The CB[n] Family: Prime Components for Self-Sorting Systems Supporting Information

The CB[n] Family: Prime Components for Self-Sorting Systems Supporting Information by Simin Liu, Christian Ruspic, Pritam Mukhopadhyay,Sriparna Chakrabarti, Peter Y. Zavalij, and Lyle Isaacs* Department of Chemistry and Biochemistry, University of Maryland College Park, MD 074 Table of Contents Pages Table of Contents. S1 Figure S1. S 1 H and 13 C NMR spectra for new compounds 3 and 11. S3 S6 Details of the x-ray structures of CB[8] 3 and CB[8] 11. S7 S1 Selected 1 H NMR spectra from the K a and K rel measurements S13 S Sample Determination of K rel for the Competition Between 5 and 1 for CB[7]. S3 Sample Error Analysis Calculation for CB[7] 4. S3 S6 S1

Figure S1. Cross-eyed stereoview of the CB[7] 11 complex from MMFF calculations. Atom colors: C, gray; N, blue; O, red; H, aqua; H-bonds, aqua-red striped. S

Figure S. 1 H NMR spectrum (400 MHz, D O, RT) of 3. S3

Figure S3. 13 C NMR spectrum (100 MHz, D O, RT) of 3. S4

Figure S4. 1 H NMR spectrum (400 MHz, D O, RT) of 11. S5

Figure S5. 13 C NMR spectrum (100 MHz, D O, 70 C) of 11. S6

Crystal Structure Information for CB[8]*3*(I) *nh O Issued by: Peter Y. Zavalij Crystal No. & ID : 0518: Lyle CB8-3 crystal#4 flat Compound name : (CB8)*(3)*(I,Cl)4*34HO Chemical formula : (C48 H48 N3 O16), (C16 H33 N), (I3.33 Cl0.67), 34.63(HO) Empirical formula : C18 H31.6 Cl0.67 I3.33 N68 O66.3 Final R 1 [I>σ(I)] : 8.6 % A colorless plate of CB8*3*I *nh O, approximate dimensions 0.09 0. 0.38 mm 3, was used for the X-ray crystallographic analysis. The X-ray intensity data were measured at 110() K on a threecircle diffractometer system equipped with Bruker SmartApex CCD area detector using a graphite monochromator and a MoKα fine-focus sealed tube (λ= 0.71073 Å) operated at 50 kv and 30 ma. The detector was placed at a distance of 5.156 cm from the crystal. A total of 59 frames were collected with a scan width of 0. in and an exposure time of 0 sec/frame using SMART (Bruker, 1999). The total data collection time was 16.51 hours. The frames were integrated with SAINT software package using a narrow-frame integration algorithm. The integration of the data using a Orthorhombic unit cell yielded a total of 101765 reflections to a maximum angle of 5.03, of which 9754 were independent (completeness = 97.9%, R int = 11.38%, R sig = 16.67%) and 16693 were greater than σ(i). The final cell dimensions of a = 5.478(3) Å, b = 5.873(3) Å, c = 6.538(3) Å, α= 90, β= 90, γ= 90, V = 17494(3) Å 3, are based upon the refinement of the XYZcentroids of 6903 reflections with. < θ < 4.9 using SAINT. Analysis of the data showed 0.07 % decay during data collection. Data were corrected for absorption effects with the Semi-empirical from equivalents method using XPREP (Sheldrick, 1997). The minimum and maximum transmission coefficients were 0.700 and 0.935. The structure was solved and refined using the SHELXS-97 (Sheldrick, 1990) and SHELXL-97 (Sheldrick, 1997) software in the space group P 1 1 1 with Z = 4 for the formula unit CB8*3*I *nh O. The final anisotropic full-matrix least-squares refinement on F with 430 variables converged at R 1 = 8.6 % for the observed data and wr = 1.11 % for all data. The goodness-of-fit was 1.001. The largest peak on the final difference map was 1.75 e/å 3 and the largest hole was -0.981 e/å 3. On the basis of the final model, the calculated density was 1.608 g/cm 3 and F(000), 8785 e. S7

