Head vs. Tail Squaramide–Naphthalimide Conjugates: Self-Assembly and Anion Binding Behaviour

: The syntheses of two squaramide–naphthalimide conjugates ( SN1 and SN2 ) are reported; the structures of SN1 and SN2 differ by the attachment of a squaramide—either at the ‘head’ or the ‘tail’ of the naphthalimide ﬂuorophore. Both compounds displayed weak ﬂuorescence due to the inclusion of a nitro-aromatic squaramide which efﬁciently quenches the emission of the naphthalimide. Both compounds were also shown to undergo self-aggregation as studied by 1 H NMR and scanning electron microscopy (SEM). Furthermore, SN1 and SN2 gave rise to stark colourimetric changes in response to basic anions such as AcO − , SO 42 − HPO 42 − , and F − . The observed colour changes are thought to be due to deprotonation of a squaramide NH. The same basic anions also result in a further quenching of the naphthalimide emission. No colour change or emission modulations were observed in the presence of Cl − ; however, 1 H NMR studies suggest that moderate H-bonding occurs between this anion and both SN1 and SN2 .


Introduction
Squaramides have attracted a significant amount of research attention in recent years due to several physical and chemical characteristics that render them an ideal scaffold on which to build charge neutral anion receptors [1,2]. Their ability to partake in bi-directional H-bonding that increases the aromatic character of the cyclobutenedione ring [3,4] has been seen to be of particular benefit for the recognition of halides [5][6][7][8]. Indeed, Amendola et al. showed that a simple squaramide-based receptor forms 1:1 complexes with halide ions that are up to 2 orders of magnitude more stable than the corresponding urea-based receptor [9]. This finding has led to several examples of their exploitation as anion transporters [10][11][12][13] with the ability to disrupt anion homeostasis in cellular models whereby the transport ability of the receptors leads to apoptosis and induced autophagy [14,15]. In parallel, the 1,8-naphthalimide fluorophore has also found significant use in the construction of anion sensors whereby the ICT character of the naphthalimides can be modulated upon interaction with anionic species [16]. Gunnlaugsson et al. pioneered their use in this area where they reported a series of studies on naphthalimide-urea conjugates [17][18][19][20][21][22][23].
Our interest in anion recognition and sensing has centred on the squaramide motif where the aforementioned bidirectional H-bonding interactions in addition to π-π stacking brought about by the aromatic cyclobutenedione ring give a useful handle on which to build effective anion sensors [24][25][26][27][28][29][30][31]. We recently reported a series of naphthalimide-squaramide conjugates that aggregated in solution and were capable of sensitive detection of bromide through an 'aggregation-disggregation approach' [32]. The aggregation behaviour of these conjugates could be modulated upon bromide recognition where the disruption of H-bonding reversed self-assembly and gave rise to a 'switch on' fluorescence response. In the current work, we have designed and synthesised a further set of squaramidenaphthalimide conjugates in which we have changed the functionality of the phenyl substituent of the squaramide. Again, we have placed the squaramide at either the 'head' or the 'tail' of the naphthalimide. We envisioned that the inclusion of a nitro group in the structures would provide a stronger H-bonding receptor while also providing interesting photophysical properties to the conjugates.
Herein, we report the synthesis of compounds SN1 and SN2 ( Figure 1) and carry out an investigation on both their aggregation behaviour and their anion recognition properties.
Chemistry 2022, 4, FOR PEER REVIEW 2 to build effective anion sensors [24][25][26][27][28][29][30][31]. We recently reported a series of naphthalimidesquaramide conjugates that aggregated in solution and were capable of sensitive detection of bromide through an 'aggregation-disggregation approach' [32]. The aggregation behaviour of these conjugates could be modulated upon bromide recognition where the disruption of H-bonding reversed self-assembly and gave rise to a 'switch on' fluorescence response. In the current work, we have designed and synthesised a further set of squaramide-naphthalimide conjugates in which we have changed the functionality of the phenyl substituent of the squaramide. Again, we have placed the squaramide at either the 'head' or the 'tail' of the naphthalimide. We envisioned that the inclusion of a nitro group in the structures would provide a stronger H-bonding receptor while also providing interesting photophysical properties to the conjugates. Herein, we report the synthesis of compounds SN1 and SN2 ( Figure 1) and carry out an investigation on both their aggregation behaviour and their anion recognition properties.

