Joule Heist search

references · updated 2026-09-25

DIY PCM Encapsulation for Secondary Glazing — Containers, Sealing, Failure Modes

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

Practical reference for encapsulating a translucent PCM as a DIY secondary-glazing panel behind an existing window. Container choice: polycarbonate (twin-wall) is the best translucent option — resists paraffin oil and saturated CaCl2/Na2SO4; acrylic is clearest but crazes under stress; PVC-U is most inert (<0.1% mass change) but not clear; avoid thin LDPE/PP/HDPE with paraffin (swells up to 6.9%). Encapsulation scale: free-liquid macro-fill is easiest/clearest but leaks; the DIY sweet spot for staying translucent AND leak-free is a CMC-gelled salt hydrate (~11 wt% CMC, 130 J/g, zero supercooling); silica-stabilized paraffin goes opaque and loses 25-50% enthalpy. Sealing: standard silicone swells in paraffin — use clamped gaskets or oil-resistant/fluorosilicone; leave >=10% headspace for expansion; a true vapor barrier (>=5000-cycle bar) for salt hydrates; no bare metal near salt hydrates. Failure modes with numbers: leakage dominates (up to 57% capacity loss), supercooling (borax 4 wt% cuts 14->4 C; overheating >40 C kills nucleation), phase segregation (thickener), water-vapor loss, UV yellowing (seal air + block UV).

The translucent-PCM glazing article established that the realistic retrofit for an existing window is a sealed PCM panel added as secondary glazing behind the sash — and that containment, not the PCM, is the hard part. This reference is the how-to layer: which container, which encapsulation scale, how to seal it, and the failure modes to design around. It is compiled from PCM–plastics compatibility tests, salt-hydrate stabilization chemistry, 700-cycle thermal-cycling data, a full-scale fill-and-seal glazing prototype, a DOE encapsulation program note, and DIY/maker practice.

Scope note: this is materials/fabrication guidance synthesized from the literature and maker sources — not a validated build. No fully documented, measured amateur translucent-PCM window build was found (BuildItSolar’s page is explicit that it has no verified build or measured result; the Endless Sphere polycarbonate-channel method is concept-stage). That is a genuine gap this project could fill.

Step 0 — pick the PCM first (it decides everything downstream)

Two DIY-viable chemistries, and the choice cascades into container, sealant, and failure modes:

  • Comfort-band paraffin (Rubitherm RT21HC–RT25HC, ~21–25 °C) — cheap, non-corrosive, no supercooling, but thermotropic: clear when molten, milky-white when frozen (which doubles as solar shading). Favor RT22-type: 700-cycle data shows solidification-side latent-heat loss of only ~5% for RT22HC versus 23% for RT25HC (melting-side ~4% across grades). Paraffin permeates and swells polyolefins, so container and sealant choice matter.
  • Salt hydrate (CaCl₂·6H₂O ~26–30 °C, or Na₂SO₄·10H₂O “Glauber’s salt”) — translucent in both phases (better for a see-through panel), higher volumetric latent heat, but needs additive stabilization, is hygroscopic, and corrodes metals. This is the chemistry GlassX uses. BuildItSolar’s target recipe is CaCl₂·6H₂O with NaCl and KCl plus ~2% SrCl₂·6H₂O as nucleator, freezing ~26 °C and melting ~30 °C.

Goia et al.’s full-scale prototype (Solar Energy 2014) is the credible proof that “fill a sealed glazed cavity” works: an 8 mm outer glass / 15 mm cavity filled with RT35 paraffin / 6 mm inner glass unit, run outdoors for six months against a reference glazing, measurably improved the indoor thermal environment, with the benefit growing with solar irradiation and daylight transmission tracking the charge state. Its lessons for a DIY panel are direct — match the melt point to the comfort band (RT35 at 35 °C is above it), expect transparency to change with charge, and seal the cavity against paraffin.

