pH controlled assembly of a self-complementary halogen-bonded dimer

Halogen bonding between an oxygen acceptor and an iodotriazole donor can be switched on an off by cycling the solution pH.


S1. Reagents and General Equipment
All reagents were purchased from commercial sources (Alfa Aesar, Apollo Scientific Ltd., Fisher Scientific UK Ltd., Fluorochem UK Ltd., TCI UK Ltd. and Sigma-Aldrich Company Ltd) unless stated otherwise, and used without further purification. Dry solvents were obtained by means of a MBBRAUN MB SPS-800 TM purification system. Flash column chromatography was performed using Geduran ® Si60 (40-63 µM, Merck, Germany) as the stationary phase and thin layer chromatography was performed on pre-coated silica gel-plates (0.25 mm thick, 60F 254 , Merck, Germany) and observed under UV light irradiation. 1 H, 13 C, 19 F NMR spectroscopic data was acquired using either a Bruker Avance (500 MHz), a Bruker Avance III (500 MHz) or a Bruker Avance II (400 MHz) spectrometer, at a constant temperature of 25 ºC unless stated otherwise. 1 H and 13 C chemical shifts are reported in parts per million (ppm) from high to low field and referenced to the literature values for chemical shift of the residual non-deuterated solvent, with respect to tetramethylsilane. 19 F NMR chemical shifts are referenced to CFCl 3 (0.00 ppm).
All melting points were measured using a Stuart SMP30 melting point apparatus.
Mass spectra were recorded on a Micromass GCT spectrometer for chemical ionisation (CI) using isobutene as the ionising gas. Electron spray ionisation (ESI) spectra were performed on a Micromass LCT spectrometer operating in positive or negative mode, m/z values are reported in Daltons.

S3. Solid state structures of 1 and solution state studies
As described in the main text, the same crystal growth method, i.e. slow evaporation from a saturated toluene solution, afforded two different batches of crystals. The acquisition of X-ray diffraction data on both crystalline materials allowed the unravelling the polymorphic character of pyridone-appended iodotriazole 1. Although the poorer quality of the second batch of crystal did not permit structure solution to an R-factor lower than 15.4%, the data was of sufficient quality to identify the presence of a homodimeric assembly within the solid state structure ( Figure S1). Figure S1. Solid state structure, determined from X-ray diffraction data, of the polymorph of pyridone-appended iodotriazole 1. The homodimeric assembly is characterised by two XB interactions 2.77 and 2.73 Å long; a third XB contact (2.73 Å) involves the second lone pair of the carbonyl oxygen atom of 1 with the iodine atom of a third iodotriazole unit.
Interestingly, the XB interactions within the homodimeric assembly slightly differ in length as a result of the bifurcated halogen bonding interaction involving only one of the two carbonyl oxygen atoms. Further attempts to crystallise this polymorph of 1 from toluene afforded only the crystalline phase reported in the main text.
The stability of dimer [1•1] in d 8 -toluene solution was determined [3] by fitting 1 H chemical shift changes to a dimerisation model using the WinEQNMR [4] program. Data for this fitting process were generated by recording the 500.

S4. Crystal structures of iodotriazoles S3 and 2-H
Given the highly crystalline character of intermediates S3 and 2-H, the crystal packing of these iodotriazole scaffolds was examined to provide a preliminary evaluation of the halogen bonding ability of this particular iodotriazole design. Figure S2 shows the crystal structure of the methyl protected phenol appended iodotriazole S3. Similarly to previously synthesised unsubstituted iodotriazoles, [1] the observed crystal packing is the result of the XB contact between the iodine atom of one iodotriazole unit with the nitrogen atom in the 3-position of the consecutive iodotriazole thus creating antiparallel halogen bonded tapes.
Evidently, the presence of the methyl group hinders the ability of the oxygen atom to perform as a XB acceptor. The replacement of the methyl group with a smaller hydrogen atom, as in 2-H, provided both a XB acceptor and a HB donor in the form of a phenol group. Despite the reported solid-state evidence of phenol groups able to engage in both noncovalent interaction at the same time reported by Desiraju [5,6] and Aakeröy, [7] the sole short contact observed in the crystal structure Seemingly, the HB donating character of the phenol group for this particular molecular design overpowers the XB donating properties of the iodotriazole.

