In high-density flux coating of low-carbon steel core wire, the transition from conventional low-pressure extrusion below 8 MPa to operation above 12 MPa is not merely an increase in pump pressure; it reflects a shift from flow-dominated filling to compaction-dominated densification. On production lines using a 50 mm single-screw extruder with a gear metering pump and hydraulic ram feeder, die inlet pressures of 12.5 MPa to 18.0 MPa are maintained for rutile-potassium silicate pastes with moisture contents between 4.5 wt% and 6.5 wt% on a dry-flux basis. The core wire, typically ER70S-6 or equivalent with a diameter of 2.5 mm to 4.0 mm, is drawn through a crosshead die at linear speeds of 2 m/s to 8 m/s. At pressures above 12 MPa, the wet flux coating density increases from approximately 2.31 g/cm³ to 2.53 g/cm³, and the residual void fraction falls below 2.0 vol% as measured by mercury intrusion porosimetry. Laser micrometer arrays positioned 450 mm downstream of the die record coating eccentricity; above 12 MPa, the acceptable total indicated runout is 0.06 mm on a 3.2 mm electrode. Pressure spikes of 2 MPa to 6 MPa at 0.5 Hz to 2 Hz are not acceptable because they correlate with slip-stick flow at the paste–die interface and with periodic binder phase separation. Operators monitor these pressure fluctuations using flush-mounted strain-gauge transducers with a full-scale range of 25 MPa and a response time below 2 ms. The high-pressure regime also exposes air entrapment defects: without vacuum deaeration of the paste at -0.08 MPa for at least 25 min, entrained microbubbles collapse unevenly during extrusion, producing after-bake blistering at 2 to 6 blisters per 100 mm of electrode.
The controlling rheological parameter in this pressure regime is not the simple apparent viscosity but the yield stress and wall-slip velocity at the polished die surface. Mineral flux pastes for covered electrodes behave as Herschel-Bulkley fluids with yield stresses between 18 kPa and 45 kPa at 25°C. A paste with a yield stress below 18 kPa at 25°C tends to extrude at die inlet pressures under 10 MPa and produces slump and teardrop-shaped coating after the die. A yield stress above 45 kPa causes pressure peaks above 18 MPa and audible cavitation in the gear metering pump. The apparent viscosity measured at a shear rate of 10 s⁻¹ on a rotational rheometer using a serrated parallel-plate fixture of 35 mm diameter is maintained between 80 Pa·s and 160 Pa·s for rutile grades; below 80 Pa·s, the paste is too fluid to retain the impressed shape, while above 160 Pa·s, the die inlet pressure rises non-linearly and the pump amperage exceeds 85% of nameplate. Sodium silicate binder with a SiO₂:Na₂O modulus of 2.2 to 2.8 and solids of 40 wt% to 44 wt% provides the necessary bridge between particle wetting and low torque. Cellulose ether additions of 0.2 wt% to 0.5 wt% of total dry flux increase low-shear viscosity and green strength, but at pressures above 12 MPa they also migrate toward the coating surface and can form skins that seal in moisture. Published data for the exact Herschel-Bulkley parameters of proprietary flux formulations is limited; however, industrial batch logs show that die inlet pressure is repeatable within ±0.7 MPa across a 250-kg batch when paste temperature is held at 25 ± 1°C and silicate solids are measured per ISO 3251:2019. The addition of 1 wt% to 2 wt% bentonite raises yield stress without increasing high-shear viscosity, but amounts above 2.5 wt% produce shear-induced aggregation and pressure oscillations at 0.8 Hz.