Figure S6. A view of a molecule of CB8*3*I *nh O from the crystal structure. Anisotropic atomic displacement ellipsoids for the non-hydrogen atoms are shown at the 50% probability level. Hydrogen atoms are displayed with an arbitrarily small radius. Table S1. Sample and crystal data for CB8*3*I *nh O. X-ray labbook No. 0518 Crystal ID Lyle CB8-3 crystal#4 flat Empirical formula C18 H31.6 Cl0.67 I3.33 N68 O66.3 Formula weight 435.48 Temperature 110() K Wavelength 0.71073 Å Crystal size 0.38 0. 0.09 mm 3 Crystal habit colorless plate Crystal system Orthorhombic Space group P 1 1 1 Unit cell dimensions a = 5.478(3) Å α = 90 b = 5.873(3) Å β = 90 c = 6.538(3) Å γ = 90 Volume 17494(3) Å 3 Z 4 Density, ρ calc 1.608 g/cm 3 Absorption coefficient, μ 0.743 mm -1 F(000) 8785 e S8

S9

Table S. Data collection and structure refinement for CB8*3*I *nh O. Diffractometer Bruker SmartApex CCD area detector Radiation source fine-focus sealed tube, MoKα Generator power 50 kv, 30 ma Detector distance 5.156 cm Detector resolution 8.33 pixels/mm Total frames 59 Frame size 51 pixels Frame width 0. Exposure per frame 0 sec Total measurement time 16.51 hours Data collection method ω scans θ range for data collection 1.10 to 5.03 Index ranges -30 h 30, -30 k 30, -31 l 31 Reflections collected 101765 Independent reflections 9754 Observed reflection, I>σ(I) 16693 Coverage of independent reflections 97.9 % Variation in check reflections 0.07 % Absorption correction Semi-empirical from equivalents XPREP (Sheldrick, 1997) Max. and min. transmission 0.935 and 0.700 Structure solution technique direct Structure solution program SHELXS-97 (Sheldrick, 1990) Refinement technique Full-matrix least-squares on F Refinement program SHELXL-97 (Sheldrick, 1997) Function minimized Σw(F o - F c ) Data / restraints / parameters 9754 / 1633 / 430 Goodness-of-fit on F 0.997 Δ/σ max 0.001 Final R indices: R 1, I>σ(I) 0.086 wr, all data 0.111 R int 0.1138 R sig 0.1667 Weighting scheme w = 1/[σ (F o ) + (0.08P) +.P], P = [max(f o,0) + F o ]/3 Absolute structure parameter 0.49() Largest diff. peak and hole 1.75 and -0.981 e/å 3 R 1 = Σ F o - F c /Σ F o, wr = [Σw(F o -F c ) /Σw(F o ) ] 1/ S10

Crystal Structure Information for Complex CB[8] 11 Issued by: Peter Y. Zavalij Crystal No. & ID : 114: Isaacs Compound name : CB[8] 11 Chemical formula : (C48 H48 N3 O16), (C17 H1 N7 O), I.4, (H O)13., (H3 O)0.4 Final R 1 [I>σ(I)] : 6.38 % A colorless pyramid of C65 H96.54 I.41 N39 O30.56, approximate dimensions 0.185 0.60 0.80 mm 3, was used for the X-ray crystallographic analysis. The X-ray intensity data were measured at 193() K on a three-circle diffractometer system equipped with Bruker Smart1000 CCD area detector using a graphite monochromator and a MoKα fine-focus sealed tube (λ= 0.71073 Å) operated at 50 kv and 40 ma. The detector was placed at a distance of 4.990 cm from the crystal. A total of 1868 frames were collected with a scan width of 0.3 in ω and an exposure time of 3 sec/frame using SMART (Bruker, 1999). The total data collection time was 15.6 hours. The frames were integrated with SAINT software package using a narrow-frame integration algorithm. The integration of the data using a Orthorhombic unit cell yielded a total of 4731 reflections to a maximum θ angle of 5.00, of which 43080 were independent (completeness = 99.9%, R int = 0.00%, R sig = 5.36%) and 30509 were greater than σ(i). The final cell dimensions of a = 1.5160(1) Å, b = 7.964(3) Å, c = 5.43() Å, α= 90, β= 90, γ= 90, V = 8898.(14) Å 3, are based upon the refinement of the XYZ-centroids of 746 reflections with.3 < θ < 4.0 using SAINT. Analysis of the data showed 0.16 % decay during data collection. Data were corrected for absorption effects with the Semi-empirical from equivalents method using SADABS (Sheldrick, 1996). The minimum and maximum transmission coefficients were 0.70 and 0.860. The structure was solved and refined using the SHELXS-97 (Sheldrick, 1990) and SHELXL-97 (Sheldrick, 1997) software in the space group Pna 1 with Z = 4 for the formula unit C65 H96.54 I.41 N39 O30.56. The final anisotropic full-matrix least-squares refinement on F with 1138 variables converged at R 1 = 6.38 % for the observed data and wr = 14.81 % for all data. The goodness-of-fit was 1.000. The largest peak on the final difference map was 0.461 e/å 3 and the largest hole was -0.403 e/å 3. On the basis of the final model, the calculated density was 1.657 g/cm 3 and F(000), 457 e. Comments: - non-merohedral twinning: rotation around [011] in 0.57:0.43 ratio; - pseudo-centrosymmetric: CB8 shows centrosymmetric arrangement but diamine, iodide and most of the water don't. - disorder: diamine is disordered inside CB8 in about :1 ratio; - occupation disorder of water - modeled with partially occupied oxygen atoms. S11