Container — the translucent envelope

OptionClarityParaffinSalt hydrateDIY verdict
Twin-wall polycarbonatetranslucent (clear/opal)resists paraffin oilresists sat. CaCl₂/Na₂SO₄Best all-round — ready-made channels, greenhouse-cheap, seal the ends yourself
Acrylic / PMMA boxclearestOK if clean aliphatic waxOKBest optics, but crazes under stress + solvent-cut PCM — keep stress-free
Glass IGU cavitymax clarity, inert✓ (Goia)✓Not hobbyist-fabricable (needs sealed-unit making)
PVC-Unot clear<0.1% mass change<0.1%Most inert but opaque — only for hidden trays
HDPE / PP rigidtranslucent-to-opaqueswells up to 6.9%mild (≤1.8%)Fine for salt hydrate; paraffin uptake saturates after ~4 weeks but threatens seals
LDPE/PP film pouchtranslucentswells / permeatesmildCheap heat-sealable pouch for salt hydrate, not paraffin
PET bottle/tubeclearoil-attacked (mineral-oil “C” rating)okBuildItSolar’s 1-L bottles, walls ≥0.035 in; watch oily paraffin grades
10-mil polycarbonate film (Lexan)clear✓✓BuildItSolar’s see-through glazing-cavity option

Headline pairing: twin-wall polycarbonate is the DIY default — manufacturer chemical-resistance charts rate it resistant to both paraffin oil and saturated CaCl₂/Na₂SO₄. The one caveat: additives/plasticizers in solvent-cut paraffin oils can migrate into PC and craze it under stress, so use a clean wax-grade PCM. Acrylic wins on clarity but only if the panel is kept free of assembly stress (no over-tight clamps, relieve molded-in stress), because environmental stress crazing needs three things at once — tensile stress, a crazing agent (alcohols, ketones, esters, solvent-cut PCM), and a susceptible polymer — and daily melt-expansion cycling supplies the stress. The Rathgeber immersion tests underpin the polyolefin rows: no plastic showed visible crazing or dissolution, but HDPE and PP gained up to 6.9% mass in paraffin (most in the first four weeks, then saturating) against ≤1.8% in salt-hydrate PCM. The general rule: organic (paraffin) PCM plasticizes and permeates polyolefins; inorganic (salt-hydrate) PCM does not.

One thermal consequence of plastic containment, from the DOE note: polymer walls conduct only ~0.15 W/m·K, so a plastic-encapsulated panel charges and discharges slowly — acceptable for passive glazing, but it means a thick PC channel adds to the PCM’s own low conductivity (the Endless Sphere builder noted the same).

Encapsulation scale — free-liquid vs shape-stabilized

The core tension: staying translucent vs not leaking.

  • Free-liquid macro-fill (pour liquid PCM into a sealed channel/box) — easiest DIY, best clarity, but leak-prone and, for salt hydrates, prone to phase separation. This is Goia’s approach and the maker-forum method: silicone-seal both ends of extruded greenhouse multiwall polycarbonate, fill each channel ~90% with liquid paraffin leaving 10% for expansion (concept-stage, no measured result).
  • Shape-stabilized / gelled — PCM held in a matrix so it never flows even when melted. Here the chemistry splits sharply on clarity:
    • CMC-gelled salt hydrate = the DIY translucent sweet spot. Glauber’s salt thickened with carboxymethyl cellulose forms a 3D gel that pins salt and water together, killing the phase separation that ruins free-liquid salt hydrates. Molecular weight matters: low-MW CMC-90 at ~11 wt% gave the best result — 130 J/g retained after cycling, zero supercooling, still translucent — while high-MW CMC broke apart on salt addition. Recipe: heat stoichiometric water to 50 °C, stir in CMC at 800 rpm for 30 min, add salts sequentially, cool, refrigerate overnight. The latent-heat penalty is modest (~108–130 J/g against ~180 J/g theoretical for pure Glauber’s salt).
    • Silica-/graphite-stabilized paraffin = disqualified. About 9 wt% hydrophobic fumed silica stops leaks at 60–90 wt% PCM loading and melt-mixing is DIY-feasible with off-the-shelf materials — but the composites are fully opaque and deliver only 50–75% of theoretical enthalpy, because an immobilized layer at the pore walls never melts. Fine for a hidden thermal-mass panel, useless for a window.

