Residual catalytic ash in capacitor-grade isotactic polypropylene originates from the MgCl₂-supported fourth-generation Ziegler-Natta catalyst architecture and from alkylaluminum cocatalyst decomposition products. A typical polymerization recipe contains a solid catalyst comprising 0.1–0.5 wt% Ti on MgCl₂ crystallites, an internal donor such as dibutyl phthalate or 9,9-bis(methoxymethyl)fluorene, triethylaluminum as cocatalyst, and an external alkoxysilane donor such as dicyclopentyldimethoxysilane or cyclohexylmethyldimethoxysilane. The reactor powder, before deashing, contains total ash levels between 80 ppm and 300 ppm when measured by ISO 3451-1:2019 at 600 °C; after liquid-phase washing with polar alcohols, neutralization of chloride-containing residues, and mechanical dewatering, deashed capacitor-grade grades typically fall between 5 ppm and 25 ppm total ash. Residual elements include magnesium, titanium, aluminum, chlorine, and silicon, with magnesium and chlorine usually dominating because the MgCl₂ support is not fully removed by alcohol washing. The particle size distribution of residual ash is skewed toward sub-5 µm fragments because the catalyst support fractures during polymerization and subsequent processing; when these fragments enter a 3.0 µm dielectric film, they act as high-permittivity inclusions with local electric field enhancement factors that can exceed 1.5 under direct-current stress. Chloride residues are particularly damaging in humid environments because hydrolysis generates hydrogen chloride and ionic conductivity within the film, accelerating electrochemical degradation of metallized electrodes. Therefore, the first control point for residual catalytic ash is not melt filtration alone but the deashing stage: a poorly washed reactor powder with 40 ppm total ash can overwhelm a 10 µm melt filter within 150 h of continuous extrusion, whereas a deashed powder with 12 ppm total ash may permit stable pressure drop for 800 h or longer on the same filter configuration.

Which Filtration Thresholds Reduce Macro-Gel Counts Below 10 Particles per kg?

Melt filtration for ultrathin PP dielectric film is configured not as a single absolute screen but as a graded depth-filtration train that sequentially removes large agglomerates, gel particles, and partially deactivated catalyst fragments. In a representative production line, a co-rotating twin-screw extruder with a 40:1 L/D ratio feeds a gear pump that discharges into a continuous screen changer equipped with breaker plates and candle filter elements. The first stage uses a 20 µm nominal stainless-steel mesh or sintered fiber web to protect the downstream layers; the second stage uses a 10 µm sintered fiber felt; the third stage, if installed, uses a 5 µm sintered powder cartridge. The practical limitation is that filtration is not a sharp size exclusion process: depth media retain particles by interception, inertial impaction, and adsorption within tortuous void networks, so a 5 µm nominal element may only remove 50–70% of 2 µm particles at typical melt temperatures of 230 °C to 250 °C, while 1 µm particles pass through without significant retention. For capacitor film, the key quality threshold is not the removal of all sub-micron ash fragments but the elimination of gel particles larger than approximately 20 µm, because these form surface protrusions and pinhole defects during biaxial stretching. Production-scale data from continuous extrusion lines indicate that a staged 20 µm/10 µm screen pack reduces macro-gel counts from approximately 80 particles kg⁻¹ to 15 particles kg⁻¹, while adding a 5 µm third stage lowers counts below 5 particles kg⁻¹ but increases screen-changer pressure from 120 bar to 170 bar and shortens filter life from 600 h to 200 h. However, published data for this specific configuration is limited; accepted best practice is to qualify filter media using a cast film trial rather than relying solely on nominal retention ratings. Melt pressure upstream of the screen pack is monitored by a strain-gauge pressure transducer; a pressure rise above 200 bar is commonly used as the change-out criterion to avoid rupture of the filter medium and release of accumulated debris into the melt stream.