Figure S7. A view of a molecule of [CB8*11]I*18HO from the crystal structure showing the numbering scheme employed. Anisotropic atomic displacement ellipsoids for the non-hydrogen atoms are shown at the 50% probability level. Hydrogen atoms are displayed with an arbitrarily small radius. Table S3. Sample and crystal data for [CB[8] 11]I.4*0.4H3O*13.HO. X-ray labbook No. 114 Crystal ID Isaacs Empirical formula C65 H96.54 I.41 N39 O30.56 Formula weight 19.18 Temperature 193() K Wavelength 0.71073 Å Crystal size 0.80 0.60 0.185 mm 3 Crystal habit colorless pyramid Crystal system Orthorhombic Space group Pna 1 Unit cell dimensions a = 1.5160(1) Å α = 90 b = 7.964(3) Å β = 90 c = 5.43() Å γ = 90 Volume 8898.(14) Å 3 Z 4 Density, ρ calc 1.657 g/cm 3 Absorption coefficient, μ 0.953 mm -1 F(000) 457 e S1

Table S4. Data collection and structure refinement for [CB[8] 11]I.4*0.4H3O*13.HO. Diffractometer Bruker Smart1000 CCD area detector Radiation source fine-focus sealed tube, MoKα Generator power 50 kv, 40 ma Detector distance 4.990 cm Detector resolution 8.33 pixels/mm Total frames 1868 Frame size 51 pixels Frame width 0.3 Exposure per frame 3 sec Total measurement time 15.6 hours Data collection method ω scans θ range for data collection 1.78 to 5.00 Index ranges 0 h 14, 0 k 33, -30 l 30 Reflections collected 4731 Independent reflections 43080 Observed reflection, I>σ(I) 30509 Coverage of independent reflections 99.9 % Variation in check reflections 0.16 % Absorption correction Semi-empirical from equivalents SADABS (Sheldrick, 1996) Max. and min. transmission 0.860 and 0.70 Structure solution technique direct Structure solution program SHELXS-97 (Sheldrick, 1990) Refinement technique Full-matrix least-squares on F Refinement program SHELXL-97 (Sheldrick, 1997) Function minimized Σw(F o - F c ) Data / restraints / parameters 43080 / 340 / 1138 Goodness-of-fit on F 1.0 Δ/σ max 0.009 Final R indices: R 1, I>σ(I) 0.0638 wr, all data 0.1481 R int 0.0000 R sig 0.0536 Weighting scheme w = 1/[σ (F o ) + (0.04P) + 18.15.P], P = [MAX(F o,0) + F o ]/3 Absolute structure parameter 0.33() Largest diff. peak and hole 0.461 and -0.403 e/å 3 R 1 = Σ F o - F c /Σ F o, wr = [Σw(F o -F c ) /Σw(F o ) ] 1/ S13

