Globoflexia, or balloon twisting, is a popular form of entertainment, especially for children; a skilled balloon twister can make an evocative, if cartoonish, sculpture of an animal or person within a few seconds to a minute. Extremely elaborate sculptures are feasible over a few hours; because the material is so light, getting all of its compressive strength from air, sculptors can easily build and manipulate sculptures far larger than themselves. Because the balloons leak, the sculptures are ephemeral, lasting at most a few days.
Much to my surprise, there’s a world globoflexia conference every two years at which teams from dozens of countries compete.
What if you could use globoflexia as a medium of expression for more permanent ideas? Obviously you can photograph the sculptures, thus making images or videos of them, but the underlying three-dimensional form of the object is lost.
So I’ve been thinking about three different ways to do this: photogrammetry, spray foam, and papier-mâché.
If you can 3-D scan balloon sculptures into a computer, you can use them as a means for telling the computer what to do; this could be, as in Dynamicland, a real-time interactive process of shaping computations with your hands, with real-time projected feedback, or it could be more a kind of batch data-entry thing, for example for designing three-dimensional shapes for later tweaking and automated fabrication, whether at the same scale, a larger scale, or a smaller scale.
Existing photogrammetry methods do not work for balloons. But the balloons in question are not inherently algorithmically difficult: each is a well-controlled solid color, displaying gradients of intensity corresponding to local degree of stretch and illumination, with well-controlled specular highlights. The images generally only contain edges at the edges of the balloon silhouette, at wrinkles around twists, around specular highlights, and outside the balloons. These highlights give a fairly precise read on the surface angle and curvature at a particular point, as do the silhouette edges.
Moreover the balloon sculptures’ shapes are themselves well-behaved: the surface at most points has a relatively smooth curvature determined mostly by the gauge pressure and the tensions in two directions. Rubber’s complex pseudoelastic thermodynamic behavior is not so complex as to make this a very difficult problem.
Further information can be obtained by looking at the balloon sculpture from different angles, as is normally done in photogrammetry, thus scanning the specular highlights and silhouette contours over the surface.
Given this information, it remains to optimize a model of the balloon sculpture to account for the observed photos as parsimoniously as possible, using standard methods like finite element analysis, Markov-chain Monte Carlo, gradient descent, and genetic algorithms.
What if you fill the balloons with a hardening foam instead of air?
Conventional polyurethane expanding spray foam insulation has been available for decades. You spray it as a thixotropic liquid foam, which accommodates itself to the container it’s in before slowly polymerizing into a light, hard, thermally and electrically insulating foam with substantial mechanical strength. Some formulations form waterproof closed-cell foams, while others form lighter-weight open-cell foams. There are formulations that ship as pairs of liquids to be mixed in a gun, for high-volume applications, and other formulations that you just squirt out of a can.
The materials that form polyurethane foams are fairly reactive until they’ve finished forming the foam, and that may be a fatal flaw for squirting them into rubber balloons: they may corrode the balloons and pop them before the foam has hardened.
Polyurethane is not the only possible hardening foam. Latex foam is widely used for pillows, mattresses, and theater special-effects makeup (“prosthetics”), in which last use it is typically cast in molds before curing. Protein foam is a popular dessert, both as meringue (with air whipped into it) and as gelatin foam in so-called “molecular gastronomy” or “modernist cuisine”, where the gelatin gel is mixed with nitrous oxide under high pressure and low temperature, like canned whipped cream. Gelatin foam is also widely used for makeup, sometimes whipped like meringue, but sometimes foamed with non-double-acting baking powder (for example baking soda with cream of tartar) or even yeast.
There’s also been a lot of work in recent years on foamed concrete, sometimes called “aircrete”. This consists of portland cement, water, a surfactant (Suave shampoo is reputed to work well, though there are also specific surfactant mixes from companies like Drexel), possibly some foam stabilizers (I suspect gelatin might work well for this), and a great deal of gas. Sometimes sand is used, but rocks are never used. The original 1920s process for foaming concrete (“autoclaved aerated concrete”) used aluminum powder mixed into the concrete mix. After molding the concrete was heated in an autoclave to react the aluminum with some of the lime in the concrete, thus foaming the concrete with hydrogen gas, as well as accelerating the formation of the calcium silicate hydrates that bond the concrete. The more common method nowadays is more like meringue: air is mechanically mixed into some of the water before mixing in the wetted cement.
You probably can’t foam any traditional concrete with anything similar to baking powder, because the strongly basic nature of portland cement, lime, Sorel cement, and refractory calcium aluminate will destroy the baking-powder acid without producing any gas. Non-traditional concrete binders like low-alkalinity waterglass or molten sulfur might be less corrosive, but probably also are not a realistic way to fill balloons. And I suspect that at room temperature aluminum powder will not produce hydrogen fast enough.
