Team:NEFU China/Hardware




Previously, a device based on biology to detect landmines has been designed, which employed bacteria to detect the vapors emitted by landmines for the purpose of landmine detection. However, the system has serious deficiencies, such as high cost and difficulty in large scale production. Therefore, we proposed a new design concept and applied it to the practical landmine detection.


Our project aims to detect landmines based on synthetic biology, and then apply our device to the real world. So far, we have developed a convenient, low-cost and user-friendly device carrying engineered bacteria generated by us to detect landmines. Compared with the previously reported device, our device is cheap, flexible and easy to operate.

In our device, 2,4-Dinitrotoluene (DNT) released by landmines can trigger the engineered bacteria to synthesize the EGFP protein. When exposed to the excitation light, the bacteria can fluoresce and the fluorescent signal can be captured by a photoresistor. The signal will be converted to digital signal and transmitted to a smartphone. Based on the signal from the bacteria, an algorithm of a designated app will be used to locate the landmines.

Our device is mainly divided into four modules: air collection, biological reaction, light sensing and signal processing.

Air Collection

This module is designed to gather air from the ground surface into our device. As shown in the picture, a miniature fan is installed at the bottom for the purpose of promoting airflow through the device and increasing the amount of collected DNT.

Fig.1. The DC Motor Circuit

Fig.2. The Inside of the 3D Printing Device

Meanwhile, there is a gap between the device chassis and the mini-fan bearing surface. When the fan is spinning, air can enter the device through this gap. To avoid the interference of external light, we designed a "gas channel" that only allows gas to get through, but outside light cannot access to stimulate the engineered bacteria. Thus, our device can operate in daytime.

Biological Reaction

After the airflow containing DNT enters our device, the next step is to trigger our engineered bacteria. When DNT in the air gets contacted with the bacteria, it will activate the constructed gene pathway to produce EGFP and make bacteria emit green fluorescent light (see our Design section for details). In addition, our engineered bacteria need to be immobilized in the device. For this purpose, we used the sodium alginate embedding method to attach the bacteria to the interior surface of the device. Sodium alginate has excellent permeability, which allows the bacteria to maintain their normal metabolic activities when exposed to the airflow. It also contains sufficient nutrients to help the engineered bacteria survive for several days so that the bacteria have enough time to collect DNT from landmines and emit fluorescence.

Light Sensing

When our bacteria sense the DNT, they will express EGFP. Under the excitation light, the bacteria can fluoresce, which can eventually be captured by the photoresistor in a circuit. However, the excitation light could also be sensed by the photoresistor, which could falsely trigger the device. To present this from happening, we use a filter that only allows the emitted light to pass through, so that the photoresistor can only sense the fluorescent light generated by the engineered bacteria. Meanwhile, the opacity of the device can prevent the interference of external light.

At the first time, we adopted the method of reflecting excited light, hoping to reduce the volume of the device. However, we found this method lost a lot of excited light and could not accurately measure the fluorescence intensity of EGFP, so we improved it into the second-generation detection device.

Fig.3. The First Generation of Detection Device.

Fig.4. The Second Generation of Detection Device.

Signal Processing

To make the signal readable, the light generated by the engineered bacteria needs to be transformed into electrical signal. We use the firefly luciferase as the reporter gene at the beginning, but the intensity of its emitted light is difficult to be detect by any photoresistor. Thus, we switched to EGFP that needs to be excited by a light of certain wavelength. For this purpose, a lamp generating the required excitation light is an essential part of the device.

First, once bacteria have expressed EGFP, the photoresistor will detect the light emitted by EGFP after being exposed to its excitation light. We used filters to avoid the interference of excitation light. A filter of 510 nm can filter out the excitation light and allow the photoresistor to only detect the fluorescent light from EGFP. In the whole process, we use the photoresistor to convert the light signal into digital signal, and thus the fluorescence intensity of EGFP can be quantified by detecting the resistance value of the photoresistor. In addition, we also build a connection of our device with a smartphone to allow us operate easily and remotely.

