Designing an In-Situ Soil Conductivity Monitoring Sensor for Precision Agriculture and Water Management


Soil Conductivity, the measure of soil salinity, offers farmers important information on potential crop yield and proper watering levels. However, most soil EC sensors and measurement tools are time-consuming, expensive, or difficult to operate. I decided to design a low-cost soil EC sensor using a Microbial Fuel Cell. The first step was to find a method for calculating soil EC using voltage measurements from an MFC. The equation for soil EC, k=ΔL/(ΔRs)*S. I can calculate the resistance of the MFC, as R=V/I. I found that I could find the estimated MFC current using charge rate over time, (Δv/Δt)(∫v(t) dt)/T, and using Ohms Law, found R=V((Δv/Δt)(∫v(t) dt))/T. Thus I found a method for obtaining soil EC measurements using an MFC, k=(ΔL*(Δv/Δt)*(∫v(t) dt))/(V*T*S). Now that I had a method for calculating soil EC, I needed to design the sensor. My sensor is a soil based multi-anode MFC, with variable electrode depths (L) to satisfy my soil EC equation. I was able to successfully record measurements within + 4 mS/m, as well as accurately measure a changing soil EC. My solution is useful in precision agriculture, as it offers a more cost-effective method for measuring soil EC. In the future, I plan on expanding my project to include a low voltage grid that records voltage readings from nearby MFC sensors and plots them on a field map, offering the user an instant and continuous reading on soil health over the entire field. 

I linked a short 5-minute video explaining my project below.





Question / Proposal

Can I use an MFC to create a fast, simple, and affordable sensor for measuring Soil Conductivity?

MFC technology is a growing field of science that aims to harness the metabolism of microbes to create electricity. However, since most MFC's produce a low current, their application is limited. So I started researching for a useful application of this technology, and along the way, I came across the field of precision agriculture. I found my problem: how can I make a fast, simple, and affordable sensor for measuring Soil Conductivity? Soil Conductivity [EC] monitoring and management are integral subsections of precision farming that allow farmers to increase crop yield, plant sustainability, nutrient availability, as well as prevent overwatering crops. While soil EC does not measure specific ions of salt compounds, it does offer a good generalization of the amount of nitrogen available for plant growth by measuring soil salinity. Since nitrogen concentration has a direct relationship with plant growth, finding and maintaining optimal soil EC offers farmers the opportunity to create ideal environments for plant growth. Since Soil conductivity can be measured by the resistance, distance, and surface area of electrodes, I realized that I could use data from an MFC to measure the soil conductivity. I predicted that by planting an MFC into the soil, I could create a low cost, accurate, and fast sensor. Furthermore, farmers could use the data from the sensor to create a map of watering needs across their farm and help prevent over-watering.


Soil conductivity [EC], measured in mS/m, is the measure of the salinity of the soil and affects crop yield, sustainability, and plant nutrient availability. While soil EC does not directly measure specific ions of salt compounds, it does offer a good generalization of the amount of nitrogen available for plant growth. Keeping the amount of nitrogen at an optimum concentration is paramount in ensuring larger crop yield and measuring soil EC is one of the best methods for doing so.


Currently, methods for measuring soil EC are cumbersome and don’t offer constant readings. This can be problematic as soil EC is a dynamic property of soils that changes from soil mineral concentrations, climate, temperature fluctuations, soil dampness, and other factors. The best method for measuring soil EC is by using a contact sensor. Contact sensors are typically fixed to a trailer and driven over the fields using trucks or tractors. By using an array of electrodes, normally discs or coulters, electric current is applied to the soil and received by a receiving electrode. The voltage drop is then measured and used to find soil EC measurements. While this method is extremely accurate, these sensors are very bulky, have extremely specific requirements on field sizes, and have to drive over the field to measure soil EC. For this reason, I decided to design a sensor that would be implanted into the ground and offer the user an up-to-date soil EC value of the entire field.