Inside the cast web quenching zone, residual ash and gel particles produce characteristic disturbances in the melt puddle that are visible as transverse thickness bands after biaxial orientation. The cast film process uses a slot die with a lip gap of 0.5–0.8 mm, a chill roll temperature set point between 28 °C and 34 °C, electrostatic pinning, and an air knife to stabilize the web. Volatile residues such as silane hydrolysis products, low-molecular-weight oxidized oligomers, and chloride-containing species migrate to the die lip surfaces during extrusion and condense into deposits; these deposits grow over 4–8 h of continuous running and eventually detach as black or brown specks that become embedded in the cast web. The deposits are composed predominantly of oxidized isotactic PP wax, siloxane condensation products from external donor residues, and ash clusters cemented by polar residues. Maintaining die lip plate temperature 10–20 °C above the melt exit temperature delays condensation but may accelerate oxidative discoloration if oxygen is not excluded from the die exit area. The chill roll puddle is another accumulation zone: denser ash agglomerates settle into the slower-moving boundary layer and create local variations in cast web thickness, which later appear as thickness streaks in the 2.0 µm film after machine-direction stretching of 4.5:1 to 5.5:1 and transverse stretching of 8:1 to 10:1. If an ash agglomerate larger than 5 µm lies within 200 nm of the cast web surface, it can nucleate a void during stretching because the local strain hardening around the rigid particle differs from the bulk PP matrix, producing microvoids that are unacceptable in capacitor film due to partial discharge activity. Therefore, cast web inspection is a critical early warning of downstream dielectric defects.

If Oxygen Ingress Persists Above 8 ppm in Extruder Feed, Crosslinked Gel Fractions Rise Sharply

Oxidative gel formation during PP extrusion is controlled by the oxygen concentration in the feed throat, the melt temperature profile, the residence time distribution, and the remaining activity of the phenolic and phosphite stabilizer package. At oxygen levels above 8 ppm in the feed hopper atmosphere, measurable oxidative degradation occurs within the extruder barrel before the polymer reaches the melt seal, especially when the feed zone temperature exceeds 90 °C. The reaction sequence begins with hydrogen abstraction from tertiary PP carbon atoms, followed by β-scission of alkoxy radicals, carbonyl formation, and radical recombination that generates long-chain branching and crosslinked gel networks. The gel fraction is quantified by hot xylene or decalin extraction; capacitor-grade PP typically requires an insoluble gel fraction below 0.01 wt% when filtered through a 5 µm membrane. In a twin-screw extruder with 40:1 L/D and an average residence time of 90–120 s, local melt temperatures in kneading blocks can exceed the bulk melt set point by 10–25 °C due to viscous dissipation; if the set point is 245 °C, local hot spots above 260 °C accelerate crosslinking. Nitrogen purging of the feed hopper to an oxygen concentration below 50 ppm is common, but the most effective intervention is to maintain a slight positive pressure of nitrogen in the feed zone and to use a vacuum vent at −0.8 bar gauge for volatile removal. Stabilizer selection follows the ASTM D3895-19 oxidation induction time method for polyolefins, with capacitor-grade formulations typically showing an OIT of more than 30 min at 200 °C under oxygen. A batch that exhibits OIT below 15 min is generally rejected for ultrathin dielectric film because it indicates partial stabilizer consumption or inadequate dispersion. The interaction between residual titanium and oxidative gel formation is not negligible: titanium residues can catalyze hydroperoxide decomposition, increasing radical flux, so deashing efficiency directly influences the allowable residence time and melt temperature.