S5. Deprotonation of phenol 2-H and related control experiments
In a typical experiment, 12 mg (0.02 mmol) phenol-appended iodotriazole 2-H were dissolved in 0.98 mL of CD 3   TBAOH

S6. Single point XB association constant determination
Iodotriazole S4 [1] was chosen as the model XB donor to measure the single point association constant with a phenoxide anion in order to establish the effective molarity of homodimer [2•2] 2-.
In first instance, a fresh solution of tetra-n-butylammonium phenolate was titrated in a 10 mM CD 3 CN solution of S4. Unfortunately this measurement was hampered by the tendency of iodotriazole S4 to undergo reductive deiodination in presence of the TBA phenolate salt to afford triazole S4-H. Figure S7 shows the NMR characterisation of the isolated prototriazole S4-H. Such reactivity was not observed when the titration was performed with DBUH phenolate as the XB acceptor. This allowed us to titrate a 10 mM solution of S4 with 1 to 10 equivalents of DBUH phenolate in presence of 4-fluoro-toluene as internal standard (10 mM, δ F referenced at -120 ppm). Fitting [4] of the 19 F chemical shift changes of S4 to a 1:1 binding model using the WinEQNMR program afforded an association constant of 6.6 ± 0.2 M -1 .

S8. DFT calculations
All calculations were performed using Gaussian09 [8] suite of programs -revision D.01 was used in all calculations. Calculation were performed using TPSSh [9] functional and the def2-TZVP basis set [10] of Weigend and Ahlrichs. This basis set is not standard in Gaussian09 and was introduced using the GenECP keyword using an appropriately-formatted input block for the basis set generated from data obtained from the Basis Set Exchange [11] (https://bse.pnl.gov/bse/portal). An effective core potential on iodine, [12] which replaces 28 valence electrons on each iodine atom in the structure was used in all calculations. The basis set is all-electron for all other elements.
All geometries were optimized fully in internal (keyword: opt) or cartesian (keyword: opt=cartesian) coordinates using the default optimisation protocols within Gaussian09.
Stationary points were characterised by means of a vibrational analysis (keyword: freq) and zeropoint energy corrections and other thermodynamic parameters, used in the calculation of interactions energies, were derived [13] from this analysis. Population analyses using the Natural Bond Orbital (NBO) method were performed using the NBO6 program [14,15] in a two-stage procedure. The input for NBO6 was generated using Gaussian09 (keyword: The basis set superposition error was calculated using the counterpoise method [16] as implemented within Gaussian09 (keyword: counterpoise = 2).
Halogen bonds were visualized using the NCIPLOT [17] program. SCF densities written as an extended wavefunction file from Gaussian09 (rev. D.01) and NCIPLOT used to generate Gaussian cube files from which isosurfaces could be visualised using VMD [18] with s = 0.5 and a colours (blue (attractive) to red (repulsive)) mapped on to -0.05 < ρ < 0.05.

Data Reduction
Of the 36305 reflections were collected, where 5647 were unique (R int = 0.1242); equivalent reflections were merged. Data were collected and processed using CrystalClear (Rigaku). 1 The linear absorption coefficient, µ, for Mo-Kα radiation is 10.667 cm -1 . An empirical absorption correction was applied which resulted in transmission factors ranging from 0.416 to 0.808. The data were corrected for Lorentz and polarization effects.

Structure Solution and Refinement
The structure was solved by heavy-atom Patterson methods 2 and expanded using Fourier

Data Collection
A colourless prism crystal of C 29 H 28 F 4  uniquely determine the space group to be: The data were collected at a temperature of −148 ±1 ºC to a maximum 2θ value of 50.8º. 26

Data Reduction
Of the 109125 reflections were collected, where 16446 were unique (R int = 0.2103); equivalent reflections were merged. Data were collected and processed using CrystalClear (Rigaku). 1 The linear absorption coefficient, µ, for Mo-Kα radiation is 11.207 cm -1 . An empirical absorption correction was applied which resulted in transmission factors ranging from 0.789 to 0.967. The data were corrected for Lorentz and polarization effects.

