Fertilizer Conference Proceedings
S.J. Birrell, J.W. Hummel, T.R Peck and R.G. Hoeft1
Precision agriculture is a management strategy which seeks to address within-field variability and to optimize inputs such as pesticides and fertilizers on a point-by-point basis within a field and not according to the field average. This strategy has the potential to improve profitability for the producer and also to reduce the threat of ground water or surface water contamination from agrichemicals. The concept of using established scientific principles on smaller areas is sound. However, the successful implementation of variable rate fertilizer application relies on the accurate quantification of the spatial variation of important soil factors.
The full benefit of SSM will only be realized if the spatial variation across the field is accurately determined. Some research has indicated that nutrient maps generated from soil samples collected on 2.5- and 3.3-acre grids do not accurately represent the variability that exists. Data collection on a finer spatial resolution than is feasible with manual and/or laboratory methods is required, but cost prohibitive. Sensors will allow the collection of data on a much finer spatial resolution, to more accurately characterize within-field variability. Real-time sensors can provide a sampling intensity several orders of magnitude greater than traditional methods, resulting in significant reductions in sampling errors. Therefore, a real-time soil sensor can tolerate much higher analysis errors while providing greater overall accuracy in mapping soil variability.
The objective of this work is to develop and test a real-time soil nutrient analysis system, based on ion-selective field-effect transistors (ISFETs). The development of a real-time soil nutrient sensor will allow the automated collection of soil nutrient data on a fine resolution to accurately characterize within-field variability for site-specific fertilizer application. The specific research objectives are:
Soil fertility testing, particularly in the case of immobile nutrients, is based on the assumption that the extraction procedure differentiates between the quantity of plant available nutrient in the soil and the total amount of the nutrient in the soil. A significant advantage of the proposed soil nutrient sensing system over non-invasive sensing technologies (such as NIR reflectance, LIDAR, etc.) is that this procedure essentially automates the standard soil testing methods, and the chemistry of the testing procedure is not fundamentally different from that of current laboratory methods. Therefore, the test results obtained from this sensor technology would be compatible with the body of knowledge which has formed the basis of soil fertility recommendation procedures.
Ion-selective field effect transistors (ISFETs) have inherent features such as small dimensions, low output impedance, high signal-to-noise ratio, low sample volumes and the potential for mass production, which are advantages in the development of a real-time soil sensor. However, ISFETs have the disadvantage of long-term drift, which is diminished by the use of a Flow Injection Analysis (FIA) system. In fact, FIA and ISFETs are complementary since the small sample volumes and rapid response of ISFETs allow the miniaturization of the FIA system, dramatically decreasing sample dispersion and thereby increasing both sample resolution and sample frequency. ISFETs use the same chemical principles for analysis as conventional ion-selective electrodes and pH electrodes, which are well established in soil nutrient testing.
The ISFETs used in this work were developed at Hitachi Central Research Laboratory, Hitachi Ltd., Japan (Tsukada et al., 1987). The chips were designed with integrated signal conditioning circuits and a polyamide layer with sensor wells (250 µm x 700 µm) over the gate area of each of the four ISFET sensors (Tsukada et al., 1989). The selectivity of the ISFET sensors to different ionic species is determined by the composition of the ion selective membrane deposited into the wells. Different membranes on multiple sensors on a single chip provides a capability for sensing different ionic species, or using the same membrane on multiple sensors can provide two or more sensors sensitive to one ionic species. This flexibility allows simultaneous multi-component analysis and/or redundant signals for automatic failure detection.
A multi-sensor nitrate ISFET chip was developed and integrated into a FIA system. The optimal flowrate, and injection and washout times were investigated using the FIA system and multi-sensor ISFET. Four flowrates (0.057, 0.120, 0.179, 0.236 ml/s), three washout times (2, 1, 0.75 s) and four injection times (2, 1, 0.75, 0.5 and 0.25 s) were used to determine the shortest injection and washout times, and the lowest flowrate which provided acceptable results. The FIA parameters were tested using an incomplete factorial block design, since the sample injection time was always equal to or less than the washout time.
Testing of the nitrate ISFET/FIA system for soil analysis was carried out using manually extracted samples. An injection time of 0.5 s, and washout times of 0.75 s and 2.5 s were used during these tests. The nominal flowrate was 0.17 ml/s. The multi-sensor ISFET/FIA system was calibrated before the analysis of manually extracted samples using standard solutions. The ISFET predicted nitrate concentrations of manually extracted soil solutions were compared to the actual soil concentrations determined using the cadmium reduction method (Lachat Analyzer) on the identical solutions.
There are ion-selective membranes available for most of the important soil nutrients and other ionic species in the soil (H+, K+, NH4+, NO3+, Na+, Ca2+, Mg2+, Cl-). A potassium ISFET has been prepared and preliminary tests of the response of the potassium ISFET conducted. The response of the ISFET was tested using standard solutions and manually extracted soil samples. An injection and washout times of 5 seconds were used for these preliminary tests. Future tests will be conducted on optimal flowrate and injection times for the potassium ISFET/FIA system. A pH ISFET has also been produced and tested using buffer standards.
