|
Evaluating Kaolin Clay as a Potential Substance for ISO Sprayer Cleaning System Tests
Carla Román1, Hongyoung Jeon2,*, Heping Zhu2, Erdal Ozkan3
Published in Applied Engineering in Agriculture 39(3): 347-358 (doi: 10.13031/aea.15466). 2023 American Society of Agricultural and Biological Engineers.
1Department of Food, Agricultural, and Environmental Engineering, Ohio State University, Wooster, Ohio, USA.
2Application Technology Research Unit, Agricultural Research Service, USDA, Wooster, Ohio, USA.
3Department of Food, Agricultural, and Environmental Engineering, Ohio State University, Columbus, Ohio, USA.
*Correspondence: hongyoung.jeon@usda.gov
The authors have paid for open access for this article. This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License https://creative commons.org/licenses/by-nc-nd/4.0/
Submitted for review on 20 November 2022 as manuscript number MS15466; approved for publication as a Research Article by Community Editor Dr. Joe Luck of the Machinery Systems Community of ASABE on 5 May 2023.
Mention of company or trade names is for description only and does not imply endorsement by the USDA. The USDA is an equal opportunity provider and employer.
Highlights
Abstract. ASP® 602 (kaolin clay) was evaluated as an alternate material for performing ISO 22368-1 sprayer clean out test standard. Recommendations for sample homogenization, measurement temperature and sample sizes to address potential concerns or technical challenges in assaying ASP® 602 samples were provided under relevant conditions. Linear regression models for predicting ASP® 602 concentrations from 0 to 100 ppm (mg·L-1) in sample mixtures were developed using outputs from a spectrophotometer and two different turbidimeters. Test results showed that the three instruments could measure ASP® 602 concentrations as low as 2 ppm. Validations of the developed models showed approximation errors were 0.9% to 24.6% and 0.1% to 18.4% for the spectrophotometer and the turbidimeters, respectively. However, their maximum absolute errors were less than 3.7 ppm for the spectrophotometer and 2.3 ppm for the turbidimeters within the validation range. Two agitation and cleaning systems of a sprayer were evaluated with ASP® 602 following the ISO 22368-1 clean out procedure. Residue reduction factors from the initial mixture concentration were 163 for one system which did not meet ISO 16119-2 environmental requirement (factor over 400) and 819 for the other system which exceeded the requirement. Analysis results of ASP® 602 samples from ISO 22368-1 clean out tests showed that ASP® 602 could be used as a potential testing substance to evaluate and differentiate the performance of agitation and cleaning systems of sprayers.
An effective way to protect crops from pests and retain their maximum yield potential is to apply crop protection products (CPP). Although the benefits of using CPP are compelling, its drawbacks are also concerning, e.g., human exposure and environmental contamination while mixing and applying CPP, and cleaning application equipment. Engineering solutions to minimize those risks are effective for all applicators; however, since the designs and configurations of agricultural sprayers (especially the shape of sprayer tanks and the agitation systems in the tanks) vary widely between sprayer manufacturers, applying uniform solutions to all sprayers is impractical and would run into implementation issues. Thus, establishing standards for testing the sprayer performance can be useful to sprayer manufacturers since these standards can provide design baseline for new sprayers (Tamagnone et al., 2012) and inspection guidelines for existing sprayers (Balsari et al., 2012). This approach will also be beneficial to applicators to assure that their sprayers meet safety and performance standards. It is also the reason that the European Directive 2009/127/EC (European Union, 2009a) establishes such performance standards for equipment manufacturers to produce new spray equipment. The series of ISO 16119 standard (ISO, 2013a, b, c) describes the environmental requirements that new sprayers must meet in the EU. In addition, the EU Directive on sustainable use of pesticides also requires all existing sprayers to be inspected every 3 years to apply CPPs (European Union, 2009b).
One of the critical requirements in the standard is specifically regarding the performance of the sprayer cleanout system. Since cleanout recommendations from different CPP labels vary widely, and various sprayer designs are available, having one cleanout procedure for all CPPs is impractical. However, an internal cleaning system for basic sprayer cleanout should be an essential feature for a sprayer thus, applicators can clean out CPP residues from their sprayer after applications. Although cleanout performance of internal cleaning system can be subjective since it depends on the sensitivity of crops to active ingredients for subsequent applications, a standard procedure is necessary to evaluate if a sprayer has an acceptable cleaning system in removing CPP residues from mixture tanks and spraying lines after applications.
The ISO 22368-1 standard (ISO, 2004) describes the procedures for the evaluation of sprayer cleaning systems. It recommends copper oxychloride, like other ISO standards, as the testing substance for the evaluation. However, there are concerns about using this substance due to potential health risks (Thompson et al., 2012) and, soil and groundwater contamination (European Food Safety Authority et al., 2018). In addition, using copper oxychloride for testing sprayer cleaning system is a challenge because the availability of copper oxychloride has been very limited. Also, its suspension is chemical waste which requires to contain entire suspension in a waste container during the test. As a result, equipment manufacturers are reluctant to test cleaning systems of their sprayers due to these complications (Tomagnone et al., 2016; Gil et al., 2018).
Due to the drawbacks of copper oxychoride, the efforts to find alternate substance to test sprayer cleaning system have been made. For example, Andersen et al. (2010) evaluated a continuous cleaning method in three different sprayers using natrium fluorescein, a water-soluble dye, as a testing substance for evaluating the efficiency of the cleaning system. The dye concentrations were used to determine if the cleaning method was effective enough to reduce by at least 1% the initial concentration to meet the cleaning out requirements of French Ministry of Agriculture (Ministère de l’Agriculture, 2006). However, dyes at typical concentrations for testing a sprayer cleaning system are fully soluble in water, thus the performance of sprayer cleaning systems may not influence the test outcomes. Therefore, water insoluble wettable powder would be more suitable for testing cleaning system since it would require effective agitation and cleaning system to keep suspending it in the spray tank and clean it (Ali and Ali, 2017).
Alternative testing substances of copper oxychloride to address aforementioned issues have been sought for past few years by ISO working group (WG) as well, and a new substance, ASP® 602 (Badische Anilin und Soda Fabrik (BASF) Corp., Ludwigshafen, Germany), has been identified, evaluated, and included in ISO 5682-4 (ISO, 2021). ASP® 602 is a naturally occurring processed kaolin clay. It is expected to have a more desirable toxicity profile to the environment than copper oxychloride since it has already been approved for use as an organic crop production product for controlling arthropod in the U.S. due to its favorable toxicity profiles (Puteka and Glenn, 2008).
WG 6 of subcommittee 6 of ISO technical committee 23 is currently working on revision of ISO 22668-1 as a preliminary work item, and one proposal for ISO 22368-1 revision is to use ASP® 602 as a testing substance. However, utilizing new testing substance in existing standard creates unforeseen challenges, one of which is available analytical methods to quantify ASP® 602 from test samples. While current ISO 22368-1 suggests atomic absorption spectrometry as an analytical tool for quantifying copper oxychloride, it is unknown that if an atomic absorption spectrometry can be utilized to quantify ASP® 602 until the development of an analytical method using an atomic absorption spectrometer is accomplished.
