A sheet with one freedom

A flat sheet of paper has an enormous number of ways to move. Crease it into a Miura pattern and it has one.

That reduction — from a floppy sheet to a single-degree-of-freedom mechanism — is the property everything else depends on, and it is worth understanding why it happens.

Where the freedom goes #

A degree-four vertex has four fold angles. They are not independent: fix one and the other three follow, because the four facets have to close up around the vertex.

So one vertex is a one-degree-of-freedom mechanism. Two adjacent vertices share a crease, and therefore share that crease’s fold angle — so the second vertex’s single freedom is consumed by the first vertex’s choice.

Chain that across the sheet and every vertex is determined by its neighbour, which is determined by its neighbour, back to one number. The whole tessellation has one degree of freedom however large it is.

This only works because the vertices are identical. Mixed vertices demand different values for the shared crease, the demands conflict, and the sheet locks.

Both dimensions shrink together #

The consequence is the property the pattern is famous for.

Take the folded sheet and pull it open along its length. An ordinary material would get narrower across — that is what a positive Poisson’s ratio means, and almost everything has one. The Miura gets wider.

$$\nu = -\frac{\varepsilon_{\text{transverse}}}{\varepsilon_{\text{axial}}} < 0.$$

The reason is visible in the motion. The zig-zag in each row straightens as the sheet opens, which lengthens it; and the rows separate at the same time, which widens it. Both happen from the same single parameter, and both go the same way.

Materials with negative Poisson’s ratio are called auxetic and are rare. Here it comes not from the paper but from the fold pattern, which is why the term metamaterial fits: the behaviour belongs to the geometry.

The packing ratio #

The engineering reason to care is that both dimensions shrinking together means the area falls fast.

A material that contracts in one direction and expands in the other keeps roughly constant area. One that contracts in both loses area quadratically, and that is what a deployable structure wants: large in use, small in transit, with the transition controlled by one input.

The numbers matter. A Miura array packs to a few per cent of its deployed area, and it does so along one path with no possibility of jamming, because there is only one way for it to move.

Reliability is the real product #

That last point deserves emphasis, because it is why the pattern reached orbit rather than merely a paper.

A deployable structure with many degrees of freedom can deploy incorrectly. It can bind, it can deploy partially, it can find a configuration nobody anticipated. Every additional freedom is a way for the mechanism to be wrong, and for a satellite there is no second attempt.

A one-degree-of-freedom mechanism has one path. Push it and it goes along that path or it does not go at all. There is no partial state that is not on the way to the full one, and no configuration to get stuck in.

That is worth more than the packing ratio. Several folding schemes pack better; almost none deploy as predictably.

The kinematics, briefly #

The motion has a closed form, and having it is what lets the figures be computed rather than posed.

Parameterise by the dihedral angle of one crease. For a Miura with parallelogram angle $\alpha$, the in-plane dimensions of the folded sheet are

$$\text{pitch}_x \propto \sin\alpha \cos\theta, \qquad \text{pitch}_y \propto \sqrt{1 - \sin^2\alpha\sin^2\theta},$$

with $\theta$ running from zero at flat to a right angle at fully packed, and the out-of-plane amplitude growing as $\sin\alpha \sin\theta$.

Both pitches fall as $\theta$ increases, which is the negative Poisson’s ratio in algebra. The figures on this page evaluate exactly these expressions.

The ratio, in closed form #

The two pitches are enough to write the Poisson’s ratio down, which is worth doing because this site measures it numerically elsewhere and the expression says where the numbers come from.

Write $s = \sin\alpha$. Differentiating the logarithms of the two pitches with respect to $\theta$ and taking the ratio:

$$\nu \;=\; -\,\frac{s^{2}\cos^{2}\theta}{1 - s^{2}\sin^{2}\theta}$$

Negative throughout, as the prose says, and now with a size attached.

At the flat state, $\theta = 0$, it is exactly $-\sin^{2}\alpha$. At the fully packed state, $\theta = 90°$, it is zero. So the auxetic effect is strongest where the sheet is deployed and vanishes as it closes, which is the opposite of the intuition that a more folded sheet is a more folded sheet in every respect.

Which fixes the parallelogram angle’s job #

That gives the design decision a formula rather than a direction.

A shallow angle of twenty degrees gives $\nu = -0.12$ at the flat state; forty-five gives $-0.5$; sixty gives $-0.75$. The parallelogram angle sets the auxetic effect through its sine squared, so the effect quadruples between twenty degrees and forty-five and only doubles again from there.

