The interplay between hydrogen and halogen bonding: substituent effects and their role in the hydrogen bond enhanced halogen bond

The hydrogen bond enhanced halogen bond (HBeXB) has recently been used to effectively improve anion binding, organocatalysis, and protein structure/function. In this study, we present the first systematic investigation of substituent effects in the HBeXB. NMR analysis confirmed intramolecular HBing between the amine and the electron-rich belt of the XB donor (N–H⋯I). Gas-phase density functional theory studies showed that the influence of HBing on the halogen atom is more sensitive to substitution on the HB donor ring (R1). The NMR studies revealed that the intramolecular HBing had a significant impact on receptor performance, resulting in a 50-fold improvement. Additionally, linear free energy relationship (LFER) analysis was employed for the first time to study the substituent effect in the HBeXB. The results showed that substituents on the XB donor ring (R2) had a competing effect where electron donating groups strengthened the HB and weakened the XB. Therefore, selecting an appropriate substituent on the adjacent HB donor ring (R1) could be an alternative and effective way to enhance an electron-rich XB donor. X-ray crystallographic analysis demonstrated that intramolecular HBing plays an important role in the receptor adopting the bidentate conformation. Taken together, the findings imply that modifying distal substituents that affect neighboring noncovalent interactions can have a similar impact to conventional para substitution substituent effects.


Introduction
Understanding how to modulate noncovalent interactions that are in close proximity is paramount to engineering functional materials, 1 supramolecular assemblies, 2 drugs, 3 and catalysts. 4 Foundational work has detailed the importance of substituent effects 5 to tune the electronics of noncovalent interactions. 6 Despite this pioneering work, few have looked at substituent effects involving multiple noncovalent interactions that are spatially close. Recently, Cockro, and coworkers reported rare experimental data quantifying through-space substituent effects on noncovalent interactions and presented the inadequacy of describing substituent effects using classic Hammett parameters when through-space effects dominate. 7 Additionally, Zonta, and coworkers utilized similar methods to carry out an experimental survey of aromatic stacking interactions in solution. 8 Clearly, understanding how adjacent noncovalent interactions inuence each other by substituent effects is critically lacking.
Among the myriad noncovalent interactions, HBs are privileged for their directionality and tunability. Halogen bonds (XBs) share similarities with HBs, yet contain an electron decient donor (halogen) that forms an attractive noncovalent interaction with an electron rich species. This interaction can be understood in term of electrostatics and covalency. 9 An electrostatic description suggests that due to the polarizability of the halogen, its electron density can become anisotropic. In this case, a partial positive potential develops on the halogen, opposite to the C-X s bond which has been coined the s-hole. Concurrently, an electron rich belt is generated on the XB donor which is orthogonal to the direction of the s bond. 10 The high directionality, tunability and complementarity with "so" Lewis bases, 11 makes XBing accordant with HBing. Given these unique properties, exciting applications of XBing supramolecule selfassembly, 12 molecular recognition, 13 anion binders, 14 organocatalysts, 15 and anion transporters 16 are appearing at a rapid pace. Understanding how to modulate both the strength of the XB and the structure of these molecules is of broad importance for the continued development of functional halogenated species. Using substituent effects to alter the electronics of the halogen donors remains a leading strategy. However, experimental substituent effects on XBing are limited (despite numerous computational studies 17 ). Only Taylor, 18 Diederich, 19 Erdelyi, 20 Stilinović 21 and Franz 22 have explicitly studied XBing substituent effects in solution-with different outcomes for each of their systems. Despite the critical importance of these substituent studies, the inuence of a neighbouring noncovalent interaction has not been studied.
We recently introduced a new strategy-the hydrogen bond enhanced halogen bond (HBeXB)-that directs intramolecular HBs to the electron rich belt of XB donors for preorganization and enhanced XB strength (Fig. 1). Recently, the HBeXB interaction has been used to increase anion binding affinity by nearly an order of magnitude and to improve the function of organocatalysts. 14b, 15a Similarly, Ho and his group have noted the HBeXB in biological settings and employed it to stabilize and improve a T4 lysozyme mutant. 23 These studies have been complemented by fundamental studies as well. 14b,24 Over the course of these seminal works there are implications of reciprocity between the hydrogen and halogen bond, especially when considering circumstances where augmentation and preorganization are simultaneously operating. In fact, using substituent effects to modulate the HBeXB may not be straightforward as there are potentially competing electronic effects. Making the halogen atom more electron rich may decrease the XB donor ability but will increase the strength of the adjacent HB. The subtle interplay between these two interactions naturally leads one to ask important questions. For example, when optimizing binding, is it more efficient to tune the XB strength or the HB strength with electronic effects? Additionally, how does this inuence molecular conformations? To address these questions and more, we present the rst HBeXB substituent and LFER studies in solution, gas, and solid phases.