Microbes play an important role in what is known as the soil nutrient cycle and microbe health is closely related to soil salinity and moisture content. (Yan, N). Microorganisms participate in oxidation, nitrification, ammonification, nitrogen fixation, and other processes which lead to decomposition of soil organic matter and transformation of nutrients. For this reason, one method for measuring the soil EC would be to compare voltage readings across multiple sensors in different areas on the land. By finding where microbe growth was more prominent, an understanding of comparative EC’s could be generated, however, it would not give exact soil EC values. Typically, soil EC is measured using the equation k=L/(Rs*S), where L is distance, S is the cross-sectional area, and Rs is the resistance of the soil in a simple circuit. Furthermore, the Rs, resistance, of the soil can be obtained using Ohms law, where R=V/I. The voltage can easily be obtained from the sensor, but I soon realized that the current could not be measured without using an external source. I decided to do more research and found that MFC’s are considered as constant production, which allows the current of the cell to be measured by (rate of charge)/t. In addition, the rate of charge can be measured as ∫v(t) dt, where V is voltage over the change in t, time. (McNeal, B L). Therefore, I realized that I could use the resistance measured from the charge rate in the Soil EC calculation to create the equation, k=(L*(Δv/Δt)*(∫v(t) dt))/(V*T*S).




Method / Testing and Redesign

In order to create a low cost and reliable sensor, I decided to create a multi-chambered MFC. The voltage readings from the sensor were used to calculate exact soil EC and create the final soil EC field map.

The sensor was designed to be modular in design, as to allow the user to tailor the depth of the sensor to the depth of topsoil in the soil sample. The accuracy of the sensor improves as the number of 'cells' in the sensor increase, where each cell measures at a deeper depth. Each section of the sensor has one electrode and fits into the next section, forming a multicellular MFC sensor.

In prototype one, the electrodes of the cell were all constructed using a carbon sponge, as it was conductive while also encouraging microbe fixation to the electrodes. The electrodes were then connected to wires that ran to the sensor cap and transmitted data to the computer.

The first round of testing was conducted using the first prototype design. In this testing round, the cells performed fairly well, measuring a soil EC of 49.6 mS/m. However, when fitting a linear approximation for the resistance over depth curve, the measured values were inconsistent, resulting in an R2 value of 0.56.

In order to improve this consistency, the anodic electrodes were modified to increase microbe attraction and encourage nanowire growth around the electrode. The new electrodes were treated with acetone and pressed with carbon powder. In prototype one, the electrodes were fixed to wires, which allowed for the simple circuit to be constructed. However, in order to improve the organization of the cell, the wires were replaced with copper strips, which ran across each sensor module. This helped to maintain the usability of the sensor, as no rewiring was necessary when cells were added or removed.

In the second round of testing, the sensor was tested to ensure consistent measurements. In this testing round, the sensor continuously updated the soil EC values over the 4.5 hour testing period and used voltage measurements for only the 5 most recent readings. The sensor measured an average soil EC value of 24.55 mS/m, showing fairly consistent values. In addition, the average R2 value recorded during linear approximation of Rs/D profiles increased to 0.747. This further reinforced the increase in reliability from the new electrodes.

However, a few changes were still made before the final prototype was constructed. A new air-based cathode was designed to allow for more precision in soil EC measurement. The cathodic electrode was constructed with 60 stainless steel mesh sheets, pressed together in a rolling press method. A Gas Diffusion layer, comprising of Carbon Black powder and PTFE emulsion, was then rolled to the air side. Furthermore, a Catalyst layer, comprising of Activated carbon and PTFE emulsion, was rolled onto the soil side of the electrode.

This new cathode showed increased the reliability and consistency of the soil EC values, increasing the R2 value to 0.957, and was incorporated into the final design.



My results prove the sensor to be successful in estimating soil conductivity. In the final testing, the sensor estimated a rising soil EC that rose to ~24 mS/m. This corresponds with the true soil EC value measured using a probing device. While the sensor did show some variability in the measurements, it followed a similar average trend to that of the probe, showing an overall accurate measurement. In addition, the resistance over distance plot showed an accurate trendline with an R2 value of 0.98.