General Experimental Methods
All chemicals purchased were obtained from commercial suppliers and used without further purification. DCM was distilled over CaH2 and MeCN was dried over 3 Å molecular sieves. Anhydrous DMF was purchased from Sigma Aldrich (Arklow, Ireland). TLC was performed on Merck Silica Gel F254 plates (Merck Life Science Limited, Cork, Ireland) and visualised under UV light (λ = 254 nm). Flash chromatography was performed with Merck Silica Gel 60 (Merck Life Science Limited, Ireland). Compounds were lyophilised on a Labconco Freezone 1 Dry system. NMR spectra were recorded on a Bruker Ascend 500 NMR spectrometer, operated at 500 MHz for 1 H NMR analysis and 126 MHz for 13 C analysis, both at 293 and 343 K. The residual solvent peak was used as an internal standard for DMSO-d6 and TMS for CDCl3. Chemical shifts (δ) were reported in ppm. NMR spectra were processed and stack plots produced using MestReNova 6.0.2 software. Proton and carbon signals were assigned with the aid of 2D NMR experiments (COSY, HSQC, and HMBC). Multiplicity is given as s = singlet, bs = broad singlet, d = doublet, brd = broad doublet, dd = doublet of doublets, ddd = doublet of doublet of doublets, t = triplet, q = quartet, m = multiplet as appropriate, and J values are given in Hz. Infrared spectra were obtained via ATR as a solid on a zinc selenide crystal in the region of 4000-400 cm −1 on a Perkin Elmer Spectrum 100 FT-IR spectrophotometer. LC-MS was performed on an Agilent Technologies 1200 Series instrument consisting of a G1322A Quaternary pump and a G1314B UV detector (254 nm) coupled to an Advion Expression L Compact Mass spectrometer (ESI) operating in positive mode. High-resolution ESI spectra were recorded on an Agilent 6310 LCMS TOF. UV/Vis spectra were recorded on a Varian Cary 50 Scan UV/Vis spectrophotometer and fluorescence spectra were measured using a Cary Eclipse operating Cary WinUV software. Diethyl squarate was synthesised as previously described [33].

General Experimental Methods
All chemicals purchased were obtained from commercial suppliers and used without further purification. DCM was distilled over CaH 2 and MeCN was dried over 3 Å molecular sieves. Anhydrous DMF was purchased from Sigma Aldrich (Arklow, Ireland). TLC was performed on Merck Silica Gel F254 plates (Merck Life Science Limited, Cork, Ireland) and visualised under UV light (λ = 254 nm). Flash chromatography was performed with Merck Silica Gel 60 (Merck Life Science Limited, Ireland). Compounds were lyophilised on a Labconco Freezone 1 Dry system. NMR spectra were recorded on a Bruker Ascend 500 NMR spectrometer, operated at 500 MHz for 1 H NMR analysis and 126 MHz for 13 C analysis, both at 293 and 343 K. The residual solvent peak was used as an internal standard for DMSO-d 6 and TMS for CDCl 3 . Chemical shifts (δ) were reported in ppm. NMR spectra were processed and stack plots produced using MestReNova 6.0.2 software. Proton and carbon signals were assigned with the aid of 2D NMR experiments (COSY, HSQC, and HMBC). Multiplicity is given as s = singlet, bs = broad singlet, d = doublet, brd = broad doublet, dd = doublet of doublets, ddd = doublet of doublet of doublets, t = triplet, q = quartet, m = multiplet as appropriate, and J values are given in Hz. Infrared spectra were obtained via ATR as a solid on a zinc selenide crystal in the region of 4000-400 cm −1 on a Perkin Elmer Spectrum 100 FT-IR spectrophotometer. LC-MS was performed on an Agilent Technologies 1200 Series instrument consisting of a G1322A Quaternary pump and a G1314B UV detector (254 nm) coupled to an Advion Expression L Compact Mass spectrometer (ESI) operating in positive mode. High-resolution ESI spectra were recorded on an Agilent 6310 LCMS TOF. UV/Vis spectra were recorded on a Varian Cary 50 Scan UV/Vis spectrophotometer and fluorescence spectra were measured using a Cary Eclipse operating Cary WinUV software. Diethyl squarate was synthesised as previously described [33].