The extraction of all the nutrient (or at least a consistent percentage of the total) in the soil sample must be achieved for a reliable soil nutrient analysis system. The effect of soil type and texture, soil sample volume, sample preparation (solid core or pulverized sample), rate of extraction solution injection into soil, and total extraction solution volume is being evaluated. A system has been designed to evaluate effect of sample volume, soil density and total extraction solution volume on extraction efficiency and extraction times. Tests are progressing on the evaluation of extraction times for nitrates using the system and nitrate ISFET detector. Preliminary test results suggest that the minimum extraction times are in the 2-5 s range. These extraction times meet the criteria for real-time sensing and control. Upon completion of these tests the same system will be used to test extraction efficiency and extraction times for potassium using the potassium ISFET.
The nitrate ISFET/FIA system was successfully tested. The optimal FIA injection time is the minimum injection time required such that a finite section of the sample injected reaches the ISFET sensor without dilution. The optimal FIA washout time is the minimum time required to flush the previous sample from the ISFET flow cell before the next sample is injected. The optimal injection and washout times are dependent on the flowrate.
The response of the ISFET in the FIA system to changes in injection time (0.25, 0.50, 0.75, 1.0, 2.0 s), and at different flowrates (0.06, 0.12, 0.18, 0.24 ml/s) are shown for washout times of 2.0 s and 0.75 s, respectively (Figure 1, Figure 2, Figure 3, Figure 4, Figure 5, Figure 6, Figure 7, and Figure 8). The lowest flowrate (0.06 ml/s) was insufficient for complete washout of the flow cell before the next sample was injected, with the possible exception of the 2.0 s washout time. The gradients of the ISFET response for the higher flowrates were essentially zero when the next sample was injected, which signified that complete washout had occurred, for all four injection times (graphs not shown). At the highest flowrate, the response peak height was not substantially decreased by a decrease in injection time; thus the dispersion factor at the peak maxima was approximately 1 and the sample slug at this point was undiluted. At the two intermediate flowrates (0.12 and 0.18 ml/s), response peak heights for the 0.25 s injection time were lower than for the longer injection times, showing that the maximum dispersion factor was less than 1 and the sample slug was diluted even at the peak maximum, although the response peak heights were only reduced by 25 percent of the maximum. The 0.75 s washout time showed the same trends as the 2 s washout time, with low flow rates resulting in lower response mean peak heights due to carryover and short injection times resulting in reduced response peak heights due to dispersion. These response curves demonstrated that even with an injection time of only 0.5 s, the response is satisfactory, provided that the flowrate is high enough for adequate washout of the flow cell. An increase in injection time beyond 0.5 s does not significantly affect the peak height but does increase the peak width at the baseline which increases the minimum washout time required. The precision and accuracy of the system were highly dependent on maintaining precise, repetitive sample injection times and maintaining constant flow parameters during the testing cycle. The rapid response of the system allowed samples to be analyzed within 1.25 seconds, with sample flowrates less than 0.2 ml/s.
The multi-sensor ISFET/FIA system was calibrated before the analysis of manually extracted samples. The ISFET predicted nitrate concentrations of manually extracted soil solutions were compared to the actual soil concentrations determined using the cadmium reduction method (Lachat Analyzer). The ISFET predicted concentration versus actual concentration are shown in Figure 9 and Figure 10. The plots of predicted concentration versus actual concentration show an obvious difference in the response between the first replication and last two replications. This response change was caused by a change in FIA parameters such as flowrate or the injection valve operation. There was no significant increase in soil nitrate prediction capability when the washout time was increased from 0.75 s to 2.5 s. The slope of the regression line of predicted versus actual concentration was approximately 1, for all four ISFETs (Birrell, 1995) provided the ISFETs were calibrated correctly before testing. The multi-ISFET sensor was successful in predicting soil nitrates in manually extracted soil solutions, with correlation coefficients greater than 0.9. The rapid response of the system allowed samples to be analyzed within 1.25 s, with sample flowrates less than 0.2 ml/s. The low sample volume specification is particularly apropos to real-time soil nutrient analysis, since extremely small quantities of soil and extracting solutions can be employed for each sample.