Common instruments to measure the concentrations of kaolin clay suspensions are turbidimeter and spectrophotometer. For example, Ucar et al. (2000) used kaolin clay as a testing substance to evaluate the performance of jet agitation systems in sprayer tanks and analyzed the mixture concentration using a turbidimeter. On the other hand, Puterka et al. (2000) developed a methodology to extract kaolin clay from leaves with methanol after spraying its suspension on them so the clay was suspended and homogized in water to quantify the deposition of the clay suspension with an spectrophotometer using a wavelenght of 400 nm. In addition, turbidity sensors could be used to monitor spray mixture concentrations produced from the premixing in-line injection system developed for precision variable-rate orchard sprayers (Zhang et al., 2022).
More recently, Jeon and Zhu (2022) developed a device coupled with a light-emitting diode (LED) with a peak wavelength of 700 nm and a photodiode to measure ASP® 602 concentrations of suspension samples. The device was capable to measure kaolin clay concentrations below 25 ppm, the current requirement in ISO 16119-2 (ISO, 2013b) standard for spray cleaning system, which meets the minimum specification for instruments to assay samples from ISO 22368-1 tests. However, this work was focused on the device development and sensitivity analysis guideline. Therefore, there are still uncertainties in characterization of ASP® 602 suspension samples and measurement variations with different instruments.
The objective of this work was to evaluate ASP® 602 as a potential testing substance for ISO 22668-1 standard test. The specific goals were (1) to evaluate kaolin clay suspension samples to provide guidelines for sample homogenization measurement temperature, and assay performance and variations between a spectrophotometer and two different turbidimeters, and (2) to simulate ISO 22668-1 tests using a laboratory sprayer with two different agitation and cleaning systems in a laboratory condition to evaluate the feasibility of kaolin clay differentiating the effectiveness of sprayer cleaning systems.
Materials and Methods
Kaolin Clay Suspension Samples
ASP® 602 (Al2H4O9Si2, Lot No.: 00623T 096 56335171) was used as a testing substance throughout the tests. It was formed with dried hydrous white kaolin clay particles with an average particle size of 0.6 µm, and a bulk density of 720 kg m-3 (BASF, 2020). ASP® 602 suspension samples were prepared in 125 mL wide-mouth jars (FB02911775, Thermo Fisher Scientific Inc., Waltham, Mass.) with an analytical balance (XS205DU, Mettler Toledo Co., Greifensee, Switzerland) for weighing ASP® 602 and a precision balance (Practicum1102-1x2, Sartorius AG, Göttingen, Germany) for weighing tap water (TW). Two different balances were used due to the differences in weight ranges for ASP® 602 and TW.
Sample Homogenization
Three 100-mL ASP® 602 samples at intended concentrations of 20, 50, and 100 ppm (mg·L-1) were prepared as mentioned above (table 1). After being on the bench for 24 h to precipitate, each sample was homogenized by stirring it with a magnetic stirrer (84003-71, Cole Parmer Instrument Co., Chicago, Ill.) and a magnetic stir-bar (14-513-59, Octagon Spinbar™, Thermo Fisher Scientific Inc. Waltham, Mass.). To define the best procedure for the homogenization, the samples were homogenized at three stirring speeds (300, 500, and 750 r·min-1) and were sampled at five different times (1, 2, 3, 4, and 5 min) while being homogenized at the given stirring speed. ASP® 602 samples were subsampled prior to the homogenization and at each homogenization time at each rotational speed. Absorbance of each subsample at 400 nm (Puterka et al., 2000) was measured three times with a spectrophotometer (GenesysTM 140, Thermo Fisher Scientific Inc., Waltham, Mass.) (fig. 1a). Each subsample was vigorously homogenized between measurements. Since actual concentrations of ASP® 602 samples were slightly varied, absorbance of the subsamples was normalized by the intended concentration to compare between homogenization conditions using:
(1)
| Table 1. Actual concentrations of ASP® 602 samples at three stirring speeds to evaluate sample homogenization recommendations for absorbance measurements. | |||
| Stirring Speed (r·min-1) | ASP® 602 Concentration (ppm) | ||
| 20[a] | 50 | 100 | |
| 300 | 19.1[b] | 56.8 | 106.9 |
| 500 | 20.0 | 49.9 | 110.1 |
| 750 | 19.0 | 50.0 | 102.0 |
[a]Intended concentrations of ASP® 602 in ppm. [b]Actual concentrations of ASP® 602 in ppm. | |||
Measurement Sample Size Requirements
Three ASP® 602 samples below 25 ppm (15, 21, and 23 ppm) were prepared to determine the required number of sample absorbance measurements at the wavelength of 400 nm. Three samples were subsampled 25 times and absorbance of initial 3, 5, 10, and 25 subsamples was measured to observe the variability along with the sample size. Data normality for subsamples was evaluated with Shapiro-Wilk test, as the normal distribution of the measurement data was critical to determine the limit of the detection. In addition, coefficient of variations (CV) for different subsample sizes were calculated to determine the relationships between variations and the sample size. JMP® Pro v.15 (SAS Institute Inc., Cary, N.C.) was used to analyse the normality of data with different sample sizes.
Measurement Performance and Variation of Instruments
Measurement Instruments
Three instruments were used to evaluate and compare their measurement performances and variations for ASP® 602 samples. They were the spectrophotometer, which measured absorbance of the sample (fig. 1a), a stationary fluorometer (ST) (fig. 1b, Trilogy fluorometer, Turner Designs Inc., San Jose, Calif.) with a turbidity module (7200-060) and a portable fluorometer and turbidimeter (PT) (fig. 1, AquaFluor, Turner Designs Inc., San Jose, Calif.). The two turbidimeters were used to evaluate the impacts of measurement resolution in predictions of ASP® 602 concentrations. ST and PT had resolutions of 0.05 NTU (Nephelometric Turbidity Unit, Turner Designs, 2022b) and 0.5 NTU (Turner Designs, 2022a), respectively, although they had the same detection range (0-1000 NTU).
   | ||||||
| (a) Spectrophotometer | (b) Stationary turbidimeter | (c) Portable turbidimeter | ||||
| Figure 1. Laboratory instruments to measure concentrations of ASP® 602 suspensions. | ||||||
A set of 100-mL ASP® 602 samples (2.0, 11.0, 25.0, 47.9, and 102.8 ppm) was prepared for the calibration as mentioned. After the samples were stirred at 750 r·min-1 for at least 2 min, 10 subsamples of approximately 3.5 mL were transferred from the top layer of the suspension, last place to be homogenized by the stirrer, into disposable polystyrene cuvettes (14-955-125, Thermo Fisher Scientific Inc., Waltham, Mass.) by disposable pipets. Then, the cuvettes were immediately sealed with square caps (14-385-999, Thermo Fisher Scientific Inc. Waltham, Mass.). Before analyzing the subsamples with the different instruments, the sealed cuvettes were re-homogenized by a vortex mixer (Vortex Genie-2, Scientific Industries, Bohemia, N.Y.) for 5 s and then inverted the cuvettes three times to move bubbles out of the measurement area of the cuvettes. Sample absorbance was measured by the spectrophotometer. ST and PT measured turbidity of the samples in Raw Turbidity Units (RTU) and NTU, respectively.
Averages, standard deviation and variations (CV) of absorbance and turbidity data were calculated and the averages were used to build regression models to predict ASP® 602 concentrations for each instrument using MS-Excel (Microsoft Co., Redmond, Wash.). Shapiro-Wilk test was performed to check data normality as mentioned before for sensitivity analysis.
In addition, a set of ASP® 602 samples (5.0, 15.0, 22.0, 28.0, 30.6, and 88.8 ppm) was prepared to validate the regression models of three instruments for the samples. Their accuracy was evaluated by the approximation error (eq. 2) as reported in Jeon and Zhu (2022).
