Environmental Sensing in High-Altitude Mountain Ecosystems Powered by Sedimentary Microbial Fuel Cells

The increasing need for fresh water in a climate change scenario requires remote monitoring of water bodies in high-altitude mountain areas. This study aimed to explore the feasibility of SMFC operation in the presence of low dissolved oxygen concentrations for remote, on-site monitoring of physical environmental parameters in high-altitude mountainous areas. The implemented power management system (PMS) uses a reference SMFC (SMFCRef) to implement a quasi-maximum power point tracking (quasi-MPPT) algorithm to harvest energy stably. As a result, while transmitting in a point-to-point wireless sensor network topology, the system achieves an overall efficiency of 59.6%. Furthermore, the control mechanisms prevent energy waste and maintain a stable voltage despite the microbial fuel cell (MFC)’s high impedance, low time response, and low energy production. Moreover, our system enables a fundamental understanding of environmental systems and their resilience of adaptation strategies by being a low-cost, ecological, and environmentally friendly alternative to power-distributed and dynamic environmental sensing networks in high-altitude mountain ecosystems with anoxic environmental conditions.


Introduction
A fundamental understanding of environmental systems, their resilience, and adaptation strategies drive innovative technology development in distributed and dynamic environmental sensing via novel methodologies/technologies (unconventional sensing involving living organisms). Therefore, environmental monitoring has become a global concern [1,2]. For example, the growing need for fresh water requires remote monitoring of water bodies in mountain areas. In addition, microelectronic advances enable a wide range of sensors in remote locations for monitoring habitat, climate, agricultural applications, fisheries research, health systems, air, soil, and water [1,3]. Energy sources of the monitoring systems must be renewable and reliable and not require human intervention [2]. The main concern of implementing a distributed and local sensing of environmental variables via novel methodologies/technologies is the reliability of primary energy sources. Often energy harvesting depends on environmental conditions, and energy can be available when the primary source is available. Energy harvesting or scavenging technology has recently gained attention for clean and sustainable energy [4]. Thus, some alternative energy sources depend on physical conditions, such as the availability of light, wind, heat, and others [2].
Environmental variables sensing and monitoring systems in remote areas use mainly collected energy. It is often not logistically feasible or sustainable due to the high implementation and maintenance costs. Moreover, the monitoring system requires capacitors or disposable batteries, which contain toxic, environmentally hazardous materials, so their

Sedimentary Microbial Fuel Cells
Sediment and water were collected from the Colta lagoon, coordinates −1.725873, −78.757530 at 3212 masl. Five SMFCs were implemented using cylindrical plastic containers with dimensions of 145 mm radius and 200 mm height ( Figure 1a) and filled with a 60 mm thick sediment layer. The water is 100 mm above the sediment, and the cathode electrode is installed 40 mm below the water surface. Without any treatment, the anode and cathode electrodes were made of carbon fiber cloth whose dimensions were 100 mm × 100 mm and 100 mm × 150 mm, respectively. The electrodes were attached with a 100 mm × 20 mm stainless steel mesh and connected to the outside with a 100 mm long nickel/titanium alloy wire. The anode electrode was installed 40 mm inside the substrate and connected outside the container with a cork (Figure 1). a 100 mm × 20 mm stainless steel mesh and connected to the outside with a 100 mm long nickel/titanium alloy wire. The anode electrode was installed 40 mm inside the substrate and connected outside the container with a cork (Figure 1). Each SMFC was connected to a 100 Ω resistance to improve the establishment of the biofilm at room temperature without adding additional nutrients during the 30 days. In addition, water collected from the lagoon was poured periodically to compensate for water loss due to evaporation.
The differential voltages of individual SMFC in steady-state were monitored every minute for 48 h. Subsequently, four SMFCs were connected in parallel (Figure 1b) to increase the power generated. Data were collected in both cases with a NI DAQ SB-6009 card (National Instruments Corporation, Austin, TX 78759, USA) and a LabVIEW virtual instrument. The polarization and power curves of the SMFCs in the parallel circuit were obtained with the resistance box method using variable precision resistors of 2 MΩ, 10 kΩ, and 1 kΩ. The SMFC's current was calculated using Ohm's law ( V IR = ) and the power equation ( P VI = ); these parameters were used to graph polarization and power curves. Finally, the fifth SMFC was connected to a variable resistor to obtain a percentage of its open-circuit voltage (OCV) and used in the quasi-MPPT algorithm.