Our device has a very low cost to be assembled and great stability, which makes it possible for mass production. We got very good data in the simulative detection to prove its accuracy and stability.


Here are all the parts when dividing our device. The budget is as follow.

Device Budget

Fig. 5. The Device Budget.

Code & Circuit

The code has been uploaded to GitHub.

Fig.6. The Circuit.



Ideally, we used the UAV (unmanned aerial vehicle) to pick up our device and put it into the minefield. In the actual detection, we put the device directly on the ground surface spread with a trace amount of DNT, and let it detect volatilized DNT from the soil for a certain period of time. We real-time recorded the device and detected the change of the signal. As a control group, the device was put on a ground area without DNT. We tested whether the device would emit any signal during the day to excluding the interference of external light. Due to time shortage, we only tested our device under laboratory conditions.

In order to test the stability and reliability of our homemade device, we performed two experiments:

(1)The photoresistor measures the changes in the resistance reflecting EGFP fluorescence intensity over time.
  We tested the influence of the fluorescence emitted by bacteria on the resistance value of the photoresistor under different conditions. Original DH5α was used for calibration of the resistance value. An arabinose-induced pBAD promoter was used to overexpress EGFP with 1% arabinose to check whether if our circuit worked. DH5α contained only the yqjF and DH5α contained the yqjF and yhaJ were both induced by 20 mg/L DNT.

Fig.7. The Resistance Values Affected by EGFP Expressed under Different Conditions over Time.

The photoreceptor is characterized by a gradual decrease in the resistance value with increasing light intensity. As shown in Fig.1, the induction by 1% arabinose proved that the circuit worked well. Under the induction of DNT, EGFP expression driven by the yqjF promoter decreased the resistance values, which initially proved that our device could detect EGFP biofluorescence.

(2)The enzyme-labeled instrument measures the changes of EGFP fluorescence intensity over time.
  In order to obtain an accurate relative intensity of EGFP fluorescence, we used the enzyme-labeled instrument for precise measurement under the same condition as above, and the data were collected every hour.

Fig.8. The Changes of Fluorescence Intensity under Different Conditions over Time.

As shown in Fig.8, The relative intensity of EGFP fluorescence measured by the enzyme labeling device increased in a time-dependent manner. We used DH5α bacteria as the control group, and the relative fluorescence intensity of the other three groups showed an increasing trend with time, which proves that our device can measure the EGFP fluorescence intensity of the engineered bacteria. It indicated the stability of our device.

In order to prove the accuracy of the device and obtain the mathematical correlation between the resistance value and the relative fluorescence intensity, we fit the data measured by our device to the data measured by the enzyme-labeled instrument. The results are shown below:

Fig.9. Correspondence between Fluorescence Intensity and Resistance Value.

The horizontal coordinate of the figure is the data measured by the enzyme-labeled instrument, and the vertical coordinate is the resistance value of the photoresistor.

The regression line showed an R2 value higher than 0.8, which means that the fit of our data was excellent. We also obtained the mathematical relationship between the resistance value of the photoresistor and the relative fluorescence intensity. Through the experiments above, we proved the stability and reliability of our homemade device that can measure EGFP fluorescence stably and accurately.

2.The operation of device

  The device is a fluorescence measuring device, which is composed of a NodeMcu board, Arduino uno board, light-dependent resistor and a filter of 510 nm.

(2)Operating instructions:
①Turn on the excitation light in the device through the app in a smartphone to ensure that the blue LED is on.
②Pour the bacterial solution with the inducer into a test cup.
③Check that the excitation light path of the LED lamp passes through the 510 nm filter.
④Record the data recorded by the mobile terminal.


Fig.10. The Detection Device.

Fig.11. The 3D Priting Device.

Fig.12. Our Device in the Real Environment.