Understanding gas capacity, guest selectivity, and diffusion in porous liquids

An in-depth study of porous liquids using measurement techniques, molecular simulations, and control experiments to advance their quantitative understanding.


General Synthetic and Analytical Methods
Materials: 1,3,5-Triformylbenzene was purchased from Manchester Organics, UK, and used as received. Other chemicals were purchased from Sigma-Aldrich or TCI UK and used as received, unless otherwise stated. PCP was purchased from Sigma-Aldrich. Other solvents were reagent or HPLC grade purchased from Fisher Scientific.
Methods: All reactions requiring anhydrous or inert conditions were performed in oven dried apparatus under an inert atmosphere of dry nitrogen, using dried and purified solvents (PCP and 1-tbutyl-3,5-dimethylbenzene -see section 3.3) introduced using disposable needles and syringes. All HPLC: HPLC was conducted on a Dionex UltiMate 3000 equipped with a diode array UV detector using a Thermo-Scientific Syncronis C8 column, 150x4.6 mm, 3 μm (SN 10136940, Lot 12459). The mobile phase was isocratic MeOH at a flow rate of 1 mL/min for a 10 min run time. The injection volume was 10 μL and the sample concentration was approximately 1 mg/mL. Detection for UV analysis was conducted at 254 nm.
Melting Points: Obtained using Griffin melting point apparatus and are uncorrected. FTIR: Infra-red spectra were recorded on a Bruker Tensor 27 FT-IR using ATR measurements for solid samples, and single sample transmissions for neat liquids with a Specac omni-cell demountable liquid cell with calcium fluoride (CaF 2 ) plates and a 0.05 mm PTFE insert (Scans: 16 background, 32 sample).
NMR: 1 H Nuclear magnetic resonance spectra were recorded using an internal deuterium lock for the residual protons in CDCl 3 (δ = 7.26 ppm) or d 2 -DCM (δ = 5.32 ppm) at ambient probe temperature on the following instruments: Bruker Avance 400 (400 MHz) or Bruker DRX500 (500 MHz). NMR studies were conducted using in-house made calibrated capillary of TMS in d 2 -DCM (100 μL sample from 10 μL TMS in 0.5 mL d 2 -DCM), so as to have no effect on the studies of the PL.