Power Management System
The PMS uses a bq25505 and the tps61200 IC. The bps25505 IC is from Texas Instruments (Dallas, TX, USA), an ultra-low-power battery lifter and charger specialized in high-impedance low-power sources. The tps61200 from Texas Instruments (Dallas, TX, USA) is a high-efficiency synchronous DC-DC boost converter with a configurable voltage output from 1.8 V to 5.5 V against over-voltage and current. Figure 2 shows the implemented system where the bq25505 extracts the charges from the SMFCs, allowing input voltages as low as 100 mV. The charges are stored in a 0.7 μF supercapacitor (SC). When the SC reaches 2.44 V (VBAT_OK_PROG), it activates a control signal (VBAT_OK). The signal deactivates when it falls below 2.38 V (VBAT_OK_HYST). The VBAT_OK_PROG, VBAT_OK_HYST, and other supercapacitor protection parameters are configured using Rok1, Rok2, Rok3, and Rov1, 2 external resistors ( Figure 2a). Each SMFC was connected to a 100 Ω resistance to improve the establishment of the biofilm at room temperature without adding additional nutrients during the 30 days. In addition, water collected from the lagoon was poured periodically to compensate for water loss due to evaporation.
The differential voltages of individual SMFC in steady-state were monitored every minute for 48 h. Subsequently, four SMFCs were connected in parallel (Figure 1b) to increase the power generated. Data were collected in both cases with a NI DAQ SB-6009 card (National Instruments Corporation, Austin, TX 78759, USA) and a LabVIEW virtual instrument. The polarization and power curves of the SMFCs in the parallel circuit were obtained with the resistance box method using variable precision resistors of 2 MΩ, 10 kΩ, and 1 kΩ. The SMFC's current was calculated using Ohm's law (V = IR) and the power equation (P = V I); these parameters were used to graph polarization and power curves. Finally, the fifth SMFC was connected to a variable resistor to obtain a percentage of its open-circuit voltage (OCV) and used in the quasi-MPPT algorithm.