(2)
Effect of Ambient Temperatures
The temperature effects on the measurements and predictions of the instruments, spectrophotometer and PT, were analyzed. A set of ASP® 602 samples (14.0, 21.0, 48.0, 100.0, and 250.9 ppm) was prepared, and the instruments were calibrated with the samples while maintaining their temperature at 30°C in a temperature-controlled room (Thermolinear, Cincinnati, Ohio). After the calibration, validation samples (24.9, 71.0, 149.9, and 207.0 ppm) were also prepared and sealed with paraffinic film (Parafilm® M, Bemis Co., Inc, Neenah, Wis.) and square caps to minimize concentration changes due to evaporation. The instruments and ASP® 602 samples were kept in the temperature-controlled room where intended temperature was adjusted to 10°C, 20°C, and 30°C. The room temperatures were within 5.0% to 9.0% of the intended temperatures. At each temperature, absorbance and turbidity of the samples were measured to analyze the temperature effect on measurement and prediction results. The samples and the instruments were placed in the room for at least 24 h before the measurements.
Evaluation of Kaolin Clay as Test Substance for Sprayer Cleaning system
Laboratory Sprayer
Two ISO standard sprayer cleanout tests were conducted using a laboratory sprayer (Hydra SprayerTM, Dramm Corp., Manitowoc, Wis.) to evaluate ASP® 602 as a test material. The sprayer was equipped with a 37.9-L high-density polyethylene square tank (SP00010SWSS, Ace Roto-Mold, Den Hartog Ind., Hospers, Iowa) and a closed-coupled pump (MS40, Hydra, Dramm Corp., Manitowoc, Wis.) which was driven by a sealed electric motor (056B17F5310, Marathon® Electric Motors, Regal Rexnor Corp., Wausau, Wis.) to generate a flow rate of 15.2 L·min-1. A flat-fan nozzle (XR8006, TeeJet Technologies LLC, Glendale Heights, Ill.) was used to simulate spray application.
Agitation and Cleaning Systems
 |  |
| (a) Sprayer with agitation and cleaning configuration C1 | (b) Sprayer with agitation and cleaning configuration C2 |
| Figure 2. Schematics of the laboratory sprayer agitation and cleaning systems. | |
Two agitation and cleaning configurations of the sprayer were used for the cleanout tests. Hydraulic diagrams of the sprayer with different agitation configurations are shown in figure 2. First configuration (referred as C1 hereafter) was equipped with a jet agitator and a custom made sparge bar (fig. 2a). The agitator was located at a corner of the tank and had two orifices at the end to discharge liquid jet to sweep the bottom of the tank. Custom-made sparge bar was made from a 19-mm diameter polyvinyl chloride pipe, and it was located at the center of the tank with ten 3.2-mm orifices. Five sets of two orifices on the bar discharged liquid jets to opposite directions to homogenize tank mixture on both sides at different heights. An additional orifice was at the end of the bar for more homogenization below the bar and kept the clay suspending. In addition, a tank flange (TF125AV, Banjo, Crawfordsville, Ind.) was used to attach the tank outlet to a pipe to the pump. The flange left a rim of approximately 6.5-mm around the outlet. A throttle valve (23520, TeeJet Technologies LLC, Wheaton, Ill.) controlled the flow between the agitator and the sparge bar.
The other configuration (referred as C2 hereafter) was equipped with a rinsing nozzle and a sparge bar (fig. 2b). They both were custom-made in the laboratory. A 19-mm diameter polypropylene pipe was used to make the rinsing nozzle which was installed at the upper center of the tank. The rinsing nozzle had 18 orifices (one 1.6-mm orifice every 20°) around the cylinder to discharge rinsate to most internal surface of the tank, and one orifice at the bottom. Similarly, the sparge bar was also made with the same size polypropylene pipe as in C1, except it was located at one corner of the tank. The sparge bar had four 2.4-mm orifices and a jet agitator (BXTMMP6, Bex, Ontario, Canada) at the end for tank mixture homogenization. A custom-made tank flange was used to eliminate the rim around the tank outlet. A pressure relief valve (23120A, TeeJet Technologies LLC, Glendale Heights, Ill.) controlled the flow between the sparge bar and the rinsing nozzle. Table 2 summarizes key differences between two agitation and cleaning configurations.
| Table 2. Key differences between configurations C1 and C2 for the laboratory sprayer. | ||
| Component | Sprayer with C1 | Sprayer with C2 |
| Tank outlets | Rim (approx. 6.5 mm) around the tank outlet. | No rim at the tank outlet. |
| Agitation system | Ten 3.2-mm orifices along the sparge bar and one 3.2-mm orifice at the end. Additional jet agitator (two orifice). | Four 2.4-mm orifices along the sparge bar and one jet agitator at the end. |
| Cleaning system | No cleaning nozzle. | Custom made cleaning nozzle in the tank. |
Evaluations of Agitation and Cleaning System Performance with ASP® 602
Performance of the two agitation and cleaning systems described previously was evaluated by using ASP® 602 as the test substance and by following the ISO 22368-1 test protocol. Before the test, the sprayer was thoroughly cleaned with 37.9-L tank mixture containing spray tank cleaner (Gempler’s, Janesville, Wis.) followed by triple rinses of the tank with water. After the cleaning, the sprayer was dried for 24 h before the tests were conducted. In addition, the throttle valve for the sprayer with C1 was adjusted for the jet agitator and the sparge bar to have the flow rates of 7.2 and 8.0 L·min-1, respectively. Also, similar adjustment was made with the pressure relief valve of the sprayer with C2 to have the flow rates of 7.8 and 7.4 L·min-1 for the rinsing nozzle and the sparge bar, respectively.
Clean out test and sampling procedures described in ISO 22368-1 were followed for both systems except that entire tank mixture was collected from the sprayer to weigh it with a scale (ISI10, Sartorius AG, Göttingen, Germany) at every step to calculate the mass balance of ASP® 602 during the tests. Moreover, additional rinsate sample after each rinse was collected for the same purpose (SRinse1, SRinse2, and SRinse3). After performing ISO 22368-1 test procedures, additional cleaning was performed for complete removal of ASP® 602 in the sprayer. The sprayer tank bypass line was flushed with TW and a sample was collected to collect residual ASP® 602 in the recirculation line of the sprayer (SBP). Then the tank was cleaned with a pressure washer and a rinsate sample was collected (SWash). The dead volume left inside the sprayer circuit and residual product in the inline filter cup were collected as a final sample (SDV). The volume of each sample was at least 50 mL, and the sample at each step had three replicates.
ASP® 602 of 380.01 and 380.20 g were added to the laboratory sprayer while testing the sprayer with C1 (test 1) and C2 (test 2), respectively. Thus, the initial concentrations of 10.09 and 10.66 g·L-1 should be achieved for tests 1 and 2, respectively, based on ASP® 602 amount and TW volume. The collected samples were visually assessed and divided into two groups based on their cloudiness. The samples expected to be high concentration (above 500 ppm) were placed in a laboratory oven (6925, Thermo Fisher Scientific Inc., Waltham, Mass.) for 24 h at 80°C to evaporate water from the sample to acquire the mass of ASP® 602. Total dissolved solid (TDS) in TW was measured by the same method (287 ± 11 ppm for test 1 and 295 ± 14 ppm for test 2) to compensate while calculating ASP® 602 mass.
