The Self-Sorting Behavior of Circular Helicates and Molecular Knots and Links**

We report on multicomponent self-sorting to form open circular helicates of different sizes from a primary monoamine, FeII ions, and dialdehyde ligand strands that differ in length and structure by only two oxygen atoms. The corresponding closed circular helicates that are formed from a diamine—a molecular Solomon link and a pentafoil knot—also self-sort, but up to two of the Solomon-link-forming ligand strands can be accommodated within the pentafoil knot structure and are either incorporated or omitted depending on the stage that the components are mixed.

1. General Experimental S1 2. Self-Sorting Reactions S3 2.1 Open Systems S3 2.1.1 Experimental Procedure for the Self-Sorting of Cyclic Helicates 4 and 5 S3 2.1.2 Experimental Procedure to Determine the Effects of Mixing on the Relative Yields of Cyclic Helicates 4 and 5 S4 2.1.3 Experimental Procedure to Determine the Effects of Concentration on the Self-Sorting of Helicates 4 and 5 S5 2.1.4 Thermodynamic Investigations of the Self-Sorting of Helicates 4 and 5 S7 2.2 Closed Systems S11 2.2.1 Experimental Procedure for the Self-Sorting of Solomon Link 7 and Pentafoil Knot 8 S11 2.2.2 Experimental Procedure to Determine the Effects of Mixing on the Relative Yields of Solomon Link 7 and Pentafoil Knot 8 S15 2.2.3 Experimental Procedure to Determine the Effects of Concentration on the Self-Sorting of Closed Topologies 7, 8, 9 and 10 S17 2.2.4 Experimental Procedure For the Thermodynamic Investigations of the Self-Sorting of Closed Topologies 7, 8, 9 and 10 S19 3. References S20

General Experimental
Unless stated otherwise, all reagents and solvents were purchased from Aldrich Chemicals and used without further purification. NMR spectra were recorded on a Bruker DMX 500 instrument. Chemical shifts are reported in parts per million (ppm) from low to high frequency and referenced to the residual solvent resonance.

Self-Sorting Reactions
Experiments are described in two sections, dealing with the "open" cyclic helicates 4 and 5 first, followed by the "closed" topologically complex systems 7 and 8.

Open Systems
2.1.1 Experimental Procedure for the Self-Sorting of Cyclic Helicates 4 and 5 Scheme S1. Mixing experiments conducted with aldehyde 1 (orange) and aldehyde 2 (blue) with iron(II) (purple) and hexylamine (green) to form tetramer 4 and pentamer 5.
To two separate DMSO-d 6 solutions of 1 (1.1 mg, 2.4 mol, in 0.5 mL) and 2 (1.0 mg, 2.4 mol in 0.5 mL) was added FeCl 2 (25 L of a 0.10 M stock solution, 2.6 mol). The two purple solutions were mixed thoroughly to ensure complete dissolution of both dialdehydes. The solutions were combined together in a clean NMR tube before addition of n-hexylamine (50 L of a 0.21 M stock solution, 5.2 mol). The resulting mixture was heated for 2 d at 60 ºC before being allowed to cool.
The reaction mixture was precipitated using KPF 6 (saturated aqueous solution). The resulting purple powder was collected on Celite and washed with water, a small amount of EtOH (4 is soluble) and finally Et 2 O. The product mixture was dissolved in CH 3 CN and the solvent removed under reduced pressure to give a mixture of 4 and 5 as a purple powder. The mixture was fully soluble in CH 3 CN and could be analyzed by 1 H NMR and LR-ESI ( Figure 1, main text). All results were consistent with previous reports for 4 [S1] and 5 [S2] .

