Kyoto chemists tune a porous liquid's viscosity 170-million-fold from the cage surface alone

The lever sits on the outside of the cage. Twelve flexible PEG chains hang off each rhodium metal-organic polyhedron, making a star-shaped molecule that assembles into a solvent-free fluid ^[7]. What the surface group decides is where those chains go. Dodecyloxy ends are hydrophobic, so they are pulled back toward the surface they grew from, and each molecule ends up wearing a compact folded shell; shells like that interpenetrate little and slide past one another ^[9]. Put a hydroxy group there instead and the chains reach into neighbouring cages, knitting a transient network that resists flow ^[10]. Synchrotron X-ray scattering and molecular dynamics both point at conformation rather than composition ^[9], and the series design is what makes that reading defensible: within a chain-length series the cavity, the chain count and the chain length are all held fixed, and only the surface group moves ^[8].

That control matters because the older ways of thinning a porous liquid all cost you something. Dilution drops viscosity and the concentration of porous component together ^[3]; shortening the tethers in a solvent-free type I liquid changes how much non-porous material is in the pot and can seal the pores outright ^[4].

The span is worth checking arithmetically. Zero-shear viscosity at 60 C runs from 18 pascal-seconds to 3.0 billion ^[11]; the ratio is 1.7 x 10^8, which is 8.2 orders of magnitude ^[18]. Xiang attributes the whole 170-million-fold change to a change confined to the cage surface, with polymer length and pore volume untouched ^[16]. The account does not break the range out by series, so a reader cannot see how much of those 8.2 orders one chain length delivers on its own ^[21]. And the top end is not a liquid in any everyday sense: the stiffest samples behave like solids on short timescales and only flow if you wait ^[12].

The gas side has the more surprising result. Warming the long-chain dodecyloxy liquid from 0 C to 30 C raised CO2 uptake from 0.29 to 1.25 moles per mole, a 4.3-fold increase ^[14][19], because cold chains sit still and block the cavity entrance while mobile ones swing clear ^[15]. Porous solids generally go the other way ^[14]. The hydroxy surface, whose chains stay extended, keeps the door open even when cold ^[15].

What a kilogram of the stuff holds is the number missing here. Uptake is reported per mole of porous liquid ^[14], and a mole of cage plus twelve polymer arms is heavy; the account covers that mole-based figure but stops short of a per-mass or per-volume capacity, repeat-cycle data, CO2-over-nitrogen selectivity, or a cast membrane ^[22]. Nor were the two headline properties characterised at one operating point: the rheology is at 60 C, the pore accessibility at 0 to 30 C ^[11][13][14].

So read this as a design rule for chemists rather than a fluid for a plant. Furukawa's claim is that pore volume and viscosity have been decoupled ^[17], and on the evidence shown that holds. Whether 18 pascal-seconds is pumpable in a real contactor, and whether the rule survives a metal cheaper than rhodium, are separate questions the surface-group finding does not settle.