Receptor design
We previously developed two generations of bis-ethynyl XB receptors that presented two charged pyridinium XB donors in a convergent manner (Fig. 2 top). In the 2 nd -generation receptor we introduced an amine substituent to the core that provided intramolecular N-H/I HBs to the XB donors. This innovation improved binding by nearly an order of magnitude over a control lacking the amine. 14b We coined this effect the HBeXB and showed that the enhancement was due to both preorganization of the receptor into a bidentate binding conformation and strengthening of the XB. The success of the HBeXB and our experience with this scaffold prompted us to evaluate HBeXB substituent effects using a 3 rd -generation anion receptor presented here (G3XB, Fig. 2 bottom). This latest iteration replaces the anking iodopyridinium rings with neutral aromatic rings to improve solubility in organic solvents. The redesign also reduced the number of synthetic steps to produce the diverse range of receptors needed to examine substituent effects. Specically, we prepared a series of compounds that contained substituents of varying electronic properties that were para to both the amine (R 1 -G3XB) and the iodine donors (2R 2 -G3XB). This design permitted systematic modulation of the electron density on both the HB and XB donor rings to test substituent effects. Controls with triuoromethyl substituents (the strongest electron withdrawing groups in this study) were also prepared without the amine (nHBeXB) and without the iodine XB donors (G3HB).

Experimental evidence of intramolecular hydrogen bonding
Analysis of the amine 1 H NMR resonances provided initial indication of intramolecular HBing between the amine and the electron rich belt of the XB donor (N-H/I). The analysis provided a preliminary evaluation of substituent effects and provided rare experimental evidence of HBing to larger halogens. 24a,25 Control receptor G3HB, lacking XB donors to accept HBs, had an amine 1 H chemical shi of 4.64 ppm in C 6 D 6 , whereas the chemical shi for G3XB, with iodine acceptors, was 5.53 ppm. This 0.89 ppm downeld shi is indicative of intramolecular HBing. The series of 2R 2 -G3XB derivatives (Fig. 3, le) showed that as the para substituent on the XB donor ring became more electron donating, the HBing amine proton shied downeld from 5.53 ppm to 5.80 ppm. This downeld shi occurs despite the expectation that adding electron donating substituents should shield the nuclei and would produce an upeld shi for the proton. However, the electron donating groups in the 2R 2 -G3XB compounds transfer additional electron density onto the iodine atoms (vide infra) making the iodine a better HB acceptor, resulting in the downeld shi. ‡ Overall this NMR analysis suggests that the intramolecular HBing is formed and the strength of this HBing correlates with the electron density of the halogens.

Computational evaluations
To garner preliminary insight into substituent effects of HBeXBs, in silico studies were performed. All the receptors were evaluated using Gaussian 16 at the M06-2X/def2TZVPP level of theory (see ESI † for further details). A systematic use of conformational analysis and electrostatic potential (ESP) mapping provided early insight into the synergy between the HB and XB and helped to stimulate deeper LFER analysis.

s-Hole calculations
s-Hole analysis (maximum/minimum electrostatic potentials, denoted by V s,max /V s,min ) showed several trends regarding the inuence of the intramolecular HB on XB donor strength. Consistent with our previous studies, the HB from the amine to the iodine augments the s-hole (greater V s,max ). For example,   G3XB in the bidentate conformation has a V s,max that is 4.02 kcal mol −1 greater than nHBeXB-a consequence of the HBeXB (Table 1).

Substitution on the HB donor ring
Modulation of the R 1 group helped determine if stronger HBs would correlate to greater XB enhancement. Within the R 1 series, the V s,max values on the XB donors ranged from 32.05 kcal mol −1 for the most electron withdrawing G3XB in the bidentate conformation to 28.11 kcal mol −1 for Me-G3XB ( Table  1). The data verify that stronger HB donors correlate with a more positive V s,max -conrming a notion alluded to in our previous HBeXB systems. 14b,24a Our results here highlight that tuning the HB donor with a single remote substituent can inuence the XB donor V s,max by 3.94 kcal mol −1 .

