Elucidating sensing mechanisms of a pyrene excimer-based calix[4]arene for ratiometric detection of Hg(ii) and Ag(i) and chemosensor behaviour as INHIBITION or IMPLICATION logic gates

This article reports the synthesis and characterisation of two lower rim calix[4]arene derivatives with thiourea as spacer and pyrene or methylene–pyrene as fluorophore. Both derivatives exhibit a fluorimetric response towards Hg2+, Ag+ and Cu2+. Only methylene–pyrenyl derivative 2 allows for selective detection of Hg2+ and Ag+ by enhancement or decrease of excimer emission, respectively. The limits of detection of 2 are 8.11 nM (Hg2+) and 2.09 nM (Ag+). DFT and TD-DFT computational studies were carried out and used to identify possible binding modes that explain the observed response during fluorescence titrations. Calculations revealed the presence of different binding sites depending on the conformation of 2, which suggest a reasonable explanation for non-linear changes in fluorescence depending on the physical nature of the interaction between metal centre and conformer. INHIBITION and IMPLICATION logic gates have also been generated monitoring signal outputs at pyrene monomer (395 nm) and excimer (472 nm) emission, respectively. Thus 2 is a potential primary sensor towards Ag+ and Hg2+ able to configure two different logic gate operations.


Introduction
Transition metal ions play an important role in a wide range of chemical reactions, including biological metabolism and many other natural or anthropogenic processes. [1][2][3][4] Hence, much effort has been made in order to develop chemosensors capable of their qualitative or quantitative determination. 1,[5][6][7] Silver ions are widely studied because of their antimicrobial activities, 8 applications in photography, electronics and medicinal chemistry (e.g. silver nanoparticles). 9 However, silver contamination can occur through a variety of sources including volcanoes, gold mining or fossil fuel combustion. 10 As a result of that, excess of silver ions can lead to bioaccumulation and intoxication, 11 resulting in adverse effects affecting kidney function, immune and cardiovascular systems, and inactivation of sulydryl enzymes. 12 Additionally, mercury is one of the most toxic and environmentally widespread contaminants, causing serious damage to living organisms and the environment. Consequently, when absorbed in tissues it can result in damage to the central nervous system and liver, leading to motor and cognitive disorders. 13,14 Within the different toxic species of mercury, its divalent form can lead to severe brain and kidney damage, even though it cannot cross the blood brain barrier. 15 Consequently, the development of new methods providing high sensitivity and selectivity to detect Ag + or Hg 2+ are highly welcomed in many elds.
In this context, different techniques for detection of such contaminants have already been reported. 3,[16][17][18][19][20][21][22][23][24][25] Among them, uorescence-based sensors are particularly interesting over other techniques due to their simplicity for implementation, fast response, local observation, selectivity and sensitivity. 1,5 In particular, uorescent ratiometric detection has become one of the most powerful tools for molecular sensing. Since it allows measurement of emission intensities at two different wavelengths, so that the ratios of signals will be independent of the environmental effects. Ratiometric-based chemosensors rely on the principle that it improves the dynamic response according to the change of intensity ratios, thus establishing the internal standard. Compared with conventional chemical analysis, the ratiometric method has thus a higher selectivity and more simplicity in the detection of ligand-receptor binding by observing merely the enhanced uorescence of the acceptor. [26][27][28][29] An approach generally used for this kind of molecular sensors consist of an ion recognition moiety, required for the selective binding to the substrate, coupled to uorophore unit, providing a moniterable signal. Both units are typically linked to a molecular scaffold. 30 In order to develop recognition moieties, calix [4]arenes have been widely used as molecular scaffold due to their ease and controlled functionalisation. 5,31 Furthermore, they provide a suitable platform to develop a cavity in which selectivity can thus be nely tuned. 32 For example, by the use of triazoles, ureas, amides or thioureas as linker between signalling moiety and the molecular scaffold. The latter has shown to be a highly versatile recognition moiety for metal cations, 33 as well as organic anions 34 and DNA, [35][36][37] due to the variety of possibilities to bind the substrate, i.e. the presence of nitrogen and sulphur coordinating atoms are able to bind cations by either throughbond or through-space interactions, while NH hydrogen bond donor groups are able to stabilize anions and DNA by Hbonding interactions. 33,[38][39][40] Among the wide range of uorescent moieties reported, pyrene has attained much attention due to its well-known photophysical properties such as high quantum yields, well-dened absorption and emission bands, long lifetime and the possibility to form excimers. 41,42 These unique properties have inspired the development of pyrene-based sensors due to providing several sensing modes, by monitoring either the monomer (ca. 370-390 nm) or excimer (ca. 470-490 nm) emission bands, or the ratio of both. The ratio of the intensities of excimer and monomer emission is very useful for analyte sensing because it is highly sensitive to the conformational changes of the pyrene-anchored to any molecular architecture, which are induced by the interaction mode between the analyte and the recognition moiety.
