Measurement uncertainty chemical monitoring

The above sections have highlighted the importance of data comparability and trace-ability in the context of WFD chemical monitoring. Let us now examine in detail what references need to be considered for the development of a sound metrological system. Firstly, as a reminder, traceability is defined as the property of the result of a measurement or the value of a standard whereby it can be related to stated references, usually national or international standards, through an unbroken chain of comparisons all having stated uncertainties (ISO, 1993). The ways in which these elements can be applied to chemical measurements were discussed some years ago (Valcarcel and Rios, 1999 Quevauviller, 1999 Walsh, 2000) and those discussions still continue. In this context the basic references are those of the SI (Systeme International) units, i.e. the kg or mole for chemical measurements. Establishing SI traceability of chemical measurements may, in principle, be achieved in relation to either a reference material or to a reference method (Quevauviller and Donard, 2001). The unbroken chain of comparison implies that no loss of information should occur during the analytical procedure (e.g. incomplete recovery or contamination). Finally, traceability implies, in theory,... 

Thus, the systems used to monitor and control the process are conventional. Process materials (with the exception of agents and energetics), temperatures, and pressures used in this technology package are commonly used in other industrial applications, where they are routinely monitored and controlled. The usual collection of equipment for monitoring temperature, pressure, level control, flow, and other parameters normally measured in a chemical plant is used. The analytical procedures to be used to monitor certain streams will be new, and they present the greatest uncertainty. 

In many countries, emissions are determined by direct monitoring. In these cases the measurements should be subject to quality assessment, and the sampling plan should be evaluated to estimate the uncertainty in all steps of the procedure. To assess emissions of mixtures of chemicals, concentrations of chemicals should be measured simultaneously, but the validation procedures would be the same as for single chemicals. When emissions are estimated from data on produced, processed, or used amounts of the individual chemicals in mixtures, the calculations should be validated by measurements in the field. 

In the human population the level of exposure to a chemical can, in some cases, be known precisely, for example known doses of drugs are given to patients. For industrial chemicals, and especially chemicals in the environment, however, this is much more difficult and sometimes impossible to know. Well-regulated chemical industries in highly developed countries will monitor workers by taking blood or urine samples from them at intervals. The air in the plant or factory can be monitored and workers may wear personal sampling devices that measure how much of a volatile chemical they are exposed to. This is not practised in all industries nor in less well-developed countries. In such cases, where there is environmental exposure to chemicals, the level of exposure will often have to be estimated, a process that is often fraught with uncertainty. For exposure via water or food, this can be estimated based on information... 

Methods which have been developed for the monitoring of radiation forces can be very sensitive (up to 1 mW). They are better suited in free-field and noncavitating conditions, but involve errors due to acoustic streaming. The uncertainty for such methods ranges from 2.2% at 1 MHz to 12% at 30 MHz, and accuracy depends on the shape of the ultrasonic wave. Radiation force measurement provides a good method for the calibration of transducers in specifically devised chambers, but its use in chemical reactors of defined geometry could prove to be difficult. 

Improved reactor operation. Given these modeling results, we can shorten the batch time significantly. First we switch the LC on at time [t -= 0. Then as the LC measurements become available, we estimate the initial number of moles n 40 and monitor its confidence interval As soon as the uncertainty in Hao reaches a sufficiently low threshold, we are confident how much B is required and can add the remainder in one shot. Testing this approach with many datasets at Kodak allowed us to conclude that by the time the first large addition was completed, we obtained sufficient confidence on n o that the rest of the B could be added immediately. Such a procedure reduces the batch time from about 900 minutes with the conventional approach to about half that time with the model-based operation. The new operation essentially doubles the production rate without constructing new reactor facilities, which is significant for this capacity-limited chemical. 

It seems paradoxical to try to monitor quantitatively reactions on polymer phase — which are eventually incomplete because of the heterogeneous locations of functional sites — with additional chemical reactions of similar uncertainty. This fundamental contradiction cannot be abolished by any chemical method of test employed on insoluble gel phases generating a colored reaction on polymer (see Fig. 31). Here, with an additional chemical reaction, a specific dye has to be released and measured as is the case in the quantified [115] ninhydrin [129] test and in the Schiff base formation with 2-hydroxy-naphthaldehyde [130]. Although both methods are discontinuous, time consuming, and des-... 

---