Limitations of turbidity process probes and formazine as their calibration standard

Abstract

Turbidity measurements are frequently implemented for the monitoring of heterogeneous chemical, physical, or biotechnological processes. However, for quantitative measurements, turbidity probes need calibration, as is requested and regulated by the ISO 7027:1999. Accordingly, a formazine suspension has to be produced. Despite this regulatory demand, no scientific publication on the stability and reproducibility of this polymerization process is available. In addition, no characterization of the optical properties of this calibration material with other optical methods had been achieved so far. Thus, in this contribution, process conditions such as temperature and concentration have been systematically investigated by turbidity probe measurements and Photon Density Wave (PDW) spectroscopy, revealing an influence on the temporal formazine formation onset. In contrast, different reaction temperatures do not lead to different scattering properties for the final formazine suspensions, but give an access to the activation energy for this condensation reaction. Based on PDW spectroscopy data, the synthesis of formazine is reproducible. However, very strong influences of the ambient conditions on the measurements of the turbidity probe have been observed, limiting its applicability. The restrictions of the turbidity probe with respect to scatterer concentration are examined on the basis of formazine and polystyrene suspensions. Compared to PDW spectroscopy data, signal saturation is observed at already low reduced scattering coefficients.

Keywords: Calibration standard; Formazine; Photon Density Wave spectroscopy; Process analytical technology; Turbidity probes.

Fig. 1

Reflected intensity at 860 nm and reduced scattering coefficient at 690 nm as function of time for three repeated syntheses of formazine at 25 °C, with stirring at 200 rpm. Dotted line indicates signal noise from PDW spectroscopy due to insufficient light scattering

Fig. 2

Reflected intensity at 860 nm and reduced scattering coefficient at 690 nm as function of time for two syntheses of formazine at 25 °C, without stirring

Fig. 3

Reflected intensity at 860 nm and reduced scattering coefficient at 690 nm for a synthesis of formazine at 25 °C without stirring for the first 24 h. Afterwards, the stirrer was alternately turned on (dashed line) and off (dotted line)

Fig. 4

Reflected intensity at 860 nm and reduced scattering coefficient at 690 nm for formazine syntheses at different temperatures, with stirring at 200 rpm

Fig. 5

Clouding onset time from Fig. 4 as function of reaction temperature with exponential fits and linearization with an Arrhenius approach (inset) as well as its residuals

Fig. 6

Reflected intensity at 860 nm and reduced scattering coefficient at 690 nm for formazine syntheses with different relative starting concentrations

Fig. 7

Relative stock solution concentration as function of the clouding onset time from Fig. 6 with fits and resulting 95 % confidence intervals

Fig. 8

Reflected intensity at 860 nm (squares) and reduced scattering coefficient at 690 nm (circles) and 906 nm (triangles) after 24 h for different relative concentrations with linear fits, including repeated experiments at c_rel = 100, 200, and 333 %. Relative intensity for pure water is not included into the fit

Fig. 9

Reflected intensity at 516 nm (circles) and 860 nm (squares) as function of volume fractions for a polystyrene suspension

Fig. 10

Reflected intensity at 516 nm (circles) and 860 nm (squares) and reduced scattering coefficient at 515, 690, and 906 nm (triangles) as function of volume fraction for a polystyrene suspension. Inset displays the reduced scattering coefficients in a double-logarithmic plot

Fig. 11

Reflected intensity (diamonds), absorption coefficient (circles), and reduced scattering coefficient (triangles) as function of wavelength for volume fractions of 0.07 and 0.025 of a polystyrene suspension, plus absorption coefficient for pure water [–39]

Fig. 12

Absorption coefficient at 515 nm (squares) and 982 nm (circles) as a function of volume fractions of a polystyrene suspension and absorption coefficient of pure water at Φ _PS = 0 for both wavelengths (full symbols) [–39]. Deviating absorption coefficients at low volume fractions are neglected for the fit. 95 % confidence interval resulting from fit. Inset shows the fitted values at 515 nm in detail