Joule Heist search

topics · updated 2026-09-25

Translucent Phase-Change Materials for Glazing — the Clarity Ranking

confidence: medium volatility: warm verified: 2026-09-25fresh

Which phase-change materials are translucent enough to sit in a window, and how they're packaged. Ranking: salt hydrates (CaCl2·6H2O) are the only bulk PCM naturally clear in BOTH solid and liquid phases; paraffins are a thermotropic switch — clear liquid, opaque solid (6% → 64% measured for RT21HC) — unless engineered clear via refractive-index matching or grain control; sugar alcohols melt far above comfort and fatty acids offer no clarity edge. Opacity is polycrystalline light scattering from a refractive-index mismatch; three clarity routes exist (index-matched shells/matrices, grain-size engineering, confining hydrogels), reaching 74–90% VLT. Commercially, GlassX Crystal (salt hydrate in polycarbonate) is the only mature translucent-PCM glazing — and it ships as full IGU replacements, NOT slot-in retrofits for existing frames. Numeric haze and >1000-cycle optical durability remain unreported.

The short answer. The most translucent PCM depends on whether you need clarity in one phase or both. Salt hydrates (chiefly CaCl₂·6H₂O) are the only common bulk PCMs that stay see-through as a solid and a liquid — they crystallize into transparent ionic crystals rather than light-scattering wax, which is exactly why the one mature commercial product (GlassX Crystal) uses one. Paraffins (e.g. Rubitherm RT21HC) are only thermotropically translucent: clear when molten, opaque/white when frozen — useful as a self-dimming smart window, but not a permanently clear one unless engineered. Sugar alcohols melt far above room temperature and fatty acids offer no clarity advantage. The best engineered transparent formulations — index-matched microcapsules, confining hydrogels — reach 74–90% visible transmittance, but every clarity trick trades against latent-heat density or adds cost.

This article answers a specific applied question: what could go in (or behind) an existing window and stay see-through while buffering heat? It sits at the intersection of the envelope & glazing load-reduction layer and latent thermal storage.

The clarity ranking, by chemistry

Ordered by inherent translucency (before any engineering):

  1. Salt hydrates — the clarity winner. The 2015 Springer review makes material selection optics-first and singles out high-water-content salt hydrates as transparent in both phases, with CaCl₂·6H₂O (24–29 °C, 192 J/g) sealed in clear polycarbonate as the dominant packaging — high water content is favored because its transmittance spectrum matches eye sensitivity. This is why the daylighting literature reaches for salt hydrates and why GlassX uses one. Caveat: salt hydrates supercool and phase-segregate, which clouds and degrades them over cycles unless stabilized with nucleators/thickeners or held in a gel.
  2. PEG-type polymers — the closest transparent-composite prior art outside salt hydrates is a carboxymethyl-cellulose form-stable PEG PCM with reversible transparency (a lead noted in the sugar-alcohol survey, not retrieved). Treat as promising but unquantified here.
  3. Paraffins (n-octadecane C18, RT-series) — switchable, not permanently clear. Highly transparent as a liquid, opaque/white as a solid; Heim et al. measured 6% → 64% direct-light transmittance across the melt of RT21HC. They become 74–88% clear only via engineered refractive-index matching or grain-size control (below). Cheapest and most-studied PCM chemistry otherwise.
  4. Sugar alcohols (erythritol, xylitol, dulcitol) — bio-based and very high-enthalpy (xylitol ~238 J/g, erythritol 316–353 J/g), and their optics are more interesting than “opaque like paraffin”: the state is phase-history dependent. A polycrystalline solid is cloudy, but because these polyols supercool deeply (erythritol ~47 K; xylitol at 100 g “never crystallizes after melting once”) and often vitrify to a clear glass instead of crystallizing, a freshly cooled sample can stay glassy-transparent and only clouds when it cold-crystallizes on reheating — the reverse of paraffin. This is moot for windows: melt points run ~90 °C (xylitol) to 105–118 °C (erythritol) and up to 190 °C for dulcitol/starch composites, a medium-temperature storage class far above any comfort setpoint. Ruled out for glazing regardless of clarity.
  5. Fatty acids / esters — the review notes stearic acid as a transparent-insulation candidate, but these semicrystalline organics offer no clarity edge over paraffin, and the durability survey flags oxidative yellowing as their characteristic failure.

Source disagreement, annotated. The 2015 review states that “only organic PCMs and high-water-content salt hydrates are transparent as both solid and liquid.” The primary optical measurements compiled here (Heim 2021; the Polymers 2022 silicone–paraffin study; Otaegui 2024) all show bulk paraffin is opaque as a solid. Read the review’s “transparent” as “translucent/light-passing” — a frozen paraffin layer still transmits diffuse daylight — not as see-through. For an always-clear window only the salt hydrate (or an engineered composite) qualifies.

