In industrial process laboratories, sample preparation is consistently the rate-limiting step. An XRF analyzer can measure a steel sample in 45 seconds. Preparing that sample — cutting, grinding, polishing to a flat 40 mm disc with less than 1 µm surface roughness — takes a skilled technician 4 to 6 minutes. On a steel melt shop producing one heat every 45 minutes, with 3 samples required per heat, the lab needs to process 4 samples per hour continuously across three shifts. Manual preparation at that rate requires 3 to 4 dedicated technicians. Automated preparation systems handle it with one.
Stella’s Analysis
Automated sample preparation for XRF analysis involves a defined sequence of operations, each with specific engineering requirements.
Sample reception: the robotic system receives cast or drilled samples from the furnace area, typically arriving in a pneumatic tube transport capsule. The sample is identified by barcode or RFID, linked to the heat number in the LIMS, and loaded into the preparation cell without human contact. Eliminating human sample handling at this stage removes the most common source of sample ID errors in production labs.
Milling: for cast samples (steel, iron, aluminum), the surface oxide layer must be removed and a fresh, representative surface exposed. An automated milling machine with a carbide or diamond-tipped cutter removes 0.5 to 1 mm from the surface in a controlled pass, delivering a surface roughness below Ra 3.2 µm. Cycle time: 45 to 90 seconds.
Grinding and polishing: for XRF, the surface requires roughness below Ra 0.4 µm for reliable light element (carbon, silicon, phosphorus, sulfur) analysis. Automated grinding stations use fixed-grit silicon carbide paper sequences, with robotic part loading and unloading between stages. The polishing stage uses diamond suspension on a rotating wheel with controlled pressure. Total preparation to XRF-ready surface: 3 to 4 minutes automated versus 6 to 8 minutes manual.
Shanghai Meinolf Technology’s rapid analysis systems achieve throughput of up to 100 samples per hour in fully automated configurations, reducing manual errors by 90%. This performance level requires that every step — transport, preparation, measurement, result validation — runs in parallel for queued samples rather than sequentially. Pipeline architecture rather than sequential processing is the engineering design that enables high throughput.
Industry Impact
The labor saving is quantifiable but secondary to the quality improvement. Automated preparation delivers more consistent surface quality than manual preparation, which means tighter analytical precision across the full measurement range. For steel producers tracking carbon at 0.05% in low-carbon grades, a 0.002% analytical uncertainty improvement from better sample preparation has direct impact on yield and scrap rate in the steelmaking process.
Automated systems also enable 24/7 lab operation without premium labor costs. A lab running automated preparation through the night shift generates the same quality data as the day shift — removing the asymmetry between day-shift quality control and night-shift production decisions that affects many traditionally operated labs.
Visit our industrial inspection equipment factory page for a complete overview of automated laboratory solutions from Chinese providers.
Trends
Fusion bead preparation for XRF — dissolving the sample in a lithium borate flux at 1,050°C and casting a glass disc — is being automated for non-metallic materials (cement, minerals, refractories) where milling preparation is insufficient. Automated fusion machines handling 6 to 12 samples simultaneously are now standard on Chinese cement lab lines and are entering minerals processing labs across Southeast Asia.
For related context on how Chinese automation systems are deployed in process industries, see robotics in aluminum production and innovations in China’s industrial inspection landscape.