Experimental Procedure to Determine the Effects of Mixing on the Relative Yields of Cyclic Helicates 4 and 5
Scheme S2. Self-sorting experiments conducted to determine the effect of mixing on the relative yields of 4 and 5. Reactions were carried out using stock solutions of dialdehydes 1 and 2 under identical conditions.
To two separate DMSO-d 6 solutions of 1 (2.1 mg, 4.7 mol, in 1.0 mL) and 2 (2.0 mg, 4.7 mol in 1.0 mL) was added FeCl 2 (50 L of a 0.10 M stock solution, 5.2 mol). The two purple solutions were mixed thoroughly to ensure complete dissolution of both dialdehydes. A 0.5 mL aliquot of 1+FeCl 2 was mixed with a 0.5 mL aliquot of 2+FeCl 2 in a clean NMR tube before addition of n-hexylamine (50 L of a 0.208 M stock solution, 5.2 mol). The remaining 0.5 mL aliquots of 1+FeCl 2 and 2+FeCl 2 were separately reacted with n-hexylamine (25 L of a 0.21 M stock solution, 5.2 mol, for each reaction). The three solutions were heated for 2 d at 60 ºC before being allowed to cool. The following purification procedure was applied to each sample individually: The reaction mixture was precipitated using KPF 6 (saturated aqueous solution). The resulting purple powder was collected on Celite and washed with water, a small amount of EtOH (4 is soluble) and finally Et 2 O. The product mixture was dissolved in CH 3 CN and the solvent removed under reduced pressure. The three samples were dissolved in 0.5 mL of CD 3 CN and compared by 1 H NMR ( Figure S2). Reactions were run at the same concentration with dialdehydes taken from a stock solution. As can be seen the yield of formation of pentamer 5 is unaffected by the presence of 4 whereas the yield of 4 is reduced by 50 % upon mixing.

Experimental Procedure to Determine the Effects of Concentration on the Self-Sorting of Helicates 4 and 5
To two separate DMSO-d All nine reactions were heated for 1 d at 60 ºC before being allowed to cool. The following purification procedure was applied to each sample individually: The reaction mixture was precipitated using KPF 6 (saturated aqueous solution). The resulting purple powder was collected on Celite and washed with water, a small amount of EtOH (4 is soluble) and finally Et 2 O. The product mixture was dissolved in CH 3 CN and the solvent removed under reduced pressure. The nine samples were dissolved in 0.5 mL of CD 3 CN and compared by 1 H NMR (Sample 1, Figure S3. Sample 2, Figure S4. Sample 3, Figure S5). The thermodynamics of self-sorting was probed by monitoring reactions C and D mixed at different times during the self-assembly process (Scheme S3). Two samples were prepared (see below for details); In sample C aldehydes 1 and 2 (with FeCl 2 ) were mixed prior to the addition of n-hexylamine 3 and sample D where each dialdehyde (with FeCl 2 ) was reacted separately with n-hexylamine, then after 24h these reaction mixtures were combined. Both samples were heated at 60 ºC and monitored over the course of 4 days ( Figure S6). As expected after the first day of heating sample C shows the required distribution of 4 and 5, which remains constant over the course of 4 d (see Figure S6d). Immediately after mixing sample D shows an approximately 1:1 distribution of helicates 4 and 5 ( Figure S7a), which slowly rearranges to the ~1:3 ratio seen above (see Figure S7d). Upon work up both samples C and D appear to be identical (see Figures S7a and b) therefore both reactions reach the same end point irrespective of the times mixed. Indicating that the formation of 4 is dynamic (as its abundance was observed to decrease) and that it is possible to disassemble preformed tetrameric helicate 4 by addition of the reaction mixture for the formation of 5.

Experimental Procedure for the Thermodynamic Investigations of the Self-Sorting of Helicates 4 and 5
To two separate DMSO-d   The ratio of 4:5 starts at 1:1 as expected for mixing equimolar solutions of 4 and 5. Over time the yield of 5 is depleted until after 4 days of heating the ratio matches that in Figure S6.  at 60 ºC before being allowed to cool to RT. The reaction mixture was precipitated using an excess saturated aqueous KPF 6 . The resulting purple powder was collected on Celite and washed with water, a small amount of EtOH and finally Et 2 O. The product mixture was dissolved in CH 3 CN and the solvent removed under reduced pressure to give a mixture of 6, 7, 8 and 9 as a purple powder which was analyzed by 1 H NMR (Figure 2, main text), LR-ESI ( Figure S9), HR-ESI ( Figures S10 and 11), COSY ( Figure S12) and ROESY ( Figures S13 and 14) spectroscopy.