Nevertheless, many studies that have used aforementioned supramolecular approach for the development of chemosensors do not explore in detail the different binding modes available within the cavity. Although many succeed in the determination of the host-guest stoichiometry, 43,44 or are able to provide crystal structures, 22,45 a signicant part of them only propose a most plausible binding mode. Furthermore, whether different equilibriums are reached upon increasing concentration of guest, then the determination of the binding constant turns challenging. In this context, quantum chemical methods are an alternative tool to investigate the electronic and molecular structures of host and respective host-guest complexes, unveiling the physical nature of the interactions and their impact on the uorescence response, giving a different standpoint that goes beyond the typically proposed mechanisms. Potentially, this type of studies could allow to gain a new insights into the calix [4]arenes-based chemosensors. Many quantum-chemical-based studies have focused their attention on the formation of isolated pyrene-excimer 46 and their properties as well as in the investigation on the binding of different analytes using calix [4]arene scaffolds. 47,48 More recently, density functional theory (DFT) has been employed to understand the sensing mechanism of lysine through the formation of the intermolecular pyrene-excimer. 49 Herein we present the synthesis and characterisation of two calix [4]arene derivatives composed by pyrene units connected through a thiourea-based spacer. We studied their photophysical properties, their ability to show excimer emission, and their uorescent response to different cations. Furthermore, we carry out a quantum chemical study based on density functional theory (DFT) and time-dependent (TD-DFT) frames, aimed to explain the experimental observations through the identication of the specic interactions that lead to the stabilization of different ligand conformers and their respective metal complexes. Additionally, operation of one of the sensors as INHIBITION or IMPLICATION logic gate is also described, 50 since it gives different uorescent outputs when it is treated with Ag + and Hg 2+ .

Synthesis
The synthetic route designed for chemosensors 1 and 2 takes advantage of the thiourea forming reaction within amines and isothiocyanates. Pyreneisothiocyanates and calix [4]arene 3 derivatives were prepared by reported methodologies. [51][52][53] The reactions were carried out in dry CH 2 Cl 2 /triethylamine solution, and stirred at room temperature for 48 hours (Scheme 1). An advantage of this synthetic strategy is that the nal products precipitated. By simply ltering and washing off these precipitates, products 1 and 2 were isolated in moderate yields without further purication. It is noteworthy to mention that 1 synthetic route was rstly attempted by coupling 1-aminopyrene derivative to the analogue isothiocyanate derivative from 3. Nevertheless, this route did not yield the desired products, most likely due to the low basicity of the aromatic amine. 1 H and 13 C NMR spectra were consistent with the proposed structures. In both cases, two differentiated aromatic signals related to pyrene (8.3-7.4 ppm) and calix [4]arene units (6.7-6.4 ppm) are observed. In our case, a characteristic pair of doublets at 3.3 and 2.4 ppm for compound 1 and 3.3 and 2.7 ppm for 2, whose coupling constants are around 13.0 Hz, support the cone conformation of 1 and 2. For compound 2, a singlet at 5.3 ppm is assigned to the methylene bridge between the thiourea and pyrene units. Besides, in 13 C NMR spectra, a signal observed at around 182 ppm for both molecules conrmed thiocarbonyl bond formation. Additionally, high-resolution mass spectrometry is in agreement with the proposed molecular structures (see ESI †).

