The overall goal of the following experiment is to identify flocculants that can improve the removal of dispersed particles from plant extracts. This is achieved by setting up an experimental design to identify the most effective polymers and parameters controlling the aggregation of these dispersed particles. The second step defines the work flow, which corresponds to the unit operations used in a large-scale application.
Next, leaf extracts are prepared for flocculants under various conditions and analyzed in order to measure turbidity and product recovery. The results show the optimal flocculation conditions and polymers based on a design of experiments model evaluation, which we shorten to DoE. The main advantage of this technique compared to traditional clarification using depth-filtration alone, is that the filter capacity can be increased several-fold, reducing overall process costs.
Flocculation can increase the effective diameter of particles and facilitate their removal by centrifugation or filtration. Therefore, the environmental process parameters that are relevant for the flocculation procedure must be identified first. Typically, there are several such parameters.
So a DoE approach is appropriate given the lack of mechanistic models. The DoE is set up by selecting parameters based on literature data and prior knowledge. Typical parameters include incubation time as well as polymer type and concentration.
The experimental outcomes that are used to evaluate the efficiency of flocculation, such as filtratability, are also defined before the experiment. This information is then fed into a DoE type appropriate for the number of factors to be investigated and the degree of knowledge already accumulated about the system. A work flow is developed to encompass all steps relevant for the flocculation and laboratory-scale processing of plant extracts.
The aliquot volume should be small enough to allow high sample throughput. The performance of the laboratory equipment should correlate with that of the process-scaled devices. Plants are cultivated in a manner that represents the final process.
And all utilities are prepared before starting the extraction and flocculation experiments. The flocculant stock solutions are prepared carefully taking into account any pre-dilution by the manufacturer. Choosing the same stock concentration for all flocculants simplifies subsequent pipetting, reducing the likelihood of errors.
The buffer components, pH, and conductivity for each fresh flocculant stock solution are prepared to match the conditions of the extracts. The DoE schedule is converted into a pipetting scheme setting out the volumes of stock solutions to be added to the plant extract samples. Plant extract samples are prepared from harvested tobacco leaves using the buffers matching the conditions to be tested in the DoE.
While stirring the extract, aliquots are prepared, matching the pre-defined volume. This will ensure homogeneous particle distribution in all samples. Extraction buffer is added to each aliquot so that every sample has the same volume once a flocculant stock has been added.
The volume of stock solution defined in the DoE is added to the sample sequentially as indicated by the randomized run order. The polymer is homogeneously distributed throughout the sample by shaking the tube immediately after adding the polymer, which will also ensure that the flocculation results are reproducible. The method can easily be modified to test combinations of flocculants or the repeated addition of the same flocculant.
But this must be implemented in the DoE. Floc formation typically requires three to 30 minutes incubation time. Sticking to the incubation times defined in the DoE is necessary to achieve reproducible results.
Either the settling velocity of the flocs or the turbidity of the filtrates and filter capacities can be used to evaluate the flocculant performance. Confirming that the polymers do not interfere with products such as proteins, is important too. Once the data quality is confirmed, information can be transferred into the DoE software.
The experimental data are analyzed using the tools built into the DoE software including data transformation, factor selection, and quality control tools. As well as using these tools, the final model should also agree with the basic laws of physics and thermodynamics. Once the model quality is confirmed, the model can be used to predict the optimal conditions for flocculation.
Based on the model predictions, the design space can be narrowed down to the most relevant region for which a new DoE can be set up, repeating the procedure described before. The results can then be confirmed at the pilot or production scale. In this representative study, the performance of 23 different synthetic and four natural flocculants was tested for their ability to flocculate tobacco leaf extracts using a DoE approach.
Taking into account polymer-specific parameters such as polymer concentration and type, but also process parameters, such as incubation time, temperature, buffer pH, and conductivity. With the sequence of four nested IV-optimal designs, including a total of about 200 20 mL extract aliquots, we found that Polymin P was the most effective flocculant for tobacco extracts in the pH range four to eight, as long as the temperature was between four and 30 degrees Celsius. An incubation time of 15 minutes was sufficient for flocculation, but the effectiveness declined at conductivities greater than 40 mS per cm.
However, the DoE approach showed that Catiofast VSH was complementary to Polymin P, achieving effective flocculation at conductivities greater than 40 mS per cm. When these flocculants were introduced into the downstream processing of tobacco extracts, the capacity of the deployed depth filters increased 3.2 to 5.7-fold, reducing the cost of goods for this step accordingly. We confirmed that the flocculants did not reduce the recovery of target proteins including a monoclonal antibody, a fluorescence protein, a lactin, and different vaccine candidates.
The optimal flocculation conditions were also confirmed in a 100 L pilot scale batch. The results of the DoE were also successfully transferred to tobacco extracts prepared with a screw press rather than a blade-based homogenizer underlining how broadly applicable the collected information can be. We also adapted the method to test combinations of flocculants which worked equally well.
After watching this video, you should understand how to set up flocculation experiments for plant extracts using a DoE approach.