The following spectra are representative examples of those acquired during the determination of the various K a and K rel values. In the spectra presented here only some of the resonances are integrated to focus attention on those uncluttered regions of the spectrum where good integrals can be obtained. Although we have fully analyzed and assigned all of the spectra, we have not depicted those assignments here for reasons of clarity. The resonance at 1.9 ppm is due to residual H-atoms present in the CD 3 CO D used to make the buffer. Figure S8. One of the 1 H NMR spectra used in the direct determination of K a for CB[6] 8. S14

Figure S9. One of the 1 H NMR spectra used in the direct determination of K a for CB[6] 0. S15

Figure S10. One of the 1 H NMR spectra used in the determination of K rel for CB[6] 16 and CB[7] 16. S16

Figure S11. One of the 1 H NMR spectra used in the determination of K rel for CB[7] 1 and CB[7] 17. S17

Figure S1. One of the 1 H NMR spectra used in the determination of K rel for CB[7] 14 and CB[8] 14. S18

Figure S13. 1 H NMR spectrum of a :1 mixture of 14 and CB[8] used to establish the exclusive formation of a 1:1 complex in this case. S19

Figure S14. One of the 1 H NMR spectra used in the determination of K rel for CB[8] 14 and CB[8] 3. S0

Figure S15. One of the 1 H NMR spectra used in the determination of K rel for CB[8] 11 and CB[8] 3. S1

Figure S16. One of the 1 H NMR spectra used in the determination of K rel for CB[8] 11 and CB[8] 1. S

Figure S17. One of the 1 H NMR spectra used in the determination of K rel for CB[7] 5 and CB[7] 1. S3

Sample Determination of K rel for the Competition Between 5 and 1 for CB[7]. We use equation 1 to determine K rel for the interaction of 5 and 1 for CB[7]. For this purpose, we prepared a solution containing CB[7] (40 μm), 5 (840 μm), and 1 (137.7 mm) and allowed it to reach equilibrium (Figure S17). Next, we determined the relative concentrations of 1 and CB[7] 1 by integration of the appropriate resonances in the 1 H NMR spectrum (Figure 1: -0.17 ppm; CB[7] 1: -0.93 ppm). Using the relative concentrations and the mass balance expression (equation ) allowed us to calculate [1] free = 709.4 μm and [CB[7] 1] = 130.6 μm. Equation 3 is then used to calculate [CB[7] 5] (109.4 μm) using the known value of CB[7] 1. Lastly, equation 4 is used to calculate [5] free (137.59 mm) using the known value of [CB[7] 5]. K rel = ([CB[7] 1][5] free ) / ([CB[7] 5][1] free ) (1) [1] Total = 840 μm = [1] free + [CB[7] 1] () [CB[7]] Total = 40 μm = [CB[7] 1] + [CB[7] 5] (3) [5] Total = 137.7 mm = [5] free + [CB[7] 5] (4) Substitution of the values of [CB[7] 1], [1] free, [CB[7] 5], and [5] free into equation 1 gave K rel = 31.5. These determinations were done in triplicate from independently prepared stock solutions and the average values were used in the calculations of K a and the error analysis shown below. In preparing the solutions for the above determinations we used a small excess of 1 (to ensure there is no free CB[7]) and a large excess of 5; under those conditions the errors in [5] free, [1] free are small and both [1] and [CB[7] 1] are kept in a good range for accurate measurement of their ratio by 1 H NMR. Sample Error Analysis Calculation for CB[7] 4. Since the binding constants in this paper are determined by several levels of 1 H NMR competition experiments referenced to an absolute K a measured for CB[7] 5 measured by UV/Vis titration, a proper error analysis is critical. In this section we give a sample calculation of the error analysis used to determine the uncertainty associated with the K a value for CB[7] 4. Step 1 Estimation of the accuracy of 1 H NMR methods for the determination of guest and host guest concentrations. We used 1 H NMR to repeatedly determine the concentration of samples of known concentration of guest and host guest complex by monitoring guest resonances. The 1 H NMR based method was accurate with a standard deviation of ± 3%. Step Determination of the Uncertainty Associated with a Single Level of Competition (K rel ). We propagated the above uncertainty associated with the NMR determination of concentrations (e.g. {(σ [CB[n] guest] )/[CB[n] guest]} = 0.03 and {(σ [Guest] )/[Guest]} = 0.03) of using equations 5 and equations 6 7 (Bevington eq. 4-11, pages 61 6). Equation 6 delivers the uncertainty associated with the weighted product (x) of two values (u and v) (e.g. x = ±a u v). Similarly, equation 7 delivers the uncertainty for dividing two numbers (e.g. x = ±(a u) / v). We make the assumption that the fluctuations in u and v are not correlated (σ uv = 0) which when substituted into equations 6 and 7 delivers equation 8. Rearranging slightly yields equation 9 which allows us to directly use the 3% uncertainty determined for our NMR method. K rel = ([CB[n] G][G1]) / ([CB[n] G1][G]) (5) S4

