Earthships are cooled by the sun. The greenhouse contains a skylight, which when open produces suction from the buoyancy of the sun-heated air. This sucks air through the dwelling spaces from the cooling tubes that run through the giant thermal mass behind the house. When this is not desired, the inhabitants close the skylight and the cooling tubes and open the doors to the greenhouse to let the heat in.
Thus can a tiny air pressure difference produce an enormous heat flow: the viscosity of air is low, so if an opening is wide, even an imperceptibly tiny difference in pressure can produce a mass flow as fast as the wind, which can carry with it an immense power of heat.
To be concrete, let us suppose that we have an opening of two meters square with a subtle breeze of 2 m/s flowing through it. Evidently 8 m³/s of air flows thus; at 1.225 kg/m³ this is about 9.8 kg/s, a staggering rate of 35 tonnes per hour. At typical air’s isobaric mass heat capacity of 1.012 J/g/K this works out to 9900 watts per kelvin. That is, if the air traveling through this aperture have a temperature a single degree hotter or colder than the space it enters, it brings or displaces 9900 watts. If the temperature difference be ten degrees, the power is 99 kilowatts, and if it be 40°, the power is nearly 400 kilowatts.
Air has some other remarkable advantages as a heat transfer medium, aside from its low viscosity. It’s less toxic than any other fluid. It’s relatively non-corrosive up to 300° or so, and up to over 2000° for oxides, phosphates, and fluorides. It spans a wide range of temperatures in gaseous form, from -182° (oxygen’s boiling point) to some 1600° while preserving its nontoxic and noncorrosive character, and up to an unlimited temperature if this is not important. For the time being, it’s easily available, and the cost is low.
At temperatures between 0° and 100° we can direct the flow of air with nearly any everyday solid material, including polyethylene bags, waxed or lacquered paper, styrofoam, plaster, wood, oiled muslin, PET bottles, mica, EVA glue, cardboard, clay, glass, and aluminum foil; outside this temperature range it can become more demanding. Within this temperature range the most diaphanous of materials can contain airflow without injury, needing only enough support from wires or the like to prevent its collapse. Inflatable tubes, with stagnant air in a tube alongside the moving-air tube, may provide the lightest-weight support of all.
The idea inevitably suggests itself to raise a flexible chimney with a hot-air balloon to produce suction from sun-heated air. A solid bootstrapping heating chamber provides initial warm air to initially inflate the ultralightweight balloon, made perhaps of mylar; once it achieves inflation, it can serve as its own air-heating chamber, particularly if the exterior is visibly transparent and painted or mixed with an infrared-blocking paint and/or a low-infrared-emissivity coating, while an inner mylar partition is black, re-emitting the visible light absorbed as thermal infrared into the interior of the balloon. Alternatively, the balloon can be fed from the chimney itself, which if fastened to the top of the balloon can inject any available hotter air into the balloon, displacing less-hot air out of an open balloon bottom.
The chimney itself can have a construction similar to that of the balloon, and perhaps either or both would benefit from a light springy spiral or two to prevent their collapsing or rippling, resulting in airflow obstruction.
Of course, a more conventional rigid chimney can also be used, with a greenhouse at its base and perhaps transparent panels within it to permit further heating of the air during its ascent. If it can be laid upon the slope of a hill or mountain, it would require minimal material.
Unlike chimneys intended for flue gases, such chimneys need only handle air in the range of 50°–200°, depending on the design, and can thus use non-heat-resistant materials, perhaps even PET.
Even on cloudy days, greenhouses and greenhouse-effect chimneys can achieve significant temperature rises and thus significant pressures.
Such a comparatively high-pressure, low-volume flow can be called upon to fluidically pump larger volumes of air at lower pressures, as in the Dyson bladeless fan. I’m not quite sure how to use this negative gauge pressure to pump air into spaces at a zero or positive gauge pressure, particularly without sending the hot air into those areas; some ideas follow.
By accelerating air down a wide tube toward a suction aperture, the air can be given momentum, and the air that misses the suction aperture can then be diverted through other tubes wherein it loses some, but not all, of its speed; this air, now at a positive gauge pressure (?), can then be used to direct other air at still lower speeds.
Another approach, which may sort of be the same thing, is to use the low pressure to suck a flow down a long tube, which near the end has an aperture in its wall to another similar long tube, transferring some of its momentum to the fluid in the other tube, and mixing somewhat.
Bistable fluidic valves of the well-known design using the Coandǎ effect can be used to change the direction of flows for an arbitrarily long period of time with just a puff of air.
If you’re just using the vacuum to drive a closed-circuit thermodynamic system, the problem vanishes; you can maintain the whole system at a negative gauge pressure and bring in nozzles of zero-gauge-pressure air as “compressed air” to blow the flow wherever it’s desired.
One difficulty arises if your thermodynamic system is evolving some gases you don’t want in your chimney, perhaps because they are toxic or corrosive. In some cases you can bubble them through limewater or something, but in other cases no such convenient outlet is available. In such a case some kind of pumping system like the ones I mentioned above seems essential, whether purely fluidic or merely pneumatic.
Pumping via pneumatic, rather than fluidic, methods at low pressure simplifies the problem enormously; a flexible balloon of any kind (polyethylene, cloth, origami) within a larger space serves as a gas-tight piston, while a flap of flexible material over an aperture serves as a check-valve. A balloon that inflates to obstruct an air passage provides a simple form of pneumatic switching that easily permits the construction of oscillators and the like.