Experimental Procedure to Determine the Effects of Mixing on the Relative Yields of Solomon Link 7 and Pentafoil Knot 8
Scheme S5. Self-sorting experiments conducted to determine the effect of mixing on the relative yields of 7 and 8. Reactions were carried out using stock solutions of dialdehydes 1 and 2 under identical conditions.
To two separate DMSO-d 6 solutions of 1 (2.1 mg, 4.7 mol, in 1.0 mL) and 2 (2.0 mg, 4.7 mol in 1.0 mL) was added FeCl 2 (50 L of a 0.10 M stock solution, 5.2 mol). The two purple solutions were mixed thoroughly to ensure complete dissolution of both dialdehydes. A 0.5 mL aliquot of 1.FeCl 2 was mixed with a 0.5 mL aliquot of 2.FeCl 2 in a clean NMR tube before addition of 2,2'-(ethylenedioxy)bis(ethylamine) 6 (50 L of a 0.10 M stock solution, 5.2 mol). The remaining 0.5 mL aliquots of 1+FeCl 2 and 2+FeCl 2 were separately reacted with 2,2'-(ethylenedioxy)bis(ethylamine) 6 (25 L of a 0.10 M stock solution, 2.6 mol, for each reaction). The three solutions were heated for 4d at 60 ºC before being allowed to cool. The following purification procedure was applied to each sample individually: The reaction mixture was precipitated using KPF 6 (saturated aqueous solution). The resulting purple powder was collected on Celite and washed with water, a small amount of EtOH and finally Et 2 O. The product mixture was dissolved in CH 3 CN and the solvent removed under reduced pressure. The three samples were dissolved in 0.5 mL of CD 3 CN and compared by 1 H NMR ( Figure S15). Figure S15. Effect of mixing on the yield of formation for closed topologies 6 and 7. 1 H NMR (500 MHz, CD 3 CN), (a) mixture of 6 (orange), 7 (blue), 8 (red) and 9 (not detected by 1 HNMR, see main text), (b) 7 only, (c) 6 only. * indicates peaks corresponding to aldehyde 1 or 2 resulting from hydrolysis of the respective helicate. Reactions were run at the same concentration with dialdehydes taken from a stock solution. As can be seen the yield of formation of Solomon link 6 is unaffected by the presence of 7 whereas the yield of 7 is drastically reduced with the presence of mix species 8 clearly visible. The total yield of pentafoil knot formation (7 + 8 + 9) remains comparable with the yield of 7 in spectrum (b).
All nine reactions were heated for 2d at 60 ºC before being allowed to cool. The following purification procedure was applied to each sample individually: The reaction mixture was precipitated using KPF 6 (saturated aqueous solution). The resulting purple powder was collected on Celite and washed with water, a small amount of EtOH and finally Et 2 O. The product mixture was dissolved in CH 3 CN and the solvent removed under reduced pressure. The nine samples were dissolved in 0.5 mL of CD 3 CN and compared by 1 H NMR (Sample 4, Figure S16. Sample 5, Figure   S17. Sample 6, Figure S18). See the main text for further details.
To two separate DMSO-d 6 solutions of 1 (2.1 mg, 4.7 mol, in 1.0 mL) and 2 (2.0 mg, 4.7 mol in 1.0 mL) was added FeCl 2 (50 L of a 0.10 M stock solution, 5.2 mol). The two purple solutions were mixed thoroughly to ensure complete dissolution of both dialdehydes. A 0.5 mL aliquot of 1+FeCl 2 was mixed with a 0.5 mL aliquot of 2+FeCl 2 in a clean NMR tube before addition of 2,2'-(ethylenedioxy)bis(ethylamine) 6 (50 L of a 0.10 M stock solution, 5.2 mol). The remaining 0.5 mL aliquots of 1+FeCl 2 and 2+FeCl 2 were separately reacted with 2,2'-(ethylenedioxy)bis(ethylamine) 6 (25 L of a 0.10 M stock solution, 2.6 mol, for each reaction). The three solutions were heated for 1 d at 60 ºC. The solutions of 1+FeCl 2 +6 and 2+FeCl 2 +6 were combined in a clean NMR tube and thoroughly mixed before being heated at 60 °C for a further 7 days. Both samples were monitored over time ( Figures S19 and S20). The two samples were allowed to cool and the following purification procedure was applied to each sample individually: The reaction mixture was precipitated using KPF 6 (saturated aqueous solution). The resulting purple powder was collected on Celite and washed with water, a small amount of EtOH and finally Et 2 O. The product mixture was dissolved in CH 3 CN and the solvent removed under reduced pressure. Both samples were dissolved in 0.5 mL of CD 3 CN and compared by 1 H NMR (Figure 3, main text).