Dome Cooling — Passive Cooling from Curved Geometry

A dome cools passively three ways at once: less envelope per enclosed volume (≈30% less surface area than an equivalent box), inherent stack ventilation (hot air pools high above the occupants), and curvature that accelerates wind and opens most of the surface to the sky for radiative emission. The effect is real and CFD-validated — but it is not sufficient on its own in serious summer heat, and its most promising multipliers (ultrawhite coatings, curved PCM modules) are still untested combinations.

Fuller’s 1940 discovery: the reverse chimney effect

In August 1940 Buckminster Fuller’s team, investigating cheap wartime shelter in Kansas, built the Dymaxion Deployment Unit (DDU), a suspended metal grain-silo structure. During a heat wave a visiting business partner found the interior unexpectedly cool. Fuller credited “Bernoulli’s principle”; the more accurate physics (per physicist Chris Weekes) is acceleration under conservation of energy plus adiabatic expansion, consistent with the first law — a reverse chimney effect, opposite to the conventional stack effect:

Sun-baked exterior walls drive convective updrafts along the dome’s outer surface; warmed surface air rises off the sun-facing side.
That rising air draws interior air out through short, wide openings at the base.
Make-up air is pulled inward through a narrow top vent, accelerating as it enters — converting thermal energy to kinetic energy — then expanding and cooling further in the lower-pressure interior, like the frost that forms on a wide-open compressor valve.

Design parameters from Fuller’s practice: small inlets at top and multiple lower-edge outlets with greater total area; darker exterior coloring helps (more absorption → stronger convection); moderate insulation (R-12 in his conversions); stack length and aspect ratio matter, and a tapered stack may add velocity as the air cools and contracts. The surface-area-to-volume ratio is the governing geometric variable — larger domes enclose more volume per unit skin, so the insulation needed to optimize heating and cooling falls as diameter grows.

The limits are just as instructive: the effect needs direct sun on the walls (dead at night and in shade), lateral breezes from open windows disrupt the vertical flow pattern, and the vent effect weakens as diameter grows. Fuller envisioned computer-controlled electronic vents for seasonal management but died in 1983 without systematically testing vent positioning and inlet-to-outlet ratios — an explicit invitation to the CFD work below. The source offers five reasons the idea never spread: public disbelief at a counter-intuitive effect, the night/shade limitation, incompatibility with ordinary window-opening habits, better passive strategies in some climates, and doubtful viability at very large scale.

What CFD confirms — and what it corrects

Two peer-reviewed studies of the same 3V-icosahedron dome (two floors, 9.06 m tall, 143 m² floor, 864 m³, Yazd hot-desert climate; ANSYS FLUENT 15, standard k-ε RANS, ~4.6–4.67M-element unstructured mesh with grid sensitivity verified at 5.72% medium-to-fine error, 18 measurement points at 1.2 m and 4.2 m) put numbers on the geometry. The 3V form was chosen as smoother than a 2V and less complex than a 4V.

Confirmed advantages (MDPI Computation, 2016)

~30% less surface area than a box house of equal volume → a proportional cut in envelope heat gain and material
Indoor air in domed roofs measured 8 K lower than flat-roof equivalents; total heat transfer per surface area 23% less than cubic
52% reduction in cooling energy vs. conventional prismatic buildings; annual heating loads 62.6% lower; ~90 vs. 150 kWh/m²/yr (BFI/Oregon Dome field data)
Curvature produces higher windward/leeward pressure differentials than flat roofs; wind flows smoothly around the shell; the concave interior has no corners or dead-air zones, so air circulates unobstructed and the tall interior stratifies heat above the occupied zone
Two workable configurations: wind-induced (upper roof vents exploiting the pressure differential) and buoyancy-induced (roof vents plus lower windward/leeward windows for stack-driven flow in calm periods)

Even this favorable study concludes that roof-vent ventilation alone cannot meet summer thermal requirements: winter performance is excellent, summer needs a hybrid passive-active approach.