Substitution on the XB donor ring
Next, the V s,max of the 2R 2 -G3XB derivatives in the bidentate conformation was probed to determine how substituent electronics on the XBing ring inuence XB donor strength. As expected, the V s,max becomes larger as the substituent para to the XB donor increases in electron withdrawing capacity. The G3XB derivative with CF 3 substituents represents the upper end in this series with a V s,max of 32.05 kcal mol −1 whereas, 2tBu-G3XB represents the lower end at 22.41 kcal mol −1 . These studies align with previous reports 18,19,22 showing that XB donors can be directly inuenced by electronic substituent effects. For instance, the V s,max of the XB donors in Talyor's 4-R-C 6 F 4 I studies differ by 6.9 kcal mol −1 when a para uoro substituent is changed to a piperidyl group. 18 However, it is notable that in this bisethynyl system the V s,max varies by 9.64 kcal mol −1 by direct substitution on the XB ring. In contrast to the above discussion on XB donor strength (i.e. V s,max ), we also considered the V s,min of the iodine atoms as a measure of HB acceptor capacity. As the substituent para to the XB donor becomes more electron donating, the iodine species becomes more electron-rich (Fig. 3). The V s,min of the 2R 2 -G3XB compounds (−1.11 to −11.14 kcal mol −1 , for details see ESI †) trend with the downeld NMR shiing-suggesting that R 2 electron donating groups strengthen the intramolecular HBing between the amine and the XB donor.
Conformational effects on the XB donor R 1minimal through bond effects on the XB donor. ESP maps were calculated for the R 1 -G3XB derivatives in the S conformation (Fig. 4). The S conformation contains one XB donor that accepts a HB and one that does not. This S arrangement was used to determine whether substitution on the HB donor ring has a through bond electronic effect on the XB donor. The S conformation of R 1 -G3XB derivatives all exhibit similar V s,max values (z24 kcal mol −1 ) for the iodine not accepting a HB. This demonstrates that substituents on the central HB donor ring do not directly inductively alter the electronics of the XB donor. § R 1through space effects on the XB donor. In contrast, the potency of the HB donor does have an inuence on the XB donor strength. The V s,max of the halogen accepting a HB has a greater ESP than the non-HB accepting iodine. The V s,max values for the iodine in G3XB that accepts a HB was 31.45 kcal mol −1 and 27.36 kcal mol −1 for Me-G3XB which has the most electron donating substituent R 1 . Thus, strengthening the intramolecular HB donor can modulate the V s,max of a single XB donor by 4.09 kcal mol −1 , a value similar to the bidentate assessment described above.
The through space inuence of a HB on the XB donor was analyzed by computing the difference (DV s,max ) between the two iodine donors in the S conformation (DV s,max = V s,max HBed iodine − V s,max non-HBed iodine). Here the weakest HB donor Me-G3XB derivative had a DV s,max of 3.31 kcal mol −1 while the strongest HB donor G3XB had the largest DV s,max of 6.99 kcal mol −1 . Collectively the DV s,max values of the R 1 -G3XB derivatives adhere to the trend that increasing the electron withdrawing ability of R 1 , strengthens the HB which in turn has a larger inuence on the XB donor. R 2through bond effects on the XB donor ring{. V s,max values for 2R 2 -G3XB derivatives in the S conformation were used to quantify how changing the electronics of the XB donor ring inuences the iodine s-hole. For example, the V s,max of the externally directed iodine (non-HBed iodine) generally followed the trends expected from the electronic contributions of the R 2 group (i.e. V s,max values for 2R 2 -G3XB trended in the order CF 3 > F > H > Me > tBu) ranging from 24.46 to 16.15 kcal mol −1 . The internally directed iodine atoms (accepting a HB) all had larger V s,max values ranging from 31.45 to 24.37 kcal mol −1 and generally followed the same trend. R 2through space effects on the XB donor †. The strength of the HB donor in the 2R 2 -G3XB derivatives is constant (i.e. a CF 3 group para to the central amine). Thus, the DV s,max values here are a measure of how the R 2 electronics impact the halogen as both a XB donor and a HB acceptor. The DV s,max values of 6.99, 7.58, 8.10, 7.75 and 8.22 kcal mol −1 were obtained for G3XB, 2F-G3XB, 2H-G3XB, 2Me-G3XB and 2tBu-G3XB, respectively. The trend parallels previous evaluations and shows that generally, more electron rich iodines experience a greater augmentation.
Comparing 2R 2 -G3XB and R 1 -G3XB DV s,max values provides a measure of which substituent position impacts XBing the most by a through space effect. The smaller range of values for the 2R 2 -G3XB series (1.23 kcal mol −1 ), as compared to the R 1 -G3XB derivatives (3.68 kcal mol −1 ), suggests that the inuence of HBing on the halogen atom is more sensitive to substitution on the HB donor ring. For example, altering one R 1 substituent from CF 3 to Me results in a 3.68 kcal mol −1 DV s,max difference. However, altering two R 2 substituents from CF 3 to Me results in a 0.76 kcal mol −1 DV s,max difference.
Inuence of R 1 and R 2 on conformation. To assess the role of preorganization in G3XB derivatives, their relative stabilities were assessed based on electronic energies from DFT. The difference between the S and bidentate conformation energy illustrates that intramolecular HBeXBs stabilize the bidentate conformation of the receptors ( Table 2). The bidentate conformation of G3XB contains two intramolecular HBs (N-H/I) and is more stable than the S conformation by 1.53 kcal mol −1 . The W form, lacking intramolecular HBs is 3.11 kcal mol −1 higher in energy than the bidentate conformation. For the 2R 2 -G3XB series, as R 2 becomes more electron donating, the energy differences between the bidentate and S conformation increases from 1.76 kcal mol −1 for 2F-G3XB to 2.15 kcal mol −1 for 2tBu-G3XB. This suggests a greater stabilization when the iodine HB acceptor is more electron rich. These results track with the amine 1 H NMR chemical shi analysis for 2R 2 -G3XB. In contrast, the difference between the bidentate and S conformation for the R 1 -G3XB series is comparatively attenuated; however, only a single substituent group is modied. The DE is 1.42 kcal mol −1 , 1.33 kcal mol −1 , 1.43 kcal mol −1 , and 1.36 kcal mol −1 for Cl-G3XB, F-G3XB, H-G3XB and Me-G3XB, respectively. These data indicate that conformational preference is sensitive to the electronics of both the XB and HB donor.

