Photovoltaic module lamination is performed in a vacuum membrane laminator in which a glass sheet, an ethylene-vinyl acetate encapsulant film, the cell circuit, a second EVA film, and a backsheet are stacked on a heated platen. Prior to the application of the flexible membrane pressure, the chamber is evacuated to a residual pressure in the range of 30–100 Pa and the stack is heated through the glass to a platen set point of 145–155 °C. Gas emission before the peroxide cure is established is a critical process variable because any volatile species that is not removed during the vacuum dwell or dissolved into the molten EVA will form a bubble at the glass–cell or cell–backsheet interface. The defect threshold is low; IEC 61215-2:2021, MQT 01 requires that no major visual defect such as delamination or bubble be present after lamination and after environmental testing. Production-scale laminators typically operate with a vacuum dwell of 3–6 min followed by membrane pressing at 30–70 kPa for 8–12 min. If the gas generation rate at the onset of heating exceeds the ability of the vacuum system to remove the gas before the EVA films soften and seal the cell, the remaining gas is trapped. The physical state of the EVA changes from a semicrystalline film with a melting peak near 65–85 °C to a low-viscosity melt before the peroxide decomposition accelerates at temperatures above 120 °C. This sequence defines the pre-cure outgassing window. EVA encapsulant films are typically based on ethylene-vinyl acetate copolymers with a vinyl acetate content of 28–33 wt% and are supplied as rolls with a thickness of 0.40–0.46 mm; the film contains an organic peroxide, a silane coupling agent, a co-crosslinking agent, and stabilisers. The peroxide is present at low concentration, generally below 1.5 wt%, but its decomposition products are volatile enough to contribute to gas evolution during the thermal transient. The melt flow index of EVA encapsulant grades, measured according to ASTM D1238 at 190 °C and 2.16 kg, is commonly in the range of 15–45 g/10 min, which indicates the low viscosity that permits gas bubbles to migrate during the pre-cure phase. If the gas is not removed, the bubbles are frozen into the laminate as crosslinking advances and the EVA elastic modulus increases.

What Gas Species Evolve from the Encapsulant Film Before Crosslinking Advances Beyond the Gel Point?

The principal gas species observed before crosslinking advances are water vapour, methanol from silane coupling agent hydrolysis, acetic acid from vinyl acetate hydrolysis, and low-molecular-weight oxygenated hydrocarbon decomposition products from the organic peroxide. Water is the largest volumetric contributor under most production conditions. EVA films are hygroscopic; moisture uptake after removal from sealed packaging can reach 0.05–0.25 wt% depending on ambient humidity and residence time. During heating, the water desorption rate increases rapidly between 40 °C and 110 °C, with the peak mass-loss rate in thermogravimetric analysis commonly occurring before the EVA melt is fully developed. Karl Fischer coulometric titration according to ASTM D6304-20 is used to quantify the moisture content of incoming film and backsheet. Methanol evolution arises from alkoxy silane adhesion promoters such as vinyltrimethoxysilane and 3-methacryloxypropyltrimethoxysilane; these additives hydrolyse in the presence of moisture and release methanol at temperatures in the 90–130 °C range. Acetic acid can be liberated from the hydrolysis of vinyl acetate comonomer in the EVA backbone when both heat and moisture are present, with a higher onset temperature near 120 °C. The peroxide decomposition products—primarily tert-butanol, acetone, methane, and minor alkenes—are generated from the dialkyl peroxide initiator, typically 2,5-dimethyl-2,5-di(tert-butylperoxy)hexane with a 10 h half-life temperature near 117 °C and a 1 h half-life temperature near 138 °C, as the laminate approaches the cure temperature. These decomposition products are not emitted exclusively after cure; a fraction is released during the thermal transient before the gel point is reached, which is why real-time residual gas analysers on vacuum laminators often show a pressure rise between the initial water desorption peak and the later cure exotherm. The relative proportions of these species depend on the silane type, the peroxide half-life characteristics, the moisture content of the film, and the heating rate. Published data for specific encapsulant formulations under defined lamination heating profiles is limited, but the qualitative order of evolution is consistent across EVA grades: water and trapped air first, methanol and acetic acid next, and peroxide decomposition products last before the gel point.

Vacuum Dwell, Heating Rate, and Bubble Nucleation at the Glass–Cell Interface

During the vacuum dwell phase, the gas removal rate is controlled by the chamber conductance, the pumping speed of the dry rotary vane or screw pump, and the time-dependent outgassing flux from the stack. The laminated stack is not homogeneously heated; the glass side heats faster than the backsheet side, creating a transient temperature gradient of 5–15 K across the package. This gradient means that EVA near the glass reaches its softening and melting range before the backsheet-side EVA, and gas evolved from the backsheet-side materials must diffuse through a partially melted polymer layer or travel laterally to the stack edge. If the vacuum dwell is too short, a bubble is preferentially nucleated at the glass–cell interface, where the glass surface is smooth and the cell surface provides a high-energy solid substrate. The bubble nuclei may be submicron, but they grow during the subsequent pressure step because the dissolved water vapour and peroxide decomposition products come out of solution as the temperature continues to rise and the pressure inside the membrane chamber increases. Laminator recipes for EVA encapsulants are therefore developed with a heating plate set point and a dwell time that produce a melt-state outgassing period of at least 3 min before the membrane pressure is applied. The exact window is material-specific, and published data for individual encapsulant grades under defined production heating profiles is limited; qualification runs commonly use pressure-rise monitoring during the vacuum dwell to infer the gas load. In a single-chamber laminator, the chamber is evacuated and heated in the same volume, so volatile species removed from the stack are continuously pumped away. In a two-chamber laminator, the stack is heated in one chamber and pressed in a second chamber, which separates the outgassing phase from the pressure phase and reduces the effect of gas generation during pressing. The vacuum system is typically sized for a base pressure of 5–10 Pa and pumpdown to 50 Pa in less than 60 s; the exact value depends on the chamber volume and the leak rate of the membrane seals. Heating rate is also significant: a heating plate temperature of 155 °C can generate a glass-side heating rate of 10–20 K/min after insertion, and an excessive heating rate can shorten the pre-cure outgassing window because the EVA melts before the pressure-independent gas release is complete.