Power Management System
The PMS uses a bq25505 and the tps61200 IC. The bps25505 IC is from Texas Instruments (Dallas, TX, USA), an ultra-low-power battery lifter and charger specialized in high-impedance low-power sources. The tps61200 from Texas Instruments (Dallas, TX, USA) is a high-efficiency synchronous DC-DC boost converter with a configurable voltage output from 1.8 V to 5.5 V against over-voltage and current. Figure 2 shows the implemented system where the bq25505 extracts the charges from the SMFCs, allowing input voltages as low as 100 mV. The charges are stored in a 0.7 µF supercapacitor (SC). When the SC reaches 2.44 V (VBAT_OK_PROG), it activates a control signal (VBAT_OK). The signal deactivates when it falls below 2.38 V (VBAT_OK_HYST). The VBAT_OK_PROG, VBAT_OK_HYST, and other supercapacitor protection parameters are configured using Rok1, Rok2, Rok3, and Rov1, 2 external resistors ( Figure 2a). The maximum power point (MPP) observed in SMFCs is usually around 50% of their open circuit voltage (OCV). However, when the MPPT module (bq25505) was set to 50% of the OCV, the circuit operation stopped after a few minutes and started working. This phenomenon was attributed to the slow voltage recovery of the SMFCs in open-circuit. The phenomenon was reported by Degrenne et al. [31], and Alaraj and Park [32]used an MFC model's perturbation and observation algorithm to solve this issue. These authors estimated its MPP to modulate the behavior of a bq25504 to obtain an energy harvest close to real MPP with fast convergence.
The SMFCRef was connected to a variable resistor whose output was attached to the VREF_SAMP terminal that controls the voltage at which the MPPT works ( Figure 2a). The variable resistor was calibrated to obtain at its output 50% of the OCV and to be similar to the OCV of the SMFCs with parallel connection. This process avoids estimating the MPP through the direct observation of the OCV or some parameter of the SMFCs. Instead, similar behavior of the SMFCRef and the SMFCs operating under identical conditions is assumed [33,34].
We used the multi-unit optimization method algorithm [35], with correction parameters calculated for non-identical units, to determine the MPP of an MFC with relatively slow dynamics.
The SMFCRef and the SMFCs have different internal resistances, so their MPPs are not coincident. However, previous reports suggest that their OCV present similar and stable behaviors under the same conditions. Therefore, it is an excellent approximation to determine the MPP using a SMFCRef, considering a stable and smooth behavior from the MFCs, implementing a PMS with a quasi-MPPT [11,36,37].
Once the supercapacitor reaches the charging voltage of 2.44 V (VBAT_OK_PROG), the tps61200 module regulates and raises the final voltage level to 3.3 V, activated by the VBAT_OK signal connected to the EN pin ( Figure 2).
Once the capacitor voltage drops below 2.38 V (VBAT_OK_HYST), the sensor signal (SEN) generated by the load activates an Opto-MOSFET to keep the converter active until the transmission process has been accomplished ( Figure 2a). Figure 3 shows the load block diagram consisting of a PIC 16f628A microcontroller (Texas Instruments Co., Austin, TX, USA) as a node controller. A ds18b20 sensor (Hefei Jingpu Sensor Technology Co., Hefei, China) was used as a device to acquire temperature values using the 1-wire communication protocol, with a resolution of 9 bits with an error of ±0.5 °C. In addition, a radio frequency module Xbee ZB S2C TH (Digi International, The maximum power point (MPP) observed in SMFCs is usually around 50% of their open circuit voltage (OCV). However, when the MPPT module (bq25505) was set to 50% of the OCV, the circuit operation stopped after a few minutes and started working. This phenomenon was attributed to the slow voltage recovery of the SMFCs in open-circuit. The phenomenon was reported by Degrenne et al. [31], and Alaraj and Park [32] used an MFC model's perturbation and observation algorithm to solve this issue. These authors estimated its MPP to modulate the behavior of a bq25504 to obtain an energy harvest close to real MPP with fast convergence.