Furthermore, the voltage discharge curves for all three anodes of the sensor followed a similar 'path' and were similar to the expected discharge curves for soil based MFC designs. This indicates that there was a similar microbe colony size across the depth of the sensor. In addition, the slightly increasing voltage values from the cell after full discharge prove that the microbes continue to produce electrons after a full discharge. This proves that the cell will be able to continue measuring soil EC values for longer than the measured time. 

When comparing each prototype, an improvement in the accuracy of resistance over distance trendlines was seen from prototype one to three. This improvement in accuracy and reliability is primarily due to the change in the electrode material. Since the electrodes were altered to better fix to microbes and to better attract electrons, they were able to decrease the release of electrons to the environment, increasing the reliability of measurements. 

Average Rs/D Profiles for each prototype:

When this data is represented on an R2 Value / Prototype chart, a clear trend appears showing an increase of reliability through each prototype. 

Overall, the results showed a fairly accurate sensor measurement, when a mass test was conducted with a constantly planted sensor, the results recorded were as shown below. There were some inaccuracies, as shown with the +30 mS/m outliers. However, the average soil EC was accurate. The slight inaccuracy could be due to the length of time sampled, as shown in the table below, the longer sample times tend to be closer to the true soil EC measurements, as they are able to get a more accurate charge rate measurement - average magnitude of derivative of voltage over time multiplied by the integral of the voltage-time discharge curve post discharge. Furthermore, the cell could be improved by introducing extra growth agents to facilitate a more consistent microbe layer on the electrodes. However, in order to preserve the usability and low-maintenance nature of the MFC based sensor, such a design improvement may be avoided. One last potential explanation for the variable soil EC measurements over the multiple measurement tests could be as a result of variable moisture content as a result of intermittent rain during testing, as rain would saturate the soil and create a higher soil EC, as seen in the 6th and 9th test segment. However, when taking the average soil EC across the time segments, the soil EC value was extremely close to the real EC.






Overall, my solution was successful in measuring the soil EC value of the local soil. The final prototype was able to measure the soil EC to within 6.5% error. While the sensors did have some error, they were able to provide a fairly reliable estimation, with the EC-Segment trendline lining up with data measured using a soil EC probe. The total cost of the sensor did not exceed $15, proving to remain affordable and easy to use. Furthermore, the sensor was modular in design, allowing for each sensor depth to be tailored towards the depth of the field topsoil. The final design proved that that microbe-based sensor could be a viable method for soil EC measurements, and proved that the MFC sensor provides comparable results with a conventional probe. My design offers an advantageous method for obtaining soil EC measurements. Since the sensor relies on the microbes to power the measurements, the life of the cell is determined based on the life of the surrounding fixed microbes, which self-regulate colony sizes in their environment. This allows the MFC sensor to have a long lifetime and require low-maintenance. Furthermore, if the sensor is applied to a large field, a low-voltage sensor grid can be set up to locally receive and interpret soil EC estimations from surrounding sensors, and send the estimations to a central computer. This would allow the user to remotely receive a detailed report on soil EC, eliminating the need for costly, high maintenance and time-consuming contact sensor methods. 

My solution 

 Contact method -  Existing Solution          

Image courtesy of Veris Industries - Cited in Bibliography


In the future, I plan on constructing the low voltage sensor grid that would receive and transmit soil EC readings from a large field's worth of sensors and be able to create a field map based on the results. The field map would be the final step in the solution and allow the user to gain useful knowledge of their soil health, estimated crop yield, and watering needs. I have designed a method for interpreting a sample data set of a theoretical field with MFC sensors. 

The field is first split into a grid and the sensor locations are filled with corresponding soil EC ranges.

From there, paths are drawn from adjacent grid pixels of similar soil EC. This creates a skeleton of the true field conditions. 

Lastly, the remaining pixel soil EC is estimated as an average of their surrounding pixels. The final field map is then 'smoothed' to create a gradient of each field condition. These sample field maps are the final results of the data collection and serve to show the reader information on soil health, estimated crop yield, and watering needs.