The practical split: if you want a window that is always clear, you want a salt hydrate (ideally gel-confined). If you want a window that auto-dims when it gets hot (clear in the morning, milky-white in the afternoon heat), an ordinary paraffin does that for free — the opacity is the solar-shading feature.

Why solid PCMs go opaque — the physics

Transparency is governed by refractive-index homogeneity at the scale of visible wavelengths. A single-phase liquid (or an amorphous solid) with uniform index transmits light; scattering arises wherever there are index discontinuities.

When a paraffin freezes it forms a polycrystalline solid — randomly oriented crystallites plus grain boundaries, each a Mie/Rayleigh scattering center (strongest when the feature size ≈ λ_visible); the Polymers authors state it plainly: opaque “due to their semicrystalline and polycrystalline nature.” The liquid, being isotropic and index-uniform, is clear. Heim et al. measured the index step directly for RT21HC: liquid n_D ≈ 1.434, solid n_D ≈ 1.531, glass ≈ 1.52 — the solid PCM no longer matches the glass, so it scatters; the liquid does match, so it transmits. Salt hydrates dodge this because they crystallize as transparent ionic crystals whose index stays close to the surrounding medium. Sugar alcohols show the same rule from the other side: their vitrified (amorphous) solid is clear precisely because it has no grain boundaries.

Heim’s kinetics matter for a real window: the melt is strongly non-linear. Transmittance stays below 0.10 through the long “mushy” stage, then climbs 0.1 → 0.64 in 30–40 minutes as the layer clears, with plateaus from interreflection between solid, mushy, and liquid sub-layers. Full melt of the 16 mm layer took 105–135 min at 1000 W/m² and 220–250 min at 500 W/m² — irradiance sets the speed, not the end-state transmittance. Placing the PCM in the inner cavity of the triple glazing capped transmittance at 0.57 instead of 0.64. The authors’ warning: transmittance-versus-temperature “cannot be treated as a linear function.”

Three engineering routes keep the solid transparent:

  • Refractive-index matching. Match the PCM’s index to its host/shell so the crystal↔matrix step vanishes. The ACS Omega 2024 smart window matched a liquid PCM (n ≈ 1.43) to its polymer matrix (n ≈ 1.42) for 74.5% transmittance (“like ordinary glass”) when molten; on freezing the crystalline PCM stops matching and, with a color-changing dye, transmittance falls to 0.2% — a passive-dimming window by design. The Nanomaterials 2026 n-octadecane@SiO₂ composite instead uses an index-stable silica shell to buffer the paraffin’s own index swing, holding ~83.75% VLT at 550 nm through the transition. Both papers state the same design rule: the encapsulant’s index must match the PCM or switchable transmittance is lost; and because Mie scattering dominates, capsules must be submicron.
  • Grain-size engineering. Push crystal features away from λ_visible. Otaegui et al. (ETH Zurich) found odd-carbon-number paraffins grow large grains (fewer boundaries per volume → far less scattering), and paraffin dispersed as ~30 nm nanoparticles in a host polymer stays far below λ so scattering is negligible (~80–88% VLT, 400–750 nm, through the transition). The retained transition still modulates optical phase by up to π radians — the melt is being used optically, not just thermally.
  • Confining hydrogels / shape-stabilizing matrices. A salt-hydrate PCM held in a poly(acrylamide-co-acrylic acid) hydrogel stayed transparent in both phases — 90% visible, 78% total solar transmittance — because the network constrains crystal domains and prevents leakage (J. Energy Storage 2025). Silicone–paraffin blends apply the same shape-stabilization idea to a thermotropic film: RT44HC in PDMS gives a 65–73 percentage-point clear↔opaque swing (2 mm PDMS-P5-C1 = 73.3 pp), 250 kJ/kg, with CTAB surfactant holding exudation to ~0.3 wt% — though the blend melts at 35–39 °C, a demonstrator rather than a comfort product.

Getting it into a window — and the retrofit reality

The 2015 review taxonomizes four packaging methods:

  1. Between-pane fill — PCM directly in the cavity of a double-glazed unit (Heim’s 16 mm RT21HC layer is this).
  2. Container behind glazing — PCM in sealed transparent plastic containers with ~10 mm air gaps behind existing double glazing.
  3. Glass blocks / bricks — hollow translucent blocks filled with PCM as thermal-storage wall elements.
  4. Translucent PMMA / polycarbonate panels — PCM inside a plastic panel in a glass façade (e.g. Delta-Cool 28 in PMMA).