Load System
The SMFC Ref was connected to a variable resistor whose output was attached to the VREF_SAMP terminal that controls the voltage at which the MPPT works ( Figure 2a). The variable resistor was calibrated to obtain at its output 50% of the OCV and to be similar to the OCV of the SMFCs with parallel connection. This process avoids estimating the MPP through the direct observation of the OCV or some parameter of the SMFCs. Instead, similar behavior of the SMFC Ref and the SMFCs operating under identical conditions is assumed [33,34].
We used the multi-unit optimization method algorithm [35], with correction parameters calculated for non-identical units, to determine the MPP of an MFC with relatively slow dynamics.
The SMFC Ref and the SMFCs have different internal resistances, so their MPPs are not coincident. However, previous reports suggest that their OCV present similar and stable behaviors under the same conditions. Therefore, it is an excellent approximation to determine the MPP using a SMFC Ref, considering a stable and smooth behavior from the MFCs, implementing a PMS with a quasi-MPPT [11,36,37].
Once the supercapacitor reaches the charging voltage of 2.44 V (VBAT_OK_PROG), the tps61200 module regulates and raises the final voltage level to 3.3 V, activated by the VBAT_OK signal connected to the EN pin ( Figure 2).
Once the capacitor voltage drops below 2.38 V (VBAT_OK_HYST), the sensor signal (SEN) generated by the load activates an Opto-MOSFET to keep the converter active until the transmission process has been accomplished ( Figure 2a). Figure 3 shows the load block diagram consisting of a PIC 16f628A microcontroller (Texas Instruments Co., Austin, TX, USA) as a node controller. A ds18b20 sensor (Hefei Jingpu Sensor Technology Co., Hefei, China) was used as a device to acquire temperature values using the 1-wire communication protocol, with a resolution of 9 bits with an error of ±0.5 • C. In addition, a radio frequency module Xbee ZB S2C TH (Digi International, Hopkins, MN, USA) with IEEE 802.15.4 standard support for creating point-to-point networks in transparent mode was used.  The system starts the microcontroller when it receives the 3.3 V power supply (VCC) from the PMS and immediately turns the SEN signal on to keep the PMS output activated ( Figure 3). Then it enters standby mode with consumption around 100 nA, waiting for the ASSOC signal ( Figure 3) emitted by the radio frequency module, indicating that it is ready to transmit. The ds18b20 takes a temperature value and sends it through the universal asynchronous receiver transmitter (UART) port for transmission. The transmission format follows the JavaScript Object Notation (JSON) key/value pair format, sending the node identifier and the measured temperature value {"Id": "1", "T1": "22.1"} as parameters. A total of 22 bytes are sent during each transmission.

Load System
A watchdog timer (WDT) was built to automatically reset the microcontroller when no response to the ASSOC signal was received within 2.3 s. Hence, it disconnects the PMS from the load by setting the SEN signal low and preventing energy waste if the connection with the XbeeRF module is not achieved. Table 1 shows the power consumption of the devices according to the manufacturer, with the Xbee RF module consuming approximately 92.6% of the total power.

Base Station System
The base station has an Xbee ZB S2C TH (Digi International) radio frequency module to receive the data sent by the sensor node. In addition, it works with a UDOO NEO Mini PC, intercommunicating with its UART interfaces.

Load Consumption Test
A 6.8 Ω shunt resistor was used to calculate the actual load consumption, whose voltage ℎ was measured every ms with an NI USB-6009 card when the load The system starts the microcontroller when it receives the 3.3 V power supply (VCC) from the PMS and immediately turns the SEN signal on to keep the PMS output activated ( Figure 3). Then it enters standby mode with consumption around 100 nA, waiting for the ASSOC signal ( Figure 3) emitted by the radio frequency module, indicating that it is ready to transmit. The ds18b20 takes a temperature value and sends it through the universal asynchronous receiver transmitter (UART) port for transmission. The transmission format follows the JavaScript Object Notation (JSON) key/value pair format, sending the node identifier and the measured temperature value {"Id": "1", "T1": "22.1"} as parameters. A total of 22 bytes are sent during each transmission.
A watchdog timer (WDT) was built to automatically reset the microcontroller when no response to the ASSOC signal was received within 2.3 s. Hence, it disconnects the PMS from the load by setting the SEN signal low and preventing energy waste if the connection with the XbeeRF module is not achieved. Table 1 shows the power consumption of the devices according to the manufacturer, with the Xbee RF module consuming approximately 92.6% of the total power.

Base Station System
The base station has an Xbee ZB S2C TH (Digi International) radio frequency module to receive the data sent by the sensor node. In addition, it works with a UDOO NEO Mini PC, intercommunicating with its UART interfaces.

Load Consumption Test
A 6.8 Ω shunt resistor was used to calculate the actual load consumption, whose voltage V shunt was measured every ms with an NI USB-6009 card when the load consumed energy in a temperature data transmission. The current flowing through the load was calculated using Equation (1). The voltage across the load is the difference between the voltage of the power supply and the voltage across the shunt resistor, calculated using Equation (2).