Synthesis
The syntheses of compounds SN1 and SN2 were achieved through an analogous method previously reported by us (Scheme 1) [32]. For SN1, 4-bromo-1,8-naphthalimide 1 was reacted with ethylenediamine before nucleophilic addition of the product 2 to 4-nitrophenyl squarate monoester 3 to yield SN1 in 73% yield. Similarly, SN2 was synthesised by reaction of 4-nitro-1,8-naphthalic anhydride 4 with N-Boc-ethylenediamine to give product 5. Following TFA-mediated deprotection and reduction in the presence of H 2 and Pd/C, compound 7 was reacted with 3 to yield SN2 in 67% yield. All compounds and intermediates were fully characterised by 1 H NMR, 13   We have previously noted that the 1 H NMR spectra of analogous squaramide-naphthalimide conjugates give rise to complex NMR spectra showing a large degree of peak broadening owing to self-aggregation in solution. In order to investigate whether similar behaviour would be observed for SN1 and SN2, 1 H NMR and UV/Vis absorption measurements were carried out in addition to SEM analysis.

Variable-Temperature NMR and SEM Analysis
The 1 H NMR spectra of SN1 and SN2 in DMSO-d6 were measured at both 298 K and 343 K. There were clear differences observed between both spectra where at 298 K signals are broadened, suggesting that aggregation is occurring under these conditions. Conversely, at 343 K some signals become less broad and show a much-improved resolution while others become more broadened. Figure 2 shows the 1 H NMR spectrum of SN1 at 298 K and 343 K as an example. Naphthalimide signals appear to resolve from broad to sharp for each proton while the signals associated with the 4-nitrophenyl portion display the opposite behaviour-the signal at 8.2 ppm becomes broad at 343 K and shifts upfield to 8.05 ppm while the signal at 7.45 ppm also becomes broad and shifts upfield to 7.4 ppm. Similarly, the CH2 signals in the aliphatic region resolve from broad to sharp, clearly showing the correct multiplicity. The observation of peak broadening for the nitro-phenyl substituent suggests that this portion of the molecule is undergoing rapid rotation around the C-N bond axis at elevated temperature. This behaviour is reminiscent of a molecular rotor [34] and explains the observed peak broadening at elevated temperature. Conversely, the resolution of the CH2 signals suggests that this linker is also taking part in the self-assembly behaviour. The broad to sharp signal resolution signifies the disaggregation event. Taken together, the evidence from the above study suggests a temperature-dependent aggregation of the molecule that involves predominantly contributions from the naphthalimide and squaramide portions of the assembly. We have previously noted that the 1 H NMR spectra of analogous squaramidenaphthalimide conjugates give rise to complex NMR spectra showing a large degree of peak broadening owing to self-aggregation in solution. In order to investigate whether similar behaviour would be observed for SN1 and SN2, 1 H NMR and UV/Vis absorption measurements were carried out in addition to SEM analysis.