That is the trade the essay describes, priced. A designer moving from a shallow to a strong angle is buying a factor of four in the ratio and paying for it in facet shape and in stacking — and the return diminishes sharply past forty-five degrees, which is where the packing argument starts winning.

And it agrees with the reciprocal rule #

One check falls out for free and it ties this expression to a measurement made elsewhere.

A sheet with a single degree of freedom has both dimensions as functions of one parameter, so the ratio measured the other way round is the reciprocal of this one — and their product is exactly one. At a forty-five degree angle in the flat state that is $-0.5$ one way and $-2$ the other.

So the number quoted for a Miura is meaningless without saying which pair of dimensions it is across, and the expression above says which: the $x$ pitch on the bottom, the $y$ pitch on top. Reported the other way it is $-(1 - s^{2}\sin^{2}\theta)/(s^{2}\cos^{2}\theta)$, which is never smaller than one in magnitude.

What the parameter buys #

The parallelogram angle is free, and choosing it is the main design decision.

A shallow angle gives a sheet that collapses strongly in one direction and weakly in the other — closer to an accordion, with a modest packing ratio and a mild auxetic effect.

A strong angle gives strong collapse in both directions, a much better packing ratio, and long thin facets that stack awkwardly.

Between them the designer trades packing against the thickness problem, and the optimum depends entirely on the material. A Mylar membrane can take a strong angle; a panelled array in aluminium cannot.

Where the freedom actually goes #

The claim that a whole sheet has one degree of freedom deserves a more careful statement, because as usually phrased it is slightly too strong.

A single degree-four vertex has one freedom, counted as four fold angles minus three closure conditions. Adjacent vertices share creases, so they share fold angles, and the constraint count grows with the pattern.

For a general tessellation the shared constraints would over-determine the system and the sheet would be rigid — no motion at all. The Miura escapes because its vertices are identical and periodically arranged, so the constraints they impose on shared creases are consistent by symmetry rather than by accident.

That is a delicate property. Perturb one vertex’s angles slightly and the consistency fails, the constraint system becomes over-determined, and the sheet locks. A single defect in the middle of a Miura sheet stops the whole thing moving.

Real sheets do not lock completely, because the facets bend a little. But the stiffness rises sharply, and a Miura array with a manufacturing error in one panel deploys badly or not at all.

Reading the packing curve #

The area curve on this page falls faster than either dimension, and it is worth extracting what that means for a designer.

At half-closed the sheet is about 70% of its flat width and 70% of its flat height, so about half its flat area. At nearly closed it is a few per cent of each and a fraction of a per cent of the area.

The consequence is that most of the packing benefit arrives at the very end of the motion. A deployment that stops at 90% has recovered nearly all of the area; a packing that only reaches 90% of full closure has achieved very little of the volume saving.

That asymmetry matters in practice. Deployment can be sloppy and still work; stowage cannot, and the last few per cent of closure is where the thickness of the stack becomes the binding constraint.

Bistability, and the patterns that have it #

The Miura is not bistable: it moves smoothly along its path with no preferred stopping point, which is what a deployable wants.

Other tessellations are different. The square twist has a snap — an energy barrier partway through its motion, so it sits stably open or stably closed and resists the middle.

That difference is exploited deliberately. A bistable pattern makes a switch, a latch or an energy absorber; a monostable one makes a deployable. Both come from tiling a vertex, and the distinction is in the vertex.

What it is not #

Two clarifications, because the claims about origami metamaterials are often overstated.

One degree of freedom is a statement about the ideal mechanism. Real paper bends, real panels flex, and a real Miura sheet has many low-stiffness modes in addition to the intended one. It is one degree of freedom that is much softer than the others, not the only one.

Negative Poisson’s ratio is a property of the folded configuration. It is not a property of the material at all scales — a Miura sheet is not an auxetic solid, it is a mechanism whose kinematics happen to have that sign.

Where the model stops #

Rigid facets. The kinematics assume the parallelograms do not deform. That is a good approximation for stiff panels and a poor one for paper, which absorbs a good deal of the motion by bending.

Zero thickness. The single degree of freedom is exact for a zero-thickness sheet. Real thickness breaks it, and the whole field of thickness accommodation exists to restore it.

Infinite and periodic. The properties are those of the infinite tiling. A finite panel has edges, and edges do not obey the interior kinematics.

No dynamics. Everything here is quasi-static — the sequence of shapes, with no account of the forces or the time taken. A deployment is a dynamic event and this says nothing about it.

The figure’s projection is schematic. The folded states are computed from the closed forms above and then drawn with a simple painter’s algorithm. The positions are right; the shading and the hidden-surface treatment are approximate.

