Corrected exercises: states of matter and phase diagrams (202-SN1-RE)

By increasing strength: LONDON dispersion forces, then DIPOLE-DIPOLE interactions, then HYDROGEN BONDING. London forces, caused by the instantaneous dipoles of the electron cloud, are ALWAYS present, even between nonpolar molecules: they are the only forces that exist in a noble gas or in an alkane.

Because water forms HYDROGEN BONDS and hydrogen sulfide does not. Hydrogen bonding requires the hydrogen to be bonded to $N$, $O$ or $F$, small and very electronegative atoms. Sulfur is much less electronegative and much larger, so $H_{2}S$ has only ordinary dipole-dipole interactions. The 160 °C gap therefore directly measures the energy cost of the hydrogen-bond network of water, which is what keeps water liquid under terrestrial conditions, a requirement for all of biology.

Because neither $CH_{4}$ nor $SiH_{4}$ forms hydrogen bonds: carbon and silicon are not electronegative enough. Both molecules are practically nonpolar and have only London forces. These grow with the number of electrons and the POLARIZABILITY of the cloud: $SiH_{4}$, the larger molecule, has more of them, so it boils higher. Here mass and polarizability vary in the same direction, so there is no reason for a reversal.

Both are nonpolar and so have only London forces, but iodine has far more electrons than fluorine, in shells much farther from the nucleus and therefore easier to distort. Its cloud is much more POLARIZABLE, the instantaneous dipoles are stronger, and the London forces are multiplied. That is why difluorine is a gas at room temperature whereas diiodine is a SOLID: the same kind of force, at very different strengths, produces three different states.

Linear pentane boils at 36 °C, neopentane at 10 °C. Both have exactly the same formula, hence the same number of electrons and the same overall polarizability. What distinguishes them is the CONTACT AREA: an elongated molecule can lie against its neighbour along its whole length, multiplying the points of interaction, whereas a compact, nearly spherical molecule touches only over a small cap. The more BRANCHED a molecule is, the more compact it is, the smaller its contact area and the lower its boiling point. This shape argument, independent of mass, is regularly asked on exams.

The curve rises, shows a horizontal PLATEAU at the melting temperature, rises again, shows a second, longer plateau at the boiling temperature, then rises again. Each plateau corresponds to a phase change: the heat supplied does not raise the temperature, it is used entirely to overcome the intermolecular forces. The slope of each rising segment is inversely proportional to the heat capacity of the phase concerned.

The amount of substance is $n=\dfrac{100}{18.02}=5.549$ mol. Step 1, heating the ice from $-20$ to 0 °C: $100\times 2.09\times 20=4180$ J. Step 2, melting: $5.549\times 6010=33\,352$ J. Step 3, heating the water from 0 to 100 °C: $100\times 4.18\times 100=41\,800$ J. Step 4, vaporization: $5.549\times 40\,700=225\,860$ J. Step 5, heating the steam from 100 to 120 °C: $100\times 2.01\times 20=4020$ J. TOTAL: $309\,212$ J, or about 309 kJ.

VAPORIZATION dominates by far, with 225.9 kJ out of 309.2, or 73.0 % of the total on its own. The two steps that heat the condensed phases, although spread over 120 degrees in all, together weigh only about 15 %.

Melting only LOOSENS the network: the molecules stay in contact with one another, the average distance barely changes, and the density varies by only a few percent. Only the long-range order is destroyed. Vaporizing, on the other hand, completely SEPARATES the molecules: all the hydrogen bonds must be broken and each molecule moved to a distance where it no longer interacts, which multiplies the volume by about 1600. It is therefore logical that the second process costs nearly seven times the first.

Because the steam, condensing on the skin, first gives back its 40.7 kJ/mol of enthalpy of vaporization BEFORE it even starts to cool. For the same mass deposited at the same temperature, steam therefore releases about $\dfrac{40\,700}{18.02}=2258$ J/g more than liquid water, which, relative to the heat capacity of skin, is a considerable and almost instantaneous input of heat. That is why steam burns are deep, and why autoclaves and steam workshops require specific protection.

Vapour pressure is the pressure exerted by the vapour in dynamic EQUILIBRIUM with its liquid, in a closed container: as many molecules leave the liquid as return to it. It does not depend on the amount of liquid because the equilibrium is established at the SURFACE: doubling the volume of liquid doubles the number of departures but also the number of returns, and the equilibrium pressure is unchanged. It depends only on temperature, which fixes the proportion of molecules energetic enough to escape.