S3
Data are presented as follows: chemical shift, integration, peak multiplicity (s = singlet, d = doublet, t = triplet, q = quartet, qu = quintet, sex = sextet, sept = septet, m = multiplet, br = broad, app = apparent), coupling constants (J / Hz) and assignment. Chemical shifts are expressed in ppm on a δ scale relative to δ TMS (0 ppm) or δ CDCl3 (7.26 ppm) Assignments were determined either on the basis of unambiguous chemical shift or coupling patterns or by analogy to fully interpreted spectra for structurally related compounds. 13 C NMR Spectra were recorded using an internal deuterium lock using CDCl 3 (δ = 77.16 ppm) or d 2 -DCM (δ = 54 ppm) at ambient probe temperatures on the following instruments: Bruker Avance 400 (101 MHz) or Bruker DRX500 (126 MHz). 19 F NMR Spectra were recorded using an internal deuterium lock using a sealed capillary of TMS in CDCl 3 at a nominal probe temperature of 298 K on the following instrument: Bruker Avance 400 (376 MHz). 129 Xe NMR Spectra were recorded using an internal deuterium lock using a sealed capillary of TMS in CDCl 3 at a nominal probe temperature of 298 K on the following instrument: Bruker Avance II (400 MHz -1 H) wide bore spectrometer operating at 110.64 MHz.
HRMS: High resolution mass spectrometry was carried out using an Agilent Technologies 6530B accurate-mass QTOF Dual ESI mass spectrometer (capillary voltage 4000 V, fragmentor 225 V) in positive-ion detection mode. The mobile phase was MeOH + 0.1% formic acid at a flow rate of 0.25 mL/min. TGA: Thermogravimetric analysis was carried out using a Q5000IR analyser (TA instruments) with an automated vertical overhead thermobalance. The samples were heated at the rate of 5 °C/min to 600 °C in an aluminium pan under a nitrogen flow. All materials were desolvated by heating to 90 °C in a vacuum oven overnight prior to TGA analysis. SEM: Scanning electron microscopy was conducted using a Hitachi S4800 scanning electron microscope. Powdered samples of scrambled cage were deposited onto adhesive graphite tabs mounted on 15 mm aluminium stages. S6 GC: Gas chromatography measurements were carried out using a Thermo Scientific TRACE 1310 instrument configured with an FID detector and a 2,3-di-O-methyl-6-O-TBDMS-β-cyclodextrin capillary column (Supelco Beta DEX 325; 30 m × 0.25 mm × 0.25 μm). Samples were analysed using headspace injections and were performed by incubating the sample at 100 °C for 30 minutes followed by sampling 1 mL of the samples headspace. The following GC method was used; the oven was programmed from 95 °C with 35 min hold and 10 °C/min increments to 150 °C with 4 min hold, the total run time was 44.5 min; injection temperature 300 °C; detection temperature 300 °C with hydrogen, air, and make-up flow-rates of 35, 350, and 35 mL/min respectively; helium (carrier gas) flow-rate 1 mL/min. The samples were injected in the split mode (5:1). Numeric integration of the resulting peaks was performed using the supplied Chromeleon 7.1.2.1478 (Thermo Scientific Corporation) software package.
Diffusion NMR: All measurements were carried out non-spinning on a 400 MHz Bruker Avance 400 spectrometer, using a 5 mm indirect detection probe, equipped with a z-gradient coil producing a nominal maximum gradient of 34 G/cm. Diffusion data was collected using the Bruker longitudinal eddy current delay (LED) pulse sequence (ledgp2s). In the case of highly concentrated, viscous samples, bipolar gradients were used (ledbpgp2s) to minimise artefacts in the spectrum. A diffusion encoding pulse δ of length 1-7 ms, and diffusion delay D of 0.1-0.25 s were used. Gradient amplitudes were equally spaced between 1.70 and 32.4 G/cm. Each FID was acquired using 16 k data points. All experiments were carried out at a nominal probe temperature of 298 K, with an air flow of 800 m 3 /min to minimise convection. A sealed lock tube containing TMS in d 2 -DCM (100 μL sample from 10 μL TMS in 0.5 mL d 2 -DCM) was used to facilitate a deuterium lock without affecting the chemical makeup of the PL. All measurements were carried out three times and the numbers quoted represent the mean.
Diffusion coefficients were calculated from signal intensities using the Skejskal-Tanner equation 2 : Where I is the signal intensity, I 0 is the signal intensity at a gradient strength of zero, g is the gradient strength, and D is the diffusion coefficient (D = m 2 /s). Solvodynamic radii, R S (nm), of solution-phase species were calculated from the Stokes-Einstein equation assuming molecules have a spherical geometry: Where kT is the thermal energy (k = 1.38 x 10 -23 N m/K; T = 298 K) and η is the viscosity of the solvent (η = cP = 0.001 N s/m 2 ). Viscosities were measured at a regulated internal temperature of 298 K.

Synthesis of Scrambled Cage Mixture 3 3 :13 3 -R
Three batches of scrambled cage were produced by the following procedure: To a 3 L jacketed vessel, equipped with overhead stirrer, was added 1,3,5-triformylbenzene (7.5 g, 46.25 mmol, 4 eq.) in DCM (2.25 L), followed by solutions of (R,R)-1,2-diaminocyclohexane (3.961 g, 34.69 mmol, 3 eq.) in DCM (225 mL) and 1,2-diamino-2-methylpropane (3.058 g, 34.69 mmol, 3 eq.) in DCM (225 mL). The resulting solution was stirred at 20 °C under N 2 for 3 days and the reaction completion checked by HPLC analysis before concentration in vacuo. This preparation was repeated and both crude reaction products were redissolved in DCM (250 mL), combined, and filtered to remove any insoluble solids before concentration in vacuo. The resulting yellow solid was washed with EtOAc (3 x 100 mL) and the collected solid completely dissolved in DCM before concentration in vacuo prior to drying in the vacuum oven at 90 °C overnight to afford the scrambled 3 3 :13 3 -R cage mixture as a very pale yellow solid (Batch 1: 14.71 g, 61%; Batch 2: 14.54 g, 60%; Batch 3: 9.33 g, 77% (NB preparation only conducted once for batch 3)).

Purification of Solvents
Hexachloropropene (PCP): It is worth noting that different batches of PCP often had markedly different impurity profiles. The properties of the PL could be affected by these impurities, especially smaller-sized impurities which could act as a competitive guest in the cage cavities. Therefore, the solvent was thoroughly purified and analysed before use.

Safety Note -PCP is fatal by inhalation and a lachrymator. Therefore it is important that all manipulations using this solvent are conducted in a fume cupboard, with all samples for
measurements being appropriately sealed before removal.

Standard Procedures for Preparing the Liquid Samples for Testing
Preparation of the Scrambled Porous Liquid (PL): Scrambled 3 3 :13 3 -R cage mixture desolvated in a vacuum oven at 90 °C overnight before being evacuated and refilled with N 2 on a manifold in an oven dried GC headspace vial. The scrambled 3 3 :13 3 -R cage subsequently had degassed PCP added at a concentration of 20% w cage /v PCP (e.g. 200 mg dissolved in 1 mL PCP, 600 mg in 3 mL), the N 2 line removed and was stirred or vortexed until fully dissolved to afford a pale yellow liquid.

