Solvent selection — water, supercritical CO₂, ionic liquids: Visual walkthrough

We will use only two everyday ideas to start: squeezing (pressure) and how packed the molecules are (density). Everything else we earn as we go.

Step 1 — What "pressure" and "density" even mean

WHAT. Picture a sealed box of CO₂. Two numbers describe it:

$P$ = pressure = how hard the gas pushes outward on the walls (how much you are squeezing it).
$ρ$ (the Greek letter "rho", say "row") = density = how many molecules are crammed into each cubic centimetre — how packed it is.

WHY start here. The parent note's punch line is "solvent power tracks density." To dissolve a greasy caffeine molecule, CO₂ molecules must physically surround it and tug on it. More CO₂ packed per cubic centimetre = more tugging = stronger solvent. So $ρ$ is literally "how good a solvent CO₂ is." Everything below is about controlling $ρ$ .

PICTURE. On the left, a loosely-packed gas (low $ρ$ , weak solvent). On the right, the same molecules squeezed close (high $ρ$ , strong solvent). The knob between them is $P$ .

Step 2 — The P–T map: where CO₂ is solid, liquid, gas

WHAT. Draw a map whose horizontal axis is temperature $T$ and vertical axis is pressure $P$ . Every point $(T, P)$ is one "setting" of the box. The map is carved into three countries — solid, liquid, gas — separated by border lines. (This is a phase diagram.)

WHY a map. We want to reason about changing $P$ and $T$ and watching what happens. A map lets us draw a path and see which borders we cross. Crossing a border = a phase change (boiling, freezing).

PICTURE. The three countries meet at the triple point (all three coexist). Follow the liquid–gas border upward and to the right: that is the boiling line — the pressure at which liquid turns to gas at each temperature.

Step 3 — The border line ends: meet the critical point

WHAT. Trace the liquid–gas boiling line further up. It does not go on forever — it stops dead at a single dot. That dot is the critical point $(T_{c}, P_{c})$ . For CO₂: $T_{c} = 304.25 K = 31.1 ° C, P_{c} = 7.38 MPa = 73.8 bar$

Reading the symbols where they sit:

$T_{c}$ — the critical temperature, the hottest you can be and still have a real liquid–gas border. Subscript $c$ = "critical."
$P_{c}$ — the critical pressure, the border's height at that final dot.

WHY the line ends. As you heat toward $T_{c}$ , the liquid expands (gets less dense) and the gas is compressed harder (gets more dense). The two densities crawl toward each other. At exactly $T_{c}$ they become equal — liquid and gas are now indistinguishable, so there is nothing left to separate: the border must stop.

PICTURE. The boiling line rises and terminates at the star marked $(T_{c}, P_{c})$ . Just below the star, liquid and gas densities are almost touching.

Recall Why "supercritical" is NOT "super hot"

Supercritical means past that star in both $T$ and $P$ . ::: For CO₂ that star is at a balmy $31 ° C$ — barely above a warm room. It is about phase behaviour, not extreme heat.

Step 4 — Sneak around the endpoint: no boiling, no boundary

WHAT. Here is the trick. Take liquid CO₂ at low temperature and high pressure. Instead of heating it straight across the boiling line (which would make it boil — a violent density jump), take a detour: go up in pressure, then right in temperature (staying above the star), then down. You arrive at gas-like conditions without ever crossing the border.

WHY this matters. Because you never crossed a phase boundary, there was no boiling, no sudden jump — the density changed smoothly the whole way. The region beyond the star, where $T > T_{c}$ and $P > P_{c}$ , is a single fluid that is neither purely liquid nor purely gas. That is the supercritical fluid.

PICTURE. Two paths from liquid to gas: the red straight path crosses the border (bang — boiling); the teal looping path swings around the star (smooth, no border crossed).

Step 5 — In that region, density becomes a dial

WHAT. Now sit at a fixed temperature just above $T_{c}$ and slowly turn up $P$ . Watch $ρ$ . Far from the star, pushing $P$ barely changes $ρ$ (gas is springy). But right near $P_{c}$ , a tiny extra squeeze packs in a lot more molecules — the $ρ$ -vs- $P$ curve goes almost vertical.

WHY. With no phase border to jump across, density responds continuously to pressure. And near the critical point the fluid is on the knife-edge between liquid and gas, so it is extraordinarily easy to compress — a whisper of pressure moves it a long way.

PICTURE. Density plotted against pressure at $T$ slightly above $T_{c}$ . The curve is flat at low $P$ , then rears up steeply through the critical region — the "solvent-strength dial" is that steep part.

Step 6 — Naming the steepness: isothermal compressibility

WHAT. We need a number for "how much density jumps per unit of squeeze, at fixed temperature." That number is the isothermal compressibility $κ_{T}$ (kappa, with subscript $T$ = "at constant temperature"): $κ_{T} = - \frac{1}{V} (\frac{\partial V}{\partial P})_{T} = \frac{1}{ρ} (\frac{\partial ρ}{\partial P})_{T}$

Reading it term by term, right where each symbol lives:

$V$ — the volume of the gas; squeezing shrinks it.
$(\frac{\partial V}{\partial P})_{T}$ — "how much volume changes when I nudge pressure, holding $T$ fixed." The curly $\partial$ (say "partial") means change one knob while freezing the others. WHY this tool and not an ordinary derivative? Because two knobs exist ( $P$ and $T$ ); the partial derivative is the only tool that says "wiggle $P$ only." That is precisely the experiment in Step 5.
The minus sign — squeezing raises $P$ but shrinks $V$ , so $\partial V / \partial P$ is negative; the $-$ flips it so $κ_{T}$ comes out positive (a tidy "bigness of squishiness").
$\frac{1}{V}$ — divides out the box size, so $κ_{T}$ measures fractional squishiness, fair for any amount of gas.
The second form swaps $V$ for $ρ$ (they are reciprocals: pack tighter = smaller volume), giving directly "fractional density gain per unit pressure" — exactly our solvent dial's steepness.