A potassium ISFET has been developed and evaluated for the prediction of soil potassium concentration. The sensitivity response of the ISFET to potassium is shown in Figure 11. The sensitivity response of the ISFET was 36.4 mV per decade change in concentration (R2=0.96). The ISFET buffer amplifier's gain was approximately 0.65. Therefore, the sensitivity of the potassium membrane on the ISFET sensor was 56 mV/decade which is very close to the theoretical Nernst response of 59 mV/decade. Figure 12 shows ISFET predicted soil potassium concentration versus actual soil concentration of manually extracted samples. The regression slope of predicted versus actual soil potassium was 1.05 with a correlation coefficient of 0.96. The ISFET will be tested in the ISFET/FIA system to determine the optimum FIA parameters and minimum sampling interval. However, the theoretical rate of response for the potassium ISFET is of the same magnitude as that for the nitrate ISFET, therefore, similar injection and washout times should be possible. The minimum sample interval of an ISFET/FIA system is limited by the fluid dynamics and not ISFET response speed.
A pH ISFET has been developed and tested using standard buffer solutions (Figure 13). The sensitivity response of the pH ISFET was 18.65 mV per decade change in hydrogen ion concentration (R2=0.997). The corrected pH ISFET response (28 mV/decade) is lower than the Nernst response. Future research will include the evaluation of a pH ISFET for soil testing.
A multi-ISFET nitrate sensor was developed and integrated into a flow injection system to measure soil nitrates. The sensor was successful in measuring soil nitrates in manually extracted soil solutions (r2> 0.9). The rapid response of the system allowed a sample to be analyzed in 1.25 s, which is satisfactory for real-time soil sensing. A multi-ISFET potassium sensor has been developed and successfully used to predict soil potassium concentration in manually extracted soil solutions (r2 > 0.96). The results are extremely promising and evaluation of this sensor will continue. The ability to integrate multiple sensors allows simultaneous multi-component analysis and/or redundant signals for automatic failure detection. The rapid response and low sample volumes required by the multi-sensor ISFET/FIA system make it a strong candidate for use in real-time soil nutrient sensing.
Funding for this research has been provided by the Illinois Fertilizer Research and Education Council (FREC) and the Illinois Council on Food and Agriculture (C-FAR).
Figure 1. Nitrate ISFET response peak height versus concentration for flowrate 0.06 ml/s, 2.0 second washout time and four difference injection times (0.25, 0.5, 1.0, 2.0 s)
Figure 2. Nitrate ISFET response peak height versus concentration for flowrate 0.12 ml/s, 2.0 second washout time and four different injection times (0.25, 0.5, 1.0, 2.0 s)
Figure 3. Nitrate ISFET response peak height versus concentration for flowrate 0.18 ml/s, 2.0 second washout time and four different injection times (0.25, 0.5, 1.0, 2.0 s)
Figure 4. Nitrate ISFET response peak height versus concentration for flowrate 0.24 ml/s, 2.0 second washout time and four different injection times (0.25, 0.5, 1.0, 2.0 s)
Figure 5. Nitrate ISFET response peak height versus concentration for flowrate 0.06 ml/s, 0.75 second washout time and three different injection times (0.25, 0.5, 0.75 s)
Figure 6. Nitrate ISFET response peak height versus concentration for flowrate 0.12 ml/s, 0.75 second washout time and three different injection times (0.25, 0.5, 0.75 s)
Figure 7. Nitrate ISFET response peak height versus concentration for flowrate 0.18 ml/s, 0.75 second washout time and three different injection times (0.25, 0.5, 0.75 s)
Figure 8. Nitrate ISFET response peak height versus concentration for flowrate 0.24 ml/s, 0.75 second washout time and three different injection times (0.25, 0.5, 0.75 s)
Figure 9. Nitrate ISFET predicted versus actual soil concentration for manually extracted solutions with 0.5s injection time and 2.5s washout time
Figure 10. Nitrate ISFET predicted versus actual soil concentration for manually extracted solutions with 0.5s injection time and 0.75s washout time
Figure 11. Potassium ISFET response peak height versus actual soil concentration for manually extracted soil solution with 5s injection and washout time
Figure 12. Potassium ISFET predicted versus actual soil concentration for manually extracted soil solutions with 5s injection and washout time
Figure 13. pH ISFET response peak height versus standard buffer solution concentration with 5s injection and washout time.
1 S.J. Birrel is a Visiting Assistant Professor, Dept. of Agricultural Engineering; J.W. Hummel is an Agricultural Engineer, USDA–Agricultural Research Service, Urbana, IL; T.R. Peck is a Professor, Dept. of Natural Resources and Environmental Sciences, University of Illinois; and R.G. Hoeft is a Professor, Dept. of Crop Sciences, University of Illinois.
Tsukada, K., M. Sebata, and T. Maruizumi. 1987. A multiple-chemfet integrated with CMOS interface circuits. Transducers 87:155-158.
Back to 1998 Index
of Crop Sciences
University of Illinois Extension
College of Agricultural, Consumer, and Environmental Sciences
University of Illinois at Urbana-Champaign
Illinois Fertilizer & Chemical Association
Illinois Department of Agriculture
|Design by: Crop
Sciences Computer and Web Support Group
Copyright © 2005 University of Illinois at Urbana-Champaign
Email site problems to the webmaster