We have a automatic Robolab with Pfaff divider, Hertzog mills & press with Phillips 2400 XRF. We are seeing a drop in our Fe₂O₃ in our raw meal &kilnfeed samples of 0.2 per cent, it occurs about one time in 10. It also occurs in a stock sample with similar frequency. If the low Fe₂O₃ pellet is analysed ten times the same drop can be seen with a SD of 0.01. This suggests xrf is not the problem! The problem never occurred in fused beads. We have tried to contaminate the sample with cement in the dividing station to replicate the problem, this also failed. Contamination of sample with limestone going to press had no significant effect. Using more or less grinding aid had no significant effect on Fe₂O₃. Using different weights of sample to both milland press had no effect on the iron.
This problem left for 12 months and is now back. The problem never occurred in 12 years with a Hertzog manual mill, manual press and phillips 1606XRF.Fls Has looked at the problem for us with no answers. Has anybody else had a similar drop in Fe₂O₃ in sample while other oxides remain constant.

This a question we are going to have to put on the web (anonymously) as I do not have an answer. It must be something systematic in the sample preparation or measurement, but you have tried everything obvious?

We intend to increase the strength of our pozzolan (from natural pozzolan) cement to 30MPa (28 days). But we could hardly reach it. What would the best thing to do?

There are two possible options. First you need to maximise the strength potential of the cement clinker you are blending with the natural pozzolan to make the pozzolanic cement. This means stabilising the clinker mineralogy and boosting the alite content. You can do that by lifting LSF or SM dependent on your current clinker chemistry. Another alternative would be to add calcium fluride to the raw mix. Secondly you could add a cement conditioner to activate the pozzolanic reaction in the cement. Possibly alkali carbonates or potassium sulphate.