The correction (Soleimani et al., 2018, all 12 months): a roof vent alone acts as an “air curtain” — it captures the approaching flow and redirects it to the leeward wall, forming a large recirculation on the upper floor that supplies fresh air but fails to exhaust warm air. Ventilation rate was 7.7 l/s per m² at 3 m/s wind, upper-floor air velocity 28% higher than lower, and the lower floor averaged 3.14 °C warmer than the upper; the coolest spot (17 °C) sat under the vents and the hottest (37 °C) in a low-airflow corner. In 38–41 °C summer conditions every PMV value sat above +3 (“Hot”): natural roof-vent ventilation cannot carry the summer load by itself. The authors’ fixes are exactly Fuller’s unfinished agenda: add lower openings for through-flow, divide the vent into inlet/outlet zones, add inlet/outlet shafts — plus thermal mass with night cooling, and evaporative cooling, heat pipes, PCMs, or desiccant dehumidification for hot-humid sites.

A companion study of arched-roof greenhouses (Li et al., 2020) shows what vent optimization buys on a curved roof: a rolling-shutter vent at a 75° position angle moved 2.66× more air than the baseline pivoting window (whose air-curtain behavior mirrors the dome roof vent), gave 1.5 °C additional cooling within ten minutes, and cut temperature inhomogeneity 33%; 85° was optimal for most configurations. Vent type and placement on curved roofs is a high-leverage variable that the dome studies call for but do not perform.

Multiplier 1: reflective/radiative coatings

A dome’s larger curved surface is a double-edged sword — it absorbs ~20% more beam and ~30% more total solar radiation than a flat roof of the same base area (Tang et al. 2003), but offers more area for convective shedding (wind accelerates over the curve) and radiative emission, and most of a dome’s surface has an excellent sky view factor ((1+cos β)/2 per element). The one direct study — Faghih & Bahadori (2011), a thermal-network model of the School of Theology dome in Yazd with glazed tiles at absorptance 0.4 versus ordinary material at 0.8 — found:

Configuration (August, Yazd)	Max indoor air temp
Flat roof, ordinary material (α = 0.8)	37.30 °C
Dome roof, ordinary material (α = 0.8)	37.09 °C
Dome roof, glazed tiles (α = 0.4)	36.04 °C

The coating multiplied the dome’s advantage over a flat roof by 6× (0.21 °C → 1.26 °C indoor reduction): reflectivity suppresses the absorption penalty while the geometry keeps the emission advantage. The LBNL steady-state roof balance — (1−a)·I = ε·σ·(T_s⁴ − T_sky⁴) + h_c·(T_s − T_a) + U·(T_s − T_in) — applies to a dome with a per-element view factor in the radiative term and a wind-boosted convective term, but has never been run for a dome with a modern coating. Flat-roof cool-roof data gives the per-area baseline: 10–50% cooling savings in real buildings, ~314 kWh/yr per 100 m² of roof in Phoenix, with most albedo loss (~0.15) in year one and washing restoring 90–100%. Nobody has yet put an ultrawhite radiative paint (98.1% reflectance, ε = 0.95 BaSO₄ — see Radiative Cooling) on a geodesic dome; the predicted effect is multiplicative — per-area coating benefit × ~30% more emitting area — while attacking the dome’s main thermal weakness.

Source disagreement, annotated. Fuller’s practice says a darker dome cools better (stronger wall-driven convection powers the reverse-chimney vent); Faghih & Bahadori and the cool-roof literature say a reflective dome cools better (less absorbed load). Both can be true because they describe different mechanisms: the dark shell buys ventilation only while the sun is on the walls, whereas the reflective shell cuts the absorbed load all day. No source tests the two against each other; the modeled 1.26 °C coating gain is the only quantified figure, and Fuller’s darker-is-better claim is anecdotal.

