Hydrogels

A solid that is mostly water

Squeeze a contact lens and it feels like a tiny pane of soft rubber, yet it is almost 40 % water by weight. A diaper crystal the size of a grain of sugar swells into a wobbling bead a hundred times its size. Both are hydrogels: networks of long polymer chains tied together at scattered points, soaking in water they will not let go of and will not dissolve in.

The trick is the cross-link. A lone polymer chain dotted with water-loving groups — hydroxyls, amides, carboxylates — would simply dissolve, dispersing into solution like sugar. But if you stitch the chains together at intervals with permanent bridges, you convert millions of separate molecules into a single molecule that spans the entire sample. That giant molecule can stretch and drink water, but it can never come apart and drift away. The result sits in a strange middle ground between liquid and solid: it holds its shape like a solid yet transports small molecules almost as freely as the surrounding water.

Two numbers describe any hydrogel. The cross-link density sets how often the chains are tied — every few hundred monomers in a tight gel, every several thousand in a loose one. The swelling ratio Q reports how much it expands: the swollen mass divided by the dry mass. The two trade off against each other, and almost everything a hydrogel does follows from where on that trade-off it sits.

Why it swells, and why it stops

Swelling is a tug-of-war between two free-energy terms. The pull comes from mixing: water molecules are entropically and enthalpically happy to surround the hydrophilic groups, so water floods in and the network expands. As the chains stretch, the resistance comes from elasticity: a stretched polymer chain has fewer available conformations than a relaxed one, so its conformational entropy drops, generating a rubber-elastic retractive force that fights further expansion.

Paul Flory and John Rehner captured the balance in 1943. At swelling equilibrium the two contributions to the osmotic pressure cancel:

Π_mix + Π_elastic = 0

−[ ln(1−φ) + φ + χφ² ]  =  V₁·ν·( φ^(1/3) − φ/2 )

       mixing term                elastic term

Here φ is the polymer volume fraction in the swollen gel, χ is the Flory–Huggins polymer–solvent interaction parameter (small χ means water-loving), V₁ is the molar volume of water (18 cm³/mol), and ν is the cross-link density (moles of elastic chains per unit volume). Read it qualitatively and the whole behaviour falls out: make the polymer more water-loving (lower χ) and φ at equilibrium drops — it swells more. Add more cross-links (raise ν) and the elastic term grows, so the gel swells less.

For ionic gels there is a third, often dominant, term. Fixed charges on the chains — say carboxylate –COO⁻ groups after neutralisation — trap mobile counter-ions inside the network. Those trapped ions create a Donnan osmotic pressure that can dwarf the simple mixing term, which is exactly why sodium polyacrylate swells hundreds of times its weight while a neutral gel of similar density swells only a few times. It is also why salt kills the effect: dissolved Na⁺ and Cl⁻ outside the gel screen the fixed charges and erase the osmotic imbalance.

How the network gets built

There are two routes to the cross-links, and the choice defines almost every property that follows.

Chemical (covalent) gelation. The most common laboratory route is free-radical chain polymerisation of a vinyl monomer in the presence of a small fraction of a difunctional cross-linker. Acrylamide plus a few mol % N,N′-methylenebisacrylamide, kicked off by an ammonium persulfate / TEMED redox initiator, gives the polyacrylamide gels used for electrophoresis. The cross-linker has two double bonds, so it gets built into two growing chains at once and ties them together. The bridges are real C–C and C–N covalent bonds, roughly 350 kJ/mol, so the gel is permanent: it cannot be melted back into a liquid.

  n CH₂=CH–CONH₂   +   x CH₂=CH–CO–NH–CH₂–NH–CO–CH=CH₂
   (acrylamide)            (bis-acrylamide cross-linker)
            │  persulfate / TEMED, radical
            ▼
   ─[CH₂–CH]─[CH₂–CH]─[CH₂–CH]─   chain A
         │        │        │
        ...      cross    ...
                  │  (–CO–NH–CH₂–NH–CO–)
   ─[CH₂–CH]─[CH₂–CH]─[CH₂–CH]─   chain B