Boiling occurs when the vapour pressure EQUALS the external pressure. Below that threshold, a vapour bubble formed inside the liquid would be crushed immediately; at equality, it can persist and rise. Since atmospheric pressure decreases with altitude, the required vapour pressure is lower, and it is reached at a LOWER temperature: water boils colder in the mountains.

Clausius-Clapeyron with $P_{1}=101.3$ kPa and $T_{1}=373.15$ K: $\ln\dfrac{34.0}{101.3}=-\dfrac{40\,700}{8.314}\left(\dfrac{1}{T_{2}}-\dfrac{1}{373.15}\right)$. The left side equals $-1.0917$, and $\dfrac{\Delta H}{R}=4895$. We get $\dfrac{1}{T_{2}}=\dfrac{1}{373.15}+\dfrac{1.0917}{4895}=2.9029\times 10^{-3}$, so $T_{2}=344.5$ K, or $71.3$ °C.

With $P_{2}=202.6$ kPa, the same calculation gives $T_{2}=394.0$ K, or $120.8$ °C. The pressure cooker therefore cooks at 121 °C instead of 100. The time saved comes from KINETICS, not thermodynamics: by the rule of thumb that reaction rates double for every 10 degrees, a rise of 21 degrees multiplies the rate of the cooking reactions by about four. It is also the standard sterilization temperature in an autoclave, chosen precisely because it destroys bacterial spores.

Because at 71 °C water boils but hardly cooks anything: the cooking reactions are too slow there, an egg does not set properly and food soaks up water instead of cooking. For frying, the problem shifts: the oil, whose vapour pressure remains negligible, easily reaches its usual 180 °C, but the water contained IN the food boils from 71 °C. It therefore escapes earlier and more violently, and the food dries out on the surface before the inside is cooked. The practical solution is the pressure cooker, which restores a normal pressure, or longer cooking times.

The three regions are the SOLID, at upper left, at low temperatures and high pressures; the LIQUID, in the centre and at upper right; and the GAS, at lower right, at high temperatures and low pressures. The three curves are the MELTING curve, between solid and liquid; the VAPORIZATION curve, between liquid and gas; and the SUBLIMATION curve, between solid and gas, below the triple point. Each curve is a locus of EQUILIBRIUM: the two phases coexist on it.

The TRIPLE POINT is the one and only combination of temperature and pressure at which the THREE phases coexist in equilibrium. For water it is 0.01 °C and 0.611 kPa; it is so reproducible that until 2019 it was used to define the kelvin. The CRITICAL POINT ends the vaporization curve: beyond it, the distinction between liquid and gas disappears and we speak of a SUPERCRITICAL fluid. Note that the melting curve never ends: a substance can always be solidified by compressing it enough.

The triple point of $CO_{2}$ is at 518 kPa, or about 5.1 atmospheres, so CLEARLY ABOVE the atmospheric pressure of 101.3 kPa. In the open air we are therefore below the triple point, in the region where the sublimation curve directly separates solid from gas: there is no temperature, at 1 atm, at which liquid $CO_{2}$ is stable. The solid therefore goes directly to the gas state, hence the name DRY ice, and hence its use for cooling without wetting. For water, on the contrary, the triple point is at 0.611 kPa, far below 1 atm, so the liquid exists under ordinary conditions.

A negative slope means that INCREASING the pressure at constant temperature takes us from solid to liquid: compression makes ice MELT. By Le Chatelier's principle, a compressed system shifts toward the DENSER phase; if that is the liquid, then ice is LESS dense than liquid water. This is the anomaly of water, a direct consequence of the hydrogen-bond network of ice, which is open and leaves large hexagonal voids. Almost every other substance, $CO_{2}$ included, has a melting curve with a positive slope.

Beyond the critical point, only a single FLUID phase remains, neither quite liquid nor quite gas. When a liquid is heated in a closed container, the density of the liquid decreases and that of the vapour increases, until they become EQUAL at the critical point: the meniscus that separated the two phases fades and disappears. Beyond it, there is no longer any interface, no surface tension, and no experiment can tell which side we are on. The supercritical fluid then combines the density of a liquid, hence good solvent power, with the viscosity of a gas, hence easy penetration.