Solution studies
NMR titrations and association constants 1 H NMR anion binding titrations were performed to quantify HBeXB substituent effects in solution. Titrations were conducted in C 6 D 6 with tetra-n-hexylammonium iodide (THAI) as the guest to ensure all complexes remained in solution. The addition of THAI resulted in downeld shis for nearly all of the 1 H NMR signals on the receptors, (except for a center core singlet of nHBeXB). Bindt 27 was used to t the changes in the  Table 3 Measured association constants and binding energies for The K a values are reported as the average of three titration experiments. All titrations were performed in C 6 D 6 ; two signicant gures are reported and errors are estimated at 10%. Tetra-nhexylammonium iodide was used and titrations were performed at 25°C. Bindt was used to t changes in chemical shi to a stepwise 1 : 1 host-guest binding model. The free energy of binding (DG binding ) was calculated from the association constant. 1 H NMR signals to a 1 : 1 binding model. Iterative and simultaneous renement of multiple isotherms provided association constants (K a ) for all scaffolds ( Table 3).

Role of intramolecular HBing on anion binding
G3XB (all R groups -CF 3 ) had the strongest binding (420 M −1 ) which was nearly 14 times greater than the isostructural no XB control (G3HB) (30 M −1 ). The substantially lower binding affinity of G3HB suggests that the amine doesn't signicantly HB to the iodide guest; the amine of the G3XB derivatives largely forms intramolecular HBs with the iodine XB donors. The considerable inuence of the intramolecular N-H/I HBs is evident when comparing G3XB to the control lacking an amine (nHBeXB). nHBeXB exhibited very weak binding in solution with a K a = 10 M −1 . The nearly 50-fold difference in K a between G3XB and nHBeXB demonstrates the striking impact that a weak intramolecular N-H/I HB can have on receptor performance. The HBeXB enhancement is far greater than our original HBeXB study using a dicationic receptor where only a 9fold increase was observed. 14b The greater HBeXB inuence in this study could be due to the iodine XB donors of G3XB being more electron rich (i.e. neutral receptor) than the iodopyridinium donors previously evaluated-allowing for stronger HBeXB and greater preorganization. It could also be attributed to solvent effects as the two studies were conducted in signicantly different media (C 6 D 6 vs. 60% CD 3 NO 2 /40% CDCl 3 ). These binding studies highlight that the central amine interacts minimally with the iodide and largely operates as an intramolecular HB donor to the iodine XB donor atoms.

