Characterisation of soil properties using a simple device

1. Synopsis

Soil is a highly complex environment, where a wide range of different processes that can influence the growth of plants take place. These processes are driven by the physical and chemical properties of the soil particles, and can be a consequence of multiple environmental factors. Thus, the characterisation of soil is often complicated, involving expensive and specialised equipment.

We developed a simple and low-cost device for the physical and chemical characterisation of soil, that could help in the selection of an appropriate substrate. Among all the different factors that can have an impact on the growth of plants, we decided to focus on 3 different properties; Particle size, soil composition and presence of organic matter, and base saturation. Whilst the particle size can be related with the compaction, which can hinder the growth of roots, the presence of carbon, and the base saturation of soil influences the availability of nutrients on the plants. Thus, these measurements could offer a complete for the assessment of soil samples.

2. Equipment design

For the measurement of soil particle size, we designed a system for the determination of the sedimentation time. This value represents the speed at which the particles in a solution settle due to gravity. As such, it is dependent on the density and radius of the particles. For the measurement of this speed, we employed a tcs34725 colour sensor and a RGB LED, and the sedimentation time was measured on water. The tcs34725 sensor was placed 1 cm below the water level, and both devices were connected to an arduino microcontroller for the data acquisition as shown below. Within this system, the light intensity produced by the LED was measured by the colour sensor after having travelled through a transparent bottle containing the soil sample suspended in water. Since the soil samples are not transparent, the amount of light that could travel through the bottle was dependent on the amount of soil particles suspended in the solution. As such, the proportion of particles in the solution could be correlated with the measured light intensity.

The whole system was contained within a standard cardboard box to block light interferences from the environment. To make a portable system, easy to transport, we included the sensing devices and the microcontroller within the same box, and separated both using an opaque methacrylate sheet as shown below. Both the colour sensor and the LED were connected to the arduino microcontroller for a better control of the experimental conditions, and the white light was produced by using the three RGB LEDs contained within the light source.

To allow a good positioning of the sample container, and reduce the amount of light interferences from the environment, we placed a hole on the cardboard box, and attached a small box onto this space. The hole had the exact size of the bottles to limit the movement of the bottle, and the small box covered the transparent sections.

We also added a hole to the lid of the bottle, which allowed us to connect a tube and the necessary electronics for the measurement of the base saturation of the samples. This parameter was measured using a simple resistance meter and a peristaltic pump connected to a saturated solution of NaCl in water. Since the electrical resistance of a solution decreases when the concentration of ions increase, the concentration of NaCl in the water used for the sedimentation experiment could be indirectly evaluated. To determine the maximum amount of ions that the soil particles can adsorb, the water sample used for the sedimentation experiment was spiked with a saturated solution of NaCl using a peristaltic pump as shown below.

The peristaltic pump was controlled by using a MOSFET Driver Module, which was connected to a 12V power supply. This module allowed us to switch on and off the pump using the arduino microcontroller. In addition, the volume delivered by this pump as a function of the time was calibrated, which allowed us to quantify the volume of the saturated NaCl solution employed.

The electrical resistance of the solution was measured using 2 metal electrodes, separated by 1 cm, and using a 200 kOhm resistance as the reference. This reference could be used to estimate the actual resistance on the solution by measuring the obtained potential through one of the analog pins. A representation of the resistance meter is provided below.

Finally, for the control of the system, we included a potentiometer as a switch, to trigger and stop the experiments. The final system with all these elements could be contained within the cardboard box.

3. Arduino code

The code that we employed for the control of the system we report here is shown below. The experiment was only triggered once the potentiometer was set as maximum (and, therefore, the analog signal received was about 1024). After triggering the switch, we started receiving the colorimetric signal from the tcs34725, and the time was also measured using the micros () function. At this point, the soil sample was added to the water sample, and the system was incubated for at least 50 mins while the different light intensities were recorded. The light intensity was recorded approximately every 1 sec.

Once the light intensity measurements were finished, the electrical resistance of the system was recorded. Initially, the resistance of the solution after introducing the soil was measured. The peristaltic pump was then used to increase the concentration of NaCl in the solution. We operated this pump for 5 s, which added a volume of 3 mL from a saturated solution of salts. The electrical resistance was then measured for 10 mins.

4. Analysis of the colorimetric measurements for estimating soil composition

The determination of the colour intensities measured by the tcs34725 could provide information about 2 different parameters of soil. As mentioned within the equipment design section, the intensity of the light can be used for estimate the particle size, whilst the colour absorption could be related with the composition of the soil itself. To test the system we designed, we studied 4 different soil samples; Pristine coco substrate before growing any plants in it, coco substrate where a plant had been grown, and 2 different clayey soil samples from the field. These soil samples presented different textures and compositions ad can be observed below.