Analytical verification of ash and gel particle loadings in capacitor-grade PP relies on a sequence of dissolution, filtration, and spectroscopic methods that separates defects by size and chemical composition. Total ash is determined by combustion of a 5 g specimen in a muffle furnace at 600 °C according to ISO 3451-1:2019; this method captures noncombustible catalyst fragments but may underestimate volatile metal chlorides. Elemental analysis of magnesium, titanium, aluminum, silicon, and chlorine in the ashed residue or pressed film is performed by X-ray fluorescence spectrometry per ASTM D6247-18 or by microwave-assisted acid digestion followed by inductively coupled plasma optical emission spectrometry. Gel particle counts are measured on cast film or compression-molded sheet using optical imaging systems described in ASTM D7310-20, with a typical rejection threshold of fewer than 10 particles kg⁻¹ above 25 µm. Xylene solubles are determined by ISO 16152:2005, and intrinsic viscosity is measured in decalin at 135 °C per ISO 1628-3:2010. The analytical matrix is summarized in Table 1.

High-Temperature GPC and Thermal Fractionation of Polypropylene Gels

Molecular origin of gel particles in ultrapure PP dielectric film is investigated by high-temperature gel permeation chromatography in 1,2,4-trichlorobenzene at 150 °C with refractive index, light scattering, and viscosity detectors. The technique separates the high-molecular-weight tail that contributes to gel formation from the bulk polymer; capacitor-grade PP typically exhibits a weight-average molecular weight between 300 kg mol⁻¹ and 500 kg mol⁻¹ and a polydispersity index below 4.0. A bimodal or high-molecular-weight shoulder above 1,000 kg mol⁻¹ is often associated with long-chain branching or microgel formation, especially when the Mark-Houwink plot shows negative deviation from the linear calibration. Thermal fractionation methods such as temperature rising elution fractionation are used to separate the highly isotactic crystalline fractions from the less stereoregular material that may contain atactic PP and oxidized chains. Atactic PP has a low melting point and is prone to exudation to the film surface, where it contributes to die lip deposits and blocking. The xylene soluble fraction specified in ISO 16152:2005 for reactor powder is usually below 3 wt% for high-isotacticity capacitor grades, while the decalin-insoluble gel fraction is independently controlled below 0.01 wt%. The presence of gels larger than 50 µm cannot be explained by high-molecular-weight tail alone; these are generally attributed to localized oxidative crosslinking or carbonized contaminants from upstream equipment. Batch-to-batch variance in gel count is frequently traced to feed hopper cleaning cycles, extruder screw wear, or dead spots in the melt channel where residence time exceeds 15 min. When gel counts rise without a corresponding increase in total ash, the root cause is usually oxidative degradation rather than catalyst residue.

Electrically, the most severe failure mode in 3.0 µm capacitor film is not uniform intrinsic aging but statistical weak-point breakdown initiated at conductive or semi-conductive inclusions. Dielectric strength is measured according to IEC 60243-1:2013 using a 50 V s⁻¹ or 500 V s⁻¹ ramped voltage ramp with cylindrical or spherical electrodes. Defect-free BOPP film of 3 µm thickness typically exhibits characteristic breakdown fields between 350 V µm⁻¹ and 550 V µm⁻¹ at room temperature, but the presence of a single embedded ash agglomerate larger than 2 µm can create a low-field failure population near 150 V µm⁻¹. Failure statistics for capacitor-grade film are described by a two-parameter Weibull distribution; clean film shows a shape parameter β between 8 and 15, while contaminated film shows a mixed distribution with a distinct low-field failure mode. Under direct-current stress, ionic residues from MgCl₂ and AlCl₃ can migrate toward electrode surfaces, increasing leakage current and reducing insulation resistance measured per ASTM D257-14. The interaction between gel particles and breakdown is more complex: a soft gel may deform during orientation and produce a thickness disturbance that enhances electric field locally, while a hard carbonized gel acts as a conducting protrusion. Table 2 summarizes the observed defect classification and the corresponding dielectric consequence for 3 µm capacitor-grade BOPP film.