2R 2 -G3XB substituent effects on anion binding
The 2R 2 -G3XB (R 2 ]CF 3 , F, H, Me, tBu) series of molecules were used to quantify how substituents para to the XB donor inuence the HBeXB. Varying these substituents resulted in association constants for THAI ranging from 50 M −1 to 420 M −1 (Table  3). Having a stronger electron withdrawing group para to the XB donor increases the XB strength and for this series of compounds this generally holds true. G3XB (R 2 ]CF 3 ) maintained the greatest affinity followed by 2F-G3XB with a K a = 170 M −1 , which is 60% less than G3XB. 2H-G3XB binding was further diminished to a K a of 70 M −1 . The most electron rich 2tBu-G3XB exhibited similar iodide binding with the 2Me-G3XB derivative (50 vs. 50 M −1 , respectively). Nevertheless, the general trend in substituent effects matches previous studies for XB derivatives. 18-22

R 1 -G3XB substituent effects on anion binding
While studies have evaluated the inuence of functional groups on the XB and HB independently none have considered their interplay. Here binding studies of R 1 -G3XB derivatives represent the rst experimental consideration of this. The binding data for R 1 -G3XB derivatives (R 1 ]CF 3 , Cl, F, H, Me) highlights that stronger intramolecular HBing enhances the XB receptor binding affinity. By increasing the electron withdrawing capacity of the substituent para to the amine (strengthening the HB donor) from a Me group to a CF 3 group the binding increased 2.5-fold. Me-G3XB, the most electron rich of the R 1 -G3XB series, had the lowest association constant (170 M −1 ), while the most electron decient G3XB had the highest (420 M −1 ). This result reveals an effective way to increase the overall XB binding ability in a system which includes an intramolecular HB to XB donor-electronically tuning the HB with substituents rather than the XB. R 1 and R 2 interplay: receptor performance As noted above, the receptor performance can be modulated by either changing the substituents para to the XB donor or para to the HB donor. To further quantify the substituent effects on binding DG binding was calculated for each receptor. The binding energy for 2R 2 -G3XB derivatives can be tuned by 1.3 kcal mol −1 simply by changing out the two CF 3 groups to Me groups on the XBing rings. Intriguingly, altering only one substituent (from CF 3 to Me) on the center ring can elicit a 0.5 kcal mol −1 change. This suggests that binding can be modied by a comparable amount with a smaller structural change to the receptor. These small energetic changes can have large implications, as previously demonstrated in a study of XB catalyst transition state binding. 28 R 1 and R 2 interplay: linear free energy relationships Linear free energy relationships (LFERs) are gaining importance in understanding substituent effects on noncovalent interactions like HBing, XBing, chalcogen bonding, cation-p and pp. 17,29 There are surprisingly few studies that have experimentally examined substituent effects by evaluating LFERs on the XB. [18][19][20][21][22] Taylor and coworkers adeptly used this approach to evaluate the XB between para-substituted tetrauoroiodobenzene and tributylphosphine oxide (Fig. 5a). These studies showed the best correlation of association constants with the s meta parameter (R 2 (s m ) = 0.94 vs. s para R 2 (s p ) = 0.82) which they attributed to inductive/eld effects being more dominant. 18 In contrast, Diederich 19 and Franz 22 interestingly reported strong correlation for the s para parameters. Diederich evaluated XBing between 4-R-iodoethynylbenzene and quinuclidine (R 2 (s p ) = 0.97 vs. R 2 (s m ) = 0.82, Fig. 5b). The strong correlation with the s para parameter in this case indicated that substituents largely inuence the halogen donor through resonance in this conjugated system. Finally, Erdélyi 20 and Stilinović 21 investigated substituent effects on XB acceptors. In Erdélyi's case, the three-center [N-I-N] + XB (Fig. 5c) exhibited linear correlation (R 2 = 0.97) between the calculated natural atomic populations with s para parameters. In all cases, substituent effects were shown to have a signicant inuence on XBing. Nevertheless, each of these examples showed great correlation with different parameters for their LFERs. Despite these important studies, there are no experimental examples analyzing substituent effects on adjacent noncovalent interactions. Supramolecular contacts don't occur in an isolated environment and as such, it is essential to understand whether traditional substituent effects hold true in these situations. Herein, we evaluated substituent effects on the HBeXB and used LFERs to establish whether changing the electronics of the HB donor has the same inuence on the overall binding as changing the electronics of the XB donor.
Our initial evaluation of LFERs compared the bidentate ESP values (V s,max ) of receptors with experimental DG values from titration studies. While XBing is known to encompass both covalent and electrostatic components, 9 plots of ESP vs. DG can establish the degree of electrostatic contribution in these particular HBeXB complexes. The following results demonstrate that in this system, the substituent inuence is largely electrostatic in nature.
The 2R 2 -G3XB derivatives show strong correlation between the electrostatics (V s,max ) of the XB donor and the iodide binding in solution. The non-normalized plots (Fig. 6, top) are linear for the 2R 2 -G3XB derivatives (R 2 = 0.99). Since the modication was directly on the XBing ring, this nding was expected given previous LFER studies on the XB. 18 In contrast, the R 1 -G3XB derivatives provided original insight into how the amine HB donor augments the XB donor strength. The LFER analysis for the R 1 -G3XB series certainly suggests an electrostatic origin (Fig. 6, bottom) as the plots are again linear (R 2 = 0.99). Collectively the ESP vs. DG plots indicates that the HBeXB iodide binding of the G3XB derivatives is largely governed by electrostatics and that ESP maps accurately model the inuence of intramolecular HBs on the XBs in this system.
We extended our LFER analysis by evaluating the correlation between our experimental association constants and Hammett parameters (s meta and s para ). While Hammett parameters were originally used to model the ionization reaction of benzoic acid, they have been increasingly used to study noncovalent interactions. 6- 8 We used Hammett parameters to analyze possible inductive (or eld effects as proposed by Wheeler and Houk 30 ) and resonance effects on the HBeXB receptor bindingan approach recently used by Hunter to effectively assess HB cooperativity. 31 Normalized association constants (log(K r /K H )) and the corresponding substituent parameters (s) were t using the Hammett equation shown below. r represents the slope.

log(K R /K H ) = rs
The R 1 -G3XB derivatives show a more linear correlation with the s para parameter (R 2 (s p ) = 0.93) than with the s meta (R 2 (s m ) = 0.87), indicating resonance effects are more important than inductive effects on the HB donor of the HBeXB (Fig. 7, top). Electron withdrawing R 1 substituents enhance the amine donor strength, thus leading to stronger intramolecular N-H/I HBing. A stronger HB in turn makes the XB donor more electron decient. Thus, the resonance of the R 1 substituent work in concert with both the HB and XB donors, resulting in a linear correlation with the s para parameters.
Curiously, normalized Hammett plots of K a values for 2R 2 -G3XB resulted in similar linear ts with both the s meta and s para parameters (Fig. 7, bottom). The t with the s meta parameters is R 2 (s meta ) = 0.95 while the t with the s para parameters is R 2 (s para ) = 0.92, implying that inductive effects may play a modestly more important role on substituent effects. As noted above, substituent effects on the XB have been found to be attributed to either inductive 18 or p-resonance effects. 19,20 It is atypical in LFER studies to obtain linear ts for both the s para and the s meta parameters simultaneously. 32 Unlike previous LFER studies, in this case there are two noncovalent interactions involved (N-H/I HBing and C-I/I − XBing). The halogen here functions as both a XB donor and a HB acceptor. The electronics of the R 2 substituents can have competing inuences on the noncovalent interactions. Thus, increasing the electron density on the halogen should weaken the XB donor but strengthen the intramolecular HBs with the amine. This competing effect produces good correlation with both the s para Fig. 6 Non-normalized plots of the ESPs and binding energies for 2R 2 -G3XB (top) and R 1 -G3XB (bottom). and the s meta parameters. Thus, when competing electronic inuences are present, selecting an appropriate para substituent for a XB donor could be tricky if relying on Hammett parameters. We also looked for linear correlations between the association constants and other parameters including Ta's s I , s R , 6d sEDA and pEDA. 33 However, no linear correlations were obtained (see ESI †). The combined LFERs herein, suggest that choosing an appropriate substituent (R 1 ) to tune adjacent noncovalent interactions could complement the common strategy of directly altering the electronics of the XB donor (R 2 ) to modulate binding.
Classically, the slope r in the Hammett equation describes the susceptibility of the reaction to substituents. While studying noncovalent interactions, r provides a measure of how sensitive the interaction is to substituent effects as compared to the ionization of benzoic acid. The r value of the R 1 -G3XB Hammett plot (rR 1 = 0.59) is between 0 and 1 indicating that the binding is sensitive to electronics (although not as much as the ionization of benzoic acid). More importantly, the r value for R 1 -G3XB can be compared to 2R 2 -G3XB (statistically taking into consideration the number of substituents); thereby, determining quantitatively which substituent position has a greater inuence on receptor performance. There are two R 2 groups affecting the anion binding per receptor, so half the r value was used for comparison. The 1/2r value for R 2 -G3XB (1/2rR 2 = 0.79) is only modestly higher than the one obtained for the R 1 -G3XB series (rR 1 = 0.59). This is notable as the R 1 substituents are much further from the binding site and maintain minimal through bond electronic inuence on the XB. The data here indicate that altering the electronics of an HB donor within the context of the HBeXB can have a similar effect on XB strength as traditional substituent effects.
For a more thorough picture of how resonance contributes to the interaction, multivariable linear regression of the normalized association constants (log(K r /K H )) with the eld (F) and resonance (R) parameters in the Swain-Lupton equation 5b were conducted in Matlab. The percent resonance contribution (%R) in this equation affords a simple and meaningful way to assess the relative importance of eld (F) and resonance (R) effects.
As shown in Table 4, %R = 49% for 2R 2 -G3XB which suggests that the binding is slightly more governed by inductive/eld effects from the substituents. The %R for 2R 2 -G3XB reveals the similarity in the impacts of resonance and inductive/eld effects on the XB donor ring's R 2 substituents. This similarity aligns with the 2R 2 -G3XB Hammett plots, which exhibit linear relationships with both the s para and s meta parameters due to the competing inuences from the electronics of the R 2 . In contrast, the substituent effects in R 1 -G3XB are more dependent on resonance where %R = 54%. However, it should be noted that the Hammett studies employed used relatively few parameters which limits the statistical strength of this Swain Lupton analysis.

Crystal structures
HBeXB impact on solid-state features. Crystal structures of several receptors were obtained to further evaluate substituent effects. While previous studies have evaluated the HBeXB in the solid-state, 11a,14b,23a,24a the HBeXB bidentate receptors reported here are the rst structures considered the structures within the context of substituent effects.
The initial solid-state assessment of the neutral receptors focused on three species (G3XB, G3HB, nHBeXB) to identify the inuence of the HBeXB. Previous generations were charged and thus in the solid state could be more inuenced by induced t binding. This series further conrms that the amine HBs promote a bidentate conformation.  As shown in Fig. 8 (le), the G3XB structure displays convergent bidentate XBing conformations promoted by the intramolecular HBing with N-H/I distances and angles of 3.12(4) Å, 163(3)°and 3.20(3) Å, 165(4)°. G3XB crystalized in the monoclinic space group P2 1 /c with a single molecule in the asymmetric unit. The XB donors are directed towards an iodine atom of an adjacent molecule. One of the iodine atoms forms a XB with C-I/I distances and angles of 3.9379(6) and 167.38(11)°(R II = 0.97). 34 The other halogen while directed at an iodine, is too far to XB with a C-I/I distance of 4.2007(11) and 173.56(9)°.
Notably, the exchange of the central amine group for a CH proton results in a molecule that adopts the W conformation ( Fig. 8 right), the form energetically favoured in the computational analysis. nHBeXB crystalizes in P2 1 /c with a single molecule in the asymmetric unit. One of the iodine atoms forms a XB with a symmetrically equivalent species on an adjacent molecule (C-I/I of 160.91(16)°and 3.8065(6) Å (R II = 0.93)). The other iodine has type II halogen contacts 35 with disordered uorine atoms of an adjacent molecule, with contacts that are less than the sum of the van der Waals radii.
Two crystal structures were obtained of G3HB, a triclinic and orthorhombic polymorph. Neither structure adopted a bidentate conformation. The triclinic (P 1) form has a single molecule in the asymmetric unit, and does not adopt one of the planar forms (i.e. bidentate, S, or W). The central amine does not play a signicant role in the packing. The orthorhombic (Fdd2) structure of G3HB has two molecules in the asymmetric unit. While each molecule adopts a W conformation, one molecule is much less planar than the other. The distortion of one species in the asymmetric unit may come from the arms having to deect to maintain a head-to-tail HBing chain.
The systematic changes of XB donors and HB donors in the G3XB, G3HB, nHBeXB series highlights that the N-H/I intramolecular HBing plays a key role in the receptor adopting the bidentate conformation. This is further demonstrated in the structures of derivatives containing different substituents on the anking iodine containing arms.
2R 2 -G3XB in the solid state. To further consider the interplay between the XB and HB we obtained crystal structures of species that modulate the electron density on the XB donating arms. Altering the functional group para to the iodine donor permitted us to evaluate two different features related to HBeXB within this system. First, the ability of the N-H/I HBing to bias the bidentate conformation was evaluated. Second, the rst solid state evaluation of substituent effects with relation to the HB acceptor capacity of iodine donors was investigated.
The planarity of the molecules did not follow the expected trend. For example, despite having shorter HB contacts 2H-G3XB was more distorted than G3XB. Angles between anking rings and central rings of the two unique 2H-G3XB species were 1.68°and 15.54°for one receptor and for the other was 4.66°a nd 17.53°. A partial explanation for this is that the two structures have different long range packing features (see ESI Fig. S90 †).
The angle formed by the centroids of the three rings (i.e. anking-core-anking angle) provides another structural measurement that indicates HBing strength between the amine hydrogen and more electron rich iodine atoms. For example, a smaller angle would suggest a stronger attraction distorting the alkyne bonds linking the core to the arms. Comparing G3XB, and 2H-G3XB, the increasing electron density on the iodine led to greater distortion with angles of 118.94°for G3XB vs. 117.18°for 2H-G3XB. R 1 -G3XB in the solid state. Several R 1 -G3XB derivatives (Me-G3XB, F-G3XB, and G3XB) that modulate the electron density of the central amine were also crystalized to evaluate the effects that these substituents have on the solid-state structures.
Similar to the 2R 2 -G3XB series, initial analysis compared the HB distances. G3XB maintained the shortest N-H/I HB contacts (see above for distances and angles) as suggested from the electron withdrawing nature of the CF 3 group. In contrast, the longest N-H/I HB contacts came from the Me-G3XB derivative. This methyl species crystalized in the monoclinic space group P2 1 /c with a single receptor in the asymmetric unit that adopts the bidentate conformation. The N-H/I HB contacts were 3.16(6) Å, 154(7)°, and 3.48(5) Å, 148(5)°. The longer HB contacts are attributed to the electron donating nature of the methyl group. Unfortunately, F-G3XB adopted the S conformation making comparisons across the series irrelevant. However, comparison of the bidendate G3XB and Me-G3XB suggests that the electronics of the aniline core may have some structural impact on the receptor. For example, the weaker HBs of Me-G3XB resulted in a less planar receptor. For Me-G3XB the angles between the planes of the core and anking arms were 5.29°and 8.75°as compared to G3XB which was 1.06°and 8.59°. Another parameter evaluated was the angle formed by the centroids of the three rings. For the electron rich Me-G3XB this angle is 123.16°whereas for G3XB the angle is 118.94°. This suggests that the amine donors are stronger in G3XB causing a slight distortion due to the stronger HBs between the amine hydrogen atoms and the iodine atoms.
G3XB and derivatives as cocrystal-salts. To probe receptor binding in the solid-state the various derivatives were crystalized with tetraalkylammonium chloride salts. Despite our efforts only four successful cocrystal-salts were obtained (G3XB$Cl − , 2H-G3XB$Cl − , 2Me-G3XB$Cl − , and H-G3XB$Cl − ). Unfortunately, the different tetraalkylammonium salts present and the crystal packing differences, made comparisons difficult (see ESI †). Despite this, all the structures maintain a bidentate conformation. Additionally, the XB contacts are quite strong with reduction ratios #0.85.

Conclusions
In this work, we reported the rst LFER studies for substituent effects on the HBeXB interaction. Electrostatic surface potentials (ESP) were used to assess the electrostatic contribution to the interaction. A strong correlation between computational ESP values and solution binding data illustrated the electrostatic nature of this cooperative interaction. Hammett plots constructed with iodide association constants for R 1 -G3XB and 2R 2 -G3XB showed that the electronics of both the HB and XB are critically important to the binding. Resonance effects of electron withdrawing R 1 substituents strengthened both the HB and XB and enhanced the overall HBeXB binding. Electron withdrawing groups of R 1 substituents generated a more potent HB donor which better polarized the XB and further promoted the bidentate conformation. In contrast, the electronics of the R 2 substituents had competing effects on the HB and XB. Specically, electron donating groups para to the iodine atoms (R 2 ) decreased the XB donor ability but made the halogen a better HB acceptor. From a design standpoint, this implies that when modulating electron density on the halogen one can enhance the preorganization of a receptor by increasing electron density on the halogen (improved HB acceptor capacity) at the expense of a slightly weakened XB. Our X-ray crystallography studies further demonstrated the role of the HBeXB on preorganizing molecular structure. Combined, the solution experiments, computations and crystallography provided a rare example of how substituents affect proximal noncovalent interactions. The results from this study also provides important insights for the design of receptors or catalysts-altering remote substituents which electronically inuence adjacent noncovalent interactions (instead of direct para substitution) can have similar impact to traditional substituent effects and should be considered for molecular design.

Data availability
The corresponding data can be found in the ESI. †

Author contributions
JS and OBB conceptualized the project. JS and EAJ conducted synthesis and characterization. VSB conducted computational studies. JS conducted the solution studies. DAD conducted solid-state studies. JS and DAD wrote the paper. OBB supervised the investigation and provided editorial assistance during manuscript preparation. All authors examined the data, results, and conclusions presented here.

Conflicts of interest
There are no conicts to declare. provides more exibility for the only N-H/I HB thus producing a stronger HB. However, when considering receptor design, this small difference would likely be compensated by allowing two convergent XBs to be available in the bidentate conformation.