Why Polysaccharides Hold Water

Advanced Chapter 4 treated stabilizers as materials that bind water. But it is a curious thing. Locust bean gum and guar gum go into a formulation at 0.2 to 0.4 percent at most. How can so little govern the behavior of several hundred times its own weight in water? The answer comes down to one point: a stabilizer is not merely a molecule but a polymer.

Hydration by hydrogen bond

A polysaccharide is a long chain of thousands to tens of thousands of single sugars joined together. Along its backbone stand countless hydroxyl groups (−OH). These −OH make hydrogen bonds with water molecules, drawing water in around the chain and holding it fast. This is hydration (idratazione).

The amount of water bound directly by hydrogen bond, though, cannot account for the dramatic thickening that so little stabilizer gives. Hydration is only the doorway. The heart of it is what shape the chain takes in solution.

The water a coil holds inside

A single polymer chain does not stretch out like a rod in solution. Turning freely about countless single bonds, the chain takes up a loosely folded random coil. What matters is that this coil occupies a space out of all proportion to its own mass. Water enters the inside of the coil, is caught by the chain, and loses its motion.

So a polysaccharide occupies a huge effective volume. Not only the water it hydrates directly, but all the water shut inside the coil with it. A slight mass governs a wide volume. That is what "a little goes a long way" really is.

This bulk can be caught in one number. Against the volume the chain itself occupies, how many times larger is the effective volume that has taken solvent in? This is the volume ratio R_V.

For a molecule gathered into a sphere, as many globular proteins are, the effective volume is little different from that of the chain. R_V stays around 1. For an extended chain the figure leaps. Polysaccharides used as thickeners run from about five thousand to two million in molecular weight. Against that, R_V spreads from around 1 to billions. Within the one word "polysaccharide", the bulk differs by orders of magnitude.

That thickening falls to polysaccharides rather than proteins comes down to this alone. Polysaccharides have a large molecular weight and an extended chain, so their R_V is high. And so they can be used at a far lower concentration.

What should be noted here is that what tells is not the molecular weight itself but the volume the chain occupies in solution. At the same molecular weight, a chain that runs straight gives a large R_V, and a branched one a small. There are indeed polysaccharides of very great molecular weight that, being branched and rolled up, give a low viscosity and behave like a Newtonian liquid. Molecular weight is a necessary condition and no more. Only an extended chain makes a thickener.

And one more thing stretches a chain: charge. Set charges of the same sign along a chain and they repel, and the chain is drawn out. Stretched, the R_V rises. Polysaccharides carrying sulfate or carboxyl groups tell in ways the count of functional groups alone will not explain, because charge changes the shape itself.

The critical point of "a little goes a long way"

Raise the concentration and at some place the viscosity leaps. In a thin solution the coils drift apart, each on its own. But once the concentration passes the overlap concentration c*, the spheres the chains sweep out begin to touch.

Where does this critical concentration fall? The R_V above settles it.

c* runs inverse to R_V. The bulkier the molecule, the lower the concentration at which it reaches the critical point. Where R_V runs to thousands or tens of thousands, c* becomes an extremely small figure. Why a formulation of 0.2 percent can govern several hundred times its weight in water is gathered into that one line of inverse proportion.

Raise the concentration further and it reaches the entanglement concentration. The chains run into one another and entangle, and can no longer slip past. This is why the viscosity climbs so steeply.

So which side is the stabilizer of a gelato on? Here is a device that runs through the whole process. The mix and the unfrozen phase are not the same place.

In the liquid mix, the stabilizer sits between the two boundaries. It has passed the overlap concentration but has not reached the entanglement concentration. The molecules act on one another while each is still surrounded, for the most part, by solvent — the region called semi-dilute. So the mix is still a liquid that flows. Freeze it, though, and the story changes. At the serving temperature some eight tenths of the water has been drawn off as ice (Advanced Chapter 7). In the unfrozen phase that remains, the solutes grow several times stronger. Nothing has been added to the stabilizer, and yet its concentration leaps. So the stabilizer is carried across to the far side of the entanglement concentration. And indeed, in the unfrozen phase of a finished gelato the stabilizer usually sits in that entangled region.

The 0.3 percent written on the formulation sheet is 0.3 percent in the mix, and nothing more. The concentration where the work is done is a different figure. Designing with stabilizers is the work of setting the number at the entrance with that journey in view.

What must "not be overdone" not cross?

Advanced Chapter 4 called stabilizers a material not to be overdone. So what is the line that must not be crossed? Not entanglement — entanglement, as seen above, is where the work happens. The danger lies beyond it.

Two kinds of sticking have to be told apart here. Entanglement is topological. The chains have merely run into one another and cannot move; they are not tied anywhere. So pull hard and the chains are drawn out of one another and come loose — and that is what shear thinning is. That the mix flows in the machine and softens in the mouth rests on this coming loose.

Where chains join chemically, though, everything changes. Let intermolecular forces act — hydrogen bonding among them — and the viscosity runs higher than entanglement alone would predict. And where those forces are strong enough, the chains are fixed into a cross-linked net. This is a gel. Cross-links do not move, so the chains cannot slip past one another. It does not flow under force, and returns to its shape when the force is taken away: not a liquid but an elastic solid.