Spectroscopic properties
Photophysical properties of 1 and 2 were studied by UV-vis and uorescence spectroscopy in acetonitrile/DMSO (99 : 1) solutions ( Fig. 1). Calix [4]arene 1 exhibits a well-dened absorption band at 280 nm and two overlapped bands at 332 and 347 nm instead of pyrene ne structure. In 2, sharper bands are observed at 265, 276, 314, 327 and 344 nm. Bands in the range from 265 up to 280 nm are most likely related to calix [4]arene derivatives, while the bands in the range from 314 up to 347 nm are related to pyrene moieties, retaining its typical absorption ne structure. 54 Although there is a slight difference between chemical structures of 1 and 2, uorescence spectra show very different emission proles when excited at 340 nm. In the case of 1, the characteristic emission spectrum of pyrene monomer is replaced by a featureless prole; a broad band with a maximum at 406 nm is observed. This emission is most likely related to the pyrene-thiourea unit, as it can be seen in previous reports that study analogous uorophore structures, plausibly due to some rotational degree of freedom. 18,55,56 This similarity in emission proles suggests that calix [4]arene core has negligible effect in the emission properties of pyrene. On the other hand, 2 presents three main emission bands. Two of them appear at 377 and 395 nm, and are related to the pyrene emission as noted earlier in its emission spectrum. 54 The third one, at 472 nm, can be attributed to the relaxation of a pyrene-excimer formed, most likely due to CH 2 bridge. This type connectivity brings an additional degree of freedom to pyrene, favouring the formation of excimers as seen in previous reports. 19,[57][58][59][60] In order to demonstrate that pyrene units are forming an intramolecular excimer rather than an intermolecular one, we performed dilution studies for 2 in the concentration range from 1 Â 10 À5 to 1 Â 10 À7 M (Fig. S1 †). Fluorescence intensity ratio of excimer/pyrene (I 472 /I 395 ) keeps constant at 0.90 suggesting that the excimer emission observed arises from an intramolecular interaction of pyrenes.
Both sensors display uorescence changes upon addition of Ag + , Hg 2+ and Cu 2+ ( Fig. 2a and b). In the case of 1, deactivation of the uorescence at 406 nm is observed for the aforementioned cations. 61,62 However, the observed response does not allow for ratiometric detection, since no new band is observed. Furthermore, the similar response to the three cations may limit the applicability of 1 as selective chemosensor. Conversely, chemosensor 2 shows a completely different response. Upon addition of Ag + the emission intensity of pyrene excimer band (472 nm) decreases (Fig. 2b), while addition of Hg 2+ results in an increase of intensity of excimer emission and a decrease in the monomer emission. When treated with Cu 2+ a slight decrease of intensity of all bands is perceived. In order to see whether higher selectivity of 2 within these cations can be obtained, ratiometric variation (I 472 /I 395 ) of uorescence intensity of 2 with the studied cations is represented in Fig. 2c. Interestingly, under this analysis it can be seen that Cu 2+ no longer is a signicant cation to be detected by 2, showing only a slight perturbation (<5%) similar to that of Pb 2+ , Ni 2+ or Zn 2+ . 27,28 Interestingly, probe 2 shows no signicative interference when a mixture of all the cations at the same concentration are present (Fig. S7a †) and, the interference studies using Hg 2+ and Ag + interacting with each other, show that the Hg 2+ output is predominant (Fig. S7b †), which is in agreement with the observed Log K app .