A molecule in the bulk of the liquid is surrounded by neighbours in every direction and the forces it feels cancel. A molecule at the surface has neighbours on one side only: it feels a net force directed inward. Enlarging the surface therefore means bringing molecules up against this force, which costs energy, and the liquid spontaneously adopts the shape of smallest surface: the sphere. Surface tension measures this cost per unit area, so it directly follows the strength of the intermolecular forces. Hexane has only weak London forces: 18.4 mN/m. Ethanol adds one hydrogen bond per molecule, but its long chain dilutes it: 22.3 mN/m. Water, a small molecule able to form four hydrogen bonds, reaches 72.8 mN/m, a remarkably high value for a molecular liquid. Mercury, at 487 mN/m, is in another category because what acts is no longer intermolecular forces but metallic bonding itself, where the atoms share a sea of electrons. That is why spilled mercury forms perfect beads.

$h=\dfrac{2\times 72.8\times 10^{-3}\times 1}{1000\times 9.81\times 2.00\times 10^{-4}}=\dfrac{0.1456}{1.962}=0.0742$ m, or 7.4 cm. Two forces compete in this phenomenon: cohesion, which holds the molecules of the liquid together, and adhesion, which attracts them to the wall. For water and glass, which is rich in $Si-OH$ groups capable of hydrogen bonding, adhesion wins: the liquid spreads on the wall, the meniscus is concave and the column rises. For mercury, metallic cohesion is overwhelming and adhesion to glass is negligible: the liquid shuns the wall, the meniscus is convex, the contact angle exceeds 90°, its cosine becomes negative and the formula then predicts a DROP. That is indeed what is observed, and it is also why a mercury barometer is read at the top of the meniscus and a water burette at its bottom.

Viscosity measures how hard it is for molecules to slide past one another, hence the number and strength of the catches between them. Glycerol carries three hydroxyl groups per molecule: each molecule is held by a dense network of hydrogen bonds and also catches by its shape, hence a factor of 1410 relative to water, whose molecules have far fewer such contacts. Hexane has only London forces and flows three times as easily as water. When the temperature rises, thermal motion gives the molecules the energy needed to break these catches more often and more quickly: the viscosity drops, often dramatically. This is the everyday experience of warming honey or oil. Surface tension also decreases with temperature, for the same underlying reason, but much more slowly.

A paper clip does not float through buoyancy, which would be far too small since steel is eight times as dense as water: it is held up by the surface itself, which behaves like an elastic membrane and whose deformation exerts a vertical force large enough to balance its weight. It must therefore be laid flat and without piercing the surface. Detergent is a surfactant: its amphiphilic molecules insert themselves between the water molecules at the surface, where their nonpolar tail replaces hydrogen bonds with mere London forces. The surface tension collapses from 72.8 to about twenty mN/m, the membrane supports nothing any more, and the paper clip sinks immediately. Biological manifestation: water striders walk on water by the same mechanism, and conversely our lung alveoli would be collapsed by surface tension if they did not secrete a surfactant, whose absence in a very premature baby causes respiratory distress. Technical manifestation: capillary action draws sap up the thin vessels of plants, fills a lamp wick, and is also what makes moisture climb walls through the pores of concrete.

For the same mass, the volume is inversely proportional to the density. The volume ratio is $\dfrac{V_{ice}}{V_{water}}=\dfrac{0.9998}{0.917}=1.0903$: ice therefore occupies about 9.0 % MORE volume than the water it came from.

In liquid water, each molecule forms on average 3.4 hydrogen bonds, but they break and re-form constantly, which lets the molecules pack into the gaps. On freezing, each molecule engages exactly FOUR hydrogen bonds, directed toward the corners of a tetrahedron, since oxygen is $sp^{3}$ with two lone pairs and two hydrogens. This geometric constraint imposes an OPEN hexagonal network, pierced by empty channels. Ice is therefore more ordered AND less compact: it is the directionality of the hydrogen bond, not its strength, that produces the anomaly.

The equilibrium is written ice $\rightleftharpoons$ liquid water. Increasing the pressure pushes the system toward the phase of smaller molar volume, hence higher density: here the LIQUID. Compressing ice at constant temperature therefore makes it melt, which means that the melting temperature DECREASES as the pressure increases. That is exactly a negative slope on the diagram. For an ordinary substance, whose solid is denser, the reasoning is reversed and the slope is positive.

