Solar Thermal — Collectors That Drive Cooling

Solar thermal is the heat supply for sun-driven cooling. In a joule-heist strategy it does the job the burner does in a boiler: deliver 60–95 °C (or much hotter, concentrated) water to desorb the adsorbent in an adsorption chiller — or, at high concentration, to run power and process heat. Its defining virtue for cooling is coincidence: peak insolation lines up with peak cooling demand, so the energy arrives when the cold is wanted.

The Zotero registry’s solar-thermal and black-body shelves (a flat-plate overview, Flamant’s Concentrating Solar Thermal Energy, Wang’s Design of Solar Thermal Power Plants, the NREL/GSA honeycomb-collector reports, and the black-body/CNT-absorber references) cover the two ends of the collector spectrum:

Collector class	Temp range	Role in cooling
Flat-plate collectors	~40–90 °C	The match for silica-gel/water adsorption and low-temp sorption cooling; cheap, robust, building-mountable. The NREL High-Performance Flat Plate evaluation and the GSA Honeycomb collector brief target raising flat-plate output (honeycomb transparent insulation cuts convective loss) into the useful adsorption-driving band.
Compound parabolic concentrators (CPCs)	moderate (low-to-mid concentration, no tracking)	Non-imaging, non-tracking mid-concentrators that accept light across a range of angles and still gather diffuse/cloudy-day light — bridging flat-plate and tracked CST. Attractive for adsorption-driving heat where tracking hardware is unwanted. See below.
Concentrating solar thermal (CST/CSP)	high (100s of °C)	Parabolic trough, power tower, linear Fresnel, and dish systems (Flamant; Wang) — for power generation and high-temperature process heat; relevant where higher-regeneration-temperature sorbents (zeolite) or combined power+cooling are wanted.

Flat-plate collectors and the honeycomb upgrade

The honeycomb solar thermal collector (HSTC) evaluated by NREL for the GSA Green Proving Ground is a flat-plate collector with a transparent-insulation honeycomb polymer layer between glazing and absorber. The honeycomb suppresses convective loss while admitting solar radiation; an air gap between honeycomb and absorber cuts conductive loss; and a heat-pipe overheat-protection device limits stagnation (measured maximum 152 °C), which matters because a collector that sits idle on a hot day with no load must survive it. Field results at two federal sites (Indianapolis and Auburn, WA — one of the weakest solar resources in the US):

Measured efficiencies matched the manufacturer’s within ~2 % (NREL’s acceptance target was ±10 %).
For ordinary domestic hot water the HSTC was up to 8 % more efficient than a standard flat plate, with little difference between hot and cold climates.
Its real advantage is at high inlet temperature — reheating an already-warm recirculation loop, process heat, space heating — where it held ~40 % daily efficiency even at Auburn and competes in the niche historically served by costlier evacuated tubes. That is exactly the regime an adsorption chiller imposes: the hot loop returns from the desorber still warm, so the collector spends its day at elevated inlet temperature.

The NREL report also supplies the standard collector performance forms (European optical-efficiency and US quadratic F_R(τα), F_R·U_L fits from SRCC lab ratings) for predicting delivered energy from rated parameters — the tool for sizing a collector field to a chiller’s drive-heat demand.

Economics. Two lessons transfer directly to solar cooling. First, collectors are only ~20 % of installed system cost, so a pricier, better collector barely moves the total; life-cycle cost, not collector efficiency, should drive selection. Second, payback depends on load, energy price, and resource far more than on hardware: at ~$46/ft² installed and a 500-gal weekday load, Phoenix reached a solar fraction of 0.71 (SIR 2.2), Denver 0.44 (2.0), Indianapolis 0.38 (1.7, 9.3-year payback), while low-load Seattle cases never paid back (SIR 0.26–0.56) and Honolulu’s $0.35/kWh electricity made every case pay (SIR up to 7.5). Large, consistent, centrally served loads, roofs with 20–25 years left, and electric (not gas) reheat are the targets; collector lifetime ~25 years. Annual resource across the test cities ranged 5.0–8.5 GJ/m²·yr.

Compound parabolic concentrators (CPCs) — concentration without tracking

A plain parabola focuses only light arriving at exactly 90° to its aperture onto a single point, so it must track the sun and is useless on a cloudy day (off-axis = no focus, like a magnifying glass aimed wrong). But a real cooling target is a finite object — a steel pipe carrying water — not an infinitesimal point. Once you only require light to land somewhere on that pipe, the ideal mirror shape changes: you can accept light from a range of input angles at once.