At pressures exceeding 12 MPa, the radial compressive stress on the wet flux coating collapses residual air voids and forces moisture redistribution from the die-wall region toward the core-wire interface. The highest green densities are obtained when the die land pressure is not the only compaction mechanism; post-die calibrating rolls operating at a linear force of 3 N/mm to 5 N/mm of electrode length remove surface defects and close microcracks. Bulk green strength reaches 1.8 MPa to 2.4 MPa in three-point bending when tested immediately after extrusion on a texture analyzer at a span of 40 mm and a crosshead speed of 1 mm/min. Under these conditions, the coating adheres to the core wire through mechanical interlocking with the roughened wire surface, which is typically electrochemically cleaned to a surface roughness Ra of 0.8 μm to 1.2 μm. A smoother surface below Ra 0.6 μm reduces adhesion and produces concentricity loss after drying, while a rougher surface above Ra 1.5 μm traps fine particles and creates corrosion initiation sites at the wire–flux interface. Core wire preheating to 35°C to 45°C before entering the crosshead die reduces moisture condensation on the wire surface and improves the initial tack of the silicate binder. Higher preheat above 55°C is not used because it accelerates silicate skin formation and increases die exit temperature beyond 45°C. The compaction regime also increases the apparent coating density but does not eliminate all interconnected porosity; the connected pore fraction remains at 0.5 vol% to 1.5 vol%, which is sufficient to permit controlled moisture escape during the first 20 min of the baking cycle.
Die geometry in high-pressure flux extrusion is selected to balance three competing requirements: pressure drop, coating concentricity, and surface finish. For a 3.2 mm core wire with a 1.0 mm radial coating, the land length is set between 8 mm and 12 mm, giving a land length-to-coating thickness ratio of 8:1 to 12:1. The pressure drop through the annular die land can be approximated by ΔP = (4τ_yL)/D_h + (32μ_pVL)/D_h² for a Bingham plastic, where τ_y is the bulk yield stress, μ_p is the plastic viscosity, V is average velocity, L is land length, and D_h is the hydraulic diameter of the annular channel. At an average velocity of 4 m/s, a τ_y of 30 kPa, a μ_p of 90 Pa·s, and a hydraulic diameter of 2.0 mm, the first term contributes approximately 12 MPa to the die inlet pressure, which explains why small changes in land length or yield stress produce immediate pressure excursions. The die entry angle is also constrained: a reduction from 30° to 15° lowers the entry pressure loss but increases the dead-zone volume at the die shoulder, where stagnant flux can crosslink under shear and contaminate subsequent material. Concentricity is controlled by three centering screws arranged at 120° intervals and by a floating core tube that allows 0.15 mm radial adjustment; the total indicator reading on the core wire at the die exit is held within 0.03 mm. Back-taper of 0.5° over the last 3 mm of the land reduces the die-exit pressure spike and minimizes coating swelling. Die inserts made from tungsten carbide with a hardness of 88 HRA to 90 HRA and a polished surface finish of Ra 0.15 μm to 0.25 μm are specified for runs exceeding 500,000 m, because rutile and ilmenite particles generate die wear of 0.03 mm to 0.06 mm per 100,000 m on hardened tool steel. When the die land roughness exceeds Ra 0.35 μm, surface tearing of the coating appears as longitudinal striations and the average pressure drops by 1.0 MPa to 1.5 MPa because wall slip increases, but concentricity deteriorates.
Because the green coating must retain a moisture content of 4.5 wt% to 6.5 wt% on a dry-flux basis, determined at 105°C for 3 h using a halogen moisture analyzer calibrated against ISO 3251:2019, the extrusion line must compensate for both ambient humidity and raw-material variation. Above 6.5 wt%, the coating shows slumping and the wet density falls below 2.38 g/cm³ at die pressures above 12 MPa. Below 4.5 wt%, the paste is too stiff and die inlet pressure exceeds 19 MPa, which can trigger the hydraulic over-pressure relief set at 20 MPa and cause core wire breakage in strands of 2.5 mm diameter. Ambient relative humidity above 65% is known to increase the absorbed moisture of the dry flux by 0.8 wt% to 1.4 wt% within 30 min during open hopper transfer, so predrying of hygroscopic minerals at 120°C for 2 h is required when the incoming moisture exceeds 0.3 wt%. The green electrode is dried in a multi-zone oven with an initial zone at 80°C to 100°C for 45 min, a second zone at 120°C to 150°C for 60 min, and a final bake zone at 180°C to 200°C for 45 min under forced air flow of 4 m/s to 6 m/s. Heating rates above 8 K/min in the first zone create internal steam pressure that causes hairline cracks at 3 mm to 8 mm intervals on the coating surface. The final baked coating moisture is held below 0.4 wt% to meet the diffusible hydrogen requirements of ISO 3690:2018 for basic and rutile electrodes; electrodes with residual moisture above this limit are re-dried but are limited to one re-dry cycle because repeated baking above 180°C oxidizes ferromanganese and reduces arc stability.