Variable-Temperature NMR and SEM Analysis
The 1 H NMR spectra of SN1 and SN2 in DMSO-d 6 were measured at both 298 K and 343 K. There were clear differences observed between both spectra where at 298 K signals are broadened, suggesting that aggregation is occurring under these conditions. Conversely, at 343 K some signals become less broad and show a much-improved resolution while others become more broadened. Figure 2 shows the 1 H NMR spectrum of SN1 at 298 K and 343 K as an example. Naphthalimide signals appear to resolve from broad to sharp for each proton while the signals associated with the 4-nitrophenyl portion display the opposite behaviour-the signal at 8.2 ppm becomes broad at 343 K and shifts upfield to 8.05 ppm while the signal at 7.45 ppm also becomes broad and shifts upfield to 7.4 ppm. Similarly, the CH 2 signals in the aliphatic region resolve from broad to sharp, clearly showing the correct multiplicity. The observation of peak broadening for the nitro-phenyl substituent suggests that this portion of the molecule is undergoing rapid rotation around the C-N bond axis at elevated temperature. This behaviour is reminiscent of a molecular rotor [34] and explains the observed peak broadening at elevated temperature. Conversely, the resolution of the CH 2 signals suggests that this linker is also taking part in the self-assembly behaviour. The broad to sharp signal resolution signifies the disaggregation event. Taken together, the evidence from the above study suggests a temperature-dependent aggregation of the molecule that involves predominantly contributions from the naphthalimide and squaramide portions of the assembly. To further probe the aggregation characteristics of both compounds, attempts were made to grow crystals from concentrated DMSO solutions. Unfortunately, in our hands the compounds did not crystallise and instead formed what appeared to be amorphous solids. The morphological features of these solids were thus analysed by scanning electron microscopy (SEM), and as shown in Figure 3, the SEM images exhibit interesting and distinct morphology on the nanoscale. SN1 appeared as an amorphous solid with no defined structure to speak of. Conversely, SN2 formed a porous sponge-like structure with a large degree of uniformity throughout the material. At higher magnifications, the material appeared to be composed of small 'nanoflake' type material. Although firm conclusions cannot be drawn on the exact material properties from these results, it is worth noting the difference in morphology brought about by the position of the 4-nitrophenyl-squaramide moiety, i.e., whether it is positioned at the head or the tail of the 1,8-naphthalimide. This behaviour was also seen with the 3,5-bis(trifluoromethyl)phenyl derivatives where this minor structural change gave rise to a significant change in material morphology as determined using SEM. To further probe the aggregation characteristics of both compounds, attempts were made to grow crystals from concentrated DMSO solutions. Unfortunately, in our hands the compounds did not crystallise and instead formed what appeared to be amorphous solids. The morphological features of these solids were thus analysed by scanning electron microscopy (SEM), and as shown in Figure 3, the SEM images exhibit interesting and distinct morphology on the nanoscale. SN1 appeared as an amorphous solid with no defined structure to speak of. Conversely, SN2 formed a porous sponge-like structure with a large degree of uniformity throughout the material. At higher magnifications, the material appeared to be composed of small 'nanoflake' type material. Although firm conclusions cannot be drawn on the exact material properties from these results, it is worth noting the difference in morphology brought about by the position of the 4-nitrophenyl-squaramide moiety, i.e., whether it is positioned at the head or the tail of the 1,8-naphthalimide. This behaviour was also seen with the 3,5-bis(trifluoromethyl)phenyl derivatives where this minor structural change gave rise to a significant change in material morphology as determined using SEM.
