Evaluation of Three Automated Pretreatment Workstations for Free Silica Determination.

BACKGROUND

In China, pneumoconioses accounted for 64.23% of newly reported occupational disease cases in 2024. Silicosis, one of the most common and severe forms of pneumoconiosis, is caused by inhalation of free crystalline silica (SiO₂), highlighting the continued importance of accurate free silica determination. The pyrophosphoric acid (PPA) method (GBZ/T 192.4-2007) predominates in China, but manual operation exposes analysts to hazardous reagents and operator-dependent variability. No comparative evaluation of automated workstations has been published.

OBJECTIVE

To evaluate three automated free silica pretreatment workstations and their agreement with the manual method.

METHODS

Three workstations (A, B and C) were evaluated with a common protocol in terms of repeatability and trueness on reference and quality control materials, inter-day precision over five days, and agreement with the manual method on field samples by Bland-Altman analysis.

RESULTS

Repeatability RSDs (n = 10-12) were 3.5-10.1% at 2.3-5.8% or 9.1%, 2.9-4.8% at 13.0-20.4%, 2.2-2.5% at 58.6-71.6% and 0.5-0.8% at ≥ 99.76% (quartz sand). HorRat exceeded 1.3 in 8 of 13 repeatability or inter-day combinations below 50% free silica and in none at ≥ 60%. Inter-day RSDs were 1.4-4.2% at 15-67% and 0.2-0.8% at 99% free silica. Means lay within the certified value or reference ranges but 0.4-0.9 points below the quartz sand's nominal content. On field samples, mean differences from the manual method were -1.5 to + 0.5 percentage points (not significant), and all 95% limits of agreement lay within ±8 percentage points.

CONCLUSION

All three workstations are safe, reliable alternatives to manual PPA digestion and agreed with the manual method on field samples.

HIGHLIGHTS: To our knowledge, the first evaluation of three automated free silica pretreatment workstations with a common protocol; repeatability and trueness characterized over 2-100% free silica; no significant bias relative to the manual method on field samples.