Biological consequence: ice FLOATS, since it is less dense. A lake therefore freezes from the top down, and the layer of ice thermally insulates the water below, which stays liquid all winter. Fish and other organisms survive there. If ice sank, lakes and oceans would freeze from the bottom up and never thaw completely, making aquatic life at temperate and polar latitudes practically impossible. Material consequence: water that freezes in a crack in rock or in a pipe expands by 9 % and exerts considerable pressure, bursting the pipe or fracturing the stone. This is the mechanism of frost wedging, the main driver of erosion in Quebec, and the reason outdoor pipes are drained before winter.

This imposes a particular STRATIFICATION on the lake. As it cools, the surface water becomes denser and sinks, which mixes the lake. But this mixing stops at 4 °C: below that, water that keeps cooling becomes LESS dense and stays at the surface. The bottom of the lake therefore settles at 4 °C, its densest temperature, while the surface keeps cooling to 0 °C and freezes. The result is a lake whose bottom stays at 4 °C all winter, a perfectly livable temperature, under an insulating layer of ice. Without this density maximum at 4 °C, the mixing would continue until freezing and the lake would freeze through its whole depth.

Molecular solid: the sites are occupied by whole molecules, held together by intermolecular forces: London, dipole-dipole or hydrogen bonds. Ice and iodine belong here. Ionic solid: the sites hold alternating ions, held by the very strong electrostatic attraction of the lattice; sodium chloride and magnesium oxide belong here. Covalent network solid, also called macromolecular: there are no molecules, the whole crystal is a single network of covalent bonds; diamond and quartz belong here. Metallic solid: cations occupy the sites and are bathed in a sea of delocalized electrons; sodium and tungsten belong here.

Melting point. Molecular solids melt low, from 0 to 114 °C here, because it is enough to overcome intermolecular forces without touching the internal covalent bonds. Ionic solids melt high, 801 and 2852 °C, the gap between the two coming from the double charges of magnesium oxide. Covalent network solids melt highest, because melting means breaking real covalent bonds. Metals cover the whole range, from 98 °C for sodium to 3422 °C for tungsten, depending on the number of valence electrons contributed to the electron sea: this shows that metallic bonding has no fixed strength. Hardness. Covalent network solids are the hardest; ionic solids are hard but brittle, because a slight slip brings ions of like charge face to face and the crystal cleaves cleanly; metals are ductile and malleable, since layers of cations can slide without ever breaking the electron sea; molecular solids are soft. Electrical conductivity. Only metals conduct in the solid state, thanks to their free electrons. Ionic solids conduct once molten or dissolved; covalent network and molecular solids are insulators.

In solid sodium chloride, the ions do exist but they are locked in place in the lattice: no charge can move, so there is no current. Melting destroys the order of the lattice while keeping the ions, which are then free to migrate toward the electrodes: the liquid conducts. In metallic sodium, it is not the cations that carry the current but the electrons of the electron sea, and these are delocalized and mobile from the solid state on; melting therefore changes nothing about their mobility. The important conclusion is that the charge carriers are not of the same kind: they are IONS in an ionic conductor and ELECTRONS in a metallic conductor. This distinction underlies all of electrochemistry, where the circuit has an electronic stretch in the wires and an ionic stretch in the solution.

In diamond, each carbon is $sp^{3}$ hybridized and bonded to four neighbours by strong covalent bonds, in all three directions of space: the whole crystal is a single rigid molecule, hence its extreme hardness and its melting point of 3550 °C. All the valence electrons are engaged in localized bonds, none is left mobile, and diamond is an insulator. In graphite, each carbon is $sp^{2}$ hybridized and bonded to only three neighbours, forming flat sheets of hexagons. The fourth $p$ orbital of each carbon stays perpendicular to the sheet, and all these orbitals combine into a $\pi$ system delocalized over the whole plane: the electrons move freely there, so graphite conducts, but in the plane only. Between the sheets there are only very weak London forces: the planes slide over one another under the slightest pressure, which explains pencil lead and the use as a lubricant. Graphite is therefore a hybrid case, covalent within the plane and molecular between planes, and it is the best example that the classification into four families is a filing tool, not a law of nature.