The blunt finding for “use in existing window frames”: true slot-into-an-existing-frame retrofit PCM products do not exist — confirmed by a patent + product sweep. Transparent PCM ships as complete factory-sealed IGUs/modules (GlassX) or as PMMA/polycarbonate panels mounted in a façade — replacements, not frame inserts. The patent record splits cleanly into non-PCM retrofit glazing on one side and full PCM-IGU replacements on the other:

  • US 4,813,203 “Retrofit glazing system” (Newman et al., NPH Partnership, 1989) — retrieved in full; a purely mechanical aluminum sub-frame plus a factory “miniframe” silicone-bonded to the glass, exterior-installed, with no PCM at all. This settles the earlier open lead: it is not a PCM design.
  • US 9,797,187 B2 (Carnegie Mellon) — the closest genuine “translucent PCM at a window without replacing the IGU”: transparent hollow vessels (PP/polycarbonate/acrylic/glass, internal conductive mesh) that are deliberately see-through but modular and relocatable (louvers, honeycomb networks, venturi-tube-with-fan, cable-hung containers), not a frame insert.
  • US 2010/0244495 A1 (GM) — automotive PCM window treatments (paraffin, salt hydrate, poly-alcohols, eutectics) attached by suction-cup housings, roller shades, pivoting visors, with three containment options (between two films, self-encapsulated beads, beads embedded in glass/plastic); abandoned, but the clearest articulation of “attach a PCM layer to already-installed glass.”
  • Fang et al. 2013 (Energy & Buildings) tests real retrofit secondary glazing for single-glazed frames (plastic film, magnetically attached sheet, plain and low-E glass; R 0.15 → 0.34–0.57 m²K/W) — but the field stops short of adding PCM.

The realistic retrofit is therefore a DIY/bespoke composite of these: method 2 — a sealed, PCM-filled polycarbonate/acrylic panel added as a secondary glazing layer behind the existing sash, borrowing the magnetic/clip/suction attachment of non-PCM secondary glazing and film-enclosed or encapsulated-bead containment. Leak-proof sealing and freeze-thaw durability are the design problem, not the PCM. The container/sealant/chemistry choices and failure modes for building one are worked out in DIY PCM Encapsulation for Secondary Glazing.

Containment and durability recur as the real constraints: salt-hydrate supercooling and phase segregation (need nucleators/thickeners — the review records one salt-hydrate formulation unchanged after 5,650 cycles), leakage prevention via sealed polycarbonate/PMMA channels or micro-encapsulation-in-matrix, incomplete nighttime discharge, and low thermal conductivity (0.15–0.5 W/m·K) that caps how much of the latent capacity can actually charge and discharge in a day.

GlassX Crystal — the one mature product

GlassX Crystal (GlassX AG, Switzerland, ~15 years of development; licensed to Saint-Gobain, manufactured in Austria/Germany; N.A. distributor Greenlite Glass Systems, Vancouver) is the flagship translucent-PCM glazing. A translucent salt-hydrate PCM sealed in polycarbonate channels, melt/freeze 26–30 °C. A four-part sandwich pairs the PCM inner IGU (low-e coatings, low-conductivity gas fill) with an outer IGU carrying a suspended prismatic (Fresnel-like) filter that reflects high-angle summer sun and admits low-angle winter sun. Direct-beam transmission runs up to 45% (PCM liquid) → ~28% (crystallized); heat storage 1,185 Wh/m² (the review’s 4,268 kJ/m² is the same figure; ~9 in of concrete); U ≈ 0.48 W/m²·K; ~79 mm thick, 95 kg/m²; the review records a claimed 30–50% load reduction. It ships as full factory-sealed modules up to 110×59 in (not a retrofit insert), with 25+ European installs by 2010 (housing, offices, retirement homes), $560–970/m² and a 5–10 yr payback. BuildingGreen’s open question — efficiency over thousands of cycles — is still the open question below.