The strongly basic nature of these cements might also cause them to attack the balloons.
Most resins can be foamed in a way similar to polyurethane spray foam. Radio-controlled airplane hobbyists commonly mix a secret “foaming agent” from R&G into two-component epoxy to get an epoxy foam, for example. One publication on the use of polysilazane for this purpose suggests that powdered aluminum mixed with soda lye is the usual foaming agent!
Whatever foam is chosen, whether one of the above or something else, the idea is simply to fill the balloons with the foam or incipient foam rather than just air. Then you twist the balloons into the right shape, carrying the foam along with them, and leave them there until the foam has hardened. You may want to spray some kind of adhesive onto the joints, since otherwise the foam in the different segments of the balloons will only be connected together through the balloon rubber, which may not be very stable.
An alternative, and possibly complementary, approach is to put something on the outside of the balloons that hardens there, forming a hollow, tubular, continuous version of the shape you have made with the balloons. The traditional material for this is strips of paper dipped in wheat paste, but there are many possible variations on the papier-mâché theme.
In addition to wrapping the balloons tightly, you can use the balloons themselves merely to form a frame over which sheets of adhesive-soaked fiber reinforcement are draped.
For the adhesive, rather than wheat paste, you could use:
For the fiber reinforcement, instead of paper, you could use:
If the adhesive is less flexible than the fiber reinforcement (e.g., has a higher Young’s modulus), then the fiber reinforcement may just weaken the binder instead of strengthening it, although it can produce some “strain hardening” behavior where the adhesive cracks but the fiber keeps the adhesive cracks from opening wider and thus continuing to propagate. Still, even weak fibers can hold the adhesive in position until it sets, and for some purposes the adhesive alone will be strong enough without any help from the “reinforcement”.
The fiber reinforcement may have other purposes as well, other than shaping or strengthening; for example, if the adhesive is transparent, decorative or informative images can be printed on the fiber reinforcement; conductive fiber reinforcement can provide Faraday-cage protection against EMI more cheaply and flexibly than sheet metal; gold leaf or aluminum foil can provide high reflectivity; and so on.
It may be worthwhile to also include other additives in the adhesive, whether inert fillers or reactive; for example:
In some cases you will want to start with a lightweight, fast-hardening system such as gauze and plaster of Paris, then overlay it with a heavier system that the balloons alone wouldn’t be able to support. There are many other reasons you might want to use multiple layers, including making sandwich panels with a light, weak inner core and stronger faces, and allowing earlier layers time to dry.
In cases where the first layer has no fiber reinforcement, it might be useful for that first layer to be sprayed onto the balloons rather than placed there by hand. This would allow it to be applied more rapidly and easily, and it could perhaps be strong enough once hardened to support significantly more weight than the balloons themselves. Spray foam seems particularly appealing for this application.
If you combine this process with the foam-filling process you can get shapes built from tubes with strong, rigid, hard surfaces braced by a weaker foam within.
Also, of course, most of these processes can be used on top of a form produced by some other method than globoflexia; for example:
In some cases it will be most convenient to apply the adhesive to the fiber reinforcement after it is already in position, but in other cases, especially with porous adhesives, it will be most convenient to combine them ahead of time.
During the process of adding layers of fiber reinforcement and (possibly filled) adhesive, it may be convenient to embed other elements in the object being constructed, in non-random positions. For example, you can embed sensors, heating elements, LEDs or other lights, antennas, pancake coils, and wires to feed all of these. For some purposes it is best to cover these with a layer of adhesive and/or fiber, for example to prevent abrasion or electrical short circuits, while for other purposes exposed electrodes or other actuators may be useful.
The above outlines a large design space of processes, a few of which are already in use:
However, the systematization suggests many new promising combinations. For example:
The initial form is produced by globoflexia and wrapped in three layers of gauze strips dipped in fresh plaster of Paris. Fifteen minutes later, steel window screens are draped over the plaster frame, and a thick mix of slaked lime, water, quartz construction sand, and chopped fibers is troweled onto the screens. Two more such layers of screen and lime plaster are immediately applied. A few hours later, the outer surface is painted with sodium silicate to increase its resistance to abrasion; sufficient air can still enter through the porous plaster and lime cement to cure the piece over the next 24 hours. Filling the final piece with foamed portland concrete, made in the usual way, is optional.
A balloon for twisting is inflated but not twisted. It is wrapped in five layers of paper towels dipped in wet portland cement, sand, and chopped basalt fiber, leaving the balloon’s ends exposed. The entire resulting concrete tube is wrapped in stretch plastic wrap to keep it from drying out. After 48 hours, the balloon is popped, exposing the inner surface of the pipe, which is then painted with a solution of potassium silicate to waterproof it.