Simple cubic: 8 corners each counting for $\dfrac{1}{8}$, so $8\times\dfrac{1}{8}=1$ atom per unit cell, and each atom touches its 6 neighbours along the three axes, hence a coordination number of 6. Body-centred cubic: $8\times\dfrac{1}{8}+1=2$ atoms per unit cell, and the central atom touches the 8 corners, hence a coordination number of 8. Face-centred cubic: $8\times\dfrac{1}{8}+6\times\dfrac{1}{2}=1+3=4$ atoms per unit cell, with a coordination number of 12.

Simple cubic: the spheres touch along the edge, which holds two radii, so $a=2r$. Body-centred cubic: they do not touch along the edge but along the BODY DIAGONAL of the cube, which passes through a corner atom, the central atom and the opposite corner atom, that is, four radii. Since this diagonal measures $a\sqrt{3}$, we have $a\sqrt{3}=4r$, so $a=\dfrac{4r}{\sqrt{3}}$. Face-centred cubic: contact occurs along the diagonal of a FACE, which also holds four radii; since this diagonal measures $a\sqrt{2}$, we get $a\sqrt{2}=4r$, so $a=2\sqrt{2}\,r$. Identifying the right contact direction is the decisive point of the whole exercise: choosing the wrong diagonal spoils everything that follows.

Simple cubic: $\tau=\dfrac{1\times\frac{4}{3}\pi r^{3}}{(2r)^{3}}=\dfrac{4\pi}{3\times 8}=\dfrac{\pi}{6}=0.524$, or 52.4 %. Body-centred cubic: $\tau=\dfrac{2\times\frac{4}{3}\pi r^{3}}{\left(\frac{4r}{\sqrt{3}}\right)^{3}}=\dfrac{\frac{8}{3}\pi r^{3}}{\frac{64r^{3}}{3\sqrt{3}}}=\dfrac{\pi\sqrt{3}}{8}=0.680$, or 68.0 %. Face-centred cubic: $\tau=\dfrac{4\times\frac{4}{3}\pi r^{3}}{\left(2\sqrt{2}r\right)^{3}}=\dfrac{\frac{16}{3}\pi r^{3}}{16\sqrt{2}r^{3}}=\dfrac{\pi}{3\sqrt{2}}=0.740$, or 74.0 %. The radius cancels each time, which was to be expected: packing efficiency is a purely geometric property of the arrangement, and depends neither on the element nor on the scale.

By increasing packing efficiency: simple cubic 52.4 %, body-centred cubic 68.0 %, face-centred cubic 74.0 %. This last value is no accident: 74.05 % is the absolute maximum that can be reached when stacking identical spheres, a result conjectured by Kepler in 1611 and proved only at the end of the twentieth century. The face-centred cubic structure, like the hexagonal close-packed structure that reaches the same value, thus achieves the densest possible packing, which explains why so many metals adopt it: copper, aluminum, silver, gold, lead. The simple cubic structure, on the other hand, leaves nearly half the volume empty and offers only 6 neighbours to each atom, hence few metallic bonds per atom: it is energetically poor and occurs in nature only in polonium. Body-centred cubic, at 68 %, remains very common, notably for iron, chromium and tungsten, because maximum packing is not the only criterion and the nature of the metallic bonding also plays a part.

The face-centred cubic unit cell contains $Z=4$ atoms. The edge is $a=361.5$ pm $=3.615\times 10^{-8}$ cm, so $a^{3}=4.724\times 10^{-23}$ cm³. Then $\rho=\dfrac{4\times 63.55}{6.022\times 10^{23}\times 4.724\times 10^{-23}}=\dfrac{254.2}{28.45}=8.94$ g/cm³. The deviation from the measured value of 8.96 g/cm³ is only 0.3 %, which validates at once the unit cell model, the value of the edge taken from diffraction, and the Avogadro constant. Such good agreement between a macroscopic measurement, weighing a piece of copper, and a measurement at the atomic scale is far from obvious: it is one of the finest confirmations of the reality of atoms.

In a face-centred cubic unit cell, the atoms touch along the diagonal of a face: $4r=a\sqrt{2}$, so $r=\dfrac{361.5\times 1.4142}{4}=127.8$ pm, to be compared with the tabulated value of 128 pm. For iron, the unit cell is body-CENTRED cubic, so $Z=2$ and contact occurs along the body diagonal: $4r=a\sqrt{3}$. Density: $a=2.866\times 10^{-8}$ cm, $a^{3}=2.354\times 10^{-23}$ cm³, hence $\rho=\dfrac{2\times 55.85}{6.022\times 10^{23}\times 2.354\times 10^{-23}}=\dfrac{111.7}{14.18}=7.88$ g/cm³, in excellent agreement with the measured 7.87 g/cm³. Radius: $r=\dfrac{286.6\times 1.7321}{4}=124.1$ pm. Using $Z=4$ or the face diagonal for iron would have given results wrong by more than 40 %: the type of unit cell governs the whole calculation.

Isolate $N_{A}=\dfrac{Z\,M}{\rho\,a^{3}}=\dfrac{4\times 63.55}{8.96\times 4.724\times 10^{-23}}=\dfrac{254.2}{4.233\times 10^{-22}}=6.005\times 10^{23}$ /mol. The deviation from the accepted value is $\dfrac{6.022-6.005}{6.022}=0.3\%$, which is remarkable for a method that requires only three quantities: a molar mass, a density and a unit cell edge. This is indeed how Avogadro's number was long determined with the best precision, and a refined version of this method, applied to a nearly perfect sphere of silicon, was used to fix the exact value of $N_{A}$ at the 2019 redefinition of the kilogram.

Three factors are involved: the number of atoms per unit cell $Z$, the molar mass $M$ and the volume $a^{3}$. Iron has a somewhat lower molar mass, 55.85 against 63.55, and a smaller unit cell volume, which would favour a higher density. But its body-centred cubic unit cell contains only 2 atoms against 4 for copper: it is this factor of 2 on $Z$ that dominates, and it is not offset by the ratio of the volumes, which is only 2.01 in favour of copper. In short, iron is less dense because its packing is less compact, 68 % against 74 %, more than because of its atomic mass. To apply the method of c) to any element, three independent measurements are needed: the density by weighing and volume measurement, the unit cell edge by X-ray diffraction, and the type of unit cell, which is also deduced from the diffraction pattern since certain families of planes are extinguished depending on the lattice type. The molar mass is read from the periodic table.

You must exceed SIMULTANEOUSLY the critical temperature, 31.0 °C, and the critical pressure, 7.38 MPa, or about 73 atmospheres. The temperature is remarkably low, barely above ambient, which is the great asset of $CO_{2}$: no significant heating is required. The pressure, on the other hand, requires strong equipment, but 73 bar is perfectly common in chemical engineering. It is the combination of the two that makes the process industrially viable, where a fluid with a critical temperature of 300 °C would be unusable on a food product.

A supercritical fluid has a density close to that of a LIQUID, hence good solvent power, and a viscosity and diffusivity close to those of a GAS, so it penetrates the coffee beans as a gas would. Decisive advantage: at the end of the extraction, you simply RELEASE THE PRESSURE and the $CO_{2}$ becomes a gas again and evaporates completely, leaving a product with no solvent residue at all. An organic solvent such as dichloromethane always leaves traces and raises toxicity and reprocessing problems. $CO_{2}$ is also nontoxic, nonflammable and recyclable in a closed loop.

The path has two stages. First a FREEZING at ordinary pressure: we move vertically down in temperature into the solid region, typically to $-40$ °C. Then we LOWER THE PRESSURE sharply, below 0.611 kPa, which takes us below the triple point, and we heat very gently: the representative point then crosses the SUBLIMATION curve, and the ice goes directly to vapour, which is pumped away and recondensed elsewhere. The liquid region is thus bypassed entirely.

The pressure must stay below that of the TRIPLE POINT, 0.611 kPa, throughout the sublimation phase. Above it, the curve met on heating would be the MELTING curve and not the sublimation curve: the ice would melt, the product would end up soaked, and we would have ordinary drying, with all the degradation it brings. This is the critical parameter of every freeze-dryer.

Because drying with heat imposes two aggressions that freeze-drying avoids. First HEAT, which denatures proteins: a vaccine is precisely a protein or a biological fragment whose conformation carries the activity, and heating it destroys it. Then the passage through the LIQUID state, which causes the porous structure to shrink and collapse as it dries, and the solutes to migrate to the surface. Sublimation, on the other hand, removes the ice where it sits and leaves an intact porous skeleton in place: the product keeps its shape, volume and structure, which explains why a freeze-dried coffee crystal redissolves instantly and a freeze-dried vaccine is reconstituted in a few seconds with sterile water, while keeping for years without strict cold-chain storage.