About me

I love to challenge myself and find new problems to solve. My passion for solving problems has led me to become extremely involved in robotics programs and in local government. I have spent over 7 years competing in FIRST robotics programs, and I am currently mentoring an FTC team and compete on an FRC team. In addition to robotics, I have also become involved in my local government. I love to find real problems in my community and have found a way to address these problems through our local Mayors Youth Advisory Board. But beyond solving problems, I also enjoy challenging myself to try new things, whether picking up a new sport or joining a club. I currently compete on my school's racquetball team and enjoy going on long walks with my dog. 

When I was younger, in kindergarten, I first learned about science and STEM-related fields. We had an after-school teacher, Mr. Orange, who would do small science experiments in our class. For me, being able to make a mini volcano or mini bottle rocket always felt like magic and inspired me to continue learning and experimenting with science. Since then, I have always been drawn to STEM-related activities. 

For me, being able to use creativity and imagination in science and in life is extremely important. I love to design and turn my ideas into reality. I am looking at the Google Science Fair as an opportunity to meet with like-minded people and share and exchange ideas and experiences. 

Health & Safety

I completed all testing in a soil environment and did not culture any bacterium. However, when working with microbes, even in low concentrations, precautions are required. The cells were treated under BSL-1 safety regulations and electrodes were not reused after each test. I did not culture any bacterium or used any culturing methods, such as nutrient broths etc. All microbes used were present in the soil before and after experimentation and all cells removed from testing were immediatly sealed and autoclaved. 

Bibliography, references, and acknowledgements

I received some mentorship from Dr. Hong Liu at OSU. I was able to discuss MFC research with her over the phone and received information on the projects that they do at their lab. 

I also reached out to Dr. Dongwon Ki and Dr. Cesar Torres from ASU; Dr. Than and Dr. Kovscek of Stanford; and Dr. Lou Durlofsky for guidance on my project. 

Contact Method Sensor Image - [Image of Contact Method Soil EC Sensor: Soil EC 3100]. (2017, September). Retrieved December, 2018, from

Logan, B. (n.d.). Bruce Logan Research - Microbial Fuel Cells. Retrieved November 7, 2014, from http://

N.A - A Basic Overview of Fuel Cell Technology. (n.d.). Retrieved November 7, 2014, from http://

Bacteria: Energy Source of the Future. Retrieved November 7, 2014, from bacteria-energy-source-of-the-future/

Suling, Y. (n.d.). - Microbial Fuel Cells. Retrieved November 7, 2014, from publication.html

Nice, K., & Strickland, J. (n.d.). How Fuel Cells Work - HowStuffWorks. Retrieved November 7, 2014, from http:// cell.htm

Nice, K., & Strickland, J. (n.d.). Polymer Exchange Membrane Fuel Cells - How Fuel Cells Work. Retrieved November 7, 2014, from http:// cell2.htm

Rabaey, K., Lissens, G., & Verstraete, W. (n.d.). - Microbial Fuel Cells: Performances and Perspectives. Retrieved November 7, 2014, from Publications/LabMET/Rabaey%20et%20al.pdf

Lovley, D. (20014, October 4). - Microbial Fuel Cells. Retrieved November 23, 2014, from https:// Microbial_fuel_cells#Geobacteraceae_spp.

Logan, B. E., Dr. (2008). Microbial fuel cells. Hoboken, NJ: Wiley-Interscience.

He, Z. (2007). Microbial fuel cells: their application and microbiology.

Wilkinson, M. (2011). Microbial fuel cells: electricity from waste? University of Liverpool.

Bennetto, H. P. (1984), inLife Chemistry Reports, vols., 2, no. 4, Michelson, A. M., and Bannister, J. V., eds., Harwood Academic, London, pp. 363–453.

Habermann, W. and Pommer, E. H. (1991),Appl. Microbiol. Biotechnol. 35, 128.

Sell, D., Kramer, P., and Kreysa, G. (1989),Appl. Microbiol. Biotechnol. 31, 211.

Tanisho, S., Kamiya, N., and Wakao, N. (1989),Bioelectrochemistry and Bioenergetics 21, 25.

Tanaka, K., Kashiwagi, N., and Ogawa, T. (1988),J. Chem. Tech. Biotechnol. 42, 235.