Preparation of PCP:
An oven dried GC headspace vial was evacuated and refilled with N 2 on a manifold before the addition of degassed PCP and removal of the N 2 line.

Preparation of Reduced-Scrambled Liquid (Red-PL):
Prepared according to the method for the PL but using the reduced scrambled cage mixture red-3 3 :13 3 -R and desolvation carried out at 80 °C. were measured using a liquid cell, and the IR of the solid scrambled 3 3 :13 3 -R cage mixture was measured using an ATR module.

Standard Procedure for Density Calculations
All measurements were repeated three times to calculate an average density with corresponding standard deviation.

PCP:
To a pre-weighed oven dried 1 mL volumetric flask with lid was added 1 mL PCP (freeze-pumpthaw degassed purified material) and the weight recorded.
PL: Three batches of 20% w/v PL were made-up according to the standard procedure (200 mg in 1 mL) and 1 mL of each was added to oven dried 1 mL volumetric flasks and the weights recorded. The pore volume can also be calculated using the measured density of the PL (20% w/v). For a sample containing 200 mg 3 3 :13 3 -R in 1 mL PCP the total mass will be 1.9127 g, with an overall volume of 1.19 mL. In this sample there will be 1.92x10 -4 mol cage, and therefore 1.159x10 20 cage molecules. Assuming a cavity size of 5 Å, the pore volume in a single cage will be 6.545x10 -23 mL, leading to a total pore volume of 7.586x10 -3 mL. This leads to a percentage pore volume of 0.64%.

Computational model and molecular dynamics simulations
The Amorphous Cell module in Materials Studio 3 was used to generate a 74.44 Å amorphous box of scrambled 3 3 :13 3 -R. In total, 30 cages were inserted and for simplicity one of the most prevalent isomers in the mixture (see Supplementary Fig. 4), a 3 3 :13 3 -R isomer, was selected (see Fig. 3a)-this removes the difficulty of positional isomerism, and it is assumed this would not affect the results.
DL_FIELD 4 was then used to solvate the system with PCP, such that the ratio of cage:PCP was 1:36.
The system was then checked to see whether PCP was inserted in the cage cavity. This is important, as the movement of PCP with respect to the cage can be accurately monitored. Supplementary Fig.   18 shows this process.
Supplementary Fig. 18: Scheme illustrating the setup of the PL.
MD simulations were subsequently carried out using DL_POLY_2.20. 5,6 A potential cut-off of 10 Å was used and electrostatic interactions were calculated using the partial charges from OPLS FF. An NVT ensemble (constant number of moles, volume and temperature) at 1 atm and 298 K was used with the Hoover barostat and thermostat, 7 and both had a time constant of 0.5 ps. A timestep of 0.5 fs was used, with the system first equilibrated for 50 ps with temperature scaling every 5 fs, followed by a production run of 10 ns, with a frame output every 2.5 ps.
Once complete, the system was analysed using an in-house script to determine the centre of mass for each cage and PCP molecule. Vectors between these were calculated for each frame of the MD simulation, such that it was possible to determine the distance between the centre of the cages, and each PCP. To ascertain whether a PCP molecule had in fact entered a cage, the following criteria were used:  It was also possible to monitor the simulation density, to make sure it agrees well with that observed experimentally. The average density of the system was maintained at 1.54 g/cm 3 post equilibration.

Optimisation and Reproducibility of CO 2 Uptake
Sample Preparation: Batches of the PL (20% w/v) were made-up according to the standard procedure in either non-degassed or freeze-pump-thaw degassed pure PCP (200 mg in 1 mL). The PL then had CO 2 bubbled through at either 5-10, 50-60 or 200-220 mL/min (6)(7)(8)(9)(10)(11)(12)(13)(14)(44)(45)(46)(47)(48)(49) or 95-100 respectively on Gilmont flowmeter scale with a stainless steel float) with an 18 gauge needle as an outlet for 2 h under a range of conditions including ice-cooled or at rt (17-25 °C), with or without the addition of 1.0 eq. water, sonicated prior to CO 2 addition, with pre-wet cage or using cage a day after desolvation, and with duplicates conducted and different batches of scrambled cage used to test reproducibility. Control samples were conducted on 1 mL PCP under the same conditions.