Performance numbers at a glance

Material / systemMelt °CVLT solid → liquidLatent heatNotes
Salt hydrate in hydrogel (J. Energy Storage 2025)32.8~90% in both phases (78% solar)133.3 J/gClearest; <10 J/g loss / 500 cyc
n-Octadecane@SiO₂ microcapsules (Nanomaterials 2026)~28~83.75% @550 nm through transition155.3 J/gSubmicron index-stable shell; 100-cycle data
Paraffin nanoparticles in polymer (Otaegui 2024)—~80–88% through transition—~30 nm particles; odd-carbon grain growth
Index-matched polymer PCM (ACS Omega 2024)35–42 (pure TD)0.2% → 74.5%161.9 J/gSelf-dimming; ~1000 cyc
Silicone–paraffin film RT44HC (Polymers 2022)35–39 (blend)65–73 pp swing250 kJ/kgAbove comfort; leak-resistant demonstrator
Paraffin RT21HC in triple glazing (Heim 2021)216% → 64%190 kJ/kgPrimary measured swing; ~10×
GlassX Crystal (salt hydrate, commercial)26–30~28% → 45% (direct beam)1,185 Wh/m²Full IGU; prismatic filter
CaCl₂·6H₂O in polycarbonate (2015 review)24–29translucent both phases192 J/gDominant packaging; 5,650-cycle formulation
Rubitherm RT18HC–RT28HC (paraffin reference)18–28switchable (opaque solid)165–260 kJ/kgComfort-range melt anchors; HC grades +25–30%

The Rubitherm datasheet fixes the comfort-band anchors the transmittance papers cite by name: RT18HC 260, RT21 165 / RT21HC 190, RT22HC 190, RT24HC 200, RT25 180 / RT25HC 230, RT28HC 250 kJ/kg. “HC” grades carry 25–30% more latent heat and melt in a narrower band; RT21HC is the most-studied glazing PCM.

The ACS Omega work also quantifies two effects a builder should expect. Encapsulation shifts and widens the transition: pure PCM melted 35–42 °C and crystallized 28–36 °C, but in the polymer matrix the ranges widened to 27–49 / 20–37 °C through altered nucleation kinetics. And additives trade clarity for function: dye particles add heterogeneous nucleation sites that cut supercooling but scatter light, so raising photothermal dye loading dropped clear-state transmittance 74.5% → 23.3%.

Haze and cycle durability — what the data does (and doesn’t) say

A dedicated survey of these two axes confirms both are thinly reported. Numeric ASTM D1003 haze for PCM glazing is genuinely sparse — the field reports luminous/solar transmittance, and explicit haze values appear almost only in thermochromic/electrochromic (non-PCM) windows. As benchmarks: a biomass bistatic (electrochromic, non-PCM) window achieved haze <1% with 1000 cycles “fairly constant” — the clarity ceiling PCM systems aspire to — and a thermoresponsive (non-PCM) hydrogel switched T_lum 92% → 0.4% but was tested only 50 cycles. Two paywalled papers would supply real haze tables: the n-octadecane@SiO₂ paper and an optical-modulation window reporting ~95% haze at 95.8 J/g.

Optical cycle durability is measured but shallow (typically 50–1000 cycles, often qualitative):

  • n-Octadecane@SiO₂ microcapsules: the paper abstract reports 0.35% attenuation after 100 cycles, while the survey extracted 0.93% transmittance variation over the same 100 cycles — two raw notes on one paper disagree on the figure; either way it is sub-1%, and the RI-matched shell is engineered specifically against phase-transition scattering.
  • Salt-hydrate hydrogel: <10 J/g loss and minor melt-point shift after 500 cycles, 90% VLT stable in both phases.
  • Passive-dimming polymer PCM (ACS Omega): stable across ~1000 cycles on a single stability curve; no haze, no multi-cycle transmittance retention reported.
  • Silicone–paraffin films: exudation plateau measured over one week; no multi-hundred-cycle optical data.

Failure modes to design against: salt-hydrate container cracking from freeze-expansion stress (the Heim work also recommends ≤16 mm PCM layers to limit mushy-state inhomogeneous scattering), and fatty-acid oxidative yellowing. Long-term (>1000-cycle) optical-clarity retention and standardized haze remain the real open gaps in the literature.

Bottom line for an existing window

  • Want it always clear + buffering heat: a gel-confined salt hydrate (or GlassX-style salt-hydrate-in-polycarbonate) is the translucency leader. As a retrofit, add it as a sealed secondary glazing panel behind the existing sash.
  • Fine with auto-dimming: a plain comfort-range paraffin (RT21HC–RT25HC) in a sealed polycarbonate cassette is cheap, self-shades when hot, and is the most-studied option — accept that it goes milky-white on hot afternoons and that it takes hours of sun to clear.
  • Melt temperature is the real spec: pick it to your comfort setpoint (RT21–RT25 for ~21–25 °C rooms), not for clarity — clarity is then an engineering choice (index-matched shell or gel) layered on top. Note that the two clearest lab systems (hydrogel 32.8 °C, index-matched polymer 35–42 °C) currently melt above comfort.
  • Expect containment work: the hard part of any retrofit is leak-proof sealing and freeze-thaw durability, not the PCM itself.

See also

Sources