What a negative Poisson’s ratio is worth #

The property is often quoted as though it were self-evidently useful, and it is worth being concrete about what it does.

Packing. Both dimensions contract together, so the area falls as the product rather than staying roughly constant. That is the deployable application and it is the obvious one.

Conforming to a curved surface. An ordinary sheet wrapped over a dome has to be cut, because stretching it one way narrows it the other and it will not lie flat. An auxetic sheet widens as it stretches, so it conforms. That is why auxetic textiles are studied for wound dressings and for composite layup over curved moulds.

Indentation resistance. Press a point into an auxetic material and the surrounding material flows toward the indenter rather than away, so it resists locally. Conventional foams do the opposite.

Synclastic bending. Bend an auxetic sheet one way and it curves the same way in the other direction, forming a dome rather than a saddle. That is the opposite of what an ordinary sheet does and it matters for anything that has to bend into a shell.

Not all of those apply to a Miura sheet, which is a mechanism rather than a bulk material. But they are the reasons the property is pursued, and folded sheets are among the cleanest ways to obtain it.

The defect problem #

A property worth flagging because it constrains manufacturing more than anything else.

The single degree of freedom depends on every vertex being identical. A sheet with one vertex out of place — a manufacturing error, a damaged panel, a hinge with the wrong angle — has a constraint system that no longer resolves consistently, and the mechanism stiffens sharply in the region around the defect.

For paper this is a nuisance; the sheet bends a little more and carries on. For a panelled structure it can be fatal, because rigid panels have no compliance to absorb the inconsistency and the mechanism binds.

That is why tolerances on folded deployables are tight in an unusual way: it is not the absolute accuracy of any panel that matters but the uniformity across the array. A sheet where every panel is one per cent oversized deploys perfectly; one where a single panel is one per cent oversized may not deploy at all.

Stiffness, and what “one freedom” means physically #

The kinematic statement is that the sheet has one degree of freedom. The physical statement is softer and more useful.

A real Miura sheet has enormously many degrees of freedom, because the facets can bend. What distinguishes the folding mode is that it costs almost nothing — the motion happens at the creases, which have very low stiffness — while every other mode requires bending a facet, which costs the bending stiffness of the material.

The ratio between those is large. For thin sheet the folding mode may be thousands of times softer than the nearest facet-bending mode, and to any practical purpose the sheet moves along its one path.

That is what makes the idealisation useful and also what limits it. Load a Miura sheet across the fold direction and it is stiff; load it along the fold direction and it collapses; and the ratio of those two stiffnesses is the property that engineering actually specifies.

So “one degree of freedom” should be read as “one very soft mode”, and the design question is how soft it is compared with everything else — which depends on the material and is invisible to the geometry.

The motion is the product #

It is worth separating two things that get conflated when this pattern is described.

The folded state is a compact stack. That is what the packing ratio measures and it is what fits in the rocket.

The motion is a single continuous path from flat to folded with no branches. That is what makes the deployment trustworthy, and it is a property of the whole family of configurations rather than of any one of them.

Most discussion of origami engineering concentrates on the first and the second is what actually distinguishes a usable pattern. A great many arrangements pack well; very few move along one path. The intersection of rigid and flat foldability is small precisely because the motion is the binding requirement.

A useful test of whether an account of folded engineering knows what it is talking about is which of the two it emphasises.

Why this pattern rather than another #

A closing question: given that many tessellations fold, why is this the one everybody uses?

It has one degree of freedom, which almost none do. It is rigid-foldable, which almost none are. It packs in both directions at once, which is what makes the area ratio good. Its vertices are all identical, which is what makes it a mechanism. And its facets are parallelograms, which is the easiest shape to manufacture.

Five properties, each independently uncommon, and the Miura has all of them. That is why the same crease pattern appears in a 1970 paper, a 1995 spacecraft, a road map and a laboratory metamaterial — not because the field lacks alternatives but because the conjunction is rare.

The ladder from here #

Later rungs: the degree-four kinematics derived. Poisson’s ratio computed from the closed forms. The parallelogram angle as a design variable. Bistability and snap-through. Other tessellations and their motions. Rigid-foldability, and why the Miura has it. Thickness accommodation. Deployment dynamics. Edge effects in finite panels. And the question of what other patterns have exactly one degree of freedom, which is not fully answered.

The whole behaviour follows from every vertex being the same. Change one vertex in the middle of a Miura sheet and it stops being a mechanism entirely — a single defect locks the whole thing.