The spectrophotometer was used for assaying samples with low concentrations (below 500 ppm). The instrument was calibrated (fig. 3) and validated before analyzing the samples. The regression models had high R2 values, ranging from 0.9995 to 0.9996 with approximation errors of 0.5% to 5.2% and 0.2% to 8.1% for the test 1 and 2, respectively.
Cleaning performance of the sprayer with each agitation and cleaning system was evaluated with (ISO, 2004):
Click or tap here to enter text. (3)
where, F is the percentage of mean concentration after rinsing, LRef is the mean of measured initial sample concentrations (SRef), and LARis the mean of measured sample concentrations after rinsing the internal sprayer system (SAR).
Results and Discussion
Sample Homogenization
Table 3 shows the absorbance changes of ASP® 602 samples over homogenization time at different stirring speeds. The samples with low ASP® 602 concentrations tended to have more variations regardless of stirring speed and time. For example, measurement CVs of 20-ppm samples were 3.3% to 15.6%, 4.3% to 16.5%, and 5.7% to 14.2% with stirring speeds of 300, 500, and 750 r·min-1, respectively. However, the CV ranges for 50-ppm samples were decreased to 3.0% to 7.0%, 0.7% to 14.2%, and 3.8% to 6.2% with stirring speeds of 300, 500, and 750 r·min-1, respectively. This tendency might be associated with achieving the uniform distribution of ASP® 602 particles in the sample at the low concentration might be challenging due to small numbers of the particles in the samples. However, although measured CVs were relatively high (>10%), the average value of absorbance for most ASP® 602 samples was stabilized with less than 10% CVs after stirring the sample more than 1 min.
 |  |
| (a) Model for assaying samples from test 1 | (b) Model for assaying samples from test 2 |
| Figure 3. Linear regression models to predict ASP® 602 concentrations from ISO 22368-1 test. | |
Although the results of the tests conducted may suggest that the samples with precipitated ASP® 602 might require different homogenization intensity and time based on sample concentrations, such guidelines would be impractical. Therefore, stirring the sample for at least 1 min or longer at the stirring speed of 750 r·min-1 was sufficient based on our results since it achieved reasonably uniform ASP® 602 suspension (CV < 15%) with average absorbance similar to converged concentrations. In addition, this guideline was used to homogenize the samples with precipitated ASP® 602 for other evaluations.
| Table 3. Normalized absorbance of ASP® 602 samples at different stirring speeds (300, 500 and 750 r·min-1) sampled at different homogenization times (0 to 5 min). | |||||||
| Intended Concentration (ppm) | Sample Stirring Speed (r·min-1) | Homogenization Time (min) | |||||
0 | 1 | 2 | 3 | 4 | 5 | ||
| 20 | 300 | 0.008[a] (13.6)[b] | 0.022 (15.6) | 0.023 (6.0) | 0.023 (5.2) | 0.026 (5.7) | 0.025 (3.3) |
| 500 | 0.008 (16.5) | 0.024 (7.1) | 0.023 (10.4) | 0.025 (10.9) | 0.025 (4.6) | 0.023 (4.3) | |
| 750 | 0.008 (11.1) | 0.024 (14.2) | 0.025 (5.7) | 0.025 (8.3) | 0.026 (5.8) | 0.025 (6.5) | |
| 50 | 300 | 0.013 (4.0) | 0.046 (7.0) | 0.050 (4.0) | 0.048 (3.0) | 0.050 (3.8) | 0.049 (3.7) |
| 500 | 0.014 (14.2) | 0.051 (0.7) | 0.050 (4.0) | 0.049 (3.2) | 0.051 (5.1) | 0.052 (5.6) | |
| 750 | 0.012 (5.3) | 0.060 (6.2) | 0.055 (3.8) | 0.055 (5.5) | 0.057 (4.3) | 0.056 (5.3) | |
| 100 | 300 | 0.022 (4.6) | 0.092 (2.0) | 0.089 (2.7) | 0.094 (6.1) | 0.094 (2.7) | 0.098 (3.3) |
| 500 | 0.017 (3.1) | 0.102 (4.1) | 0.111 (2.8) | 0.108 (3.1) | 0.108 (4.1) | 0.102 (2.8) | |
| 750 | 0.026 (3.8) | 0.115 (7.2) | 0.110 (7.2) | 0.116 (3.8) | 0.110 (5.9) | 0.112 (4.9) | |
[a] Measurement units for the spectrometer was AU (absorbance unit). [b] Coefficient of variation (%) of absorbance measurements. | |||||||
Measurement Sample Size Requirement
| Table 4. Average absorbance of ASP® 602 suspensions with different measurement populations and their probabilities for normal distribution. | ||||||||||||||
| Concentration | 15 ppm | 22 ppm | 23 ppm | |||||||||||
| Sample Size | 3 | 5 | 10 | 25 | 3 | 5 | 10 | 25 | 3 | 5 | 10 | 25 | ||
| Average | 0.0201 (2.17)[a] | 0.0198 (2.94) | 0.0200 (2.99) | 0.0199 (5.27) | 0.0271 (3.24) | 0.0270 (3.79) | 0.0268 (2.43) | 0.0270 (3.95) | 0.0284 (1.17) | 0.0279 (2.48) | 0.0281 (2.09) | 0.0283 (2.61) | ||
| P- value | 0.0219[b] | 0.6575 | 0.9401 | 0.9148 | 0.6780 | 0.7452 | 0.4373 | 0.1247 | 0.0288 | 0.5953 | 0.5895 | 0.2743 | ||
[a]Measurement coefficient of variation (%). [b]Probabilities from Shapiro-Wilk normality tests indicate normally distributed data or highly deviated data with the p-values greater than 0.05 data and less than 0.05, respectively. | ||||||||||||||
Table 4 shows the probabilities from Shapiro-Wilk normality tests and the average of 3, 5, 10, and 25 absorbance measurements of ASP® 602 samples with concentrations of 15, 22, and 23 ppm. The maximum differences of the average absorbance were 0.0003, 0.0003, and 0.0005 AU (absorbance units) for 15, 22, and 23 ppm samples, respectively. These absorbance differences among the different sample sizes were from 0.70% to 1.66% of the average absorbance of 25 measurements.
The absorbance differences tended to increase as the sample concentrations increased potentially due to the increase of the output magnitude as the concentrations increased. Measurement variations (CV) also increased as the sample size increased as expected because larger sample sizes were more closely to represent the sample. However, the number of measurements might influence the average absorbance by less than 2%, and its impact on prediction results would vary with calibration models.
Results of Shapiro-Wilk normality tests indicated that the absorbance data with a sample size of at least five to be normally distributed samples, while the sample size of three tended to have highly deviated distribution. This implies that the sample size for building calibration curves or calculating a limit of quantification (LOQ) for an instrument should be at least 5 or larger to have normally distributed measurements (Armbruster and Pry, 2008). The data normality trend with the sample sizes was consistent over the tests with three ASP® 602 samples below 25 ppm.