Multiplier 2: curved PCM modules and the thermal-battery dome

Concrete domes already behave as thermal batteries: each pound of concrete stores ~1 BTU over a 5 °F swing, so a 50-ft residential dome (~200,000 lb) holds ~200,000 BTU and a 200-ft assembly dome ~2.5 million, enough that rapid temperature change is physically impossible. The Monolithic Dome Institute reports HVAC downsized 60–75%, power bills 50–70% lower, a 50-ft Serenity Dome running under $950/yr total energy (2012, a third of the previous conventional home), one owner heating through winter on five 100 W bulbs, engineered R-values “in excess of 80,” and night-air flushing of the shell mass as standard practice — vendor data, treat as indicative.

Phase-change materials extend the same principle from sensible to latent storage. Huluka & Muthulingam (Scientific Reports 15:39845, 2025) split a 1.325 m square, 150 mm reinforced-concrete roof slab into four thermally isolated zones — plain concrete, empty forged-aluminum (Al 1200) spheres, and spheres carrying 354–358 g of OM35 or OM37 organic PCM at 75% of diameter — and instrumented it in subtropical Rupnagar, India, in April:

Zone	Cooling load reduction	Surface temp decrease	Decrement factor
Plain concrete (control)	—	—	0.56
Empty spheres	59%	—	0.50
OM35 spheres	68%	up to 11.1 °C	0.38
OM37 spheres	69%	avg 10.2 °C (min 9.2)	0.37

Two readings matter. First, empty spheres alone cut the load 59% — the voids in the slab do most of the work, and the PCM adds the last ~10 points plus the large surface-temperature damping. Second — the geometrically relevant finding — spherical/cylindrical PCM encapsulation is ~70% more energy-efficient than rectangular. The modules self-regulate: melting absorbs daytime heat, solidifying releases it at night, the same cycle as night-flushing a dome shell. No published study has yet combined PCM with dome or curved-shell geometry; the predicted synergy is curved encapsulation on a curved shell, less total PCM thanks to the lower surface-to-volume ratio, dome convection patterns cycling the modules uniformly, and sensible shell mass plus latent PCM as one combined battery. See Translucent PCM Glazing and Envelope & Glazing for the PCM-in-envelope thread.

Honest scorecard

Claim	Status
Less envelope per volume (≈30%)	Solid geometry, CFD- and field-corroborated
52% cooling energy reduction vs. box	Peer-reviewed CFD (one climate, one dome)
Reverse chimney vent cooling	Real but conditional — needs sun, dies at night, breeze-sensitive
Roof vent alone handles summer	No — “air curtain” failure, PMV > +3; hybrid needed (both CFD studies agree)
Vent type/placement optimization	2.66× shown on an arched greenhouse roof; not yet done for a dome
Coating × dome synergy	One 2011 modeling study (6× multiplier); modern coatings untested; conflicts with Fuller’s dark-shell preference
PCM × dome synergy	Untested; curved-encapsulation advantage (70%) and flat-slab results (59–69%) are suggestive
Vendor thermal-battery numbers (R>80, 50–75% savings)	Marketing-grade, no independent verification

Sources

Buckminster Fuller’s Chilling Domes: The Physics — the 1940 DDU discovery, reverse-chimney mechanism, design parameters and limits, why it never spread
Geodesic Dome CFD Cooling Analysis (MDPI Computation, 2016) — 30% less surface, 8 K cooler, 52% cooling-energy cut, two ventilation configurations, CFD method
Geodesic Dome Vent CFD Analysis (Soleimani et al., 2018) — the 12-month air-curtain failure, PMV > +3, recommended fixes, Li et al. arched-roof vent optimization
Dome Geometry + Radiative Coating: Combined Evidence — Faghih & Bahadori 6× multiplier, sky view factor, LBNL roof balance, flat-roof cool-roof data
Spherical PCM Modules and Dome Thermal Battery Performance — Huluka & Muthulingam slab experiment, curved-encapsulation advantage, Monolithic Dome vendor data, the untested PCM × dome combination