Y. Lu, H. Kong, F. Wen, S. Zhang and X. Zhang, Chem. Commun., 2013, 49, 81.
23. (a) D. Liu, M. Liu, G. Liu, S. Zhang, Y. Wu and X. Zhang, Anal. Chem., 2010, 82, 66; (b) P. Wu, L.-N. Miao, H.-F. Wang, X.-G. Shao and X.-P. Yan, Angew. Chem., Int. Ed., 2011, 50, 8118.
24. F. Vogt and B. Mizaikoff, Anal. Chem., 2003, 75, 3050.
25. (a) I. M. Raimundo Jr. and R. Narayanaswamy, Sens. Actuators, B, 2003, 90, 189; (b) D. Mikami, T. Ohki, K. Yamaji, S. Ishihara, D. Citterio, M. Hagiwara and K. Suzuki, Anal. Chem., 2004, 76, 5726.
26. M. L. Griffiths, R. P. Barbagallo and J. T. Keer, Anal. Chem., 2006, 78, 513.
27. H. Komatsu, D. Citterio, Y. Fujiwara, K. Minamihashi, Y. Araki, M. Hagiwara and K. Suzuki, Org. Lett., 2005, 7, 2857.
28. H. Komatsu, T. Miki, D. Citterio, T. Kubota, Y. Shindo, Y. Kitamura, K. Oka and K. Suzuki, J. Am. Chem. Soc., 2005, 127, 10798.
29. L. A. Greene and A. S. Tischler, Proc. Natl. Acad. Sci. U. S. A., 1976, 73, 2424.
30. P. Kaur, D. Sareen and K. Singh, Dalton Trans., 2012, 41, 8767.
31. N. Singh, R. C. Mulrooney, N. Kaur and J. F. Callan, Chem. Commun., 2008, 4900.
32. N. Kaur and S. Kumar, Chem. Commun., 2007, 3069.
33. N. Kaur and S. Kumar, Tetrahedron Lett., 2008, 49, 5067.
34. D. Y. Lee, N. Singh and D. O. Jang, Tetrahedron Lett., 2010, 51, 1103.
35. D. Y. Lee, N. Singh and D. O. Jang, Tetrahedron Lett., 2011, 52, 3886.
36. P. Kaur and D. Sareen, Dyes Pigm., 2011, 88, 296.
37. V. K. Bhardwaj, P. Saluja, G. Hundal, M. S. Hundal, N. Singh and D. O. Jang, Tetrahedron, 2013, 69, 1606.
38. (a) M. Schmittel and H.-W. Lin, Angew. Chem., Int. Ed., 2007, 46, 893; (b) A. P. de Silva, Nature, 2007, 445, 718.
39. N. Singh, N. Kaur, C. N. Choitir and J. F. Callan, Tetrahedron Lett., 2009, 50, 4201.
40. Q. Shu, L. Birlenbach and M. Schmittel, Inorg. Chem., 2012, 51, 13123.
41. M. Schmittel and S. Qinghai, Chem. Commun., 2012, 48, 2707.
42. K. Chen and M. Schmittel, Analyst, 2013, 138, 6742.
43. E. Berni, I. Gosse, D. Badocco, P. Pastore, N. Sojic and S. Pinet, Chem. – Eur. J., 2009, 15, 5145.
44. L. Tang, F. Li, M. Liu and R. Nandhakumar, Spectrochim. Acta, Part A, 2011, 78, 1168.
45. L. Tang, F. Li, M. Liu and R. Nandhakumar, Bull. Korean Chem. Soc., 2011, 32, 3400.
46. H. Ye, F. Ge, X.-C. Chen, Y. Li, H. Zhang, B.-X. Zhao and J.-Y. Miao, Sens. Actuators, B, 2013, 182, 273.
47. M. Wang, F. Yan, Y. Zou, L. Chen, N. Yang and X. Zhou, Sens. Actuators, B, 2014, 192, 512.
48. L. Shi, Y. Li, Z.-P. Liu, T. D. James and Y.-T. Long, Talanta, 2012, 100, 401.
49. K. Chen and M. Schmittel, Chem. Commun., 2014, 50, 5756.
50. E. Tulyakova, S. Delbaere, Y. Fedorov, G. Jonusauskas, A. Moiseeva and O. Fedorova, Chem. – Eur. J., 2011, 17, 10752.