Under these combined conditions, the soluble silicate binder can undergo shear-induced gelation at the die wall, producing intermittent high-torque transients and a coating with a brittle, glassy skin. Potassium silicate with a modulus of 2.8 to 3.2 has a higher SiO₂ concentration and a lower free alkali content than a modulus of 2.2 to 2.5, which increases green hardness but reduces the paste pot life from 8 h to 4 h at 25°C. Above a die inlet pressure of 13.5 MPa, the increase in local viscous dissipation raises the paste temperature at the die wall by 5°C to 10°C above the bulk temperature; shear rates in the 0.5 mm annular gap can reach 120 s⁻¹ to 300 s⁻¹ depending on line speed. At these conditions, the high-modulus silicate begins to deposit colloidal silica on the polished die surface, reducing the effective flow channel and causing the pressure to drift upward at a rate of 0.3 MPa to 0.6 MPa per 10 min of continuous operation unless the die is cleaned. The resulting coating exhibits surface porosity and a glossy outer layer after baking; cross-sectional microscopy shows a layered structure with binder enrichment at the outer 0.05 mm to 0.08 mm. To maintain stable operation, the line is limited to a maximum die inlet pressure of 14.0 MPa when the silicate modulus exceeds 2.8, and a cooling jacket on the crosshead die is operated at 18°C with a 25% aqueous ethylene glycol coolant. In addition, the flux formulation must include 0.1 wt% to 0.2 wt% of a high-molecular-weight polyethylene glycol release agent with a molecular weight of 8,000 g/mol to 20,000 g/mol to reduce die-wall adhesion. The practice of increasing silicate modulus above 2.8 should be limited to rutile-heavy formulations with at least 55 wt% titania on the total mineral fraction; in cellulosic or basic flux systems, the higher modulus causes premature reaction with moisture and produces undispersed silicate lumps.
The following comparative data from a pilot line using a 50 mm single-screw extruder with a 1.0 mm radial die land of 10 mm and a core wire diameter of 3.2 mm illustrate the effect of silicate modulus and solids on extrusion pressure and coating quality. All formulations contained 40 wt% rutile, 22 wt% ilmenite, 12 wt% calcium carbonate, 6 wt% fluorspar, and 4 wt% ferroalloy powder; the balance was binder and extrusion aids. Extrusion pressure was recorded with a flush-mounted strain-gauge transducer calibrated to ±0.5% of full scale, and green density was measured by mercury-free helium pycnometry on samples taken 500 mm after the die.