That is the compound parabolic concentrator (CPC), a non-imaging reflector optimized to intercept ~100 % of light entering within a chosen acceptance angle, concentrating it onto the target. Its virtues map directly onto the solar-cooling problem:

No sun-tracking. A fixed CPC trough collects all day within its acceptance window — removing the motors/controls that make tracked CST costly and failure-prone at building scale.
Diffuse-light capture. Because it accepts a wide cone, it gathers scattered light on cloudy days, when a tracked parabola has “nothing to aim at.”
Tunable trade-off. A 20–40° acceptance angle is the practical band: tighter = higher concentration, better for direct sun; wider = more diffuse/cloudy tolerance, lower peak temperature. (A flat 180° does not work.)

The NightHawkInLight source demonstrates a math-free “string construction” (speed square, guide rail, low-stretch braided line unwrapped from a plug cut to the target’s cross-section) to draw the exact CPC curve for any target shape — tubular absorbers (3-D target suspended above the trough) or a flat absorber at the trough base (a third line between the acceptance lines sets the absorber width) — making the geometry buildable from a wooden frame lined with aluminum flashing or peel-and-stick mirror film, with a clear plastic cover for a greenhouse effect and wind shielding, all without optics software. A crude cloudy-day comparison put the CPC target at ~91 °F vs ~72 °F for an ordinary parabola (≈60 °F ambient) despite the CPC carrying far more thermal mass — illustrative, not a calibrated figure, and cloudy-only: the host notes a sun-aimed parabola reaches far higher peak temperatures. The CPC’s edge is no-tracking and diffuse-light performance, not peak concentration. (The cited CPC review papers and the Power From The Sun ch. 9 treatment are the rigorous follow-ups.)

Where it fits the cooling chain: the CPC is a concentrator-side front-end that lifts collected water into (or above) the adsorbent desorption band without tracking hardware, sitting between cheap flat-plate (no concentration) and tracked CST (high concentration, high complexity). It pairs naturally with the absorber-surface upgrades below — a selective or geometrically-trapping absorber at the CPC focus stacks concentration gain with absorptance gain.

Concentrating solar thermal — what the power-plant handbook says

Wang’s design handbook (built around China’s 1 MW Badaling tower pilot) is a reference pointer for the high-temperature end, and a few of its numbers matter even at building scale:

The concentrator field is the dominant capital item — 45–70 % of primary investment — and the annual-mean field efficiency is 58–72 %; a 1 % gain in system efficiency cuts levelized electricity cost ~8 %. Whatever the scale, the optics are where the money goes.
Parabolic trough (commercial since the 1980s, plants to ~80 MW) has a low concentration ratio of 70–80, which caps fluid temperature; its ~4 m receiver tubes (≈70 mm steel absorber inside a 115–125 mm evacuated glass shield with getter and glass-metal seals) are the industrial ancestor of the evacuated-tube collectors sold for solar cooling. Dish-Stirling reaches concentration ratios of 600–3000 and 750 °C at 5–50 kW per unit; tower receivers see 350 kW/m² at the focal point.
Land: a trough field occupies ~2.5× its aperture area, a tower field 4–6× its heliostat area.
Soiling is measurable and large. Clean heliostat reflectance is ~0.93–0.94; at Badaling dust cut it from 94.6 % to 45.5 % in seven weeks, and one rain restored it to 82.1 %. Any outdoor optical surface — mirror, glazing, or textured absorber — needs a cleaning plan (see the soiling caveat in the bed-face thesis).

The design lever for cooling

For driving adsorption cooling, the useful figure is delivered water temperature at adequate flow across the real operating day, not peak efficiency at noon. The flat-plate sources emphasize reducing top-loss (honeycomb/transparent insulation, selective absorbers) to hold the collector in the adsorbent’s desorption band for more hours — which directly extends chiller run-time, and is why the honeycomb collector’s high-inlet-temperature advantage is the relevant one. This is the same surface-physics that the radiative / black-body thread treats from the absorption side: a good solar absorber is a near-black surface in the solar band.

Boosting capture with black-body / selective absorber surfaces (experiment thread)

The collector’s output-per-area is set at the absorber surface — so the surface is the first lever to pull to get more drive heat from the same roof footprint. Two ideas from the black-body thread are worth experimenting with on the collectors that drive 601’s adsorption chiller:

Near-black absorbers maximize solar absorptance across the spectrum. The registry’s vertically aligned single-walled carbon nanotube absorber (Mizuno et al., 2009) is the extreme case (among the blackest materials known); practical analogues are high-absorptance black coatings and textured/structured surfaces.
Spectrally selective absorbers go further: high absorptance in the solar band (0.3–2.5 µm) but low emittance in the thermal IR, so the hot absorber doesn’t re-radiate its gains back out. This is the standard high-performance evacuated-tube trick and the highest-leverage surface upgrade.

