Execution of an Experiment

Some of the practice involved in developing experimental skills occurs within a single experiment and some occurs as the superposition of mat r experiments. In carrying out any given experiment, it may be necessary to make several runs , i.e., sets of measurements. On some occasions, it is worthwhile to cany out a practice run as a warmup test of the apparatus and your own skills. Such a practice run can be made with a known standard material or standard set of conditions, or it can be a first try at a real set of measurements to get a feel for the operation of the equipment. 

The design of the experiment will require a decision about the number of data points to be taken in each run and the number of runs needed see the next subsection on data col-lectioa Often it is desirable to have some indication of the reproducibihty of a typical data point. In this case, multiple measurements of some initial value can be helpful for estimating the random errors as well as providing practice with the apparatus. 

Be alert to possible external disturbances—such as high humidity, mechanical vibrations, stray electric fields, voltage fluctuations, unusual local heating—and internal misbehaviors—such as vacuum leaks, nonhnear meter readings, hysteresis. Some of these may cause equipment failure or result in noisy, erratic readings. Some can introduce subtler troubles in the form of systematic errors, where the data look fine but are not. See Chapter II for a detailed discussion of errors. 

Data values are subject to both random and systematic errors. The effect of random errors can be reduced by increasing the number N of observations, but this is an inefficient process since, as shown in Chapter II, the error in the average of N values varies as 1 /. Thus one must quadraple the number of measurements in order to cut the random error in half A much better approach is to improve the design and sensitivity of the apparatus. Even then, you must guard against systematic errors and personal bias in recording the data. 

In most cases, data are collected sequentially rather than in random order. For example, chemical kinetics requires measuring the concentration of reactant or product sequentially as a function of time, and the measurement of pH during a titration must be made sequentially as a function of the volume of acid or base added. Thus you should be alert to the possibility that some uncontrolled variable (e.g., ambient pressure or temperature) may be undergoing a monotonic change that introduces a small but systematic effect into your data. 

---

From the results presented in Table 2 one can see that the greatest conversion was obtained at the upper limit of all process variables. Table 4 reveals that, as in the case of Lipozyme IM, the addition of water led to inhibition of the reaction. The enzyme concentration, the temperature, the oil ethanol molar ratio, and the interactions temperature-oil ethanol molar ratio and temperature-water addition had a positive effect on the production of biodiesel. Concerning temperature, the result obtained confirms the fact that the optimum temperature for this enzyme is about 70°C. As expected, the enzyme concentration, in the experimental range investigated, had a positive effect on the reaction conversion. Note also that for this system no alcohol inhibition was verified. The optimization for this system led to the following process variables values T = 65°C, [E] = 20 wt/wt%, [ W] = 0 wt/wt%, and R = 1 10, with a predicted maximum conversion of 82% in 6 h. The execution of the experiment resulted in an experimental value at these conditions of 81.4%, which agrees very well with the value predicted from the experimental model. 

When the screening experiment is conducted in an ongoing process, it is usually a good idea to choose the current operating conditions as the center point. This can provide assurance to the operating personnel that at least some of the runs in the experiment will be conducted under familiar conditions for which a satisfactory product should be manufactured. Runs at these center points could also be used to check for unusual conditions during the execution of the experiment by comparing the response at the center points to historical process performance,... 

Statistical experimental design, also called design of experiments (DoE), is a well-established concept for planning and execution of informative experiments. DoE can be used in many applications. An important type of DoE application refers to the preparation and modification of mixtures. It involves the use of mixture designs for changing mixture composition and exploring how such changes will affect the properties of the mixture [32],... 

Biochemical reactions are normally carried out in aqueous solution at a constant pH value. This is achieved by using a buffer solution, in which the H+ ions or OH- ions released in the reaction combine respectively with the anions of weak acids, or the cations of weak bases. Biochemical reactions usually show very pronounced dependences on pH. Hence, the selection of the right buffer and the accurate measurement and adjustment of the pH are of critical importance for the execution and reproducibility of an experiment (Perrin and Dempsey 1974 Beynon and Easterby 1996 Chambers 1993 Stoll and Blanchard 1990). 

As in any chromatography technique, one can break down the separation process to its two most fundamental aspects adsorption and elution. On a more practical level, the execution of an IMAC experiment involves hve discrete steps which can be readily automated column equilibration (charging of the gel), sample loading, removal of unbound material (washing), elution, and regeneration. 

Definitions (Figure 5) During a validation study, each participating laboratory generally carries out a number of separate executions of an alternative method on each material in the RSTS. Each of these independent executions is a repeat experiment . 

Probe tuning is necessary for a number of reasons. Other than the fundamental requirement for maximising sensitivity, it ensures pulse-widths can be kept short which in turn reduces off-resonance effects and minimises the power required for broadband decoupling. A properly tuned probe is also required if previously calibrated pulse-widths are to be reproducible, an essential feature for the successful execution of multipulse experiments. 

The rate of phosphate binding to EPSP synthase was measured by a single-turnover experiment in the reverse reaction shown in Fig. 10. The experiment was initiated by mixing an excess of enzyme and EPSP with a trace of labeled phosphate (<1 fiM). Under these conditions, the rate of formation of PEP was limited by the rate of phosphate binding to the enzyme-EPSP complex. However, in the absence of unlabeled PEP, the reaction did not go to completion. Successful execution of this experiment required the addition of the unlabeled reaction product (PEP) in order to ensure that the release of radiolabeled PEP was irreversible. In the absence of unlabeled PEP, the reaction came to equilibrium short of complete conversion of radiolabeled phosphate to PEP. The addition of unlabeled PEP pulled the reaction to completion by dilution of the radiolabeled PEP. Computer simulation was required to analyze quantitatively the reaction time course. Conventional data fitting to the time dependence of the reaction gives a rate of approximately 0.035 sec . The simple interpretation would then lead to calculation of a second-order rate constant for phosphate... 

An experimental design is a formal plan for execution of the experiment. It includes the choice of response, factors, designation of levels, and assignment of blocks as well as application of treatment on experimental units. 

Gathering data. With the job separated into pieces, usually a number of the pieces depend on careful data collection from the literature or from the design and execution of careful experiments. The students first impulses are often to model mathematically, but an efficient and effective solution often depends on simple experimentation or library work. We call this failure to resort to empirical work or extant data a failure of Galileo or Bacon 101 because of these individual s contribution to the creation of systematic empirical science. 

Undoubtedly a product as innovative as the 787, embracing as it does entirely new materials and technology, would always face significant challenges. However, beyond this, Boeing s experience highlights the fact that whilst companies might outsource the execution of an activity they should never outsource its control. 

The feed-back design (Figure 6.3.3 on the next page) was a 2-level, 6-variables central composite plan that required 2 = 64 experiments for the full replica. A 1/4 replica consisting of 16 experiments was made with an additional centerpoint. This was repeated after every 3 to 4 experiments to check for the unchanged condition of the catalyst. The execution of the complete study required six weeks of around the clock work. In the next six weeks, mathematical analysis and model-building was done and some additional check experiments were made. 

---