Tangled thread comes loose when pulled. Knotted thread does not. That is the difference.

And awkwardly, a good many of the stabilizers used in gelato can form a gel given the conditions. Locust bean gum, pectin, sodium alginate, κ-carrageenan — every one of them can set a thing solid if it chooses. When Advanced Chapter 4 warned against overdoing it, it was not the entanglement concentration it meant. Stay in entanglement that comes loose, and do not step into cross-linking that does not — that is what those words meant.

Why heat is needed — a difference of backbone

Tab. 4-1 in Advanced Chapter 4 recorded that carob (locust bean gum) needs heat while guar gum hydrates in cold water. The difference can be accounted for mechanically, from the backbone.

Both are galactomannans — a mannose backbone with galactose side chains — but the density of the side chains differs. Guar carries a galactose side chain on nearly every other unit. The side chains get in the way, so the backbones cannot associate. Water penetrates easily and it hydrates in cold water. Carob has fewer side chains, and stretches of "bare", unsubstituted backbone remain. These smooth stretches associate with one another and keep water out, so breaking that association takes heat. The substituted parts stand up like a brush. The difference between the two is like that between a fine comb and one missing teeth.

So here it is whether or not the chains associate with themselves that settles the need for heat. That is what lies behind the "heat required" column of Tab. 4-1. It is not the only deciding factor, though. The sheer number of hydrophilic groups can tell as well. A polysaccharide rich in sulfate groups dissolves in cold water. One poor in them wants warm water, and one with almost none wants hotter still. How readily a thing hydrates is a balance of two factors: the number of hydrophilic groups, and the strength of association between chains.

Shear thinning, and holding ice crystals back

An entangled polysaccharide solution is not a Newtonian liquid. At rest the viscosity is high, but under a strong shear stress the tangles come loose and it flows more easily. Because of this shear thinning, a stabilized mix holds water firmly while standing and yet flows smoothly under the shear of the mantecazione. That machinery can handle a mix at all rests on this property (equipment is Advanced Chapter 8; viscosity in general is Approfondimento 10).

But "high viscosity at rest" is not a property for the machines alone. When a fat globule tries to rise slowly through the mix, the shear stress it puts on the liquid around it is very small. Which is to say that a particle trying to rise meets the high-viscosity end of the shear-thinning curve. So its rise is strongly held back. Shear thinning tells hardest when nothing is being moved. What keeps the fat from rising during aging is this asymmetry.

Polysaccharides do not always hold the rise back, though. At certain concentrations they can actually promote it. The literature puts this down to a separate mechanism, depletion flocculation, and this piece does not go into it. Here too, more is not better.

Shear thinning has a face in time as well. Keep shearing at a constant rate and the apparent viscosity falls with time — the behavior called thixotropy. Its molecular origin is put down to the same things as shear thinning: alignment of the chains, disentanglement, and the breaking of weak interactions. One mechanism shows two faces — a response to shear rate and a response to time. In a step like the mantecazione, where a steady shear runs for a long while, the side in time tells as well.

And then the question is whether the structure comes back once the shear is taken away. If the chains lose their alignment, entangle again, and associate anew with their neighbors, the system returns to its former viscosity and properties — reversibly. If not, irreversibly; if only in part, partly reversibly. The literature notes that the speed of that return can itself be the practical point. To choose a stabilizer is not to choose a height of viscosity but to choose how it comes loose and how it comes back.

And shear thinning works in the mouth too. A liquid that does not shear-thin is felt as slimy in the mouth. Smoothness, meanwhile, wants some viscosity. So what divides the mouthfeel is not the height of the viscosity but whether it comes loose under the shear stress of the mouth. A product with too much stabilizer feels gummy less because the viscosity is high than because it no longer comes loose. The "do not overdo it" of Advanced Chapter 4 held two meanings. One is physical: stopping short of cross-linking. The other is sensory: staying with what comes loose in the mouth.

And holding ice crystals back can be accounted for as viscosity as well. The higher the viscosity of the unfrozen phase, the slower water molecules diffuse to the crystal surface and let the crystal grow. The rate of Ostwald ripening runs in proportion to the diffusion coefficient (Approfondimento 3). So it follows as physics that raising the viscosity slows the coarsening.

But something has to be set down honestly here. Whether stabilizers actually work by that route is not settled.

That some stabilizers slow the recrystallization of ice has been confirmed. But the mechanism is not fully understood, and research continues. Suppression of diffusion by viscosity is a strong explanation, and it agrees with the physics followed here. But there is not yet evidence enough to call it the principal one. There are also accounts suggesting direct adsorption onto the crystal surface.

"Raise the viscosity and the ice coarsens less" is supported by experience and by physics alike. But why it should be so is not yet fully known. This is as far as a textbook can state.

Synergy — why a blend is the rule

Advanced Chapter 4 said that stabilizers are used as a blend of several rather than singly, as a matter of course. This too can be understood through the interaction of polymers. Put xanthan with locust bean gum, or carrageenan with locust bean gum. A viscoelasticity and a gel strength appear that exceed the sum of each alone. A particular region of one chain associates selectively with the other chain, forming a firmer net. That a small amount gives steady properties is because this synergy is being put to use.