Many water treatment works use online instrumentation measuring UV254 or color to quantify organics in raw water for the purposes of coagulation control. These instruments have a method for removing the effect of
turbidity on the
UV254 or
true color measurement. Without this feature, the UV254 and color readings would always reflect the absorbance from the organics in the raw water and light lost due to diffraction by particles seen as turbidity. These two effects—absorption and diffraction—can’t be separated by measuring a single wavelength only, a turbidity compensation measurement is always necessary.
For accurate measurement of organics in raw water, a robust turbidity compensation mechanism must be used.
Conventional Approaches to Turbidity Compensation
Most online instruments for coagulation control in raw water measure the absorbance of the water at two wavelengths: the wavelength of interest (254 nm for UV254 and typically 400 nm for color) and a reference measurement at a different wavelength to measure the turbidity. Often, this reference wavelength measurement is in the visible region of the electromagnetic spectrum (e.g. 550 nm). The software in the instrument then corrects the measurement absorbance for turbidity in the sample, measured at the reference wavelength. The more turbidity in the water, the larger that correction will be.
However, this approach is not perfect, as turbidity does not affect all wavelengths equally. It is well understood when referring to turbidity measurement that the measurement wavelength is important and different wavelengths will give different results for turbidity. For example, using a longer wavelength (such as the near infra-red wavelength used in the ISO7027 turbidity standard method) has the potential to miss smaller particles.
When it comes to absorbance, the same principle applies: the shorter the wavelength used for measurement, the more the turbidity affects the absorbance. When discussing online measurement of UV254 and color, an instrument manufacturer must decide how they will apply their reference turbidity compensation measurement to the measurement wavelength itself.
Do they assume that whatever the absorbance is at the reference measurement wavelength should just be subtracted from the measurement wavelength absorbance? Or do they try and apply a factor to account for the fact that the absorbance at a lower measurement wavelength may be different to that at the reference measurement due to the turbidity?
 |
|
The effect of different levels of turbidity on different wavelengths. The absorption spectrum of several different turbidity standards is shown above and it is clear that turbidity does not affect all wavelengths equally. There are no other absorbing species in the sample.
|
These graphs show how a simple subtraction approach to turbidity compensation, as recommended by DIN 38404-32 is not sufficient. At 10 NTU for example, absorption at 254 nm is 18.4 Abs/m, but absorption at 550 nm is 1.29 Abs/m. Manufacturers must do something more complex in determining the effect of turbidity at the measurement wavelength. Using the above data, it could be argued that by simply multiplying the reference measurement by 15, you can establish the correct factor to apply to the measurement wavelength.
However, at higher levels of turbidity, this relationship breaks down.
 |
|
At 100 NTU, the absorbance at the measurement wavelength (254 nm) is around 80 Abs/m, but at the reference measurement (550 nm) it is 7.3 Abs/m, the compensation factor has decreased, using our x15 correction factor now is not accurate, x10 would be more appropriate.
|
In real raw water samples, the variety of scattering of different-sized and shaped particles complicates the relationship further.
Alternative Approach to Turbidity Compensation
The alternative approach to using a single wavelength turbidity compensation mechanism is to use the entire UV-Vis spectrum absorbance and apply an algorithm to this spectrum to account for the contribution to it from turbidity. This allows the generation of a turbidity-compensated spectrum from which multiple turbidity-compensated parameters can be defined simultaneously.
This method uses a mathematical equation that describes the relationship between scattering caused by turbidity and wavelength as a function of the particle diameter. The effect was first described by Gustav Mie in 1908 and is known as Mie Scattering. Practical implications of this effect for the measurement of turbidity were published by Huber and Frost in 1998 and include the well-known spectral shape caused by suspended solids, which depends upon both the wavelength as well as on the particle diameter itself.
Different applications tend to have different particle types/size distributions, so it is necessary to have different algorithms for different applications. By using this approach, more accurate readings can be obtained for parameters such as UV254 (post-sample filtration) and true color.
The s::can spectro::lyser V3 spectrometer probe, initially designed for remote monitoring in surface water applications, was the world’s first fully submersible dual-beam online UV-Vis spectrophotometer sensor. Widely deployed across the globe in a wide range of applications, the sensor can measure anything in water that can be determined by UV or Visible spectroscopy, including nitrate, organics (dissolved organic carbon and total organic carbon), true color and more. The sensor platform utilizes a multiple-wavelength turbidity compensation mechanism rather than a single-wavelength turbidity compensation approach. This mechanism applies to all s::can sensors that utilize the spectro::lyser platform, including the nitro::lyser V3 (nitrate measurement), uv::lyser V3 (UV wavelength measurement) and carbo::lyser V3 (organics measurement).
Successful UV254 Monitoring in the Field
At a water treatment works where a spectro::lyser V3 monitors the raw water for multiple parameters, an online UV254 instrument which uses the single wavelength turbidity compensation mechanism was also used.
When turbidity is low (<10 NTU), correlation between the two UV254 measurements mirror one another, the only exceptions being when instrument X appears to drift away from the spectro::lyser reading before being brought back into line via a calibration. These can be seen by the step changes in the Instrument X trend.
When turbidity increases to greater than 10 NTU, instrument X generally reads higher than the UV254 reading from the spectro::lyser. This is because of the turbidity compensation mechanism used in Instrument X. For every peak in turbidity, instrument X reads UV254 higher than the spectro::lyser and the larger the turbidity peak, the bigger the difference between the two measurements. When there is low turbidity, the two instruments correlate.
The accuracy of the spectro::lyser and turbidity compensation approach when measuring UV254 was verified by comparing the results against the laboratory method. The lab sample was filtered through a 0.45 micron filter to remove turbidity before being measured with a UV spectrophotometer at 254 nm. The correlation R2 value between the two methods over several samples taken over a period of time was 0.9896.
The above solids compensation technique has also been proven on raw water monitoring to produce accurate true color readings even when the levels of turbidity reach above 100 NTU.
Conclusion
When deciding on a UV254 or color instrument for coagulation control, it is important to consider the accuracy of the reading during a turbidity event. Calibration of a UV254 sensor using standard solutions such as potassium hydrogen phthalate does not demonstrate accuracy during turbidity events. Overestimation of UV254 can lead to overdosing of the coagulant. Underdosing can lead to an increased risk of disinfection by-product formation potential. During large turbidity events, this may be deemed as doing little harm to combat the increased coagulant demand associated with the increased turbidity. However, the consistent overestimation of UV254 even during relatively small turbidity increases can only lead to additional chemical dosing costs.
When choosing a UV254 or color monitor, it is often worth understanding the turbidity compensation used by the instrument and verifying it is accurate when you most need it.