51. C. Bhaumik, S. Das, D. Maity and S. Baitalik, Dalton Trans., 2011, 40, 11795.
52. M. Yuan, W. Zhou, X. Liu, M. Zhu, J. Li, X. Yin, H. Zheng, Z. Zuo, C. Ouyang, H. Liu, Y. Li and D. Zhu, J. Org. Chem., 2008, 73, 5008.
53. L. Yuan, W. Lin, Y. Xie, B. Chen and S. Zhu, J. Am. Chem. Soc., 2012, 134, 1305.
54. B. Rout, L. Unger, G. Armony, M. A. Iron and D. Margulies, Angew. Chem., Int. Ed., 2012, 51, 12477.
55. C. Lu, Z. Xu, J. Cui, R. Zhang and X. Qian, J. Org. Chem., 2007, 72, 3554.
56. Z. Xu, K.-H. Baek, H. N. Kim, J. Cui, X. Qian, D. R. Spring, I. Shin and J. Yoon, J. Am. Chem. Soc., 2010, 132, 601.
57. L. Xue, C. Liu and H. Jiang, Org. Lett., 2009, 11, 1655.
58. L. Xue, Q. Liu and H. Jiang, Org. Lett., 2009, 11, 3454.
59. X. Zhou, P. Li, Z. Shi, X. Tang, C. Chen and W. Liu, Inorg. Chem., 2012, 51, 9226.
60. M. Kumar, R. Kumar and V. Bhalla, Tetrahedron Lett., 2013, 54, 1524.
61. A. B. Othman, J. W. Lee, J.-S. Wu, J. S. Kim, R. Abidi, P. Thuéry, J. M. Strub, A. Van Dorsselaer and J. Vicens, J. Org. Chem., 2007, 72, 7634.
62. J. F. Zhang, Y. Zhou, J. Yoon, Y. Kim, S. J. Kim and J. S. Kim, Org. Lett., 2010, 12, 3852.
63. Y. Zhao, B. Zheng, J. Du, D. Xiao and L. Yang, Talanta, 2011, 85, 2194.
64. L. N. Neupane, J.-Y. Park, J. H. Park and K.-H. Lee, Org. Lett., 2013, 15, 254.
65. X. Liu, X. Yang, Y. Fu, C. Zhu and Y. Cheng, Tetrahedron, 2011, 67, 3181.
66. V. Luxami, Renukamal, K. Paul and S. Kumar, RSC Adv., 2013, 3, 9189.
67. E. Karakuş, M. Üçüncü and M. Emrullahoğlu, Chem. Commun., 2014, 50, 1119.
68. T. Guo, L. Cui, J. Shen, R. Wang, W. Zhu, Y. Xu and X. Qian, Chem. Commun., 2013, 49, 1862.
69. K. Chen, J. W. Bats and M. Schmittel, Inorg. Chem., 2013, 52, 12863.
70. X. Yang, Y. Guo and R. M. Strongin, Angew. Chem., Int. Ed., 2011, 50, 10690.
71. X.-F. Yang, Q. Huang, Y. Zhong, Z. Li, H. Li, M. Lowry, J. O. Escobedo and R. M. Strongin, Chem. Sci., 2014, 5, 2177.
72. J. Liu, Y.-Q. Sun, Y. Huo, H. Zhang, L. Wang, P. Zhang, D. Song, Y. Shi and W. Guo, J. Am. Chem. Soc., 2014, 136, 574.
73. J. Liu, Y.-Q. Sun, H. Zhang, Y. Huo, Y. Shi and W. Guo, Chem. Sci., 2014, 5, 3183.
74. Y.-J. Huang, W.-J. Ouyang, X. Wu, Z. Li, J. S. Fossey, T. D. James and Y.-B. Jiang, J. Am. Chem. Soc., 2013, 135, 1700.