Roller, S. D., Bennetto, H. P., Delaney, G. M., Mason, J. R., Stirling, J. L., and Thurston, C. F. (1984),J. Chem. Tech. Biotechnol..34B, 3.

Delaney, G. M., Bennetto, H. P., Mason, J. R., Stirling, J. L., and Thurston, C. F. (1984),J. Chem. Tech. Biotechnol. 34B, 13.

Thurston, C. F., Bennetto, H. P., Delaney, G. M., Mason, J. R., Roller, S. D., and Stirling, J. L. (1985),J. Gen. Microbiol. 131, 1393, and unpublished results.

Bennetto, H. P., Delaney, G. M., Mason, J. R., Roller, S. D., Stirling, J. L., and Thurston, C. F. (1985),Biotechnol. Lett. 7, 699.

Bennetto, H. P., Delaney, G. M., Mason, J. R., Roller, S. D., Stirling, J. L., Thurston, C. F., and White, D. R., Jr. (1983),Proc. First World Conf. Commercial Applications and Implications of Biotechnology. Biotech '83, Online, London, 655.

Bennetto, H. P., Delaney, G. M., Mason, J. R., Roller, S. D., Stirling, J. L., Thurston, C. F., and White, D. R., Jr. (1986),Alternative Energy Sources VII, vol. 4, Veziroglu, N., ed., Hemisphere, New York, p. 143.

Sheehan, J. C., and Hess, G. P. (1955),J. Amer. Chem. Soc.,77, 1067.

Lowry, O. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951),J. Biol. Chem. 193, 265.

Bennetto, H. P., Stirling, J. L., and Tanaka, K. (1985),Chemistry and Industry 695.

Southampton Electrochemistry Group (1985),Instrumental Methods in Electrochemistry, Ellis Horwood, Chicester, p. 418.

Bennetto, H. P., Delaney, G. M., Mason, J. R., Roller, S. D., Stirling, J. L., Thurston, C. F., and Tanaka, K. (1982),Abs. 8th Int. Conf. Non-Aqueous Solutions, IUPAC, Nantes, 4-2C.


Adviento-Borbe, M.A.A., J.W. Doran, R.A. Drijber, and A. Dobermann. 2006. Soil electrical conductivity and water content affect nitrous oxide and carbon dioxide emissions in intensively managed soils. Journal of Environmental Quality 35:1999-2010.


Dahnke. W.C. and D.A. Whitney. 1988. Measurement of soil Salinity, p. 32-34. In Recommended soil chemical test procedures for the North Central Region. NCR Publ. 221. Revised. North Dakota Agric. Exp. Sta. Bull. 499. Fargo, ND. Maas, E.V., and G.J. Hoffman. 1977. Crop salt tolerance-current assessment. J. Irrig. Drain. Div. Am. Soc. Civ. Eng. 103:115-134.


Patriquin, D.E., H. Blaikie, M.J. Patriquin, and C. Yang. 1993. On-farm measurements of pH, electrical conductivity, and nitrate in soil extracts for monitoring coupling and decoupling of nutrient cycles. Biol. Agric. Hortic. 9:231-272.


Smith, J.L. and J.W. Doran. 1996. Measurement and use of pH and electrical conductivity for soil quality analysis. P. 169-185 In J.W. Doran and A.J. Jones (ed.) Methods for assessing soil quality. Soil Science Society of America Spec. Publ. 49. SSSA, Madison, WI.


Drre, P., & Bahl, H. (n.d.). Microbial Production of Acetone/Butanol/Isopropanol. Biotechnology, 229-268. doi:10.1002/9783527620883.ch6


Han, T. H., Sawant, S. Y., & Cho, M. H. (2017). Development of Suitable Anode Materials for Microbial Fuel Cells. Microbial Fuel Cell, 101-124. doi:10.1007/978-3-319-66793-5_6


Tewari, A. K., Srivastava, P., Singh, V. P., Singh, P., Kumar, R., Khanna, R. S., . . . Hobza, P. (2014). Selective induced polarization through electron transfer in acetone and pyrazole ester derivatives via C–H⋯OC interaction. New J. Chem., 38(10), 4885-4892. doi:10.1039/c4nj00679h