Voltage Capacitor Test
Voltage measurements were performed at the cold start and transmission modes. The capacitor, PMS input, and output voltages were registered by a NI USB-6009 card over twenty hours, as shown in Figure 4. The PMS uses four SMFCs connected in a parallel configuration as an energy source. In contrast, the SMFC REF uses a variable resistor tuned until its output voltage equals the voltage where the maximum power point was found.
load was calculated using Equation (1). The voltage across the load is the difference between the voltage of the power supply and the voltage across the shunt resistor, calculated using Equation (2).

Voltage Capacitor Test
Voltage measurements were performed at the cold start and transmission modes. The capacitor, PMS input, and output voltages were registered by a NI USB-6009 card over twenty hours, as shown in Figure 4. The PMS uses four SMFCs connected in a parallel configuration as an energy source. In contrast, the SMFCREF uses a variable resistor tuned until its output voltage equals the voltage where the maximum power point was found.
The test was conducted to determine the behavior of the input and output voltages of the PMS, and to determine how the voltage of the capacitor changes from when it is empty to when it has enough energy for the sensor node to transmit intermittently.   The test was conducted to determine the behavior of the input and output voltages of the PMS, and to determine how the voltage of the capacitor changes from when it is empty to when it has enough energy for the sensor node to transmit intermittently.
Equations (3) and (4) calculate supercapacitor power and absorbed energy per unit of time. Equations (5) and (6) calculate the converter's efficiency. Three tests were conducted to compute the load consumption.

Sedimentary Microbial Fuel Cells
After passing the biofilm maturation stage, after around 34 days, the average voltages recorded in the four SMFCs in one day were: 0.49, 0.47, 0.53, and 0.41 V, and the average voltage of the four SMFCs in parallel configuration was 0.47 V, similar to those average voltages in oxygen-rich water in ponds at Milan (Italy) [1]. The SMFC Ref reached a 0.45 V average voltage. Figure 5 shows the polarization and power curves obtained from the SMFC s in a parallel configuration.

Sedimentary Microbial Fuel Cells
After passing the biofilm maturation stage, after around 34 days, the average voltages recorded in the four SMFCs in one day were: 0.49, 0.47, 0.53, and 0.41 V, and the average voltage of the four SMFCs in parallel configuration was 0.47 V, similar to those average voltages in oxygen-rich water in ponds at Milan (Italy) [1]. The SMFCRef reached a 0.45 V average voltage. Figure 5 shows the polarization and power curves obtained from the SMFCs in a parallel configuration.  [4,18,[38][39][40][41][42][43][44][45]46]. The MPP voltage is used to calibrate the SMFCRef as the MPPT reference value [6,21,26]. Figure 6 shows the voltage and current measured (red and orange lines, respectively) when the load performs a single transmission where the energy consumption is 162.02 mJ in a time of 1.73 s. The voltage drop is due to the measurement shunt resistor's power consumption.  Figure 6 shows the voltage and current measured (red and orange lines, respectively) when the load performs a single transmission where the energy consumption is 162.02 mJ in a time of 1.73 s. The voltage drop is due to the measurement shunt resistor's power consumption. Despite the microcontroller and the sensor consuming lower current (Table 1), the highest current values are recorded when the radio frequency module goes into operation and transmission mode. Table 2 shows the power and load energy consumption during the three operating cycles registered. It obtained an average energy consumption of 162 mJ and a power of 93.4 mW, close to the 117.6 mW calculated using the datasheets of the manufacturer's devices. Table 2. Values of energy and power in the load. Despite the microcontroller and the sensor consuming lower current (Table 1), the highest current values are recorded when the radio frequency module goes into operation and transmission mode.  Table 2 shows the power and load energy consumption during the three operating cycles registered. It obtained an average energy consumption of 162 mJ and a power of 93.4 mW, close to the 117.6 mW calculated using the datasheets of the manufacturer's devices.