Titrations, limits of detection and K app
Fluorescent titrations of the compound 2 with Ag + and Hg 2+ (0 to 3 equivalents) were performed. As depicted in Fig. 3a, addition of 0.1 to 0.5 equivalents of Ag + results in ratiometric variation of monomer and excimer bands, in which monomer uorescence (377 and 395 nm) increases while excimer band (472 nm) decreases. Further increase in Ag + concentration (0.6 to 3.0 equivalents, Fig. 3b) results in a different variation of uorescence, since all bands simultaneously decrease in intensity. On the contrary, although the addition from 0.1 to 1.5 equivalents of Hg 2+ the uorescence intensities also change ratiometrically (Fig. 3c), in this case monomer uorescence is quenched, while pyrene-excimer uorescence is enhanced. These differentiated changes allow 2 to discriminate between Ag + and Hg 2+ species. Upon further addition of Hg 2+ (1.5 to 3.0 equivalents), similarly as with Ag + , intensity of all emission bands decreases (Fig. 3d). Additionally, we also conrmed that the perchlorate anion had no effects in the uorescence (Fig. S2 †). In order to get better insight of the uorescence response as cation concentration increase 1 H NMR titrations  were attempted. However, they result unsuccessful due to the low solubility of 2 in DMSO-d 6 , CDCl 3 and CD 3 CN. In the following section this issue will be address by using computational approaches. The limits of detection (LODs) of 2 were estimated based on the uorescence titration data (Table 1 and Fig. S3 †). LODs were found to be 8.11 nM for Hg 2+ and 2.09 nM for Ag + . 63 These values result almost 10 fold lower than similar reported calix [4] arene based systems [64][65][66][67] or other uorescent chemosensors. 68,69 Comparative data are displayed in Table S1. † Titration data were also used to determine association constants of 2 with Hg 2+ and Ag + . Among the different methods normally applied to calculate them, Benessi-Hildebrand equation cannot be applicable in this case due to the observation of two different uorescent-based sensing mechanisms. 70 Nonlinear tting was also evaluated for 2 titration curves with Ag + and Hg 2+ , resulting in high deviations for similar reasons. 70 Additionally, we attempted to clarify complexation stoichiometry by using the method of continuous variation (Job's Plot, Fig. S4 and S5 †), even though we are conscious of its limitation in many supramolecular systems. 71,72 Unfortunately, maxima positions do not correlate with results showed by similar previously described calix [4]arene-based chemosensors, in which only a single binding site and/or conformation have been preferred. 63 For these reasons, apparent association constants K app for 2$Hg 2+ and 2$Ag + were calculated by linear tting according to the Hill equation, being Log K app 10.64 and 9.64, respectively (Table 1 and Fig. S6 †). 73,74 The data suggest that 2 interact stronger with Hg 2+ than with Ag + . This method has been used on other pyrenebased sensors involving two succinic imide labelled pyrenes that coordinate with Hg 2+ . This system, with a 2 : 1 stoichiometry, is similar to ours since calix [4]arene is assembling two thioureapyrene units. 73 The obtained Hill coefficients are n ¼ 1.7 and 1.5 when adding Ag + and Hg 2+ , respectively. 75,76 Quantum chemical calculations We carried out a computational study based on DFT and TD-DFT calculations aimed to nd 2 conformers and their respective Ag + and Hg 2+ complexes, together with the identication of stabilizing interactions and spectral features in order to explain the experimental sensing behaviour and propose a sensing mechanism.