Thermal Deactivation of Alkoxysilane Donors Suppresses Ionic Ash Mobility

Alkoxysilane external donor residues are an underappreciated source of ionic and polar contamination in capacitor-grade PP. During polymerization, the external donor coordinates to the active site and is partially incorporated into the polymer matrix as alkoxysilane or hydrolyzed silanol species. After deashing, residual silanols can condense to form siloxane oligomers that migrate to film surfaces and contribute to die lip deposits. Thermal deactivation of these residues is achieved by heating the deashed powder to 110–130 °C for 2–4 h under nitrogen in a fluidized-bed dryer, which converts reactive silanols to less polar siloxanes and removes residual alcohols and water. The residual silicon content after deactivation is typically below 2 ppm for highest-grade capacitor film, as determined by XRF per ASTM D6247-18. If silicon content exceeds 5 ppm, die lip deposit formation accelerates, and the cast web may show a characteristic haze band after 6 h of continuous operation. The deactivation step also reduces the hygroscopicity of residual MgCl₂ by forming partially hydrated or alkoxy-substituted chloride species, which are less conductive than anhydrous MgCl₂. However, thermal treatment cannot compensate for inadequate washing; chloride levels above 8 ppm remain unacceptable for metallized capacitor film because they promote corrosion of the aluminum metallization layer and increase dissipation factor. The deactivation dryer must be designed to avoid product stagnation, because powder exposed to hot surfaces for longer than 8 h can undergo thermal degradation and generate new gel precursors.

Inline defect detection on a biaxially oriented film line combines high-speed CCD line-scan cameras with crossed-polarizer illumination and multi-angle laser scattering to count protrusion features above 30 µm while the web is moving at 300 m/min. The inspection system maps defects by class using image analysis algorithms that distinguish elongated gel streaks from circular ash-induced protrusions and from die lip deposit flakes. Data are correlated with melt pressure upstream of the screen changer, extruder barrel temperature profiles, and chill roll puddle height. A common field observation is that gel count increases during the first 30 min after a screen changer indexing event because the new filter elements release entrapped air and initial debris into the melt stream. To minimize this transient, the filter housing is preheated and purged with polymer melt for 15–20 min before the element is brought into the main melt path. Melt pressure fluctuations during filtration also correlate with cast web thickness variability: pressure pulsations above ±5 bar cause visible transverse thickness bands after stretching, even when no discrete gel particles are detected. Therefore, the filtration system must be coupled to a melt pump with closed-loop pressure control, and the screen changer indexing sequence must be tuned to avoid pressure spikes exceeding 10% of the set point.

Does Cast Roll Puddle Stability Control Ash Agglomerate Distribution?

On a production cast film line for 3 µm capacitor film, the cast roll puddle is a narrow melt reservoir formed between the slot die lips and the rotating chill roll. Its stability is controlled by the die-to-roll gap, the melt temperature, the chill roll speed, and the electrostatic pinning force. When the puddle is too large or unstable, ash agglomerates and gel particles that would otherwise remain dispersed in the melt have time to settle, coalesce, or migrate to the puddle surface. The resulting defects are not uniformly distributed; they appear preferentially at the puddle edges where the melt velocity gradient is highest and where oxidized volatiles condense. Operators control puddle stability by adjusting the die lip gap and the chill roll speed, but the most robust intervention is to reduce the population of dense ash agglomerates upstream through deashing and melt filtration. A chill roll temperature of 28 °C to 34 °C promotes rapid quenching of the cast web and minimizes puddle disturbance, but it also increases water condensation on the roll surface at high humidity; this water can interact with chloride residues and create surface defects. Therefore, the cast roll area is typically enclosed with conditioned air at a dew point below 5 °C and a positive pressure differential to prevent ingress of particulate contamination. In humid production environments, pre-drying of the PP powder or pellets is required when the ambient absolute humidity exceeds 12 g kg⁻¹. The final tenter oven air balance also influences ash and gel defect visibility: if the oven air is not filtered to remove particles above 5 µm, airborne contaminants can be embedded into the film surface during transverse stretching, creating additional weak points. Thus, control of residual catalytic ash and gel particles is not limited to the raw material and melt filtration stages but extends through casting, orientation, and air handling to the point where the dielectric film is wound for metallization.