| Formulation code | Silicate modulus (SiO₂:Na₂O) | Silicate solids (wt%) | Extrusion pressure (MPa) | Green density (g/cm³) | Blister reject rate (%) |
|---|---|---|---|---|---|
| F-01 | 2.2 | 40 | 11.6 | 2.31 | 0.9 |
| F-02 | 2.5 | 42 | 13.0 | 2.44 | 0.3 |
| F-03 | 2.8 | 44 | 14.2 | 2.53 | 0.6 |
| F-04 | 3.0 | 45 | 16.0 | 2.57 | 2.4 |
For rutile electrodes intended for classification under ISO 2560:2009 and AWS A5.1/A5.1M:2012, baked coating moisture content below 0.4 wt% measured by loss-on-drying at 105°C for 3 h is required because residual moisture contributes directly to diffusible hydrogen in the deposited weld metal. The diffusible hydrogen test is performed according to ISO 3690:2018 at a welding current of 70% of the electrode maximum and a heat input of 1.0 kJ/mm on a 20 mm thick carbon steel plate. For a 3.2 mm E6013 electrode, diffusible hydrogen levels above 15 mL/100 g of deposited metal are considered a batch failure under AWS A5.1/A5.1M:2012. The coating line must therefore incorporate a post-bake moisture analyzer with an accuracy of ±0.05 wt% and a sampling interval of 30 min. Electrodes that pass the moisture limit but show a coating weight variation above ±3.5% of nominal are rejected because arc burn-off rate and slag fluidity depend on coating cross-section. Coating weight is checked every 1,000 m using a laser profilometer to calculate the coating area and a precision balance to determine mass per unit length. A coating factor of 1.6 to 1.8 is typical for rutile electrodes; below 1.6, the electrode burns too hot and spatter increases above 3 g/min under standard welding conditions; above 1.8, the weld pool is over-protected and slag detachability decreases.
In continuous high-pressure runs, frictional dissipation in the metering pump and die land raises flux temperature by 6°C to 12°C per 5 MPa of pressure increment. Bulk paste temperature entering the die is controlled at 23°C to 27°C; die exit temperatures above 48°C indicate that the thermal load on the organic binder may exceed its chemical stability limit. Cellulose ether binders used in flux pastes begin to degrade at 55°C to 65°C in the presence of sodium silicate, releasing methoxyl groups and lowering the paste green strength by up to 30% over a 2 h continuous run. The degradation products reduce the surface tension of the paste and contribute to die lip buildup that appears as white crystalline deposits at the die exit. To prevent this, the extruder barrel is water-cooled in two zones: the feed zone at 18°C and the compression zone at 22°C. The crosshead die is cooled with a separate recirculating chiller at 18°C using a 25% aqueous ethylene glycol solution. If the exit temperature exceeds 45°C for more than 10 min, the paste is purged and the die is cleaned with a soft brass scraper; steel scrapers are prohibited because they increase die surface roughness above Ra 0.25 μm. The interaction between thermal degradation and extrusion pressure is non-linear: a paste that maintains 14 MPa at 25°C may drop to 11 MPa at 50°C, leading to under-compaction and a 0.12 g/cm³ reduction in green density. Consequently, high-pressure operation above 12 MPa is only feasible when the heat removal capacity of the extruder is at least 60 W per kg/h of paste throughput, and when the line speed does not exceed 6 m/s for a 3.2 mm electrode.
When standard rutile formulations are extruded above 12 MPa, the operational boundary is defined by die inlet temperature below 48°C, paste moisture between 4.5 wt% and 6.5 wt% on a dry-flux basis, silicate solids between 40 wt% and 44 wt%, and die inlet pressure between 12.5 MPa and 16.0 MPa for standard rutile formulations. Exceeding these limits, either through the use of reclaimed flux fines above 15 wt% of the dry mix or through substitution of potassium silicate with sodium silicate of equivalent solids, produces measurable increases in die wall adhesion and post-bake surface cracking. The system is incompatible with amine-based wetting agents in silicate-bonded flux because the amine accelerates SiO₂ polymerization and reduces paste pot life below 2 h. Water used for paste make-up should have total hardness below 150 ppm as CaCO₃ and chloride below 50 ppm to avoid pitting corrosion of the core wire at the cut ends. Extended shutdowns beyond 30 min require die disassembly and rinsing with warm water at 40°C, because dried silicate on the die land is not removable by solvent wash and will score the next production run. Rejected electrodes exceeding 0.06 mm eccentricity are stripped and re-extruded, but rework is limited to two passes because repeated high-pressure shearing increases dead-zone volume and binder migration.