A light roof metal truss is built by bending and arc-welding together rebar. Jute burlap cloth is dipped into hot tar and laid on top of the truss, one square meter at a time, overlapping squares in a shingled pattern. Three layers should be sufficient for a rainproof roof sufficiently flexible not to suffer damage from hail. However, it is a fire menace, it will get very hot in the sun and may drip, and you cannot walk on it. Coating the top and bottom surfaces with aluminum foil will ameliorate these defects slightly.
A balloon for twisting is inflated but not twisted. A4-sized paper is dipped in low-melting paraffin and wrapped around the balloon in three layers. Once the paraffin is cool, the balloon is popped or untied and removed. The non-knot end of the balloon can be left open, forming a closed tube, or closed. To improve strength, a fourth and fifth layer of paper dipped in two-component resin rather than paraffin can be added on the outside.
The clay form of a hand is slipcast from a clay slip containing a mildly acidic flocculating additive such as vinegar, using a porous mold made of plaster of Paris. The hollow slipcast clay form is demolded and tightly wrapped in four layers of loosely woven cotton or linen cloth, richly smeared with high-temperature acetic-acid-catalyzed silicone (“red RTV”). Once the silicone has cured, the still-wet clay interior is washed out with water and a mild base such as baking soda in order to deflocculate the clay. The resulting piece, once the acetic acid has escaped, can withstand temperatures up to some 240°; substituting a cloth with higher-temperature capabilities should allow it to handle 280° continuously or brief exposures to 320°.
(Other reversible flocculants might be epsom salts, which can be deactivated by barium carbonate, and muriate of lime, which can be deactivated by soda ash or barium carbonate. I’ve also seen muriate of lime recommended as a deflocculant, presumably to throw down vitriol in the form of alabaster.)
It might be necessary to protect the cotton from hydrolysis by the acetic acid before curing is complete; this can be done either by replacing acetic-acid silicone with more expensive tin-catalyzed or platinum-catalyzed varieties, or (with lower certainty) by impregnating the cotton with a buffer of, for example, baking soda.
This is a case where a balloon form shaped by globoflexia is probably better than clay, actually, because it’s both easier to shape to the appropriate smooth blobby shape and easier to remove. But it’s important to remove it completely, because the balloon latex will probably fail at a much lower temperature.
By wrapping your twisted balloons in gauze soaked in shellac, you can get a waterproof, light, flexible material that allows significant light through, due to the gauze’s light weave. Alcohol is emitted as the shellac dries, but this is a fast process; once dry the material emits almost no VOCs.
If you stamp one or two layers of a loose steel mesh such as a window screen, impregnated with warm paraffin wax, with a die, then you have a waterproof and chemically inert plug which, at a predetermined temperature (one calibratable within 5°) will melt and allow liquid to flow through freely, while filtering out particles larger than the mesh. This could be useful for some kinds of over-temperature safety valves, for example for resin casting, where, if the resin starts to overheat, the ideal thing to do might be to dump it quickly out of the mold into something that dilutes and cools it. It is possible for this mesh to have a much larger area than the aperture it covers, which may be desirable for keeping it from getting clogged by particulates.
Under some circumstances it might be better to use injection molding to inject the paraffin around the reinforcing mesh. This would provide more consistent paraffin thickness but, I think, less consistent mesh protection thickness.
By cutting the shape of a stirrer out of, for example, styrofoam, you can get a very lightweight tool. But styrofoam is soluble in all kinds of solvents, and it’s kind of weak. By wrapping it with fiberglass cloth, as is done to construct some boats and aircraft, you can greatly strengthen it. A coating of, for example, paraffin, low-density polyethylene, teflon, epoxy, or polyester casting resin, could both firmly adhere the fiberglass reinforcement to the foam and add substantial chemical resistance.
You can pour portland cement foamed in the usual way, by mechanical aeration of a surfactant-water solution before mixing in the cement, into forms that are merely blocks. The next day, once the cement has partly set, you can sculpt these blocks into desirable shapes using hand tools like hacksaws, wood rasps, wire saws, hammers, and so on. The resulting surface will be porous and friable, and therefore not directly suitable for furniture use, and also an ugly gray unless you used super fancy portland cement. Several coats of lime mortar (slaked lime and quartz construction sand) can give it a hard shell, perhaps reinforced in key places with copper wires. The next day, a coat of polyurethane finish for heavy-duty floors can seal the lime and provide a softer, warmer surface to sit on or rest your feet on.