Power Management System Operating Cycle
Several PMS systems have been proposed with MPPT systems to extract energy from different MFC [6,7] and photovoltaic systems [21,26]. The objective is to overcome the challenges encountered when employing an integrated circuit not designed for the unique characteristics and dynamics of MFCs [32,34]. In this work, we design a PMS that extracts energy close to the MPP with a dynamic external reference implementing a quasi-MPPT. Figure 7 shows the supercapacitor's charging cycle starting its charge with the lowefficiency "cold-start" converter integrated with the bq25505. A transition occurs at values greater than 1.8 V (VSTOR_CHGEN). The "Main Boost Charger" mode of the high-efficiency converter is enabled using the MPPT. As shown in Table 3 Table 4 shows the energy values stored in the supercapacitor made in the four transmissions of the sensor node and the efficiency of the bq25505. The Xbee radio frequency module is the device with the predominant consumption    The Xbee radio frequency module is the device with the predominant consumption in the system, with a value of 92.6% of the total, followed by the temperature sensor with 2.81%. However, this requires only 93.75 ms to obtain and transmit temperature data. Table 5 summarizes the performance of the proposed PMS powered by SMFCs and the direct losses in the devices. The results in Table 5 suggest that the total efficiency can improve by decreasing the losses of the integrated circuits and the storage device's choice. The system can start up from any voltage value in the supercapacitor. Additionally, the system disconnects the load once the task has been completed to prevent energy losses. The proposed approach is similar to a configuration harvest energy with a quasi-MPPT. It presents a similar behavior to those reported in the literature, as shown in Table 6. However, it uses a configuration with a dynamic reference taken from a SMFC Ref , which makes the configuration even more straightforward. The proposed PMS system is categorized as a boost converter-type integrated circuit system with quasi-MPPT [6]. The PMS had a maximum efficiency of 71.2% at the best integrated circuit (IC) performances. The IC can operate with 100 mV (when the supercapacitor voltage has reached 1.8 V), with an output of 3.3 V with a maximum current of 1 A.
Discrete electronic components and ICs have been used to establish strategies to boost energy, harvest energy at the maximum power point, and increase the MFC power and behavior [21,28,47]. However, conventional ICs for energy harvesting do not work with high impedance and low time recovery seen in the SMFCs. Nevertheless, our PMS power by SMFCs has shown to be a viable alternative power source while working in unfavorable conditions. Despite the anoxic conditions of the water collected from the Colta Lagoon, using an integrated circuit-based system and quasi-MPPT is among the most efficient devices (Table 6). For the sake of comparison, Table 6 summarizes the primary performance indicator of PMS powered by MFCs from the literature compared with the prototype presented in this paper. It can be seen that the proposed approach showed reasonably good performance.

Conclusions
This study implemented a point-to-point wireless sensor network powered by SMFCs to intermittently transmit ambient temperature data to a base station. Our results show the feasibility of SMFC operation in the presence of low dissolved oxygen concentrations for remote, on-site monitoring of physical environmental parameters in high-altitude mountainous areas, although the production of electric current in SMFCs can be affected in the high-altitude mountain where oxygen dissolved in water bodies is usually low.
The implemented PMS continuously and effectively harvests energy from SMFCs. The PMS, the implemented quasi-MPPT, and a reference SMFC enable the extraction of energy stably, reaching an overall efficiency of 59.6%. Many factors affect the MFC's behavior: temperature, bacterial community, dissolved oxygen in water, and physical parameters of the surrounding environment. The control mechanisms prevent energy waste and maintain a stable voltage despite the MFC's high impedance, low time response, and low energy production.
Our system is low price, easy to implement, and adapted to loads with different low-power demands. Moreover, our system enables a fundamental understanding of environmental systems and their resilience of adaptation strategies by being an alternative to power-distributed and dynamic environmental sensing networks in high-altitude mountain ecosystems with anoxic environmental conditions. However, more studies are required to investigate the behavior of the system in long-term operation, field applications, and other SMFC architectures.

Data Availability Statement:
The data presented in this study are available upon request from the corresponding author.