Both SN1 and SN2 were also examined using UV/Vis spectroscopy and fluorescence emission spectroscopy. The UV/Vis absorption spectrum of SN1 in DMSO showed two absorption maxima at 280 nm and 420 nm with extinction coefficient values of 34,375 M −1 and 29,900 M −1 , respectively. Similarly, SN2 showed an almost identical spectrum to SN1 with maxima also at 280 nm and 420 nm and with extinction coefficients of 35,187 M −1 and 27,410 M −1 , respectively. Interestingly, a third absorption maximum at 545 nm was observed in the case of SN2 attributed to a naphthalimide π-π stacking band (Figure 4) [35]. This peak was shown not to obey the Beer-Lambert law and further confirms the presence of a self-association event at higher concentrations of SN2. This behaviour was not observed for SN1 and exemplifies the sensitivity of squaramide-naphthalimide conjugates to changes in molecular structure when considering their aggregation characteristics. A summary of the absorption properties of SN1 and SN2 as well as their molar absorptivities is given in Table S1. peared to be composed of small 'nanoflake' type material. Although firm conclusions cannot be drawn on the exact material properties from these results, it is worth noting the difference in morphology brought about by the position of the 4-nitrophenyl-squaramide moiety, i.e., whether it is positioned at the head or the tail of the 1,8-naphthalimide. This behaviour was also seen with the 3,5-bis(trifluoromethyl)phenyl derivatives where this minor structural change gave rise to a significant change in material morphology as determined using SEM.  Chemistry 2022, 4, FOR PEER REVIEW 7 Both SN1 and SN2 were also examined using UV/Vis spectroscopy and fluorescence emission spectroscopy. The UV/Vis absorption spectrum of SN1 in DMSO showed two absorption maxima at 280 nm and 420 nm with extinction coefficient values of 34,375 M −1 and 29,900 M −1 , respectively. Similarly, SN2 showed an almost identical spectrum to SN1 with maxima also at 280 nm and 420 nm and with extinction coefficients of 35,187 M −1 and 27,410 M −1 , respectively. Interestingly, a third absorption maximum at 545 nm was observed in the case of SN2 attributed to a naphthalimide π-π stacking band (Figure 4) [35]. This peak was shown not to obey the Beer-Lambert law and further confirms the presence of a self-association event at higher concentrations of SN2. This behaviour was not observed for SN1 and exemplifies the sensitivity of squaramide-naphthalimide conjugates to changes in molecular structure when considering their aggregation characteristics. A summary of the absorption properties of SN1 and SN2 as well as their molar absorptivities is given in Table S1. Furthermore, both compounds exhibited very weak fluorescence emission at ca. 525 nm. With the aggregation behaviour observed for the previously reported analogues containing a 3,5-bis CF3 substituent, we also undertook a thermal study to investigate if the fluorescence of SN1 could be modulated by disaggregation. In this case, the fluorescence of SN1 is already significantly quenched due to the inclusion of a NO2 group which is known to efficiently quench emission from aromatic systems, including naphthalimides [36,37]. A temperature-dependent fluorescence study was undertaken in 5% aq. DMSO ranging from 298 K to 343 K where the fluorescence of both compounds was largely unaffected by temperature (see Supplementary Materials). The 1 H NMR results above suggest that SN1 and SN2 aggregate in solution but, upon heating, disassemble; however, this process cannot be monitored with fluorescence emission spectroscopy.
With the known propensity of squaramides to bind strongly to halides and other anionic species [9,10,[38][39][40][41][42], we expected that SN1 and SN2 may also be capable of anion recognition and that this process may be monitored by both UV spectroscopy and 1 H NMR spectroscopy. Furthermore, both compounds exhibited very weak fluorescence emission at ca. 525 nm. With the aggregation behaviour observed for the previously reported analogues containing a 3,5-bis CF 3 substituent, we also undertook a thermal study to investigate if the fluorescence of SN1 could be modulated by disaggregation. In this case, the fluorescence of SN1 is already significantly quenched due to the inclusion of a NO 2 group which is known to efficiently quench emission from aromatic systems, including naphthalimides [36,37]. A temperature-dependent fluorescence study was undertaken in 5% aq. DMSO ranging from 298 K to 343 K where the fluorescence of both compounds was largely unaffected by temperature (see Supplementary Materials). The 1 H NMR results above suggest that SN1 and SN2 aggregate in solution but, upon heating, disassemble; however, this process cannot be monitored with fluorescence emission spectroscopy.
With the known propensity of squaramides to bind strongly to halides and other anionic species [9,10,[38][39][40][41][42], we expected that SN1 and SN2 may also be capable of anion recognition and that this process may be monitored by both UV spectroscopy and 1 H NMR spectroscopy.

Anion Recognition
To investigate the ability SN1 and SN2 to bind anions in solution, a series of 1 H NMR experiments were carried out. Initial qualitative measurements were undertaken using a screening experiment in which 30 equivalents. of several anions (AcO -, HPO 4 2− , SO 4 2− , F -, and Clas their tetrabutylammonium salts) were added to the receptors in solution (0.5% H 2 O in DMSO-d 6 ). These preliminary results showed significant changes to the spectra of both SN1 and SN2, where the addition of AcO -, F -, HPO 4 2-, and SO 4 2led to the disappearance of the NH signal. This change was also coupled with a stark colour change from yellow to purple (or yellow to orange for SO 4 2-) in the cases of both receptors (see Figure 5 for SN2).