Physical (reversible) gelation. Here the junctions are not bonds but reversible associations — hydrogen-bonded zones, ionic bridges, or ordered helices. Drop calcium ions into a solution of alginate (a seaweed polysaccharide of guluronate blocks) and each Ca²⁺ slots between two carboxylate-rich chain segments, forming the famous "egg-box" junction; the whole thing gels in seconds at room temperature and dissolves again if you sequester the calcium with citrate. Gelatin gels on cooling because collagen-derived chains re-form short triple-helix segments, then melt near 35 °C — which is why a gelatin dessert collapses in your mouth. Each physical junction is worth only 5–25 kJ/mol, so physical gels are gentle, reversible, and biocompatible, at the cost of being weaker.

A comparison across the hydrogel zoo

Hydrogel	Cross-link type	Water content / swelling	Typical use
pHEMA (poly-2-hydroxyethyl methacrylate)	Covalent (EGDMA)	~38 % water, Q ≈ 1.6	Soft contact lenses
Sodium polyacrylate	Covalent, ionic charges	200–800 g/g in pure water; 30–60 g/g in saline	Diapers, agriculture
Polyacrylamide + bis-acrylamide	Covalent	85–95 % water	Gel electrophoresis
Alginate–Ca²⁺	Ionic "egg-box"	95–99 % water	Cell encapsulation, wound gel
Gelatin	Physical (triple helix)	>90 % water, melts ~35 °C	Food, capsules, scaffolds
PNIPAM	Covalent, thermo-responsive	Swollen below 32 °C, collapses above	Switchable drug depots
PEG-diacrylate	Covalent (photo-cured)	80–95 % water, tunable mesh	Tissue engineering, 3D printing

Notice how the same family of materials spans from a barely-wet 38 % contact lens to a 99.9 % water cell scaffold simply by dialling cross-link density and charge. The mesh size ξ — the average gap between cross-links in the swollen state — runs from about 5 nm in a tight pHEMA gel to over 100 nm in a loose PEG scaffold, and it is this mesh that gates what can diffuse in and out.

Gels that switch: the volume phase transition

Toyoichi Tanaka discovered in 1978 that a gel does not have to swell smoothly — it can collapse discontinuously, jumping between a swollen and a shrunken state at a critical condition, much like a liquid-gas transition. The most-used example is poly(N-isopropylacrylamide), PNIPAM, which has a lower critical solution temperature (LCST) near 32 °C.

Below the LCST, the amide groups hydrogen-bond to water and the chains stay solvated and swollen. Raise the temperature past 32 °C and those hydrogen bonds break faster than they form; now the hydrophobic isopropyl side groups dominate, water is expelled, and the gel collapses — often to a tenth of its volume — sometimes within seconds. The transition is driven by the entropy of the water released from the hydrophobic hydration shells, the same hydrophobic effect that folds proteins.

      T < 32 °C                         T > 32 °C
  ╭───────────────╮                ╭───╮
  │  swollen gel  │   ── heat ──▶  │ * │   collapsed
  │   ~95% water  │   ◀─ cool ──   ╰───╯   ~30% water
  ╰───────────────╯
  water H-bonds chains          hydrophobic groups win,
  → expanded                    water expelled → shrunk

Because 32 °C sits just under body temperature, a PNIPAM solution stays a low-viscosity liquid in a cold syringe and gels the instant it is injected at 37 °C, forming a depot in place. Swap the trigger and you get other smart gels: weakly acidic gels swell as pH rises and the –COOH groups ionise to –COO⁻ near their pKa of ~4.5; glucose-responsive gels carrying phenylboronic acid swell when blood sugar climbs, the basis for self-regulating insulin patches.