Molecular geometries: 2 and 2$Ag + /2$Hg 2+ complexes. Three conformers were found for calix [4]arene derivative 2 in gasphase as displayed in Fig. 4. Two of them (labelled as 2a and 2b in Fig. 4) are the most stable as is unveiled by the relative Gibbs free energy values in solution (which is given by the sum of the electronic, thermal and entropic contributions and solvation energies and it is referred to the most stable conformation). Notice that both conformers present pyrene monomers nicely oriented to formation of parallel-displaced ground state pyrene dimer, with inter-plane distances of 3.337 and 3.353Å for 2a and 2b, respectively, which is slightly shorter than previous reports for isolated pyrene dimer (e.g. 3.45Å). 46,77 This type of orientation has been shown to play a key role in the excimer formation. Consequently, it can provide a contribution to the appearance of the broad and structureless excimer uorescence band, which is red-shied with respect to monomer uorescence. Moreover, we included the relative population (% pop) in solution following the standard Boltzmann distribution analysis. While 2a presents a 67%, 2b a 33%, therefore, 2 exist in solution as mixtures of 2a and 2b in an equilibrium ratio of 2 : 1. This result opens the possibility to explore different binding mode towards metallic cations, which can exhibit varied uorescence response in the development of 2-based chemosensor. 2c conformation showed to be an unstable one, concomitantly, the only option to excimer formation is by an intermolecular pathway due to that the pyrene units are away from each other. Three conformations were also found for calix [4]arene derivative 1 (see Fig. S14 †), but contrary to 2, it exists in solution in its most stable conformation 1a as predicted by Boltzmann distribution analysis. Notice that 1a presents a tilted T-shaped conguration among the pyrene moieties in its ground state, which contrasts with a parallel-displaced conformation that stabilizes the ground state of 2a and 2b. This feature explains the appearance of the excimer emission band in 2. Additionally, the basis set superposition error (BSSE)corrected instantaneous interaction energies values, DE int (-BSSE), for this fragments are higher in 2a (À12.55 kcal mol À1 ) and 2b (À12.70 kcal mol À1 ) than 1a (À8.45 kcal mol À1 ) by around 4.1 kcal mol À1 , unveiling that the structures 2a and 2b Fig. 4 Optimized conformers of 2 and their relative Gibbs free energy (in kcal mol À1 ) and Boltzmann population in acetonitrile.
This journal is © The Royal Society of Chemistry 2020 RSC Adv., 2020, 10, 21963-21973 | 21967 are more rigid than 1a, this fact explains the retention of ne structure in its absorption/emission spectra of derivative 2.
The corresponding 2$Ag + and 2$Hg 2+ complexes with each 2a and 2b conformations were also fully optimized (as displayed Fig. 5) and conrmed as local minima. DE int (BSSE) data are also included. The procedure followed here was to start with the same initial geometry, which consists in locating each cation inside the cavity of conformer 2a and 2b, until the fully relaxed structure is reached for each complex. It can be noted that the binding mode with 2a is quite similar for each metal centre. Contrarily, the binding mode with 2b is totally different. In accord with DE int (BSSE) values, it is found that Hg 2+ binds to 2a stronger than Ag + , being in agreement with experimental K app . Both 2a$Ag + and 2a$Hg 2+ complexes are mainly characterised by an interaction between S atoms and metal centre. The bond lengths are close to the sum of the corresponding covalent radii (2.500Å and 2.370Å for 2a$Ag + and 2a$Hg 2+ , respectively). This feature is an indicative that through-bond interactions or covalent-type (delocalization) in conjunction with the dispersion ones prevail over other interactions in the stability of 2a$Ag + /Hg 2+ complexes. Again, Hg 2+ binds to 2b stronger than Ag + . Nevertheless, the structures showed noticeable differences. While 2b$Ag + is cooperatively stabilized by Ag-S through-bond and cation/p interactions in conjunction with dispersion interactions, 2b$Hg 2+ is mainly stabilized by through-space interactions (localization) together with dispersions interactions. Since N and O atoms enclose Hg 2+ cation inside the cavity of 2b, the distances are in the range of 3.048 to 3.480Å, which are longer than the sum of their van der Waals radii (3.10Å and 3.07Å for Hg/N and Hg/O, respectively). This feature together with the high interaction energy suggests an electrostatic nature as stabilizing force in 2b$Hg 2+ complex. Overall, both metal cations interact more strongly with 2a than 2b. It must be emphasized that the pyrene units are less stacked in the metal complexes than in 2a and 2b, except in 2b$Hg 2+ , where the stacking is considerably more accentuated and plausibly more rigid due to the through-space interactions that stabilizing it.
