1. Company
2. Media
3. Press articles
4. Press Releases

01/14/2009

Tests on rinsing methods for non-refillable PET bottles Part 2

Dipl.-Ing. (FH) Antonia Schmitt
Priv.-Doz. Dr.-Ing. Hartmut Evers
Dipl.-Ing. (FH) Jan Bechtluft
KHS AG, Bad Kreuznach

Part 1 of the article "Tests on rinsing methods for non-refillable PET bottles" dealt with the topics of rinsing and statistical experimental design and described the necessary preliminary investigations into the selected standard soiling substances which were mustard and Allura red AC. Part 2 presents a discussion of the results of both test series. Experimental design In a brainstorming session, the various influencing and dependent variables in a rinsing process were compiled and the experimental design established. The decision as to which influencing variables should be varied and which kept constant in the tests was based on the importance of accuracy/reproducibility and the amount of effort required. When selecting the dependent variables, the following aspects should be considered: relevance, quantifiability, completeness, and dissimilarity [Kleppmann 2006]. In our case, the chosen factors were number of ribs on the bottle, the pressures of the air and water rinsing media, and the surfactant concentration level. Two factor levels (smooth and ribbed) were set for the number of ribs and three factor levels for each of the other factors (see Table 1). The remaining influencing variables such as rinsing program, media temperature, and nozzle design were kept constant. See Table 1 As the aim of the rinsing process is to remove dust particles, the Allura red AC dye in power form is a realistic soiling substance in contrast to the mustard. Because of the type and quantity of soiling, the rinsing process for mustard resembles cleaning more than a simple rinsing process which is why the dependent variable is given as a cleaning rate here. For Allura red AC as the standard soiling substance, the rinsing rate is the dependent variable. A D-optimal experimental design was set up for both systems using the statistical program Visual-XSel^® 10.0 (see Table 2). This involves 20 tests each with triple determinations. The individual tests are arranged in groups. Random variation within these blocks should be as small as possible and each factor level combination should occur with equal frequency in order to minimize chance variation. The tests are also run randomly (in random sequence). This prevents the assessment of the factor effects being distorted by a trend or another unidentified variation in the results [Kleppmann 2006]. See Table 2 Standard soiling substances Two different standard soiling substances have been used in order to evaluate the effectiveness of a rinsing process for PET bottles. These are mustard and Allura red AC (see Figure 1). As described in Part 1, fat, starch, and pectin are not suitable for this purpose and were therefore eliminated after comprehensive tests. See Figure 1 After rinsing, the residues of the standard soiling substances previously added to the bottles were rinsed out using a defined amount of water.  These samples were used to determine the electrical conductivity in the case of mustard and the extinction in the case of Allura red AC. These results represent the residual soiling RS. The initial soiling IS was determined in the same manner using reference bottles. The cleaning rate CR or rinsing rate RR can be calculated from these values using the following equations: See calculation Evaluation of the experimental design Cleaning and rinsing rates The rinsing tests using mustard as the standard soiling substance produced cleaning rates of 61.0% to 96.5%. This equals a range of variation of 35.5%. With Allura red AC as the standard soiling substance, the rinsing rates lie between 98.70% and 99.99%, corresponding to a range of variation of only 1.29%. If the range of variation is large, then the individual factorial effects are more strongly developed and are easier to identify. Figure 2 shows the cleaning and rinsing rates for all factors. In the case of cleaning rate, the air pressure effect is greatest (11.5%), followed by the number of ribs (8.0%), water pressure (5.5%) and surfactant concentration level (3.0%). Rinsing rate is dependent primarily on the number of ribs (0.55%) and only to a minor degree on air and water pressure (both 0.05%), while no significant effect is demonstrable for the surfactant concentration level. There is an interaction between the number of ribs and the water pressure for both the cleaning rate and rinsing rate. The maximum calculated is obtained with the lower setting for number of ribs and surfactant concentration and the upper setting for air and water pressures. See Figure 2 See Figure 3 Figure 3 illustrates the gradient of each factor and their effects on the dependent variable. The steeper the gradient, the greater the effect. This is decisive for interpreting the data. If the gradient is negative, then the factor has a negative affect on the dependent variable. The bottle shape plays a decisive role effect for rinsing. Cleaning and rinsing rates are inversely proportional to the number of ribs. If the bottle has a smooth surface, the water can flow off evenly and rinse away the dirt. The ribs present an obstacle to this process by slowing down the rinsing medium and therefore reducing the mechanical cleaning effect. Between the ribs are dead areas where rinse water does not penetrate well. In addition, the surface area of ribbed bottles is larger. In order to obtain the same cleaning effect for ribbed bottles as for smooth ones, a larger quantity of rinse medium would be necessary i.e. it would be necessary to lengthen the rinsing times or increase the pressure. Under identical conditions, it is obviously