The usual kind of pottery is fragile. The ceramic fibers used in foundry blankets are much less fragile, and some of them can be used up to 1600°. You could perhaps take segments of refractory-fiber cloth like these foundry blankets, dip them in a clay slip, and drape them over forms (for example, blow-molded from thermoplastic) to make a shape of two or three millimeters of thickness. Once the clay slip was plastic, but before it became leather-hard, you could add another millimeter of clay to the inside and outside. After drying and biscuit-firing these pottery pieces in the usual way, the clay should be sintered into a solid body; you can get a good biscuit fire out of at least some ball clays in 6 hours at 1020°, at which temperature some ceramic fibers are still quite inert. So they should remain embedded as fibrous reinforcement in the finished ceramic, making it dramatically less fragile.
However, care must be taken to ensure that the chemistry of the clay is compatible with that of the blanket. Pure zirconia fiber (or yttria-stabilized zirconia fiber) would probably be perfectly safe, but I think everybody includes at least alumina and usually silica in their ceramic foundry blanket fiber. (Vendors of pure zirconia fiber say it can be used up to 2200°.) I suspect that any low-firing clay would be able to flux and dissolve silica out of part-silica fiber, and maybe alumina too. The end product might still be stronger than ordinary ceramics, though.
Silicon carbide fibers are more widely available than zirconia fibers; four companies already sell them commercially, at least two since the 1980s (under the names Nicalon, Tyranno, Sylramic, and Ultra SCS). I think they are not attacked by clays at common pottery-firing temperatures, and they are already in use for reinforcing ceramics — but I think primarily ceramics otherwise made of sintered silicon carbide, not fired clay.
If desired, a second glaze firing can glaze the pieces to give them waterproof surfaces and provide protection against abrasion and crack initiation. However, this poses the risk that the more aggressive fluxing of the glaze might attack the fiber reinforcement; this is the reason for the extra protective layer of clay without fiber in it. If this is a problem, a possible alternative to traditional glazing is waterglass allowed to dry on the ceramic and then crosslinked by, for example, exposure to calcium chloride.
The usual ferrocement recipe uses iron (and consequently a little iron oxide), portland cement, and quartz, none of which is very friendly to temperatures above 1000° or 1500°. Calcium aluminate cement can replace the portland cement, and olivine, sapphire, or carborundum can replace the quartz, but what can replace the iron?
Refractory metals like tantalum and niobium are well known, but very expensive. Ceramic fibers like those mentioned above (zirconia, alumina, carborundum) might be adequate; the “ceracement” structure won’t need flexurally-stiff reinforcement to hold it up, since it can hold itself up once the cement is set.
At even higher temperatures calcium aluminate fails and needs to be replaced with higher-temperature castable refractory binders such as aluminum phosphate.
Modern synthetic grinding stones have a variety of compositions: sapphire, silicon carbide, cubic boron nitride, etc., bonded with rubber, thermoset resin, waterglass (mostly historically), Sorel cement, and so on. But they tend to fail in a brittle fashion rather than a ductile fashion, which frequently kills people when they are spinning fast around people.
Cutoff discs are like thin grinding wheels, but they are usually reinforced with a fiber, typically fiberglass, I think.
Perhaps grinding wheels could be made with much heavier fiber reinforcement to encourage them to fail in a more ductile fashion. High-energy-capacity fibers like rubber, nylon, or music wire might work better for this than high-modulus fibers like fiberglass and basalt fiber.
Coat a roll of cotton scrim fabric with a low-temperature nonpolar thermoplastic adhesive like EVA. Heat the cloth and run it through a pile of premixed quick-setting dry lime cement and construction sand, which sticks to the EVA and coats the cloth. Allow the cloth to cool before spooling it onto the takeup roll. Seal the finished roll hermetically in a reclosable container.
The resulting tape can be torn by hand like duck tape, although gloves are advised. Once a form is wrapped with it to a few millimeters thick, and flexed into the desired shape, you can moisten the tape around the form to start the cement setting. Water can soak through it easily, and it will amalgamate into a cotton-reinforced mortar mass.
Perhaps such tape can be laid between bricks or stones to hold them together, rather than troweling in mortar.
Other cements can be substituted, such as portland cement or calcium aluminate, which would give stronger results. There may be faster-setting high-strength cement formulations that are not in traditional construction use and that would activate the tape even faster. Using plaster of paris instead of the cements suggested would provide much faster results (and perhaps this is already in use) but much lower strength.
One particularly interesting possibility is using dissolved sodium-silicate waterglass as cement, which is somewhat tacky immediately and will set up hard when dried; but a variety of things will cause it to set up immediately and become water-insoluble, such as carbon dioxide gas or, I think, calcium chloride or magnesium sulfate. So you could perhaps spray solutions of those on the tape, once it is applied, from a spray bottle.
Steel wire mesh would be a stronger alternative to cotton scrim, and might still be possible to tear by hand.