CO 2 Saturation Test
PCP: Six samples were prepared according to the standard procedure (1 mL) and CO 2 was bubbled through at a flow rate of ~50-60 mL/min (44-49 on Gilmont flowmeter scale with a stainless steel float) with an 18-gauge needle as an outlet for a set time period prior to IR analysis.
PL: Six batches of 20 % w/v PL were prepared according to the standard procedure (200 mg in 1 mL) and CO 2 was bubbled through at 50-60 mL/min (44-49 on Gilmont flowmeter scale with a stainless steel float) with an 18-gauge needle as an outlet for a set time period prior to IR analysis.
IR analysis: IR analysis of CO 2 uptake conducted immediately after gas introduction (0,5,10,20,30,60 and 120 min) with the neat liquid samples in a Specac omni-cell demountable liquid cell with calcium fluoride (CaF 2 ) plates and a 0.05 mm PTFE insert.
The absorbance IR spectra were integrated using Origin with the CO 2 signal integrated from 2300.9509-2368.4558 cm -1 and the PCP signal integrated from 1500.5363-1579.6134 cm -1 . All CO 2 uptakes were made relative to the PCP to allow for comparisons. Saturation curve plotted showing relative CO 2 uptake over time for both PCP and the PL with both reaching saturation after as little as 5 minutes of CO 2 addition via bubbling at 50-60 mL/min.

S37
Scalability Test: 200 mg Scrambled 3 3 :13 3 -R cage in 1 mL PCP found to be saturated after 5 min at a flow rate of 50-60 mL/min, therefore the same experiment was subsequently conducted on 2 mL and 3 mL of the PL with CO 2 introduced for 10 min and 15 min respectively to check scalability. Supplementary Fig. 26: Saturation scale-up test measuring the PL CO 2 uptake at 50-60 mL/min for 5 min per 1 mL of PCP used. (a) Overlaid transmission IR of 1, 2, and 3 mL of the PL after CO 2 addition for 5, 10, and 15 min respectively. (b) Calculated relative CO 2 uptake for the different scales confirming that CO 2 can be reproducibly introduced for 5 min at 50-60 mL/min per 1 mL of PCP used in the PL.

Study of CO 2 Retention
PCP: Two samples were prepared according to the standard procedure (1 mL) in 4 mL oven-dried glass vials and CO 2 was bubbled through for 5 min at a flow rate of ~50-60 mL/min (44-49 on Gilmont flowmeter scale with a stainless steel float) with an 18-gauge needle as an outlet before analysis by FTIR to confirm saturation. The sample was then left in the capped vial and stored at either rt or at -20 °C in the freezer with daily analysis of the CO 2 content by FTIR.
PL: Two samples of 20% w/v PL were prepared according to the standard procedure (200 mg in 1 mL) and CO 2 was bubbled through for 5 min at 50-60 mL/min (44-49 on Gilmont flowmeter scale with a stainless steel float) with an 18-gauge needle as an outlet before analysis by FTIR to confirm saturation. The sample was then left in the capped vial and stored at either rt or at -20 °C in the freezer with daily analysis of the CO 2 content by FTIR.
IR analysis: IR analysis of CO 2 uptake conducted immediately after gas introduction and after 1, 2, and 3 days with the neat liquid samples in a Specac omni-cell demountable liquid cell with calcium fluoride (CaF 2 ) plates and a 0.05 mm PTFE insert.
The absorbance IR spectra were integrated using Origin with the CO 2 signal integrated from 2300.9509-2368.4558 cm -1 and the PCP signal integrated from 1500.5363-1579.6134 cm -1 . All CO 2 uptakes were made relative to the PCP to allow for comparisons.

CH 4 Uptake by 1 H NMR
All samples (0.6 mL) had a 1 H NMR spectra recorded using the same sealed d 2 -DCM/TMS capillary as an internal standard (capillary 1), both prior to CH 4 addition and after (3 min bubbling at 50-60 mL/min, 32-36 on Gilmont flowmeter scale with a stainless steel float). N.B. As a precaution, an NMR lid with a hole in was used to avoid build-up of pressure due to gas release. PCP: 0.6 mL freeze-pump-thaw degassed pure PCP.

PL:
Conducted on 20% w/v PL (120 mg dissolved in 0.6 mL) prepared according to the standard procedure. Therefore, calculation of the PL concentration using the calibration curve is possible using y = a + b*x so for each capillary the equation is:

Supplementary
S43 Supplementary This allows a quantitative CH 4 uptake in the PL and PCP to be calculated.

CH 4 Saturation Test
Both the PL (20% w/v-120 mg dissolved in 0.6 mL, prepared according to the standard procedure) and PCP (0.6 mL freeze-pump-thaw degassed) had a 1 H NMR spectra recorded using the same calibrated d 2 -DCM/TMS capillary as an internal standard (capillary 1), both prior to CH 4 addition and after each set addition (3 or 9 min bubbling at 50-60 mL/min, 32-36 on Gilmont flowmeter scale with a stainless steel float).
NB. As a precaution, an NMR lid with a hole in was used to avoid build-up of pressure due to gas release.
Analysis of all the 1 H NMR spectra using the same method as shown in Supplementary Tables 7-9 enabled a saturation curve of CH 4 uptake over time graph to be plotted for both PCP and the PL.