impossible to obtain the same results as with a smooth bottle. During the tests with Allura red AC it was apparent that numerous water drops remained on the ribs. Although the dye particles were mostly removed from the surface of the bottle, they were not completely rinsed out of it. This led to a considerable reduction in the rinsing rate. The cleaning and rinsing rates are proportional to the air and water pressure, although the effects are of differing magnitude. Higher pressure in the rinsing medium increases turbulence which accelerates the removal of the dirt. An increase in pressure also produces an increase in the flow rate. With constant rinsing times, this means that more water is injected into the bottles, enabling a larger quantity of soiling to be removed. However, from a particular point onwards this is counterproductive because the water flow gurgles in the bottle neck and cannot run out. See Figure 4 The higher the surfactant concentration level of the water used for rinsing, the lower the cleaning rate. Surfactants are used in practice to reduce the surface tension. This should improve the wetting of the PET and therefore the effectiveness of rinsing. Because the entire surface is soiled with mustard in these tests and at the start of the rinsing process there is no contact between the bottle and the rinsing medium, this effect does not occur. The surfactant used causes slight foaming. This foam reduces the rinsing effect and also the power of the jet from the nozzle. This is a common effect with high amounts of foam. Using surfactants to lower the surface tension results in a reduction in drop size [Keck 2005]. According to Loncin [1977], the kinetic energy and thus the effectiveness per kilogram of cleaning solution is lower for smaller drops. When rinsing out dye particles, no significant effect of surfactant concentration could be determined. This can be explained by the fact that the negative effects of additional surfactant described are compensated by the improved wetting. See Figure 5 Interactions An interaction between two factors means that the effect of one is dependent on the value of the other. Figure 5 shows the significant interactions. The blue lines refer to the number of ribs and the red ones to the water pressure. (+) indicates the upper factorial level in each case and (-) the lower one. For the cleaning rate it can be seen that the effect of water pressure for smooth bottles is 10 times as high as for a bottle with 18 ribs. The cause of this interaction could be the decline in mechanical cleaning power on the ribs and the dead areas between them. This means that an increase in the water pressure for ribbed bottles only produces a minor improvement in the cleaning effect. The rinsing rate is not dependent on water pressure in smooth bottles. One explanation might be that, as a result of the lower degree of soiling, the optimum for rinsing out the dye particles is already achieved at the lower water pressure level and therefore no further improvement can be made. In the case of ribbed bottles, increasing pressure produces a better rinsing result. Because a higher pressure is necessary for the same rinsing rate, the optimum must therefore lie at a higher level. see Figure 6 The rinsing tests reveal that the dirt is rinsed out from top to bottom. The largest amount of residual soiling always occurs in the lower part, when looking at the bottle upside down in the rinsing position, while the upper part (the base) is mostly visually clean (see Figure 6). When the jet hits the base of the bottle, this has the greatest mechanical cleaning energy and can remove a large amount of dirt. When running down the wall of the bottle this energy declines and so no longer has the same effect. Conclusion and comparison of methods The test series using mustard as the standard soiling substance produces more marked differences in the mechanical effects of rinsing than that using Allura red AC. In addition, a visual assessment of the rinsing results can be carried out, allowing critical areas on the bottle to be identified. Please note: mustard is certainly not a realistic soiling substance. In order to be able to evaluate the results of the tests, this soiling must be so persistent that a measurable residual soiling always remains in the bottle. One drawback is that the bottles must be prepared six days before the tests can be carried out. However, it would be possible to shorten the drying time if the rinsing program were to be reduced. One advantage of Allura red AC is that bottles can be prepared quickly and easily, after which the tests can be carried out immediately or at a later date. Furthermore, the dye particles constitute a realistic soiling versus mustard. Mustard is suitable for examining the mechanical effects of rinsing and for carrying out spray shadow tests. On the other hand, Allura red AC could be used for evaluating different bottle shapes and real rinsing programs. A positive result would then require that an as yet undefined limit value not be exceeded. In addition, Allura red AC could be used if quick results are required. See Table 3 The cleaning rate (mustard) is calculated according to equation 1, the rinsing rate (Allura red AC) according to equation 2. These are listed in the chapter on standard soiling substances. No target values are given because the results depend on many different factors (e.g. rinsing program, bottle size, and bottle shape). Bibliography Keck, D.: Einflussfaktoren auf die Zerstäubung, Lechler GmbH, Metzingen, 2005, pp. 4–21 Kleppmann, W.: Taschenbuch Versuchsplanung, Carl Hanser Verlag, Munich and Vienna, 2006, pp. 10–39 Loncin, M.: Modelling in cleaning, disinfection and rinsing, Proc. Symposium Mathematical modelling in food processing, 1977, pp. 301–335