When Backsheet Outgassing Overlaps EVA Melt Flow Before Peroxide Decomposition Becomes Significant

The contribution of the backsheet to pre-cure gas emission is often underestimated because its mass is lower than the glass and EVA, but its outgassing occurs on the side opposite the heated platen and is released after the glass-side EVA has already softened. In a single-chamber laminator, the backsheet side faces the flexible membrane and is not directly heated by the platen; heat transfer through the stack delays backsheet-side gas release until the membrane pressure is about to be applied. Backsheet constructions such as PVF/PET/PVF, PVDF/PET, or PET with a weatherable coating contain a polyester terephthalate core that absorbs water at a level on the order of 0.2–0.5 wt% at 50% RH; this water is released during the heating ramp and joins the water flux from the EVA films. Solvent-borne primer and adhesive layers between the backsheet films may retain residual ethyl acetate or toluene if the coating and lamination drying steps were incomplete; these solvents are released at temperatures below the peroxide cure onset and can create localised bubble defects if the backsheet-side EVA has already sealed an edge. Fluoropolymer outer films such as polyvinyl fluoride and polyvinylidene fluoride have lower moisture permeability than PET, but the backsheet as a composite remains a measurable outgassing source. The gas emitted from the backsheet is particularly problematic when the vacuum dwell time is already marginal or when the backsheet has been stored in a high-humidity environment. Pre-drying of backsheet rolls at 80–100 °C for 4–12 h is sometimes performed when lamination bubble rates exceed the production specification, but this operation is limited by the thermal shrinkage of the backsheet and the potential for blocking of the roll. The moisture content of the backsheet before lamination is measured by the same Karl Fischer method used for EVA, and a typical control limit is below 0.5 wt% for PET-containing constructions. If the backsheet outgassing flux overlaps with the EVA melt flow before the peroxide decomposition becomes significant, the gas is dissolved or dispersed in the melt; as the membrane pressure increases, the solubility of water in polyethylene-like segments decreases and the gas can be rejected at the cell metallisation interface, producing a line of edge bubbles. This failure mode is observed on production lines as a function of backsheet batch and storage history, not solely as a function of encapsulant quality.

Pre-cure gas emission is measured by coupling thermogravimetric analysis with mass spectrometry or Fourier-transform infrared spectroscopy; a representative method uses a heating rate of 10 K/min from 25 °C to 200 °C under nitrogen with a mass spectral scan from m/z 1 to m/z 100 to track water at m/z 18, methanol at m/z 31, and acetone/tert-butanol fragments at m/z 58. Thermogravimetric analysis is performed according to ISO 11358-1:2022. Headspace gas chromatography–mass spectrometry according to ISO 16000-6:2021 is used for identification and quantification of the higher-boiling volatile organic compounds emitted from the encapsulant film. Total mass loss and collected volatile condensable material are assessed by ASTM E595-15 when the laminate materials are screened for vacuum compatibility; the standard specifies a total mass loss of 1.0% and collected volatile condensable material of 0.1% as the typical acceptance thresholds for spacecraft materials, which are not directly applicable to photovoltaic laminates but provide a comparative index of film cleanliness. In the lamination facility, a residual gas analyser mounted on the vacuum line can record the partial pressure of water, oxygen, nitrogen, and organic fragments during the dwell phase. The pressure-rise test is performed by isolating the chamber at the end of the vacuum dwell and recording the rate of pressure increase; a significant pressure rise after the first evacuation indicates that the laminate continues to evolve gas, and the dwell time is extended until the pressure-rise slope falls below the equipment-specific control limit. The table below summarises the species evolved before peroxide curing, their analytical detection, and their process relevance.

The quantitative interpretation of pre-cure gas emission data is complicated by the fact that the laminator environment is not an equilibrium system. Published data for specific backsheet outgassing rates under production lamination heating profiles is limited, and the relative contribution of each species varies with the moisture exposure history of the rolls, the silane formulation, the peroxide loading, and the laminator vacuum conductance. The process window for EVA lamination is narrow; a platen temperature deviation of ±5 °C can shift the gel content measured by ASTM D2765-16 from a target range of 75–90% to an undercured or overcured state, and the same temperature deviation changes the pre-cure gas generation sequence sufficiently to alter the bubble defect density. Laminator qualification therefore includes not only mechanical and electrical tests but also lamination runs with deliberately aged films and backsheets to verify that the vacuum dwell and pressure recipe can remove the gas load before crosslinking locks the bubbles into the module.