75. M. Tian, H. Ihmels and K. Benner, Chem. Commun., 2010, 46, 5719.
76. S. H. Kim, J. S. Kim, S. M. Park and S.-K. Chang, Org. Lett., 2006, 8, 371.
77. F. Qian, C. Zhang, Y. Zhang, W. He, X. Gao, P. Hu and Z. Guo, J. Am. Chem. Soc., 2009, 131, 1460.
78. Z. Li, L. Zhang, L. Wang, Y. Guo, L. Cai, M. Yu and L. Wei, Chem. Commun., 2011, 47, 5798.
79. Y. Chen and J. Jiang, Org. Biomol. Chem., 2012, 10, 4782.
80. X. Li, M. Yu, F. Yang, X. Liu, L. Wei and Z. Li, New J. Chem., 2013, 37, 2257.
81. L. Zang, D. Wei, S. Wang and S. Jiang, Tetrahedron, 2012, 68, 636.
82. H. Lu, Z. Xue, J. Mack, Z. Shen, X. You and N. Kobayashi, Chem. Commun., 2010, 46, 3565.
83. R. Martínez, A. Espinosa, A. Tárraga and P. Molina, Org. Lett., 2005, 7, 5869.
84. K. Tsubaki, D. Tanima, Y. Kuroda, K. Fuji and T. Kawabata, Org. Lett., 2006, 8, 5797.
85. L.-L. Li, H. Sun, C.-J. Fang, J. Xu, J.-Y. Jin and C.-H. Yan, J. Mater. Chem., 2007, 17, 4492.
86. J. Tan and X.-P. Yan, Talanta, 2008, 76, 9.
87. M.-M. Yu, Z.-X. Li, L.-H. Wei, D.-H. Wei and M.-S. Tang, Org. Lett., 2008, 10, 5115.
88. R. Joseph, B. Ramanujam, H. Pal and C. P. Rao, Tetrahedron Lett., 2008, 49, 6257.
89. B. N. Ahamed, I. Ravikumar and P. Ghosh, New J. Chem., 2009, 33, 1825.
90. T. Ábalos, D. Jiménez, R. Martínez-Máñez, J. V. Ros-Lis, S. Royo, F. Sancenón, J. Soto, A. M. Costero, S. Gil and M. Parra, Tetrahedron Lett., 2009, 50, 3885.
91. T. Ábalos, D. Jiménez, M. Moragues, S. Royo, R. Martínez-Máñez, F. Sancenón, J. Soto, A. M. Costero, M. Parra and S. Gil, Dalton Trans., 2010, 39, 3449.
92. F.-J. Huo, J. Su, Y.-Q. Sun, C.-X. Yin, H.-B. Tong and Z.-X. Nie, Dyes Pigm., 2010, 86, 50.
93. H. Li, L. Cai, J. Li, Y. Hu, P. Zhou and J. Zhang, Dyes Pigm., 2011, 91, 309.
94. X. Zhang, Y. Xu, P. Guo and X. Qian, New J. Chem., 2012, 36, 1621.
95. V. Bhalla, R. Tejpal and M. Kumar, Tetrahedron, 2011, 67, 1266.
96. R. Pandey, P. Kumar, A. K. Singh, M. Shahid, P.-Z. Li, S. K. Singh, Q. Xu, A. Misra and D. S. Pandey, Inorg. Chem., 2011, 50, 3189.
97. Y. Chen, L. Wan, X. Yu, W. Li, Y. Bian and J. Jiang, Org. Lett., 2011, 13, 5774.
98. H. Yang, H. Song, Y. Zhu and S. Yang, Tetrahedron Lett., 2012, 53, 2026.
99. A. Sikdar, S. S. Panja, P. Biswas and S. Roy, J. Fluoresc., 2012, 22, 443.
100. X. He and V. W.-W. Yam, Org. Lett., 2011, 13, 2172.
101. R. Satapathy, Y.-H. Wu and H.-C. Lin, Org. Lett., 2012, 14, 2564.
102. R. Guliyev, S. Ozturk, E. Sahin and E. U. Akkaya, Org. Lett., 2012, 14, 1528.
103. V. Luxami and S. Kumar, RSC Adv., 2012, 2, 8734.
104. J. Wang, W. Lin and W. Li, Chem. – Eur. J., 2012, 18, 13629.
105. P. G. Sutariya, N. R. Modi, A. Pandya, B. K. Joshi, K. V. Joshi and S. K. Menon, Analyst, 2012, 137, 5491.
106. P. G. Sutariya, A. Pandya, N. R. Modi and S. K. Menon, Analyst, 2013, 138, 2244.
107. P. G. Sutariya, A. Pandya, A. Lodha and S. K. Menon, Analyst, 2013, 138, 2531.
108. H.-W. Rhee, S. J. Choi, S. H. Yoo, Y. O. Jang, H. H. Park, R. M. Pinto, J. C. Cameselle, F. J. Sandoval, S. Roje, K. Han, D. S. Chung, J. Suh and J.-I. Hong, J. Am. Chem. Soc., 2009, 131, 10107.
109. J. Oh and J.-I. Hong, Org. Lett., 2013, 15, 1210.
110. Z.-B. Zheng, Z.-M. Duan, Y.-Y. Ma and K.-Z. Wang, Inorg. Chem., 2013, 52, 2306.

---

Footnote

† These authors contributed equally.

---