We first proceeded to determine the changes in the total light intensity, measured as the sum of the 3 colour output intensities from the tcs34725. In the case of clay soils, the light intensity decreased greatly initially upon adding the soil samples, and it slowly increased over time due to the sedimentation of the soil at the bottom of the bottles. However, when the coco soil samples were used, given their low density, this pattern could not be observed since the soil only floated on the water, and we only obtained a small signal. For comparison, we also measured the signal of a sample that contained only water, which showed an approximately constant signal.

To estimate the particle size of the soil, we relied on the equation obtained after balancing the forces from the gravity with the buoyancy effects due to the particle density, and the resistance due to the viscosity of the liquid. This equation is shown below and requires the density of water (ρf) the density of the soil particles (ρ), the viscosity (η), the acceleration from gravity (g), and the speed of particles (U). To determine the speed of a given particle sedimenting at a given time, we assumed that the light absorption was proportional with the amount of particles, and we divided the distance of the colour sensor to the water level by the measured time. The corresponding particle size that corresponds to each sedimentation time was then plotted against the proportional amount of particles, calculated from the light intensity. In this case, each value of light intensity was indicative of the cumulative amount of particles sedimenting, since this intensity increased as the number of particles in solution was higher. Thus, by subtracting the previous intensity value on each point of the graph, and ponder the results by the total amount of light intensity obtained, we could estimate the proportion of particles of each size.

One of the limitations of this model is the need of a density value for the soil sample employed. Since we could not measure this density directly using our device, we decided to use an approximate value. Since the predominant soil in the area where the measurements were taken is clayey (as shown by this map by the Cranfield Soil and Agrifood Institute), we used a value of 1.6 g/mL as the density. The rest. Since the manual calculation of these parameters could be tedious, they can be incorporated in the arduino code in further versions of the device to simplify the process. Using this system we could determine that the average particle size of sample 2 was significantly higher than in the case of sample 1.

The colorimetric information obtained by the sensor could additionally be used to assess the amount of organic matter on the soil particles. Multiple studies have determined the relationship between the coloration of soil samples. Whilst this test is not quantitative, since we cannot accurately determine the concentration of organic molecules on the soil particles, we can use it to compare between samples. We measured the red intensity, which is indicative of the presence of oxides within the soil sample, primarily iron oxide. These oxides have been observed to retain organic compounds. Thus, we correlated the proportion of the red signal respect to the green and blue ones with the concentration of organic compounds in soil, and we could analyse the relative abundance of these components on each particle size fraction. This test could also be used in the cases of soil with low density, as can be observed below.

We observed that the majority of the organic matter in soil was contained within the large particles (>10 μm), and the red signal decreased at low particle sizes. This test could be applied to the coco substrates, where a small content of oxide materials and high signal of green was observed. As a control, we measured the signal of water alone, which showed an approximately even signal from all the three colours. Thus, this sedimentation method using a simple colour sensor can be useful for the physical-chemical characterisation of the soil samples

5. Base Saturation

To further improve the capabilities of the system that we designed, we also incorporated a system for the measurement of the base saturation. This parameter is correlated with the amount of ions that the soil can exchange, which have an impact on the nutrient availability of the substrate. As shown in the previous sections, we incorporated this method by using a resistance meter in the water solution. The conductivity of a solution increases at higher amount of electrolytes. Thus, the amount of ions being adsorbed by the soil particles are reflected on the difference between the theoretical conductivity that the sample should have before adding the soil, and the actual conductivity. This parameter could be measured in all the soil samples that we studied using the resistance meter. To the solution used for the sedimentation studies, we added 3 mL of a 1 g/L concentration solution containing NaCl. The changes in the resistance were then measured.

A decrease in the resistance was observed due to the adsorption of ions on the soil samples. This method cannot be used to directly assess the base saturation directly, since the value of cation exchange capacity of the soil samples is needed. However, we could establish the amount of cation uptake by the different soil samples as shown below. The soil samples were weighted, and the decrease in the water resistance with respect to the sample with only water were determined.

The highest absorption of ions within the tested range was the used coco substrate. This value showed a lower concentration of ions in this soil, since it adsorbed the largest amount, especially compared to the pristine coco soil. As such, it could indicate a need of fertilisers to enhance the plant growth on this soil type.

6. Limitations

We have shown that the system we report in this project can determine some of the crucial parameters of soil. However, the accuracy of the results can be compromised due to different factors indicated below:

• We rely on the density values of soil from other sources for calculating the particle size, which could not be accurate for specific types of soils.
• We cannot estimated the particle size in low density soil substrates, since we rely on the sedimentation time, and the quantity of extremely small particles cannot be estimated likewise, due to other processes such as Brownian motion, that lead to high sedimentation times.
• The calculation of the Base saturation is not accurate since the cation exchange capacity of the soil is needed, and the amount of NaCl solution provided might not be enough to saturate the soil particles with electrolytes. In addition, we do not regulate the pH, which would be needed for an accurate estimation of this parameter.

Nevertheless, this method could be a low-cost and easy option for the comparison of soils from a specific field, and predict the need of further processes to improve crop production (i.e. fertilisers, loosening compact soil).