Measurement Performance and Variations of Instruments
Table 5 shows average measurement outputs from three instruments for the calibration samples. Two of the instruments had measurement issues with blank (TW) samples. For example, the spectrophotometer was sensitive enough to vary its output for TW samples and consequently, its measurements for the blank sample had a large CV. On the other hand, PT was less sensitive to detect the variations from TW samples and outputted zero NTU for all samples. All three instruments had the measurement variations less than 10% within the calibration range (0-100 ppm), however, variations of the instrument outputs were relatively higher at low ASP® 602 concentrations as mentioned before. These variations might also imply more prediction errors for samples at low concentrations.
| Table 5. Average measurement values of ASP® 602 calibration samples from three instruments (sample size=10). | |||||||
| Instrument | Parameter | ASP® 602 Concentration (ppm) | |||||
| 0 | 2.0 | 11.0 | 25.0 | 47.9 | 102.8 | ||
| Spectrophotometer | Average measurement (AU) | 0.000 | 0.007 | 0.017 | 0.031 | 0.050 | 0.097 |
| Standard deviation | 0.000 | 0.001 | 0.001 | 0.001 | 0.001 | 0.001 | |
| Coefficient of variation (%) | -40.7 | 8.2 | 3.5 | 2.5 | 1.5 | 1.3 | |
| P-value from Shapiro-Wilk test | 0.3996[a] | 0.3132 | 0.2494 | 0.6734 | 0.5258 | 0.4300 | |
| Stationary turbidimeter | Average measurement (RTU) | 1135.9 | 3816.1 | 7738.3 | 13883.4 | 21940.6 | 44062.8 |
| Standard deviation | 21.6 | 310.8 | 483.9 | 747.5 | 1020.5 | 2443.5 | |
| Coefficient of variation (%) | 1.9 | 8.1 | 6.3 | 5.4 | 4.7 | 5.5 | |
| P-value from Shapiro-Wilk test | 0.3943 | 0.1105 | 0.7708 | 0.2131 | 0.9289 | 0.9640 | |
| Portable turbidimeter | Average measurement (NTU) | 0.0 | 2.3 | 7.2 | 14.4 | 23.0 | 50.4 |
| Standard deviation | - | 0.2 | 0.4 | 0.5 | 0.8 | 1.0 | |
| Coefficient of variation (%) | - | 6.9 | 6.0 | 3.5 | 3.7 | 1.9 | |
| P-value from Shapiro-Wilk test | - | 0.2787 | 0.0421 | 0.1280 | 0.3266 | 0.9331 | |
[a]Probabilities from Shapiro-Wilk normality tests indicate normally distributed data or highly deviated data with the p-values greater than 0.05 and less than 0.05, respectively. | |||||||
Linear regression models developed from outputs of the instruments to predict ASP® 602 concentrations had high coefficients of determination (R2) ranging from 0.9938 to 0.9973 (fig. 4). This implies that the regression models had a good fit with the calibration data and would have good predictability of ASP® 602 concentrations with respect to outputs from the instruments. Although all regression models were linear and had similar trends, the slopes and intercepts of the models were different due to the differences of the instrument sensitivity and output magnitudes.
 |
| (a) Spectrophotometer |
 |
| (b) Stationary turbidimeter |
 |
| (c) Portable turbidimeter |
| Figure 4. Linear regression models to predict ASP® 602 concentrations using outputs from three different instruments with the prediction range from 0 to 100 ppm. |
Sensitivities were analyzed for each instrument. Since PT had the average measurements of zero for the blank, its Limit of Blank (LOB) could not be determined. For the spectrometer and ST, the guideline from Armbruster and Pry (2008) to calculate LOB could not lead to meaningful LOB values, as their outputs for the blank samples were substantially lower than measurements for the sample with the lowest concentration of ASP® 602. Limit of Detections (LODs) for the instruments were estimated using linear regression models and instrument outputs of the lowest concentration (2 ppm) as shown in figure 5. The measurement data of 2-ppm and blank samples were separated by more than 6 standard deviations (6·s) for all instruments. Since measurement data for both samples was normally distributed, the blank measurements were unlikely within 2-ppm sample data with more than 99.99% probability (Goh, 2002). In addition, since measurement CVs of the calibration samples were less than 10% (table 5), the LOQ for the instruments was determined as 2 ppm which was equivalent to LOD (Armbruster and Pry, 2008).
 |
| (a) Spectrophotometer |
 |
| (b) Stationary turbidimeter |
 |
| (c) Portable turbidimeter |
| Figure 5. Estimation of limit of detection from the outputs of three different instruments for the calibration ASP® 602 suspension samples. |
Prediction results of ASP® 602 concentrations for the validation samples using linear regression models are given in table 6. Variability of the measurements (CV) was high with the 5-ppm sample (11.9%, 19.3%, and 12.6% for the spectrophotometer, ST and PT, respectively), but decreased under 10% for samples of 15 ppm and greater. Approximation errors were between 0.9% to 24.6%, 0.1% to 14.8%, and 0.1% to 18.4%, for the spectrophotometer, ST and PT, respectively, and they were relatively large when the sample concentrations were low. For example, the regression model for the spectrophotometer had the approximation errors of 11.8% to 24.6% for the concentration range from 5 to 22 ppm, however the error decreased to less than 10% for the concentration range from 28.0 to 88.8 ppm. Similar trend of the errors was observed from ST and PT regression models. Thus, as expected, variations in the prediction results were higher when the sample concentrations were low, likely due to insufficient sample homogenization at low concentrations.
Table 7 shows the overall summary of regression models, approximation errors, and LOD for three instruments. The prediction results for validation samples showed ST regression model had better predictions of ASP® 602 concentrations. This suggests that the model validation must be performed to determine best regression model for the instrument in addition to evaluate coefficient of determination (R2). For example, although all regression models for the instruments had R2 of 0.9938 to 0.9973, and the model for PT had highest R2, the ST model had least approximation errors during the validation process. In addition, determining the calibration range for a regression model would be critical. Our results were obtained from the calibration range from 0 to 100 ppm. However, the results could be different when a calibration range is relatively wider. This is because building a regression model would be influenced by R2 which would be more influenced by prediction results for higher concentration samples rather than lower concentration samples as the errors of high concentration predictions would have more influences on the sum of squared errors for R2 (Almeida et al., 2002).
Our results showed that obtaining statistically different measurements to predict ASP® 602 concentrations below 2 ppm would be challenging with the tested instruments considering the position of 6·s on the regression lines and measurement results of blank samples (fig. 5), although further study would be required to confirm this observation. LOQs from these commercial instruments were substantial improvements over the device developed by Jeon and Zhu (2022). This improvement should be sufficient for analyzing samples from ISO 22368-1 tests to verify if a sprayer could meet the ISO 16119-2 requirement. In addition, the test results show the three instruments used in this research could determine cleaning fractions much higher than 400 (or 25 ppm), the current ISO 16119-2 requirement. For example, with LOQ of 2 ppm and an initial concentration of 10 g·L-1, the largest fraction possible would be 5000 although more studies would be needed for the confirmation.