WHY define it. $κ_{T}$ is the mathematical name for the steep part of the Step-5 curve. Big $κ_{T}$ = the dial is sensitive = a little $Δ P$ buys a lot of solvent power.

PICTURE. The same curve as Step 5, with the slope $\partial ρ / \partial P$ drawn as a tangent arrow at three points — gentle far out, near-vertical at the critical region.

Step 7 — The punch line: $κ_{T}$ blows up at the critical point

WHAT. At the critical point the pressure barely changes as volume changes — the flat spot means $(\frac{\partial P}{\partial V})_{T} = 0$ . Flip it over: since $κ_{T}$ contains the reciprocal of that slope, $(\frac{\partial P}{\partial V})_{T} \to 0 ⟹ κ_{T} = - \frac{1}{V} \frac{1}{( \partial P / \partial V ) _{T}} \to \infty$

$(\frac{\partial P}{\partial V})_{T} \to 0$ — "changing the volume does almost nothing to the pressure" — the tell-tale flatness at the star.
Dividing $1$ by something heading to $0$ sends the whole thing to infinity — $κ_{T}$ blows up.

WHY this is the entire trick. Near-infinite $κ_{T}$ means near-infinite sensitivity: an infinitesimal pressure change produces a giant density change, hence a giant swing in solvent power. This is why engineers set scCO₂'s "how much do I dissolve" simply by choosing the pressure — the machine practically has a solvent-strength volume knob.

PICTURE. $κ_{T}$ plotted against pressure along the near-critical isotherm: a sharp spike stabbing upward right at $P_{c}$ , tailing off on either side.

Step 8 — The degenerate & edge cases (never leave a gap)

WHAT & WHY, case by case:

Case A — far below the critical point (ordinary gas). $κ_{T}$ is small and roughly $1/ P$ (ideal-gas behaviour). The dial is dull: big pressure changes barely move density. So scCO₂'s trick fails here — you are just compressing a normal gas.

Case B — right at the boiling line (below $T_{c}$ ). Density jumps discontinuously (liquid ↔ gas). There is no smooth dial at all — you get a sudden bang, not a tunable knob. This is exactly the border scCO₂ avoids by living above the star.

Case C — exactly at the critical point. $κ_{T} \to \infty$ : the theoretical maximum sensitivity, but also where the fluid becomes cloudy and sluggish (critical opalescence). In practice you operate just past the star, not on it.

Case D — very far above the critical point (high $T$ and $P$ ). The spike has flattened; $ρ$ is high and stable but no longer super-tunable. Good dense solvent, poor dial. You have spent energy to leave the sweet spot.

The sweet spot for real processes (like decaffeination) is just above $(T_{c}, P_{c})$ : mild temperature protects flavour, and $κ_{T}$ is still huge so the dial is live.

PICTURE. The $P$ – $T$ map with four flags planted: A (dull gas), B (boiling jump), C (the spike, on the star), D (flattened, far past) — and a shaded "operate here" sweet-spot patch just NE of the star.

The one-picture summary

Everything on one canvas: the $P$ – $T$ map with the boiling line ending at the star $(31 ° C, 74 bar)$ ; the borderless supercritical region beyond it; the density-vs-pressure dial curve; and the $κ_{T}$ spike — all wired to the plain-English claim "tiny squeeze → huge density → strong, tunable, residue-free solvent."

Recall Feynman retelling — the whole walkthrough in plain words

Think of CO₂ molecules as a crowd. Density is how tightly packed the crowd is, and a tight crowd can grab and dissolve a greasy caffeine molecule. On a temperature-vs-squeeze map, CO₂ has three neighbourhoods — solid, liquid, gas — with fences between them. The fence between liquid and gas doesn't run forever: it stops at a spot called the critical point, $31 ° C$ and $74$ bar. Why does it stop? Because as you heat things up, the liquid loosens and the gas tightens until they're the same packing — no fence needed. If you sneak around that endpoint instead of jumping the fence, the crowd never suddenly boils; its packing changes smoothly. And right near that endpoint the crowd is so easy to squish that a gentle push crams in tons more molecules — the packing (and so the dissolving power) shoots up. The math word for "how much packing you get per push" is $κ_{T}$ , and at the endpoint it goes to infinity because pushing changes the volume almost not at all. That runaway sensitivity is the whole invention: near the critical point you dial solvent strength with the pressure knob, and when you're done you just let go — the CO₂ puffs away as gas, leaving zero residue.

Parent: Solvent selection — water, supercritical CO₂, ionic liquids (index 5.5.3) · Prerequisites: Phase Diagrams & Critical Point, Hydrophobic Effect · See also Principles of Green Chemistry, Catalysis & Catalyst Recovery, Life Cycle Assessment (LCA).