Study of CH 4 Retention
The saturated PL and PCP samples from the CH 4 saturation study in Supplementary Fig. 31 were left at rt in the NMR tubes, capped with a lid with a hole in to avoid build-up of pressure due to any gas release, and analysed by 1 H NMR spectroscopy after 1, 2, 5, and 7 days using the same calibrated d 2 -DCM/TMS capillary as an internal standard (capillary 1).
Analysis of all the 1 H NMR spectra using the same method as shown in Supplementary Tables 7-9 enabled the CH 4 content to be plotted for both PCP and the PL over time.

CO 2 Uptake
Control Non-PL: A sample of 20% w/v control non-PL was prepared according to the standard procedure (200 mg in 1 mL) and CO 2 was bubbled through for 5 min at 50-60 mL/min (44-49 on Gilmont flowmeter scale with a stainless steel float) with an 18-gauge needle as an outlet before analysis by FTIR.
FTIR analysis and calculation of relative CO 2 uptake conducted using the same procedure as for the PL and PCP (see section 5.1).

CH 4 Uptake
All samples (0.6 mL) had a 1 H NMR spectra recorded using the same calibrated d 2 -DCM/TMS capillary as an internal standard (capillary 1), both prior to CH 4 addition and after (3 min bubbling at 50-60 mL/min, 32-36 on Gilmont flowmeter scale with a stainless steel float). N.B. As a precaution, an NMR lid with a hole in was used to avoid build-up of pressure due to gas release.

PL:
Conducted on 20% w/v PL (120 mg dissolved in 0.6 mL) prepared according to the standard procedure.
Control Non-PL: Conducted on 20% w/v control non-PL (120 mg dissolved in 0.6 mL) prepared according to the standard procedure. Supplementary Fig. 34 Fig. 34) showing that in order for an increased gas uptake to be observed, shape-persistent cages are required over flexible ones.

Xe and SF 6 Uptake
All samples (0.6 mL) had either a 129 Xe or 19 F NMR spectra recorded, using the same calibrated d 2 -DCM/TMS capillary as an internal lock, after 3 min bubbling of either Xe or SF 6 gas at 50-60 mL/min (60-66 or 62-69 respectively on Gilmont flowmeter scale with a stainless steel float). N.B. As a precaution, an NMR lid with a hole in was used to avoid build-up of pressure due to gas release.

PL:
Conducted on 20% w/v PL (120 mg dissolved in 0.6 mL) prepared according to the standard procedure.
Control Non-PL: Conducted on 20% w/v control non-PL (120 mg dissolved in 0.6 mL) prepared according to the standard procedure.

Sonication as a Gas Release Mechanism
Sample Preparation: Two samples of the PL (20% w/v) were made-up according to the standard procedure (600 mg in 3 mL, and 400 mg in 2 mL) and subjected to cycles of CO 2 addition (bubbled through for 15 and 10 min respectively at 50-60 mL/min, 44-49 on Gilmont flowmeter scale with a stainless steel float) and release by sonication (30 min sonication per cycle, not heated). The 3 mL sample was connected to the gas collection setup during sonication and the uptake-release cycle repeated 5 times with the amount of evolved gas being measured each time, whereas the 2 mL sample was used for IR analysis.
IR analysis: IR analysis of CO 2 uptake conducted immediately after each gas addition and after each sonication, over 3 uptake-release cycles, with the neat liquid samples in a Specac omni-cell demountable liquid cell with calcium fluoride (CaF 2 ) plates and a 0.05 mm PTFE insert. The absorbance IR spectra were integrated using Origin with the relative quantity of CO 2 calculated as discussed previously (see Fig. 6b and 6d).
Controls: Two samples of PCP (3 mL) were prepared according to the standard procedure, and one had CO 2 added (bubbled through for 15 min at 50-60 mL/min, 44-49 on Gilmont flowmeter scale with a stainless steel float) prior to both undergoing release by sonication (30 min sonication per cycle, not heated) whilst connected to the gas collection setup.

Gas Evolution Measurements
This guest selectivity of the PL allows gas evolution measurements to be conducted allowing estimated gas uptakes to be calculated without resorting to techniques such as FTIR or NMR.
Standard procedure for gas evolution studies of the porous liquid via small solvent release: The desired amount of scrambled cage (typically 600 mg, 0.5772 mmol) was added to a pre-weighed GC headspace vial ( Excess CHCl 3 : The gas flow was removed and the cap was rapidly changed for a new one with an unbroken septum. Using a syringe with a 21 gauge needle an equivalent volume of CHCl 3 (typically 3 mL) was carefully layered onto the PL whilst connected to the gas collection setup via a needle/tubing cannula, thereby also ensuring no air-locks remained and setting a start-point which was marked. The sample was stirred allowing the layers to mix and gas evolution was measured by displacement of water in an inverted Rotaflo stopcock 10 or 25 mL burette (0.1 mL graduations) in a beaker of water connected to the GC vial via the needle/tubing cannula.
1.0 eq. CHCl 3 : The gas flow was removed and the cap was rapidly changed for a new one with an unbroken septum. Using a syringe with a 21 gauge needle the gas being tested was used, whilst connected to the gas collection setup via a needle/tubing cannula, to set a start-point and ensure no air-locks remained. After marking the start-point, 1.0 eq. CHCl 3 (relative to the amount of cage present, typically 46 μL) was carefully added so as not to disturb the PL before the sample was stirred allowing the CHCl 3 to fully mix with the PL. Gas evolution was measured by displacement of water in an inverted Rotaflo stopcock 10 or 25 mL burette (0.1 mL graduations) in a beaker of water connected to the GC vial via the needle/tubing cannula. Once the gas evolution had stopped a further 1.0 eq. CHCl 3 was added to ensure no further gas could be displaced.

S56
Supplementary Fig. 37: Photo of the gas collection setup used in the gas evolution studies. The sample for gas evolution studies is made up in a GC headspace vial (22 mm x 45 mm screw top, 10 mL, Fisher Scientific) and sealed with an undamaged GC screwtop cap with septum. This is connected by a needle/tubing cannula to a water filled inverted Rotaflo stopcock 10 or 25 mL burette (0.1 mL graduations) in a beaker of water. The cannula and burette were tested for leaks using an empty GC headspace vial by adding a known amount of air into the system via a syringe.
Prior to testing a known amount of the gas in use, or the excess of chloroform, was used to set the start point and ensure no air-locks remain. After piercing the septum and removing the needle vacuum grease can be used to ensure the cap is still completely sealed and that any gas released will be solely collected in the burette.
Measurements were repeated a total of three times-two times on one batch and another on a different batch of cage material, at the same temperature to obtain a mean gas evolution and corresponding standard deviation.

Gas evolution studies of control samples via small solvent release:
PCP: A GC vial dried in the vacuum oven at 90 °C overnight was sealed with a septum cap and connected to a manifold via a needle through the septum. The vial was evacuated and refilled with N 2 before the addition of freeze-pump-thaw degassed pure PCP (2 mL) prior to gas evolution testing via small solvent addition (31 μL, 1.0 eq. CHCl 3 -calculated from the amount of cage that would have been used in 2 mL i.e. 400 mg) using the standard procedure for PLs above.
PCP + Gases: Method as for PCP but the desired gas was bubbled through the solution at a rate of 50-60 mL/min for 5 min per 1 mL of PCP prior to testing (see standard procedure above for the PL).

PL + Bulky Additive:
Conducted on 20% w/v PL (600 mg dissolved in 3 mL) prepared and saturated with CO 2 according to the standard procedure, but 1-t-butyl-3.5-dimethylbenzene (108 μL, 1.0 eq. relative to the amount of cage used) used as a test displacement solvent prior to CHCl 3 (46 μL, 1.0 eq. relative to the amount of cage used) using the standard procedure for gas evolution testing of the PL.

Control Non-PL + CHCl 3 :
A sample of 20% w/v control non-PL was prepared according to the standard procedure (600 mg in 3 mL) and saturated with CO 2 prior to gas evolution testing via small solvent addition (119 μL, 1.0 eq. CHCl 3 relative to the amount of cage used) using the standard procedure for gas evolution testing of the PL.

S58
Calculation of maximum gas release based on a 1:1 stoichiometry of gas:cage: By assuming a maximum occupancy of 1:1 gas to cage at 1 bar, i.e. that for a 20% w/v sample PL (200 mg cage in 1 mL PCP) the maximum gas uptake will be 0.1924 mmol, and by using the Ideal Gas Equation, the maximum theoretical volume of gas that could be evolved by displacement can be calculated: Therefore, the maximum theoretical gas release will be 4.6 cm 3 per 200 mg of cage in 1 mL PCP. This equates to a maximum of 13.8 cm 3 being evolved for a 600 mg in 3 mL sample of PL.

Calculation of Gas Uptake:
By re-arranging the Ideal Gas Equation it is possible to convert the measured volume of gas evolved (cm 3 ) into n(mmol) of gas evolved per 600 mg cage or per 600 mg cage in 3 mL PCP i.e. per 5.7381 g PL (cage = 600 mg, PCP = 5.1381 g, 3 mL, measured ρ = 1.7127 g/mL (see Supplementary Table 2)).

( ) =
This can be converted into an estimated gas uptake in terms of either mmol/g cage , allowing comparison to other sorption data on solid cage species, or mmol/g PL allowing comparison to uptake in alternative porous liquids.

Gas
Mean Gas Evolution ± SD cm 3 Excess CHCl 3 1.0 eq. CHCl 3 N 2 1.57 ± 0.08 2.10 ± 0.20 CH 4 5.60 ± 0.30 6.33 ± 0.21 CO 2 4.25 ± 0.13 7.53 ± 0.47 Supplementary Fig. 38: PL gas evolution measurements-average volumes of gas collected by displacement with either an excess of chloroform, or one molar equivalent based on cage. Using 1.0 eq. CHCl 3 proves to be more efficient in the gas evolution tests and was chosen as the optimal displacement method, presumably because large volumes of CHCl 3 partially dissolve the gas being displaced, which is particularly evident in the case of CO 2 .
S60 Supplementary Table 14: Volumes of gas evolved from the PL (20% w/v) and PCP using 1 molar equivalent of CHCl 3 relative to cage (see Fig. 5a) and subsequent conversion into percentage occupancies and estimated gas uptakes. Assuming a maximum occupancy of 1 gas molecule per host cage, and by using the Ideal Gas Equation, the measured amount of evolved gas can be compared to the theoretical maximum amount of evolved gas (13.8 cm 3 for the 20% w/v PL from 600 mg in 3 mL PCP) to obtain a percentage occupancy (see Fig. 5a). Calculation of n(μmol) gas evolved is possible from the average volume collected (cm 3 ) using the Ideal Gas equation, and can be converted into the uptake of each gas in terms of either μmol/g cage or equivalents of gas per cage -allowing comparison to other sorption on solid cages, or μmol/g PL -allowing comparison to uptake in alternative porous liquids (see Fig. 5b).  Fig. 39: Comparison of CO 2 gas evolution measurements on CO 2 -saturated control samples. Demonstrates that the cage is playing a role in the increased uptake of CO 2 (PCP vs PL vs control non-PL) and that the PL exhibits guest selectivity by size-exclusion-chloroform fits inside the cages displacing the gas, whilst the bulky additive 1-t-butyl-3,5-dimethylbenzene does not.

Chiral Selectivity -Enantioselective Adsorption of 1-Phenylethanol
Procedure for GC sample preparation: Solid Scrambled 3 3 :13 3 -R cage: Six samples of solid scrambled 3 3 :13 3 -R cage (50 mg) were desolvated at 90 °C in the vacuum oven in GC headspace vials (22 mm x 76 mm screw top, 20 mL, Fisher Scientific) before being allowed to cool to rt. To 3 of the samples was added 1 eq. of 1-PhEtOH (5.8 μL) and to the remaining samples was added 2 eq. of 1-PhEtOH (11.6 μL) before the vials were sealed.
PL: Six samples of 20% w/v PL (200 mg scrambled 3 3 :13 3 -R cage in 1 mL PCP) were prepared according to the standard procedure in GC headspace vials (22 mm x 76 mm screw top, 20 mL, Fisher Scientific). To 3 of the samples was added 1 eq. of 1-PhEtOH relative to cage (23.2 μL) and to the remaining samples was added 2 eq. of 1-PhEtOH relative to cage (46.4 μL) before the vials were sealed. The same method was also used to prepare an equivalent six samples of 5% w/v PL (50 mg in 1 mL) with 5.8 μL and 11.6 μL of 1-PhEtOH for 1 and 2 eq. relative to cage respectively.

Non-PL:
Two samples of 20% w/v control non-PL (200 mg SIC in 1 mL PCP) were prepared according to the standard procedure in GC headspace vials (22 mm x 76 mm screw top, 20 mL, Fisher Scientific). To one of these samples was added 1 eq. of 1-PhEtOH relative to the control molecule (59.5 μL) and to the other was added 2 eq. of 1-PhEtOH relative to the control molecule (119 μL) before the vials were sealed. The same method was also used to prepare an equivalent two samples of 5% w/v control non-PL (50 mg in 1 mL) with 14.8 μL and 29.7 μL of 1-PhEtOH for 1 and 2 eq. relative to the control molecule respectively.
All samples were vortexed at rt for 18 h at 200 rpm to allow them to reach equilibrium prior to the chiral selectivity measurements.

Chiral Selectivity Measurements:
The enantioselectivity of the solid scrambled 3 3 :13 3 -R cage, PL (20% w/v) and control non-PL (20% w/v) were measured for rac-1-phenylethanol following the previously described method. 9 Briefly, using the equilibrated samples the proportion of each enantiomer of 1-phenylethanol adsorbed in the host was measured by chiral GC analysis. The 1phenyethanol in the chromatograms is representative of what is left in the solution phase after adsorption into the host. For example, solid scrambled 3 3 :13 3 -R cage has adsorbed more (S)-1phenylethanol than (R)-1-phenylethanol. This experiment was repeated at two different guest:host ratios for the solid 3 3 :13 3 -R cage, the PL and the non-PL. Two concentrations of PL and non-PL were S62 tested; 5% w/v and 20% w/v. Each combination of host and guest were prepared and measured three times and repeat injections of each sample were also run to ensure reproducibility of the headspace injection. Representative chromatograms of each host with two equivalents of guest are shown in Supplementary Fig. 40.
The solid scrambled 3 3 :13 3 -R cage shows chiral selectivity and an ee of 14%. The ee is consistently 14% in different samples and is the same with both 1 or 2 eq. of rac-1-phenyethanol. The 5% and 20% w/v PL do not show any enantioselectivity with either 1 or 2 eq. of rac-1-phenylethanol.
Likewise the 5% and 20% w/v control non-PL doesn't show any enantioselectivity with 1 or 2 eq. of rac-1-phenylethanol. This suggests that the solid scrambled 3 3 :13 3 -R cage has potential as a material for chiral separations, whereas the PL and control non-PL currently do not.

Porosity in porous liquids vs porous organic cages
By conducting sorption of the tested gases on the solid scrambled 3 3 :13 3 -R cage and CC1α, at the same temperature as the gas evolution tests on the PL (see Fig. 7b and 7c), it is possible to compare the gas uptakes as mmol/g cage and as equivalents per cage for the two solids and the PL (20% w/v).

Study of Cage Aggregation in the Porous Liquid
Diffusion NMR was carried out on the PL at a range of concentrations (2.5% to 20% w/v) to assess whether cages aggregate at high concentration: Sample Preparation: Five samples of varying % w cage /v PCP PL (2.5%, 12.5 mg in 0.5 mL; 5%, 25 mg in 0.5 mL; 10%, 50 mg in 0.5 mL; 15%, 75 mg in 0.5 mL; 20%, 200 mg in 1 mL) were prepared according to the standard procedure in small vials and each sample was analysed by diffusion NMR and had its viscosity measured at 298 K. Changes in hydrodynamic radii were found to be negligible with increasing concentration (see Fig.   8a), and were consistent with a cage 1.5 nm in size.

Study of the Host-Guest Chemistry in the Porous Liquid
To study the host-guest chemistry within the PL, diffusion NMR was carried out to determine the diffusion coefficients of host (scrambled 3 3 :13 3 -R cage) and guest (CH 4 , CHCl 3 and 1-t-butyl-3.5dimethylbenzene) molecules within the PL (20% w/v). To assess whether these values correlated to guest-binding, diffusion coefficients of the same guests in neat solvent (PCP) were measured. To confirm that guest binding was a function of the cavity, and not the chemical structure of the cage, experiments were also repeated with 1,1',1''-(benzene-1,3,5-triyl)tris(N-cyclohexylmethan-imine) as the host in the control non-PL (20% w/v).

Sample Preparation:
The following samples were made up in vials according to the standard procedures before transferring to an NMR tube for diffusion measurements:
Guest uptake was established by comparing the solvodynamic radii (R S ) of guests in neat solvent (PCP), PL (20% w/v) and without the scrambled 3 3 :13 3 -R cage, but in the presence of a small imine control molecule in the non-PL (20% w/v). An increase in apparent R S is consistent with a small guest forming a host-guest complex and diffusing at the speed of the host, making it appear larger. An apparent increase in size was observed for both CH 4 and CHCl 3 in the PL compared to their unbound state in PCP, albeit not to the same size as the host cage molecule, suggesting a host-guest complex has been formed but is dynamic in nature where the guest can rapidly exchange between its unbound and bound state. No apparent increase in size was observed for the bulky 1-(t-butyl)-3,5dimethylbenzene guest suggesting it is excluded from the cage cavity. No apparent increase in size was observed for any of the guests in the control non-PL, suggesting the apparent sizes increases seen in the PL are due to the guests binding with the cage cavity.

Calculation of Percentage Occupancies
The estimated amount of bound guest molecules in the host cage cavities within the system (f b ) was estimated using the method previously described by Hermans et al. 10 To account for the viscosity difference between neat PCP and the PL, each diffusion coefficient was normalised by the viscosity.