| Table 6. Predicted concentrations of ASP® 602 validation samples using linear regression models for different instruments. | |||||||||||
| Instrument | Parameter | ASP® 602 Concentrations (ppm) | |||||||||
| 5.0 | 15.0 | 22.0 | 28.0 | 30.6 | 88.8 | ||||||
| Spectrophotometer | Average prediction (ppm) | 4.4 | 18.7 | 25.3 | 30.8 | 32.6 | 89.6 | ||||
| Coefficient of variation (%) | 11.9 | 2.8 | 2.5 | 1.7 | 2.2 | 1.7 | |||||
| Approximation error (%) | 11.8 | 24.6 | 15.5 | 9.8 | 6.7 | 0.9 | |||||
| Stationary turbidimeter | Average prediction (ppm) | 5.1 | 17.2 | 23.8 | 29.1 | 30.2 | 89.0 | ||||
| Coefficient of variation (%) | 19.3 | 6.6 | 4.1 | 6.0 | 5.7 | 3.9 | |||||
| Approximation error (%) | 1.3 | 14.8 | 8.6 | 3.8 | 1.1 | 0.1 | |||||
| Portable turbidimeter | Average prediction (ppm) | 4.1 | 17.3 | 23.8 | 28.8 | 30.5 | 88.0 | ||||
| Coefficient of variation (%) | 12.6 | 5.0 | 4.3 | 5.2 | 5.5 | 2.6 | |||||
| Approximation error (%) | 18.4 | 15.4 | 8.4 | 3.0 | 0.1 | 0.9 | |||||
Table 7. Summary of regression models for three instruments for their accuracy, approximation errors and limits of detection. | |||||||||||
| Instrument | Calibration Range (ppm) | Regression Model[a] | Approximation Errors (%) | Limit of Detection (ppm) | |||||||
| Spectrophotometer | 0-100 | C = 1085.3720 × AU – 4.9934 (0.9938) | 0.9-24.6 | 2 | |||||||
| Stationary turbidimeter | 0-100 | C = 0.0025 × RTU- 6.5339 (0.9968) | 0.1-14.8 | 2 | |||||||
| Portable turbidimeter | 0-100 | C = 2.0903 × NTU – 2.4111 (0.9973) | 0.1-18.4 | 2 | |||||||
[a] ‘C’, ‘AU’, ‘RTU’, and ‘NTU’ refer a concentration of ASP® 602 sample in ppm, absorbance unit, raw turbidity unit, and nephelometric turbidity unit, respectively. Values in parentheses are coefficient of determination (R2 ). | |||||||||||
Effect of Sample and Instrument Temperatures
Summaries of linear regression models for the spectrophotometer and PT to predict ASP® 602 concentrations of samples are given in table 8. The coefficients of determination for both regression models were over 0.998, indicating strong correlation between measured values with both instruments and the actual ASP® 602 concentrations.
| Table 8. Summary of regression models for the spectrophotometer and turbidimeter at room temperatures of 30°C. | |||
| Instrument | Calibration Range (ppm) | Calibration Temperature (°C) | Regression Model[a] |
| Spectrophotometer | 0-250 | 30 | C = - 6.95980 + 890.74939*AU (0.9991) |
| Portable turbidimeter | 0-250 | 30 | C = - 3.66857 + 2.38478*NTU (0.9983) |
[a] ‘C’, ‘AU’, and ‘NTU’ refer a concentration of ASP® 602 sample in ppm, absorbance unit, and nephelometric turbidity unit, respectively. Values in parentheses are coefficient of determination (R2 ). | |||
The absorbance and turbidity values of the validation samples at different temperatures are given in tables 9 and 10, respectively. Sample absorbance decreased as the temperature of samples and the spectrophotometer decreased. For example, absorbance values of the validation samples at 10°C were reduced by 0.0065% to 0.0675 AU (12.7% to 17.6%) compared to the values at 30°C (table 9). Turbidity values of the samples also changed as the temperatures of the samples and PT changed. However, the trend was opposite: sample turbidity increased as the sample and instrument temperatures decreased (table 10). Approximation errors for regression models of the instruments were less than 5% when the absorbance and turbidity of the validation samples were measured at the same temperature. However, when absorbance and turbidity of the validation samples were acquired at different temperatures, the errors increased by as high as 19.4% and 15.3% for the spectrophotometer and PT, respectively, due to changes in their measurement values. The results indicated the importance of the temperature of the samples and the instruments at the time of tests. Additional prediction errors due to temperature differences would be unavoidable unless their temperatures were the same as the calibration temperatures.
| Table 9. Absorbance of validation samples for different ASP® 602 concentrations. Predicted ASP® 602 concentrations using the regression model built with absorbance data at 30°C. | |||||||
| ASP® 602 Concentration (ppm) | Absorbance (AU) | Predicted ASP® Concentration Using 30°C Regression Model (ppm) | |||||
| Instrument Temperature | Instrument Temperature | ||||||
| 10°C | 20°C | 30°C | 10°C | 20°C | 30°C | ||
| 24.9 | 0.0305 (10.8)[a] | 0.0326 (6.6) | 0.0370 (5.1) | 20.2 [18.9][b] | 22.1 [11.4] | 26.0 [4.4] | |
| 71.0 | 0.0721 (3.5) | 0.0795 (3.0) | 0.0849 (2.7) | 57.2 [19.4] | 63.9 [10.1] | 68.7 [3.2] | |
| 149.9 | 0.1527 (2.9) | 0.1651 (1.7) | 0.1757 (2.0) | 129.0 [13.9] | 140.1 [6.5] | 149.6 [0.2] | |
| 207.0 | 0.2063 (1.9) | 0.2214 (2.0) | 0.2397 (4.4) | 176.8 [14.6] | 190.2 [8.1] | 206.6 [0.2] | |
[a] Values in parentheses are coefficient of variations (%). [b] Values in brackets are approximation error (%). | |||||||
Performance Evaluation of Agitation and Cleaning Systems
Figure 6 shows ASP® 602 samples at the same stage of implementing ISO 22368-1 test using the sprayer with C1 versus C2 rinsing configurations referred to as test 1 and test 2. The visual assessments showed that apparent concentration reduction after each rinse, and the performance differences between the agitation and cleaning systems were different enough to have different amounts of ASP® 602 carry over after each step of the tests. As a result, the sprayer with C1 appeared to have the rinsates with relatively higher ASP® 602 concentrations compared to the sprayer with C2 at the same step of ISO 22368-1 test.