σ x = σ u σ v σ + + uv x u v uv (6) σ x = σ u σ v σ + - uv x u v uv (7) σ x = σ u σ + v x u v (8) σ x = x σ u u + σ v v (9) We break the uncertainty determination in K rel (equation 1) into three steps: 1) Multiplying [CB[n] G][G1], ) multiplying [CB[n] G1][G], and 3) dividing the two results. Substituting (σ[cb[n] G] / [CB[n] G]) = 0.03 and (σ[g1]/[g1]) = 0.03 into equation 9 gives equation 10 and an uncertainty of 4.4% for [CB[n] G][G1] (equation 11). Similarly, the uncertainty of [CB[n] G1][G] is 4.4%. σ [CB[n] G][G1] = [CB[n] G][G1] (0.03) + (0.03) (10) σ [CB[n] G][G1] [CB[n] G][G1] = 0.044 (11) With the two values of the uncertainties of [CB[n] G][G1] and [CB[n] G1][G] (4.4%) in hand we next substituted these values into equation 9 to give the uncertainty in K rel (eq. 1 13) of 6%. σ Krel K rel = σ [CB[n] G][G1] [CB[n] G1][G] [CB[n] G][G1] [CB[n] G1][G] = (0.044) + (0.044) (1) σ Krel K rel = 0.06 (13) Step 3 Determination of the Uncertainty in the K a value for CB[7] 5. We obtained an uncertainty in the value of K a for CB[7] 5 (K a = 8.07 ± 0.60 x 10 4 M -1 ) (7.43 %) from the nonlinear least squares fit of the UV/Vis titration data to a 1:1 binding model. Step 4 Determination of K a for CB[7] 1 by competition of 5 and 1 for a limiting quantity of CB[7]. We used 1 H NMR competition experiments to determine K rel = 5.5 for these two guests. Substitution of K CB[7] 5 = 8.07 10 4 M -1 and K rel into equation 14 gave K CB[7] 1 = 1.8 10 7 M -1 (equation 15). The uncertainty in K CB[7] 1 can be determined using equation 16. Substituting σ(k CB[7] 5 )/K CB[7] 5 = 0.0743 and σ(k rel )/K rel = 0.10 [Note that we are using the even more conservative 10% error in this analysis] gives the percent error in K CB[7] 1 (equation 17). Substituting eq. 15 into eq. 17 gives σ(k CB[7] 5 ) (equation 18) which can be combined with eq. 15 to give a final value for K CB[7] 1 (equation 19). K G = (K G1 )(K rel ) (14) K CB[7] 1 = 1.8 10 7 M -1 (15) σ KCB[7] 1 K CB[7] 1 = σ K CB[7] 5 K CB[7] 5 + σ Krel K rel (16) σ KCB[7] 1 = 0.146 (1.46%) K CB[7] 1 (17) S5