Anion Recognition
To investigate the ability SN1 and SN2 to bind anions in solution, a series of 1 H NMR experiments were carried out. Initial qualitative measurements were undertaken using a screening experiment in which 30 equivalents. of several anions (AcO -, HPO4 2− , SO4 2− , F -, and Clas their tetrabutylammonium salts) were added to the receptors in solution (0.5% H2O in DMSO-d6). These preliminary results showed significant changes to the spectra of both SN1 and SN2, where the addition of AcO -, F -, HPO4 2-, and SO4 2-led to the disappearance of the NH signal. This change was also coupled with a stark colour change from yellow to purple (or yellow to orange for SO4 2-) in the cases of both receptors (see Figure  5 for SN2). The disappearance of the NH signal and the colour change suggest that deprotonation of the squaramide/naphthalimide may be responsible, as has been previously observed [32]. We further confirmed this deprotonation behaviour by observation of bifluoride (HF2 − ) in the 1 H NMR spectra of SN1 and SN2 upon the addition of F -. In comparison, the changes upon the addition of Clwere more subtle. For example, with SN2, Clled to a large downfield shift of the NH signal (Ha) from 10.15 ppm to 11.5 ppm coupled with a significantly sharpened spectrum. The large downfield shift of the NH signal suggests a classical H-bonding interaction between Cland the NH protons of the squaramides. More detailed 1 H NMR spectroscopic titrations were carried out with Cl -, and the resulting data fit to a 1:1 binding model using the open access BindFit software program [43][44][45] to provide the apparent stability constants (Ka). Ka values of 273.7 M −1 (±4.1%) and 114.8 M −1 (±5.9%) were measured for receptors SN1 and SN2, respectively, in DMSO-d6 at 273 K. Figure 6 shows the changes observed in the 1 H NMR spectrum of SN2 upon the addition of Cl -(20 equivalents.). The disappearance of the NH signal and the colour change suggest that deprotonation of the squaramide/naphthalimide may be responsible, as has been previously observed [32]. We further confirmed this deprotonation behaviour by observation of bifluoride (HF 2 − ) in the 1 H NMR spectra of SN1 and SN2 upon the addition of F -. In comparison, the changes upon the addition of Clwere more subtle. For example, with SN2, Clled to a large downfield shift of the NH signal (H a ) from 10.15 ppm to 11.5 ppm coupled with a significantly sharpened spectrum. The large downfield shift of the NH signal suggests a classical H-bonding interaction between Cland the NH protons of the squaramides. More detailed 1 H NMR spectroscopic titrations were carried out with Cl -, and the resulting data fit to a 1:1 binding model using the open access BindFit software program [43][44][45] to provide the apparent stability constants (K a ). K a values of 273.7 M −1 (±4.1%) and 114.8 M −1 (±5.9%) were measured for receptors SN1 and SN2, respectively, in DMSO-d 6 at 273 K. Figure 6 shows the changes observed in the 1 H NMR spectrum of SN2 upon the addition of Cl -(20 equivalents.). Overall, both SN1 and SN2 were found to bind to Clwith moderate affinities, with SN1 showing a slightly increased affinity compared to SN2. As previously reported, the position of the squaramide at the 'tail' of the naphthalimide seems to be optimal for anion binding.
Having observed such stark colour changes while conducting the 1 H NMR titrations, we also wished to examine the interaction of SN1 and SN2 with anions using UV/Vis spectroscopy. The titrations were performed by additions of aliquots of the putative anionic guest (AcO − , Cl − , F − , HPO4 2− , SO4 2− ) as the tetrabutylammonium (TBA) salt solution (20 mM) in DMSO to a 1.2 × 10 −5 M solution of the receptor in DMSO. The changes in the UV spectra upon anion addition were monitored for each receptor. The overall changes in the ground state of SN2 with TBA SO4 2− are shown in Figure 7 as an example. Overall, both SN1 and SN2 were found to bind to Clwith moderate affinities, with SN1 showing a slightly increased affinity compared to SN2. As previously reported, the position of the squaramide at the 'tail' of the naphthalimide seems to be optimal for anion binding.