| Table 10. Turbidity of validation sample for different ASP® 602 concentrations. Predicted ASP® 602 concentrations using the regression model built with turbidity data at 30°C. | |||||||
| ASP® 602 Concentration (ppm) | Turbidity | Predicted ASP® 602 Concentration Using 30°C Regression Model (ppm) | |||||
| Instrument Temperature | Instrument temperature | ||||||
| 10°C | 20°C | 30°C | 10°C | 20°C | 30°C | ||
| 24.9 | 13.6 (4.5)[a] | 12.9 (4.1) | 12.3 (3.9) | 28.7 [15.3][b] | 27.2 [9.1] | 25.8 [3.4] | |
| 71.0 | 35.7 (3.1) | 34.2 (2.4) | 32.4 (3.7) | 81.5 [14.8] | 77.8 [9.6] | 73.6 [3.7] | |
| 149.9 | 74.0 (4.8) | 69.3 (2.5) | 65.4 (2.2) | 172.7 [15.3] | 161.6 [7.9] | 152.3 [1.6] | |
| 207.0 | 98.4 (2.2) | 92.9 (2.5) | 87.0 (2.5) | 231.0 [11.6] | 217.8 [5.2] | 203.8 [1.5] | |
[a] Values in parentheses are coefficient of variations (%). [b] Values in brackets are approximation error (%). | |||||||
  | ||||
| (a) ASP® 602 samples from test 1 (C1) | (b) ASP® 602 samples from test 2 (C2) | |||
| Figure 6. ASP® 602 samples from ISO 22368-1 cleanout tests. SRef, SRinse1,2,and3, SAR, SBP, and SWash refer to the reference liquid, samples after 1st, 2nd, and 3rd rinses, sample after rinsing, sample from bypass line, and sample after high pressure tank wash, respectively. | ||||
| Table 11. Predicted ASP® 602 concentrations of acquired samples during the cleanout tests 1 and 2. Sref, SRinse1,2,and3, SAR, SBP, SWash, and SDV refer to the reference liquid, samples after 1st, 2nd and 3rd rinses, sample after rinsing, sample from bypass line, sample after high pressure tank wash, and dead volume sample, respectively. | ||||||||
| Test 1 | Test 2 | |||||||
| Sample | Analytical Method[a] | Average ASP® 602 Concentration (ppm) | ASP® 602 Recovered (g) | Analytical Method | Average ASP® 602 Concentration (ppm) | ASP® 602 Recovered (g) | ||
| SRef | 90 | O | 8461.5 (19.8)[b] | 330.2 [86.9] | O | 10486.2 (0.7) | 367.2 [96.6] | |
| 50 | O | 8489.5 (25.8) | O | 10528.6 (5.3) | ||||
| 10 | O | 9619.5 (69.9) | O | 10546.5 (6.0) | ||||
| Average | 8856.8 (573.4) | 10520.4 (27.3) | ||||||
| S\Rinse1 | O | 7496.7 (1976.0) | 27.4 [7.2] | O | 1602.1 (3.0) | 5.8 [1.5] | ||
| SRinse2 | O | 1787.2 (41.7) | 6.8 [1.8] | O | 615.6 (20.5) | 2.1 [0.5] | ||
| SRinse3 | O | 455.7 (159.6) | 1.7 [0.5] | A | 209.8 (13.3) | 0.8 [0.2] | ||
| SAR | 90 | A | 40.6 (0.6) | 2.0 [0.5] | A | 9.8 (2.0) | 0.4 [0.1] | |
| 50 | A | 48.6 (1.7) | A | 12.2 (0.6) | ||||
| 10 | A | 74.2 (1.6) | A | 16.6 (0.6) | ||||
| Average | 54.5 (15.3) | 12.9 (3.3) | ||||||
| SBP | A | 20.7 (0.6) | 0.1 [0.0] | A | 15.7 (1.4) | 0.1 [0.0] | ||
| SWash | A | 30.9 (12.4) | 0.8 [0.2] | A | 14 (1.7) | 0.2 [0.1] | ||
| SDV | O | - | 2.4 [0.6] | O | - | 2.2 [0.6] | ||
| Total | 371.4 [97.7] | Total | 378.8 [99.6] | |||||
[a]‘O’ and ‘A’ refer analytical methods to measure ASP® 602 mass by evaporating water in an oven (O) and to measure absorbance (A) to predict ASP® 602 concentrations, respectively. [b]Values in parentheses and brackets are standard deviations and a percentage of ASP® 602 recovered from each step of the test, respectively. | ||||||||
Table 11 shows analysis results of the samples from tests 1 and 2. The sample assay results showed that the trend of ASP® 602 concentration changes was similar to the visual assessments (shown in fig. 6). The results showed that the concentrations of SRef from the sprayer for tests 1 and 2 were 83.7% to 95.3% and 98.4% to 98.9% of actual concentration, respectively. Thus, the sprayer for the test 1 was incapable of homogenizing ASP® 602 in the tank and these might imply that simply having more orifices on a sparge bar might not help the homogeneity of the tank mixture. Furthermore, the concentrations of SRef increased as the tank was drained which suggested that its agitator might create too vigorous jets, moving a portion of ASP® 602 to the surface of the tank mix along the tank wall instead of homogenizing it in the tank as approximately 50% of overall flow was dedicated to the agitator. In addition, the differences of the sprayer between tests 1 and 2 resulted in different rates of removing ASP® 602 from the sprayer. For example, the ASP® 602 concentrations of the rinsates collected after 3rd rinse from tests 1 and 2 were 455.7 and 209.8 ppm, respectively, which showed the sprayer for test 2 was more effective to remove ASP® 602 from the sprayer. As a result, the concentrations after rinsing (SAR) were 40.6 to 74.2 ppm and 9.8 to 16.6 ppm for tests 1 and 2, respectively. The results showed that the fractions of mean ASP® 602 concentrations after rinsing for tests 1 and 2 were 0.62% (=54.5/8856.8) and 0.12% (=12.9/10520.4), respectively, therefore, the residue reduction factors were determined as 162.6 (=8856.8/54.5) for test 1 and 818.7 (=10520.4/12.9) for test 2. This indicates that the sprayer from test 1 with configuration C1 would not meet the cleaning system requirement, a reduction factor of 400, in ISO 16119-2 (ISO, 2013b) while the sprayer from test 2 with configuration C2 exceeded the ISO requirement. Approximately 97.7% and 99.6% of ASP® 602 added to the sprayer was recovered from tests 1 and 2, respectively, and the unaccounted amounts of ASP® 602 were less than 3% for both tests (8.66 g in test 1 and 1.43 g in test 2) which suggest the results obtained from both tests 1 and 2 were reliable and reasonable.
The amounts of ASP® 602 recovered from the sprayer bypass line or dead volume were similar for both tests. However, the amounts of ASP® 602 recovered from internal wall of the tank and during the tests were different between two tests. Based on the SRef values, ASP® 602 recovered at each step from the test 1 was generally higher than the amount from test 2 after rinsing. This suggests that the differences in internal configurations e.g., a sparge bar, jet agitator, cleaning nozzle and tank fittings of the sprayer influenced the performance of agitation and cleaning of a sprayer. In addition, test results suggest that ASP® 602 could be used to test a sprayer with ISO 22368-1 clean out test to differentiate the performance of agitation and cleaning systems of sprayers.
Conclusions
Reliance on CPP applications for sustainable food supply will continue as rapid growth of the world population and reduction of arable land for crop production are expected. Therefore, sprayers for CPP applications would be important tools for sustainable food supply. Consequently, the sprayer designs that meet performance standards will be critical to achieve successful applications while minimizing crop responses from carryover active ingredient between CPP spray applications.
The work presented herein was regarding to the feasibility of ASP® 602 to use ISO 22368-1 test. The results provided here were intended as the guidelines for addressing the potential concerns or technical challenges e.g., sample homogenization, measurement temperature, and sample sizes if ASP® 602 was used to test sprayers with ISO 22368-1 standard test method. Our results showed precipitated ASP® 602 samples should be stirred at 750 r·min-1 rpm for at least 1 min before calibrating an analytical instrument and validating the calibration with the samples which should be carried out in the same temperature to reduce measurement variations. Furthermore, a minimum of 5 subsamples is recommended to determine LOQ or LOD of an analytical instrument.
These should be guidelines for the test agencies or equipment manufacturers to start to build their own expertise to perform spray clean out tests rather than absolute requirements for sample sizes and homogenizations. In addition, three common laboratory instruments (a spectrophotometer, a stationary fluorometer with a turbidity module, and a portable fluorometer and turbidimeter) were used to quantify ASP® 602 concentration in clean out samples. The results showed that quantifying ASP® 602 from samples as low as 2 ppm (LOQ) was possible with three instruments used in this investigation, with simple calibration and validation procedures. Their prediction errors could be as high as 20% when ASP® 602 concentration was 5 ppm, however, it represented an absolute error of less than 1 ppm.
The results of ISO 22368-1 clean out tests with ASP® 602 could quantify the performance difference of agitation and cleaning systems of the laboratory sprayer by quantifying ASP® 602 concentrations of the samples acquired during the tests using the spectrophotometer. The results which included mechanical and fluid dynamics properties of the sprayer flow only were able to differentiate the performance of the sprayers with different internal configurations. The analysis results could determine residue reduction factors of 163 and 819 for the sprayer with two different internal configurations which could determine if the sprayer meets the current ISO 16119-2 requirement or not. These results imply that ASP® 602 could be used as a testing substance for ISO 22368-1 test to quantify the performance differences between various agitation and cleaning systems. For the future studies, ASP® 602 should be tested with commercial field sprayers under ISO 22368-1 procedure to further validate it for the clean out test, and it would be beneficial for the standard users to develop a procedure to qualify a substance with similar physical and toxicity profiles as ASP® 602 in case that its availability is challenging.
Acknowledgments
The authors would like to thank Andy Doklovic, Adam Clark, Barry Nudd, Brian Phouthavong, and Eve Painter for their technical and laboratory work support, and Adam J. Barlow from John Deere Des Moines Works for providing ASP® 602. The authors express their appreciation to the Association of Equipment Manufacturers for their support for this project. This research was supported by USDA-ARS in-house project 5082-21620-001-00D.
References
Ali, B., & Ali, B. (2017). Performance of a hydraulic jet agitation system with different jet nozzle sizes in the sprayer tank. Sci. Papers - Ser. A Agron., 60, 449-454.
Almeida, A. M., Castel-Branco, M. M., & Falcão, A. C. (2002). Linear regression for calibration lines revisited: Weighting schemes for bioanalytical methods. J. Chromatogr. B, 774(2), 215-222. https://doi.org/10.1016/S1570-0232(02)00244-1
Andersen, P. G., Jørgensen, M. K., & Langenakens, J. (2010). Continuous cleaning – Time efficient sprayer cleaning with minimum water usage. Asp. Appl. Biol., 99, 55-60.
Armbruster, D. A., & Pry, T. (2008). Limit of blank, limit of detection and limit of quantitation. Clin. Biochem. Rev., 29(Suppl 1), S49-S52.
Balsari, P., Tamagnone, M., Allochis, D., Marucco, P., & Bozzer, C. (2012). Sprayer tank agitation check: A proposal for a simple instrumental evaluation. Proc. 4th European Workshop on Standardised Procedure for the Inspection of Sprayers in Europe, SPISE 4, Lana, (pp. 106-116). https://doi.org/10.5073/jka.2012.439.015
BASF. (2020). ASP® 602 Premium brightness spray dried kaolin. Retrieved from https://kaolin.basf.com/products/application/asp/ASP602
European Food Safety Authority, Arena, M., Auteri, D., Barmaz, S., Bellisai, G., Brancato, A.,... Villamar-Bouza, L. (2018). Peer review of the pesticide risk assessment of the active substance copper compounds copper(I), copper(II) variants namely copper hydroxide, copper oxychloride, tribasic copper sulfate, copper(I) oxide, Bordeaux mixture. EFSA J., 16(1), e05152. https://doi.org/10.2903/j.efsa.2018.5152
European Union. (2009a). Directive 2009/127/EC of the European Parliament and of the Council of 21 October 2009 amending Directive 2006/42/EC with regard to machinery for pesticide application. Off. J. Eur. Union, L310, 29-33.
European Union. (2009b). Directive 2009/128/EC of the European Parliament and of the Council of 21 October 2009 establishing a framework for Community action to achieve the sustainable use of pesticides. Off. J. Eur. Union, L309, 71-86.
Gil, E., Salcedo, R., Campos, J., Llop, J., Ortega, P., & Gallart, M. (2018). Orchard sprayer’s evaluation according to ISO 16119-3. Need to support manufacturer’s procedure. Asp. Appl. Biol., 137, 37-44.
Goh, T. N. (2002). The role of statistical design of experiments in six sigma: Perspectives of a practitioner. Qual. Eng., 14(4), 659-671. https://doi.org/10.1081/QEN-120003565
ISO. (2004). ISO 22368-1 Crop protection equipment - Test methods for the evaluation of cleaning systems - Part 1: Internal cleaning of complete sprayers. Geneva, Switzerland: ISO.
ISO. (2013a). ISO 16119-1 Agricultural and forestry machinery - Environmental requirements for sprayers - Part 1: General. Geneva, Switzerland: ISO.
ISO. (2013b). ISO 16119-2 Agricultural and forestry machinery - Environmental requirements for sprayers - Part 2: Horizontal boom sprayers. Geneva, Switzerland: ISO.
ISO. (2013c). ISO 16119-3 Agricultural and forestry machinery - Environmental requirements for sprayers - Part 3: Sprayers for bush and tree crops. Geneva, Switzerland: ISO.
ISO. (2021). ISO 5682-4 Equipment for crop protection - Spraying equipment - Part 4: Test methods for agitation of sprayer tanks. Geneva, Switzerland: ISO.
Jeon, H., & Zhu, H. (2022). Low-cost microcontroller-based device to measure kaolin clay concentrations for ISO sprayer cleanout tests. Appl. Eng. Agric., 38(1), 135-143. https://doi.org/10.13031/aea.14764
Ministère de l’Agriculture. (2006). Arrêté du 12 septembre 2006 relatif à la mise sur le marché et à l’utilisation des produits visés à l’article L. 253-1 du code rural et de la pêche maritime. Article 6. Retrieved from https://www.legifrance.gouv.fr/loda/id/JORFTEXT000000425570
Puterka, G. J., & Glenn, D. M. (2008). Kaolin-based particle films for arthropod control. In J. L. Capinera (Ed.), Encyclopedia of entomology (pp. 2075-2080). Dordrecht: Springer Netherlands. https://doi.org/10.1007/978-1-4020-6359-6_1606
Puterka, G. J., Glenn, D. M., Sekutowski, D. G., Unruh, T. R., & Jones, S. K. (2000). Progress toward liquid formulations of particle films for insect and disease control in pear. Environ. Entomol., 29(2), 329-339. https://doi.org/10.1093/ee/29.2.329
Tamagnone, M., Balsari, P., Bozzer, C., & Marucco, P. (2012). Assessment of parameters needed to design agitation systems for sprayer tanks. Asp. Appl. Biol.(114), 167-174.
Tamagnone, M., Balsari, P., Marucco, P., & Vai, P. (2016). Evaluation of sprayer internal tank cleaning systems according to ISO Standard 22368-3. Asp. Appl. Biol.(132), 39-44.
Thompson, T., Freestone, D., Michalczyk, A. A., & Ackland, M. L. (2012). Copper levels in buccal cells of vineyard workers engaged in various activities. Ann. Occup. Hyg., 56(3), 305-314. https://doi.org/10.1093/annhyg/mer086
Turner Designs. (2022a). AquaFluor handheld fluorometer/turbidimeter product datasheet. Retrieved from http://docs.turnerdesigns.com/t2/doc/brochures/S-0033.pdf
Turner Designs. (2022b). Trilogy laboratory fluorometer product datasheet. Retrieved from http://docs.turnerdesigns.com/t2/doc/brochures/S-0068.pdf
Ucar, T., Ozkan, H. E., Fox, R. D., Brazee, R. D., & Derksen, R. C. (2000). Experimental study of jet agitation effects on agrochemical mixing in sprayer tanks. J. Agric. Eng. Res., 75(2), 195-207. https://doi.org/10.1006/jaer.1999.0499
Zhang, Z., Zhu, H., Jeon, H., Ozkan, E., Wei, Z., & Salcedo, R. (2022). A turbidity module to measure spray mixture concentration for premixing in-line injection system. Appl. Eng. Agric., 39(1), 13-22. https://doi.org/10.13031/aea.15245