Release rates, mesh, and degradation

A drug loaded into a swollen gel can only leave by diffusing through the water-filled mesh, so the mesh size sets the clock. For a molecule much smaller than the mesh, escape follows Fickian diffusion and the cumulative released fraction grows with the square root of time:

M_t / M_∞  ≈  k · t^(1/2)        (early-time Fickian release)

Tighten the cross-links and the mesh shrinks; the diffusion coefficient inside the gel drops, sometimes by orders of magnitude, and a week-long depot becomes a month-long one. When the drug's hydrodynamic radius approaches the mesh radius, diffusion nearly halts and release is gated almost entirely by network swelling or erosion. Engineers exploit this with degradable cross-links — ester linkages that hydrolyse, or peptide bridges cleaved by enzymes — so the mesh slowly widens and release accelerates on a programmable schedule. A PEG gel cross-linked through hydrolysable ester groups can be tuned to fully dissolve anywhere from days to months by changing the number of esters per junction.

This is also how hydrogels feel like tissue. Their shear modulus, set by cross-link density through G ≈ νRT (rubber elasticity), typically lands between 0.1 and 100 kPa — overlapping fat, muscle, and cartilage. That softness is not cosmetic: a 1 kPa gel nudges mesenchymal stem cells toward neuron-like fates while a 30 kPa gel pushes them toward bone, so matrix stiffness alone can steer cell differentiation.

Where hydrogels show up

• Soft contact lenses. The original 1960 Wichterle–Lím pHEMA lens still defines the category; modern silicone-hydrogel lenses blend a hydrogel with siloxane segments to multiply oxygen permeability roughly six-fold so the cornea can breathe overnight.
• Superabsorbent diapers and agriculture. Lightly cross-linked sodium polyacrylate absorbs hundreds of times its weight in water; the same beads are tilled into soil to hold irrigation water through dry spells.
• Wound dressings. A hydrogel sheet keeps a wound bed moist — which measurably speeds re-epithelialisation — while letting oxygen through and blocking bacteria; many also carry silver ions or antibiotics that diffuse out slowly.
• Drug-delivery depots. Injectable PNIPAM or PEG gels form in place and release a payload over days to weeks; ocular and subcutaneous gel implants now stretch dosing intervals from daily to monthly.
• Tissue engineering and bioprinting. Alginate, gelatin-methacryloyl, and PEG gels serve as 3D scaffolds and printable "bioinks" that suspend living cells in a tissue-soft matrix.
• Biosensors and lab tools. Polyacrylamide slabs are the workhorse matrix of gel electrophoresis, sieving DNA and proteins by size through their tunable mesh.

Common misconceptions and pitfalls

• "A hydrogel is just a wet sponge." A sponge holds water in macroscopic pores by capillarity and squeezes dry under pressure. A hydrogel holds water at the molecular scale by osmosis and elasticity, and resists being squeezed dry because emptying it costs free energy — the water is thermodynamically bound to the network, not merely parked in holes.
• "More cross-linker always means a stronger gel." Up to a point. Past an optimum, extra cross-links make the network brittle and reduce swelling so much that the gel becomes glassy and cracks. Toughness usually peaks at a modest cross-link density, not the maximum.
• "Swelling capacity measured in pure water is what you get in practice." Salt, proteins, and pH wreck ionic swelling. A diaper polymer that drinks 500 g/g of deionised water manages only ~50 g/g of urine because dissolved ions screen the fixed charges and collapse the Donnan pressure.
• "Physical and chemical gels are interchangeable." A covalent gel is permanent and tough but cannot self-heal or be re-melted; a physical gel is reversible and gentle but creeps and can dissolve. Choosing the wrong one — say a meltable gelatin where you needed a permanent scaffold — fails in service.
• "Hydrogels are inert." Residual unreacted monomer (acrylamide is a neurotoxin) and initiator must be exhaustively washed out before any biomedical use. The cured network is benign; the leftover small molecules are not.
• "Bigger mesh always releases drug faster." Only while the drug is much smaller than the mesh. Once drug and mesh sizes are comparable, release is governed by network erosion or swelling kinetics, and a wider mesh on a slowly-degrading gel can still be slow.