Spectral features. We only computed the vertical excitation property due to that is less computationally demanding than to carry out excited state optimisations. The vertical electronic excitation energies together with absorption properties, such as oscillator strength (f) and the main molecular orbitals contribution involved in the electronic transitions were computed within the TD-DFT/SMD framework for 2a, 2b and the respective metal complexes. The corresponding data are quoted in  Table 2. Firstly, the computed transition energies agree with the experimental absorption energy of pyrene dimer (3.70 eV (ref. 78 and 79)). Our attention has been focused on this absorption band, which is associated with the electronic transition that originates the uorescence response employed to ratiometric uorescent detection of Ag + and Hg 2+ . In most of the cases, we can observe that the electronic transition involves the two highest occupied molecular orbitals (HOMOÀ1, HÀ1, and HOMO, H) and the two lowest unoccupied molecular orbitals (LUMO, L, and LUMO+1, L+1). As can be seen in Fig. S15, † both occupied and unoccupied molecular orbitals show certain bonding character in 2a, which is opposite to the character found in the isolated pyrene dimer: H is antibonding while L is bonding. This feature is plausibly due to the shorter inter-plane distance in 2a than in free pyrene dimer, which, on the other hand, is favoured by the use of calix [4]arene as molecular scaffold. An interesting spectral property that provides understanding on the absorption/emission probability is the oscillator strength. Calculated f showed to be relatively high, it can be noted that the presence of Ag + in the complexes increases f while Hg 2+ decreases the values of f with respect to the conformer 2a and 2b. 2b$Hg 2+ deserves special attention because TD-DFT calculations unveil an electronic transition at 465 nm (2.67 eV) with f of 0.0480. Its character is given from inner together with H molecular orbital to L and L+1. This feature is very signicant due to that the pyrene excimer uorescence is observed at this region. This fact could explain an enhancement of the uorescent response in the rst ratiometric uorescent probe for Hg 2+ at low concentration regime showed in Fig. 3c.
To the light of our results, we propose that the rst sensing mechanism for both cations could be originated by the formation of the complex with 2b conformer of 2. In the case of silver, upon interaction of 2b with Ag + , the p-p stacking pyrene rings breaks and results in excimer quenching and monomer enhancing. While a rigid pyrene-excimer is formed 2b$Hg 2+ due to the presence of the through-space interactions, results in an enhancement of the excimer uorescence (by localizing charge into the excited state) together with quenching of monomer emission. In the second sensing mechanism is mainly explained with the presence of metal complexes with 2a. We propose that the stabilizing through-bond interactions in both complexes are responsible to quench both the monomer and excimer uorescence, by delocalizing charge into the corresponding excited states.
DFT/TD-DFT predictions suggest future photophysical and time dependent uorescent investigations on 2b-based  chemosensor in all concentration regimes to conrm the sensing mechanism proposed here.