Having observed such stark colour changes while conducting the 1 H NMR titrations, we also wished to examine the interaction of SN1 and SN2 with anions using UV/Vis spectroscopy. The titrations were performed by additions of aliquots of the putative anionic guest (AcO − , Cl − , F − , HPO 4 2− , SO 4 2− ) as the tetrabutylammonium (TBA) salt solution (20 mM) in DMSO to a 1.2 × 10 −5 M solution of the receptor in DMSO. The changes in the UV spectra upon anion addition were monitored for each receptor. The overall changes in the ground state of SN2 with TBA SO 4 2− are shown in Figure 7 as an example. Upon titration with Cl − , very minor changes were seen in all cases. In contrast, much more significant changes were seen upon the addition of AcO − , F − , HPO4 2− , and SO4 2− . As shown in Figure 5, an increase in absorption at 545 nm was observed with a concomitant decrease at 280 nm and 420 nm and a clear isosbestic point at 460 nm. Similar changes were observed upon titration with AcO − , F − , and HPO4 2− . These results again suggest a deprotonation event at the squaramide binding site. The changes in the fluorescence spec- Upon titration with Cl − , very minor changes were seen in all cases. In contrast, much more significant changes were seen upon the addition of AcO − , F − , HPO 4 2− , and SO 4 2− . As shown in Figure 5, an increase in absorption at 545 nm was observed with a concomitant decrease at 280 nm and 420 nm and a clear isosbestic point at 460 nm. Similar changes were observed upon titration with AcO − , F − , and HPO 4 2− . These results again suggest a deprotonation event at the squaramide binding site. The changes in the fluorescence spectra of SN1 and SN2 were also monitored, but as expected, only minor changes were observed in the emission spectra of SN1 and SN2 upon titration with Cl -. Titration with AcO − , F − , HPO 4 2− , and SO 4 2− led to more significant changes where some further fluorescence quenching was observed, with a small emission decrease being observed at 430 nm and 460 nm for SN1 and SN2, respectively. We also ascribe this behaviour to the deprotonation observed in the NMR titrations where similar behaviour was also observed in the presence of all basic non-halide anions (see Supplementary Materials). Overall, the results of the anion binding experiments seem to suggest that a moderate H-bonding interaction occurs between Cl − and both SN1 and SN2 but results in very minor spectroscopic perturbation as measured by UV/Vis and fluorescence. Conversely, the interaction of either receptor with more basic anions results in a deprotonation that is accompanied by both stark colour changes as well as further fluorescence quenching.

Conclusions
In conclusion, we have reported the synthesis of two new squaramide-naphthalimide conjugates, SN1 and SN2. We have shown that these receptors undergo some degree of self-aggregation in aqueous DMSO as observed by 1 H NMR and SEM. We have also shown that both receptors are capable of classical H-bonding to Cland appear to undergo deprotonation in the presence of more basic anions such as AcO − , F − , HPO 4 2− , and SO 4 2− . Deprotonation leads to stark colour changes that are clearly observable to the naked eye and can also be monitored using UV/Vis absorption spectroscopy. Unfortunately, due to the quenching effect of the NO 2 substituent on the 1,8-naphthalimide, large changes in emission characteristics are not observed apart from minor quenching brought about in the presence of basic anions, again, thought to be a result of deprotonation. The position of the squaramide relative to the naphthalimide (i.e., at the head or the tail of the structure) seems to play a role in the strength of the association constant with Cl − . Under our experimental conditions, SN1, which has the squaramide appended to the 'tail' of the naphthalimide, showed a higher affinity when compared to SN2 where the squaramide is at the 'head' of the naphthalimide. We are currently optimising our design to take advantage of the aggregation-disassembly approach to anion recognition, and our results will be reported in due course.