Rule of thumb: for a see-through leak-free panel, gel a salt hydrate (CMC). For a self-dimming panel where the milky solid state is acceptable, free-liquid paraffin in sealed polycarbonate is simplest.

Sealing

  • Standard clear acetoxy RTV silicone swells and loses adhesion in molten paraffin. Use a mechanically clamped gasket or an oil-resistant / fluorosilicone RTV instead. (Salt hydrates don’t attack silicone, so ordinary silicone is fine on a salt-hydrate panel — the swelling problem is specific to paraffin.)
  • Rigid PC/acrylic: solvent or ultrasonic welding gives a monolithic seam; heat sealing works for PE-film pouches (but the film absorbs paraffin).
  • Always leave an expansion void (see below) — never solvent-weld a 100%-full cell.
  • Vapor barrier, not just a lid: the DOE/CMEI program names water-vapor permeation as a distinct salt-hydrate failure mode — bound water escapes through permeable coatings and energy density drops — so the coating or seal must block both liquid leakage and vapor permeation. BuildItSolar makes the same point from the maker side (salt hydrates are hygroscopic; seal against water gain and loss; it worked at ~35% RH). DOE’s durability bar for “stable” encapsulation is ≥5,000 cycles (~14 years at one cycle a day).

Failure modes to design around

FailureNumbersMitigation
Leakage (the dominant one)encapsulated paraffin samples lost up to 57% of solidification enthalpy over 700 cycles via PCM escaping the coating — not chemical breakdownSealed rigid cell, vapor-tight seams; must be leak-proof at peak summer temperature (liquid phase)
Volume expansionparaffin ~10% per melt; salt hydrates ~10%Leave ≥10% headspace (the maker method) or a flexible face; never fill 100%
Supercooling (salt hydrates)borax 4 wt% cuts CaCl₂·6H₂O supercooling 14 °C → 4 °C (and can approach 0 °C); SrCl₂·6H₂O → ~2 °CAdd a nucleator; don’t overheat >40 °C (irreversibly kills nucleation centers → permanent supercooling)
Phase segregation (salt hydrates)anhydrous salt settles under gravity each cycle, progressively killing capacityThickener — sodium polyacrylate gel or the CMC gel above
Corrosion (salt hydrates)stainless 316/304 safe (0–1 mg/cm²·yr); aluminium corrodes; copper caution (esp. CaCl₂/MgCl₂); metal fins → pinhole leaksKeep salt hydrate off bare metal — plastic/glass walls, no metal fins
UV / oxidative yellowinga sealed paraffin sample showed no yellowing and no FT-IR change at 700 cycles → yellowing is a UV/O₂-ingress problemSeal out air; add a UV-filtering or UV-stabilized outer layer on a sun-facing panel
Water-vapor loss (salt hydrates)bound-water loss drops energy density; also water content drifts with humidityTrue vapor-barrier seal; verify water content at fill

The standard salt-hydrate additive stack (from the Nano Select review): melting-point adjuster (KCl) + nucleator (borax ~4 wt%) + thickener (sodium polyacrylate or CMC). All cheap, DIY-grade. BuildItSolar’s SrCl₂·6H₂O at ~2% is the maker-community variant of the nucleator.

  1. Always-clear panel → CMC-gelled Glauber’s salt in a twin-wall polycarbonate sheet, ends clamped with gasketed caps, ≥10% headspace, borax or SrCl₂ nucleator, non-metal throughout, vapor-tight. Translucent both phases, leak-resistant, ~130 J/g.
  2. Self-dimming panel → free-liquid RT22HC paraffin in sealed twin-wall polycarbonate, channels filled ~90%, oil-resistant sealant or clamped gasket, UV-filter film on the sun side. Clear when cool, milky when hot (passive shading); simplest fabrication; the closest analog to Goia’s measured prototype.
  3. Avoid: silica/graphite-stabilized paraffin (opaque, 25–50% enthalpy loss), thin LDPE/PP/HDPE with paraffin (swells and permeates), acetoxy silicone against paraffin, and any bare aluminium/steel/copper touching a salt hydrate.

See also

Sources