Application of 2 as molecular logic gate
Fluorescent chemosensors that are able to perform logic operations have attracted much attention due to its potential applications in biosensing and diagnosis, contaminant determination or molecular computation. [80][81][82] In general, their application as molecular logic gates is possible when two or more inputs give different uorescent outputs or sensing mechanism as unveiled our results. Among the sixteen Boolean operations that can be congured, six of them are rather trivial in terms of uorescent output or molecular design. 83 Within the rest, most of reported molecular logic gates are able to perform AND, OR, XOR operations. 50,84,85 In the case of INHIBITION gates, the "1" output is generated when only one input is present without the other input, i.e. one input has the power to activate the whole system. 84 Nevertheless, molecular design of IMPLICATION gates is more challenging since uorescence output has to be "1" in the absence of both inputs and only be quenched by one of the inputs. 83 As described above, chemosensor 2 shows different response at pyrene monomer (l ¼ 395 nm) and excimer (l ¼ 472 nm) emission bands, when treated either with Ag + or Hg 2+ . Therefore, we decide to evaluate its potential application as molecular logic gate at both emission wavelengths. Whether we situate the observation parameter for the output at one of monomer emission bands (l ¼ 395 nm), and dene the threshold at uorescence intensity 200 a.u., the initial output observed is "0" (Fig. 6a and b). Upon addition of 1 equivalent Ag + (as input 1), the pyrene emission of 2 increases due to the quenching of excimer emission, resulting in "1" output. Upon addition of Hg 2+ (input 2), the output becomes "0" due to the decrease of monomer emission. In the case where both inputs are added simultaneously (input 1 ¼ input 2 ¼ 1), the resulting emission output is again "0", most likely due to chemosensor 2 stronger binding affinity for Hg 2+ . The resulting logic gate at this output corresponds to an INHIBIT-type, as summarized in Table 3 (see  truth tables for all Boolean operations in Table S1 †). When the observation parameter is situated at the excimer emission band (l ¼ 472 nm), and the threshold uorescence intensity is dened at 150 a.u., the initial output observed is "1" (Fig. 6a and  c), featuring one of the main prerequisites for IMPLICATION molecular logic gates. Addition of 1 equivalent Ag + (input 1) results in quenching of excimer emission, being the observed output "0". The output generated upon addition of 1 equivalent Hg 2+ (input 2) is again "1" and similarly, the addition of both inputs simultaneously leads to an output "1". Such combination allows for the conguration of an IMPLICATION-type logic gate (Tables 3 and S2 †).

Conclusions
In the aim of developing excimer-based uorescent chemosensors, we synthesized and characterised two pyrene-thiourea lower rim calix [4]arene derivatives. Only pyrene monomer emission was observed for 1, while 2 presented both monomer and excimer emission. This reveals the critical role that CH 2 bridge plays for this type of excimer-based chemosensors. Fluorescent screening of 1 reveals a limited selectivity towards Hg 2+ , Cu 2+ and Ag + . Same experiments for chemosensor 2 displayed ratiometric detection for Hg 2+ and Ag + . Furthermore, this ratiometric variation is inverse for each cation, which allows for their selective detection. LODs for such complexes (2$Ag + and 2$Hg 2+ ) were both in the low nanomolar range (2.09 and 8.11 nM, respectively). Triggered by the complexity in identifying the coordination mode of 2, we performed computational conformational study of 2 and with the resulting complexes (2$Ag + and 2$Hg 2+ ). Boltzman distribution of 2 revealed that two conformations are present in the ground state. Study of these conformations with Ag + and Hg 2+ revealed that two differentiated binding sites are available in 2, where the stabilizing interactions were classied as through-bond or through-space, which have a different effect in the uorescence of 2 and therefore, in the sensing mechanism. To the best of our knowledge, we have not found similar studies that address or propose these two binding sites for pyrene-calixarene derivatives, which might help further understanding of sensing mechanism involved in alike uorescent chemosensors. Determination of K app by Hill equation reveals that 2 interacts stronger with Hg 2+ than with Ag + , which is in agreement with stability of 2$Ag + and 2$Hg 2+ complexes revealed by DFT calculations.
Additionally, chemosensor 2 was able to perform INHIBITtype and IMPLICATION-type logic operations when analysing either monomer or excimer emission, being the later scarce in comparison with other congurable molecular logic gates. We additionally demonstrate that 2 logic gates are able to operate with equimolar inputs. Further work will be address on improving the solubility towards more polar organic solvents or aqueous media of 2 by additional functionalisation at calix [4] arenes upper or lower rim.
In summary, we think this work could contribute to further development and understanding of excimer-based chemosensors.

Materials and methods
Reagents. All materials including solvents and perchlorate salts were purchased from commercial suppliers (Sigma The input corresponds to 1 eq. of corresponding perchlorate salt in each case to a dissolution 1.15 mM. b The addition was tested using alternating both input 1 and input 2.