Why it matters here: raising effective absorptance/selectivity lifts the ~350–500 W·m⁻² baseline used in the 601 chiller sizing — meaning either fewer m² of collector for the same chiller drive, or more drive heat (and more cooling) from the planned array. It’s a cheap, surface-level experiment with direct downstream effect on adsorbent utilization — with the NREL caveat that collector hardware is a minority of system cost, so the gain must show up as delivered hours, not just a lab absorptance number. The inverse-surface caution from radiative cooling applies: a good solar absorber (high IR emittance, like a plain black body) and a good radiative cooler (high IR emittance to the sky) want opposite spectral behavior from a selective absorber — so test the actual spectral profile, don’t assume “blacker is always better.”

Geometric light-trapping — the razor-blade analog, scaled

There are two ways to make a surface black: material blackness (a coating that absorbs) and geometric blackness (a shape that traps light so it can’t escape) — the distinction at the heart of Geometric Blackness. The razor-blade stack is the classic demonstration of the second: viewed edge-on it is near-perfectly black because the V-shaped grooves between the sharp edges form wedge cavities, and incident light reflects multiple times down each wedge, losing a fraction to absorption at every bounce, so almost nothing escapes even though polished steel is a poor absorber on its own. It is the macroscopic cousin of the vertically-aligned CNT forest (a dense forest of microscopic light traps), of moth eyes and black-silicon cells, and of cavity blackbody standards (a small hole in a cavity is near-perfect black).

Everyday instance: a vinyl record is a V-groove array — molded ~90° grooves at very fine pitch in near-black PVC. It’s the familiar, found-object version of the geometry (handy as a cheap demo specimen), though shallow-angle and a poor production material (polymer, low conductivity, softens ~80 °C).

Scaling it to a rooftop solar-thermal absorber: replace the flat absorber plate with a macro V-groove / sawtooth (corrugated) absorber — the geometry traps sunlight by multiple bounces, boosting effective solar absorptance independent of the coating, and as a bonus increases absorber-to-fluid heat-transfer area. The razor-blade framing is the intuition for why and how aggressive to make the geometry. Best of both worlds: put a spectrally-selective coating on the groove faces so you stack geometric trapping (high solar capture, low-angle/diffuse tolerance) with spectral selectivity (low IR re-emission).

Design cautions for a real roof:

Orientation: orient the grooves to trap the sun across its daily arc; deep V-grooves especially help low-angle morning/evening and diffuse light — extending useful collection hours, which is what the chiller sizing actually values.
Debris/cleaning: grooves on a roof trap dust, pollen, and leaves — the same geometry that traps light traps dirt, and Wang’s Badaling data (reflectance halved in seven weeks) shows how fast soiling bites. Plan groove pitch/angle and cleaning access accordingly.
Thermal loss: more surface area also means more area to lose heat — pair the geometry with selectivity and good top-loss control (glazing, honeycomb, or evacuated envelope), or the radiative/convective losses eat the absorptance gain.

The build-and-test design that resolves these for 601 — panel matrix, the aperture-not-developed-area fairness rule, three test tiers, and the loss/soiling penalties — is at 601 · V-Groove Absorber Experiment. The quantified physics (the cavity absorptance formula, its emittance penalty, and the selective-coating benchmark) is worked through in the bed-face thesis, which concludes that geometry pays on the collector side, not on a direct adsorption bed.

Open-research thread — geometricblackness.org. The same 601 Delaware team runs an open-access project (geometricblackness.org) testing the specific hypothesis that 60° equilateral cavities (a Buckminster Fuller geometry) out-trap flat or arbitrary-groove absorbers — i.e. that there is a best groove angle, not just “deeper is darker.” It frames the same two-routes-to-black distinction (material vs geometric), cites vinyl records, moth eyes, razor blades, and black-silicon cells as geometric exemplars, and stages experiments by cost from ~$65 cardboard prototypes up to thermal measurement of metal absorbers, publishing failures as well as successes. For this KB it is the live experimental backing for the V-groove absorber idea above: if a coating-free 60° cavity geometry reaches high absorptance on its own, it lowers the cost of the collectors feeding the chillers.

Sources

Zotero cooling source registry — the solar-thermal shelf (flat-plate overview, Flamant CST, Wang, NREL/GSA honeycomb) and the black-body shelf (Mizuno CNT absorber, blackbody references) this article draws its collector spectrum from
Wang — Design of Solar Thermal Power Plants (2019) — CSP concentrator economics, trough/tower/dish parameters, receiver-tube construction, heliostat soiling data
NREL — High-Performance Flat Plate Collector Evaluation (2016) — honeycomb transparent-insulation collector field data, high-inlet-temperature advantage, performance equations, cost-effectiveness matrix
GSA — Honeycomb Solar Thermal Collector, GPG Findings 027 — companion brief: up to 8 % gain, 152 °C stagnation, collectors ~20 % of cost, deployment guidance
Geometric Blackness — shape-based light trapping — material vs geometric blackness, exemplars, the 60° cavity hypothesis and staged experiments
NightHawkInLight — DIY compound parabolic concentrators — CPC principle, acceptance-angle rule of thumb, string construction, build materials, cloudy-day comparison