σ K = (0.146) (1.8 x 10 7 M -1 ) =. x 10 6 M -1 CB[7] 1 (18) K CB[7] 1 = 1.8 ± 0. 10 7 M -1 (19) Step 5 Determination of K a for CB[7] 19 by competition of 1 and 19 for a limiting quantity of CB[7]. We used 1 H NMR competition experiments to determine K rel = 48.8 for these two guests. Substitution of K CB[7] 1 = 1.8 ± 0. 10 7 M -1 and K rel into equation 0 gave K CB[7] 19 = 8.88 10 8 M -1 (equation 1). The uncertainty in K CB[7] 19 can be determined using equation. Substituting σ(k CB[7] 1 )/K CB[7] 1 = 0.146 and σ(k rel )/K rel = 0.10 gives the percent error in K CB[7] 19 (15.97%, equation 3). Substituting eq. 1 into eq. 3 gives σ(k CB[7] 19 ) (equation 4) which can be combined with eq. 1 to give a final value for K CB[7] 19 (equation 5) K CB[7] 19 = (K CB[7] 1 )(K rel ) (0) K CB[7] 19 = 8.88 10 8 M -1 (1) σ KCB[7] 19 K CB[7] 19 = σ K CB[7] 1 K CB[7] 1 + σ Krel K rel () σ KCB[7] 19 = 0.1597 (15.97%) K CB[7] 19 (3) σ K = (0.1597) (8.88 x 10 8 M -1 ) = 1.41 x 10 8 M -1 CB[7] 19 (4) K CB[7] 19 = 8.88 ± 1.41 10 8 M -1 (5) Step 6 Determination of K a for CB[7] 1 by competition of 19 and 1 for a limiting quantity of CB[7]. We used 1 H NMR competition experiments to determine K rel = 37 for these two guests. Substitution of K CB[7] 19 = 8.88 ± 1.41 10 8 M -1 and K rel into equation 6 gave K CB[7] 1 = 3.31 10 11 M -1 (equation 7). The uncertainty in K CB[7] 1 can be determined using equation 8. Substituting σ(k CB[7] 19 )/K CB[7] 19 = 0.1597 and σ(k rel )/K rel = 0.10 gives the percent error in K CB[7] 1 (18.84%, equation 9). Substituting eq. 7 into eq. 9 gives σ(k CB[7] 1 ) (equation 30) which can be combined with eq. 7 to give a final value for K CB[7] 1 (equation 31) K CB[7] 1 = (K CB[7] 19 )(K rel ) (6) K CB[7] 1 = 3.31 10 11 M -1 (7) σ KCB[7] 1 K CB[7] 1 = σ K CB[7] 19 K CB[7] 19 + σ Krel K rel (8) σ KCB[7] 1 = 0.1884 (18.84%) K CB[7] 1 (9) σ K = (0.1884) (3.31 x 10 11 M -1 ) = 0.6 x 10 11 M -1 CB[7] 1 (30) K CB[7] 1 = 3.31 ± 0.6 10 11 M -1 (31) Step 7 Determination of K a for CB[7] 4 by competition of 1 and 4 for a limiting quantity of CB[7]. We used 1 H NMR competition experiments to determine K rel = 5.98 for these two guests. Substitution of K CB[7] 1 = 3.31 ± 0.6 10 11 M -1 and K rel into equation 3 gave K CB[7] 4 = 1.98 10 1 M -1 (equation 33). The uncertainty in K CB[7] 4 can be determined using equation 34. S6

Substituting σ(k CB[7] 1 )/K CB[7] 1 = 0.1884 and σ(k rel )/K rel = 0.10 gives the percent error in K CB[7] 4 (1.33%, equation 35). Substituting eq. 33 into eq. 35 gives σ(k CB[7] 4 ) (equation 36) which can be combined with eq. 35 to give a final value for K CB[7] 1 (equation 37) K CB[7] 4 = (K CB[7] 1 )(K rel ) (3) K CB[7] 4 = 1.98 10 1 M -1 (33) σ KCB[7] 4 K CB[7] 4 = σ K CB[7] 1 K CB[7] 1 + σ Krel K rel (34) σ KCB[7] 4 = 0.133 (1.33%) K CB[7] 4 (35) σ K = (0.133) (1.98 x 10 1 M -1 ) = 4. x 10 11 M -1 CB[7] 4 (36) K CB[7] 4 = 1.98 ± 0.4 10 1 M -1 (37) S7