Structure Solution and Refinement
The structure was solved by heavy-atom Patterson methods 2 and expanded using Fourier techniques. The non-hydrogen atoms were refined anisotropically. Hydrogen atoms were refined using the riding model. The final cycle of full-matrix least-squares refinement 3  (4) Goodness of fit is defined as:

Data Collection
A colorless prism crystal of C 32 H 34 F 4 IN 3 O 2 having approximate dimensions of 0.240 × 0.080 × 0.070 mm was mounted in a loop. All measurements were made on a Rigaku XtaLAB P200 diffractometer using multi-layer mirror monochromated Mo-Kα radiation.
The crystal-to-detector distance was 45.06 mm.
Cell constants and an orientation matrix for data collection corresponded to a primitive

Data Reduction
Of the 38083 reflections were collected, where 5690 were unique (R int = 0.0269); equivalent reflections were merged. Data were collected and processed using CrystalClear (Rigaku). 1 The linear absorption coefficient, µ, for Mo-Kα radiation is 10.811 cm −1 . An empirical absorption correction was applied which resulted in transmission factors ranging from 0.853 to 0.927. The data were corrected for Lorentz and polarization effects.

Structure Solution and Refinement
The structure was solved by heavy-atom Patterson methods 2 and expanded using Fourier techniques. The non-hydrogen atoms were refined anisotropically. Some hydrogen atoms were refined isotropically, some were refined using the riding model, and the rest were included in fixed positions. The final cycle of full-matrix least-squares refinement 3 on F 2 was based on 5690 observed reflections and 388 variable parameters and converged (largest parameter shift was 0.00 times its esd) with unweighted and weighted agreement factors of: The goodness of fit 4 was 1.05. Unit weights were used. The maximum and minimum peaks on the final difference Fourier map corresponded to 0.54 and −0.33 e /Å 3 , respectively. The final Flack parameter 5 was −0.024(4), indicating that the value is out of statistically acceptable range. 6 It is recommended to repeat least-squares refinement carefully until getting meaningful value, or, in the worst case, to average Friedel pairs.
Neutral atom scattering factors were taken from International Tables for Crystallography (IT), Vol.
C, were those of Creagh and McAuley. 9 The values for the mass attenuation coefficients are those of Creagh and Hubbell. 10 All calculations were performed using the CrystalStructure 11 crystallographic software package except for refinement, which was performed using

Data Collection
A yellow prism crystal of C 37 H 42 F 4 IN 5 O having approximate dimensions of 0.210 × 0.120 × 0.120 mm was mounted in a loop. All measurements were made on a Rigaku XtaLAB P200 diffractometer using multi-layer mirror monochromated Mo-Kα radiation.
The crystal-to-detector distance was 44.97 mm.
Cell constants and an orientation matrix for data collection corresponded to an I-centered

Data Reduction
Of the 44203 reflections were collected, where 6762 were unique (R int = 0.0604); equivalent reflections were merged. Data were collected and processed using CrystalClear (Rigaku). 1 The linear absorption coefficient, µ, for Mo-Kα radiation is 9.242 cm −1 . An empirical absorption correction was applied which resulted in transmission factors ranging from 0.770 to 0.895. The data were corrected for Lorentz and polarization effects.

Structure Solution and Refinement
The structure was solved by heavy-atom Patterson methods 2 and expanded using Fourier techniques. The non-hydrogen atoms were refined anisotropically. Some hydrogen atoms were refined isotropically, some were refined using the riding model, and the rest were included in fixed positions. The final cycle of full-matrix least-squares refinement 3  Neutral atom scattering factors were taken from International Tables for Crystallography (IT), Vol.
C,  (4) Goodness of fit is defined as: