A Time-Interleave-Based Power Management System with ... - MDPI

applied sciences Article

A Time-Interleave-Based Power Management System with Maximum Power Extraction and Health Protection Algorithm for Multiple Microbial Fuel Cells for Internet of Things Smart Nodes Alfredo Costilla-Reyes 1, * , Celal Erbay 2 , Salvador Carreon-Bautista 3 , Arum Han 1 and Edgar Sánchez-Sinencio 1 1 2 3

*

Department of Electrical and Computer Engineering, Texas A&M University, College Station, TX 77843-3128, USA; [email protected] (A.H.); [email protected] (E.S.-S.) TUBITAK-Informatics and Information Security Research Center, Kocaeli 41470, Turkey; [email protected] Analog Devices, Colorado Springs, CO 80920, USA; [email protected] Correspondence: [email protected]; Tel.: +1-979-985-1351

Received: 25 October 2018; Accepted: 17 November 2018; Published: 27 November 2018







Abstract: Microbial Fuel Cell (MFC) technology is a novel Energy Harvesting (EH) source that can transform organic substrates in wastewater into electricity through a bioelectrochemical process. However, its limited output power available per liter is in the range of a few milliwatts, which results very limited to be used by an Internet of Things (IoT) smart node that could require power in the order of hundreds of milliwatts when in full operation. One way to reach a usable power output is to connect several MFCs in series or parallel; nevertheless, the high output characteristic resistance of MFCs and differences in output voltage from multiple MFCs, dramatically worsens its power efficiency for both series and parallel arrangements. In this paper, a Power Management System (PMS) is proposed to allow maximum power harvesting from multiple MFCs while providing a regulated output voltage. To enable a more efficient and reliable power-harvesting process from multiple MFCs that considers the biochemical limitations of the bacteria to extend its lifetime, a power ranking and MFC health-protection algorithm using an interleaved EH operation was implemented in a PIC24F16KA102 microcontroller. A power extraction sub-block of the system includes an ultra-low-power BQ25505 step-up DC-DC converter, which integrates Maximum Power Point Tracking (MPPT) capabilities. The maximum efficiency measured of the PMS was ~50.7%. The energy harvesting technique presented in this work was tested to power an internet-enabled temperature-sensing smart node. Keywords: DC-DC power conversion; Internet of Things (IoT); microbial fuel cell array; power management system; remote monitoring; step-up converter; wastewater

1. Introduction It is projected that networked microsensor technology, in the form of Wireless Sensor Networks (WSNs), will keep playing an important role in the future of remote sensing [1]. Some of the applications for WSNs range from automation, human monitoring, equipment condition monitoring, defense, aerospace, building, structural health monitoring and agriculture [2–6]. As reported in [7], WSNs have been implemented in point-to-multipoint and local point-to-point communication networks achieving long-term field deployment but a limited sensing-data processing. However, WSNs are nowadays increasingly becoming part of a more pervasive concept, the Internet of Appl. Sci. 2018, 8, 2404; doi:10.3390/app8122404

www.mdpi.com/journal/applsci

Appl. Sci. 2018, 8, 2404

2 of 16

Things (IoT), which allows devices to communicate with each other and collect information on a much larger scale through the internet [8]. This opens new possibilities to achieve more comprehensive data processing of the information being collected. IoT has the potential to foster a large number of applications by enabling more devices to be connected to the web and make use of a wider range of pre-collected and pre-processed data. A paramount challenge that can pose a limitation for a wider adoption in the use of IoT smart nodes in areas of difficult access, is inherently associated to its battery requirements for a considerable power specification needed to perform on-field sensor measurements and to send the collected information to the cloud. A desirable feature of WSN is to minimize not only the size and cost of the energy storage element itself, but also to reduce the maintenance needed to constantly replace the sensor’s batteries. Harvesting enough energy from the environment where the WSN is physically located, can be a plausible solution to extend the battery life of a remote sensor and even fully power it for prolonged periods of time [9,10]. The sources of energy that can be harvested from the ambient include solar, kinetic, thermal gradient, radiofrequency, electromagnetic and Microbial Fuel Cells (MFCs) [11–16]. MFCs are a promising emerging source of energy harvesting that can transform organic and inorganic matter to generate electricity by using a bioelectrochemical conversion process [17–21]. MFCs can find a particular application in remote sensing [22] for places that are difficult to access routinely for maintenance but with abundant organic material available, such as sewage water and even oceans [23–25], where the use and replacement of batteries, can be costly or impractical [16,23,26–30]. A potential application that can combine IoT devices and MFC technologies is in the management and optimization of traditionally public services, such as residual water treatment plants [31]. It is important to note that while MFCs can generate power from liquid and organic sediments found in wastewater and oceans, smart nodes can monitor and transmit conditions found in water being treated, such as temperature, pH, conductivity, salinity, and organic sediment concentration, to name a few [23–25,32]. Despite these opportunities, one of the main limitations of energy harvesting from MFC technologies is the low output power levels provided by each individual cell, typically in the order of microwatts to a few milliwatts per liter [33]. For this reason, a Power Management System (PMS) is required to maximize the power conversion efficiency and increase the voltage and current level to the nominal power levels required by wireless sensors. Some implementations of PMS have previously been reported to maximize the energy harvesting efficiencies of single MFCs [23,28–30,34–36]. To reach usable output voltage levels while at the same time increasing the output current density is by connecting several MFCs in series or in parallel to increase the overall power out level. However, differences in power production from individual MFCs, as well as voltage reversal issues when connecting multiple MFCs in series, present a challenge in the overall efficiency of the system, since these variations dramatically worsen the power production of MFCs arrays [33]. The objective of this paper is to present a circuit topology for efficient energy harvesting from multiple MFCs to increase the total available energy of the system to meet the voltage and current specifications of a power-demanding application. A ranking algorithm was developed based not only on the voltage of the MFC, which does not contain information about MFCs internal resistance, but rather on the available power from each individual MFC to prioritize the charge extraction from the MFC that has the highest available power. At the same time, recovery time is a critical observed requirement for the MFC to recharge its internal capacitive elements as part of its power generation cycle, to avoid damaging the MFC, due to power over-depletion and extend its lifetime. The algorithm also includes an MFC-failure protection feature that isolates an MFC in the array when the algorithm notices a significant power drop during the energy harvesting process. The MFC output power is regulated by a step-up DC-DC converter that stores the extracted charges in an output supercapacitor. Finally, the energy harvesting technique presented in this work is tested and demonstrated through an internet-enabled smart node, capable of transmitting the collected data to the internet. A performance

Appl. Sci. 2018, 8, 2404

3 of 16

Appl. Sci. 2018, 8, x FOR PEER REVIEW

3 of 16

comparison to other power management works based on MFC is made in terms of charging time, topology features and efficiency. The paper is organized as follows: In Section 2, a brief introduction to MFC is presented and its The paper is organized as follows: In Section 2, a brief introduction to MFC is presented and its electrical model is described. Section 3 discusses the choice of top level circuit architecture for electrical model is described. Section 3 discusses the choice of top level circuit architecture for multiple multiple MFC power extraction and describes the MFC power ranking and power extraction MFC power extraction and describes the MFC power ranking and power extraction algorithms. algorithms. The experimental results are presented in Sections 4 and 5 summarizes this work. The experimental results are presented in Sections 4 and 5 summarizes this work.

2. MFC and Power Management System Specification 2. MFC and Power Management System Specification This section aims to provide accurate details of the construction of the MFCs used in this work, This section aims to provide accurate details of the construction of the MFCs used in this its electrical description to determine the proper PMS specifications and the proposed system work, its electrical description to determine the proper PMS specifications and the proposed system specifications. All these components are described in the following. specifications. All these components are described in the following. 2.1. MFC MFC Construction Construction and and Characterization Characterization 2.1. Four MFCs MFCs were were fabricated fabricated from from an an acrylic acrylic anode anode and and cathode cathode chambers. chambers. One One 1000 1000 mL mL MFC MFC Four and three 240 mL MFCs were constructed, the two different sizes were used to simulate MFCs having and three 240 mL MFCs were constructed, the two different sizes were used to simulate MFCs TM, Ion Power differentdifferent output power ProtonA Exchange MembraneMembrane (PEM) (Nafion 117 having outputlevels. powerAlevels. Proton Exchange (PEM) (Nafion 117TM , Inc., Ion New Castle, DE, USA) was used to separate the anode and cathode chambers from each other, Power Inc., New Castle, DE, USA) was used to separate the anode and cathode chambers fromwhile each selectively proton generated bygenerated bacteria to overtotocross the cathode chamber to complete other, whileallowing selectively allowing proton bycross bacteria over to the cathode chamber the electrochemical reaction to produce power. The anode was a carbon felt (Morgan, Durham, CT, to complete the electrochemical reaction to produce power. The anode was a carbon felt (Morgan, 2 USA) and the cathode was a carbon cloth containing 0.5 mg/cm of Pt catalyst on one side Durham, CT, USA) and the cathode was a carbon cloth containing 0.5 mg/cm of Pt catalyst on one (ElectroChem, Inc.,Inc., Woburn, MA, USA) forfor allall MFCs. connected side (ElectroChem, Woburn, MA, USA) MFCs.Both Bothanode anodeand and cathode cathode were were connected through a titanium wire to a 1 kΩ load resistor that was placed between them to allow the electrons through a titanium wire to a 1 kΩ load resistor that was placed between them to allow the electrons produced by bacteria to flow. The schematic diagram of the MFC is shown in Figure 1a. The electrical produced by bacteria to flow. The schematic diagram of the MFC is shown in Figure 1a. The electrical model in in Figure Figure 1b 1b is is further further discussed discussed in in Section Section 2.2. 2.2. model

* (a)

(b)

Figure 1. Microbial Fuel Cell (MFC) device device description description (a) (a) schematic schematic and and (b) (b) electrical electrical model. model.

All inoculated withwith anaerobic activated sludge collected from the Austin All MFCs MFCswere were inoculated anaerobic activated sludge collected from Wastewater the Austin Treatment Plant for the initial growth of the bacteria on anodes. Then growth medium containing Wastewater Treatment Plant for the initial growth of the bacteria on anodes. Then growth medium acetate 1.0 g/L in Nutrient/Mineral/Buffer (NMB) solution was used (10 mL/L mineral base ( ) containing acetate (1.0 g/L) in Nutrient/Mineral/Buffer (NMB) solution was used (10 mL/L mineral 1, 10 mL/L mineral base 2, and 1 mL/L nutrient base). The solution was replaced with a fresh base 1, 10 mL/L mineral base 2, and 1 mL/L nutrient base). The solution was replaced with a fresh one when the the voltage voltageacross acrossthe theload load resistor dropped below 50 mV [28,37–39]. The constructed one when resistor dropped below 50 mV [28,37–39]. The constructed twotwo-chamber MFCs are shown in Figure 2. chamber MFCs are shown in Figure 2. The in in realreal time in volts (V) against time time in seconds (s) [40](s)by[40] using The MFC MFCvoltages voltageswere weremonitored monitored time in volts (V) against in seconds by ausing digital multimeter (Fluke (Fluke 79 series Everett, WA, USA) through a multiplexer (Agilent (Agilent 34970A, a digital multimeter 79II,series II, Everett, WA, USA) through a multiplexer Santa Clara, CA, USA). Once the voltage production level was stable, power andpower voltageand at maximum 34970A, Santa Clara, CA, USA). Once the voltage production level was stable, voltage at power point were obtained by varying a testing load resistor value between the electrodes inelectrodes the range maximum power point were obtained by varying a testing load resistor value between the of Ω [41]. 1 summarizes specifications typical power of all MFCs. of in 5–20 the range of Table 5–20 Ω [41]. Table 1the summarizes the and specifications and production typical power production

all MFCs.

Appl. Sci. 2018, 8, 2404 Appl. Sci. 2018, 8, x FOR PEER REVIEW

4 of 16 4 of 16

Figure 2. 2. Two-chamber Two-chamber microbial microbial fuel cell and power management unit set up. Figure Table 1. Specifications and performance of MFCs. Table 1. Specifications and performance of MFCs. MFC 1

2 MFC 3 MFC 4 MFC 1 MFC MFC 2 MFC 3 MFC 4 Total volume (mL) 240 240 240 1000 Total volume (mL) 240 240 240 1000 Catholyte Air Air Ferricyanide Ferricyanide Catholyte 2 Air Air Ferricyanide Ferricyanide 12 12 12 50 Anode area (cm ) Anode area area(cm (cm2 )2) 12 12 1212 1212 5050 Cathode Cathode areapoint (cm2(MPP) ) 12 50 Power @ maximum power (µW) 59512 48412 435 6400 Power @ maximum power point (MPP) (μW) 595 484 435 6400

2.2. MFC Electrical Equivalent Modeling 2.2. MFC Electrical Equivalent Modeling The MFC can be electrically modeled as a voltage source with an internal resistance R MFC , MFCvoltage can be electrically as a voltage source with an internal resistance 𝑅 value , open openThe circuit VMFC , and modeled internal capacitance C MFC as shown in Figure 1b. The of circuit voltage 𝑉 , and internal capacitance 𝐶 as shown in Figure 1b. The value of 𝑅 is R MFC is composed of several different components (anode, cathode, membrane, and electrolyte composed severalThe different components cathode, membrane, andwhich electrolyte resistances)of[28–30]. value of VMFC is the(anode, MFC’s thermodynamic voltage, variesresistances) nonlinearly [28–30]. The value of 𝑉 is the MFC’s thermodynamic voltage, which varies nonlinearly depending on the solution pH, temperature, and substrate concentration. Since MFCs produce a DC depending on the solution pH, temperature, and substrate concentration. Since MFCs produce a DC voltage, C MFC is mostly ignored in similar works [16,27]. However, due to the method of locating voltage, 𝐶 is mostly ignored in through similar works [16,27]. However, toMFC the method of locating the the Maximum Power Point (MPP) the open circuit voltage due of the in the proposed PMS, Maximum Power Point (MPP) through the open circuit voltage of the MFC in the proposed PMS, C MFC is important for the proper design of a Maximum Power Point Tracking (MPPT) algorithm; thus, 𝐶 is important the proper design of circuit a Maximum Pointwas Tracking (MPPT) algorithm; it is included in ourfor model. The equivalent model Power of the MFC used for the design of the thus, it is included in our model. The equivalent circuit model of the MFC was used for the design of PMS and the dynamic MPPT in this work. the PMS and the dynamic MPPT in this work. 2.3. System Specifications 2.3. System Specifications A voltage regulation stage is required to boost the low voltages (less than 1 volt) provided by each voltage regulation stagevoltage is required to boost the proper low voltages than volt) provided by of theAfour MFCs to a nominal that can provide power(less values for1the wireless sensor each of theproperly. four MFCs a nominal voltageintegrated that can provide values for the wireless to operate Thetostep-up converter into the proper system power provides dynamic impedance sensor to and operate properly. The step-up integrated into the system provides dynamic matching low-power consumption forconverter maximum-power extraction. impedance matching and low-power consumption Since power extraction is performed to an arrayfor of maximum-power four MFCs, MFC’sextraction. health-state measurements, Since power extraction is performed to an array of four MFCs, MFC’s health-state multiplexing control capabilities and the flexibility offered by a programmable device are key features measurements, multiplexing control capabilities and the flexibility offered by a programmable device required for the control block of the proposed system. Thus, an extremely-low power microcontroller are key features required for the control blockand of the proposed system. Thus, an extremely-low power is required to perform MFCs power ranking multiplexing operations. microcontroller is required to perform MFCs power ranking and multiplexing The energy-storage system requirements for an IoT smart node requiredoperations. the incorporation of The energy-storage systemofrequirements an IoT30smart node required the incorporation an an output load supercapacitor 5 F, that can for support s of active temperature sampling andof data output load supercapacitor 5 F,the that can specifications support 30 s for of active temperature sampling andsensor. data transmission to the cloud, to of meet power an internet-enabled temperature transmission to the cloud, to meet the power specifications for an internet-enabled temperature sensor. A super capacitor allows not only a larger charge storage as compared to standard capacitors,

Appl. Sci. 2018, 8, 2404

5 of 16

A super capacitor allows not only a larger charge storage as compared to standard capacitors, but also longer charging and discharging cycles than their lithium-ion battery counterparts. The design specifications are presented in Table 2. Table 2. Specifications of the MFC-powered Internet of Things (IoT) system. Specification

Value

Number of MFCs Average Vin Vout Output supercapacitor

4 350 mV 3.3 V 5F

3. Circuit Architecture for Multi-MFC PMS In this section, a summary of previous EH techniques for multiple MFC is presented along with a proposed architecture that considers a DC-DC converter with MPP capabilities that has been tailored for MFC technology. A PMS algorithm has been embedded in a Programmable Intelligent Computer (PIC) microcontroller and explained here, its main purpose is to allow the extraction of an array of MFCs while providing health protection features. Finally, a description of the final implementation of the proposed PMS is given. 3.1. Overview of Energy Harvesting for MFC Due to the technology’s nature, the MFC output voltage is very low and non-regulated [28], which represents a major challenge for efficient power extraction. A parallel configuration of MFCs can be used to increase the output current with the disadvantage that the output voltage will remain at low levels. Another drawback of the parallel connection of MFCs is that if their output voltage deviates from each other, this effect leads to a detrimental charge transfer among cells in the array, also known as voltage reversal, preventing the stack of MFCs to fully deliver its charges to a load [26]. This results in power losses that lead to a low overall power efficiency of the array. On the other hand, a serial connection of MFCs provides higher output voltages based on the sum of individual VMFC , with a common current that flows throughout the MFC array. In [33] a voltage balancer is used along with a DC-DC converter for an array of serially connected MFCs. By doing so, the output voltage, now higher than that of a single MFC, is efficiently harvested by providing an active switching capacitor method to balance the system. However, the system can also face multiple drawbacks, including power extraction interruption, and even system failure if a single MFC in the stack is damaged or physically disconnected, due to a contact deterioration, for example. This can potentially disable the operation of the DC-DC converter block, leading to a loss of power available from the remaining properly-functioning MFCs. Previous works in PMS have mainly focused on power extraction from a single MFC [27,42]. This approach is limited to the amount of power that a single MFC can generate, which can be depleted quickly, resulting in the interruption of the power supply at the load. In addition, if the system does not integrate an automatic MPPT algorithm, achieving maximum power harvesting can become cumbersome and time-consuming if conducted manually [23,27,33,43–45]. Even more important, power extraction may not be efficient, since the MPP can shift during the lifetime of an MFC. For this reason, the integration of an MPPT is essential for the DC-DC converter to dynamically adjust to the electrical characteristics across multiple MFCs. In this work, the system proposed is intended to harvest the energy from an array of MFCs in a time interleaved fashion, allowing the inactive MFCs to recharge its internal capacitances (C MFC ) when they are not being harvested. The system also integrates an MFC-failure protection mechanism, which automatically isolates one or multiple MFCs from the system if its available power is below a minimum threshold. The MFC-failure protection feature is intended to isolate an MFC when it has been depleted, damaged or it has been physically disconnected.

Appl. Sci. 2018, 8, 2404 Appl. Sci. 2018, 8, x FOR PEER REVIEW Appl. Sci. 2018, 8, x FOR PEER REVIEW

6 of 16 6 of 16 6 of 16

3.2. Overview of the Proposed PMS Circuit Circuit for Multi-MFC Multi-MFC 3.2. 3.2. Overview Overview of of the the Proposed Proposed PMS PMS Circuit for for Multi-MFC The block diagram diagram of the the proposed system system is shown shown in Figure Figure 3. Each MFC in the array array is The The block block diagram of of the proposed proposed system is is shown in in Figure 3. 3. Each Each MFC MFC in in the the array is is ®, individually connected to a low-𝑅 N-type CMOS switch DMN1019UVT (Diodes Incorporated ® individually individually connected connected to to aa low-R low-𝑅on N-type N-type CMOS CMOS switch switch DMN1019UVT DMN1019UVT (Diodes (Diodes Incorporated Incorporated®,, Plano, TX, controlled by anby extremely low power PIC24F16KA102 (Microchip Plano, TX,USA), USA), controlled an extremely lowmicrocontroller power microcontroller PIC24F16KA102 Plano, TX, USA), controlled by an extremely low power microcontroller PIC24F16KA102 (Microchip ®, Chandler,®AZ, USA) [46]. The PMS algorithm embedded in the microcontroller Technology Inc. (Microchip Inc. AZ, , Chandler, AZ,The USA) The PMS algorithm embedded in the Technology Technology Inc.®, Chandler, USA) [46]. PMS[46]. algorithm embedded in the microcontroller allows testing the total testing power the available in each MFC inin the array, byinusing a power measurement microcontroller allows total power available each MFC the array, by using a power allows testing the total power available in each MFC in the array, by using a power measurement block, as shown in as Figure 4. in The microcontroller also regulates powerthe extracted from an measurement block, shown Figure 4. The microcontroller also the regulates power extracted block, as shown in Figure 4. The microcontroller also regulates the power extracted from an individual MFC in the stack by defining the time interleaved EH sequence and the amount of time from an individual in the by defining time interleaved EH sequence the amount individual MFC in MFC the stack bystack defining the timethe interleaved EH sequence and theand amount of time needed for eachfor MFC is harvested again. The DC-DC converter block consists of an ultraof time needed eachbefore MFC it before it is harvested The DC-DC converter block consists of an needed for each MFC before it is harvested again. again. The DC-DC converter block consists of an ultra®, Dallas, low power boost converter BQ25505 (Texas Instruments Incorporated TX, USA) [47]. Finally, ® ultra-low converter BQ25505 Instruments Incorporated Dallas, TX, USA) [47]. ®, Dallas,, TX, low powerpower boost boost converter BQ25505 (Texas(Texas Instruments Incorporated USA) [47]. Finally, the extracted charges charges are stored in an output supercapacitor that powers the IoT the sensor. Finally, the extracted are stored in an output supercapacitor that powers IoT sensor. the extracted charges are stored in an output supercapacitor that powers the IoT sensor.

Figure 3. Block diagram of the proposed Power Management System (PMS). Figure 3. Block diagram of the proposed Power Power Management Management System System (PMS). (PMS).

Figure 4. Diagram of the proposed power measurement circuit. Figure 4. Diagram of the proposed power measurement circuit.

As the output voltage level of the MFC is not high enough for the proposed topology to allow As the output voltagefrom levelthe of PMS, the MFC is not high enough for the proposed topology to allow for self-starting operation an external one-time pre-charging of an output secondary for self-starting operation from the PMS, an external one-time pre-charging of an output secondary to to begin controller operation. Multiple different approaches may be taken capacitor to to1.8 1.8VVisisrequired required begin controller operation. Multiple different approaches may be capacitor to 1.8 V is required to begin controller operation. Multiple different approaches may be to startup the system [48,49]. taken to startup the system [48,49]. taken to startup the system [48,49]. Once the system begins extracting energy from the MFC, there is no longer need for an external Once theto system extracting energy from the MFC, there is no longer need for an external power source powerbegins the PMS. PMS. power source to power the PMS. 3.3. Maximum Maximum Power Power Point Point DC-DC DC-DC Converter Converter 3.3. 3.3. Maximum Power Point DC-DC Converter Based on on the diagram presented presented in in Figure Figure 1b, 1b, the selected for for Based the electrical electrical diagram the DC-DC DC-DC boost boost converter converter selected Basedwas on the electrical diagram presented in Figure 1b, the DC-DC boost converter selected[29]. for this work programmed to set its input resistance at 50% of the V open circuit voltage MFC this work was programmed to set its input resistance at 50% of the 𝑉 open circuit voltage [29]. As thisshown work was programmed to set its input resistance at 50% of the 𝑉 when open circuit voltage [29]. As As in Equation (1), since the maximum power point is reached the input resistance of the shown in Equation (1), since the maximum power point is reached when the input resistance of the shown in Equation (1), since the maximum power point is reached when the input resistance of the PMS equals it sets sets a for the PMS equals to to the the MFC’s MFC’s input input resistance resistance characteristic, characteristic, it a minimum minimum condition condition for the power power PMS equals to the MFC’s input resistance characteristic, it sets a minimum condition for the power losses that can be tolerated in by the internal resistance of switches, and sheet resistance for the losses that can be tolerated in by the internal resistance of switches, and sheet resistance for the Printed Circuit Board (PCB). Printed Circuit Board (PCB). (1) 𝑃 =𝐼 𝑅 ;𝑅 =𝑅 (1) 𝑃 =𝐼 𝑅 ;𝑅 =𝑅

Appl. Sci. 2018, 8, x FOR PEER REVIEW

7 of 16

⇒𝑅

≤𝑅

.

Appl. Sci. 2018, 8, 2404

7 of 16

3.4. Multi-MFC PMS Algorithm microcontroller’s 10-bit to Digital Converter (ADC)and determines the interleaved power lossesAthat can be tolerated in byAnalog the internal resistance of switches, sheet resistance for the Printed extraction time per MFC. As seen in Figure 5, the algorithm defines two thresholds for the power Circuit Board (PCB). PLOSS = I 2 R LOSS ; R LOSS = R MFC extraction 𝑉 and 𝑉 as follows: (1) R PMU ≤ R MFC . is predefined at 50% of the initial 𝑉 determines when to stop⇒the EH+ PCB operation and measured 𝑉 . When 𝑉 is reached, the MFC is removed from the PMS until the voltage 3.4. Multi-MFC PMS Algorithm threshold high, 𝑉 , is reached. The recovery time is defined as the time it takes for the MFC to go from 𝑉 to 𝑉 𝑉 𝑉 Converter are (ADC) determined after the selecting the voltage A microcontroller’s 10-bit.Analog to and Digital determines interleaved power thresholds time that yield the most optimum power extraction anddefines recovery time for energy extraction per MFC. As seen in Figure 5, the algorithm two thresholds forharvesting, the power without over-depleting the MFC from a series of threshold testing of an adaptive algorithm. extraction Vth− High and V th− Low as follows:

Figure 5. Key algorithm reference points of MFC power extraction waveform. Figure 5. Key algorithm reference points of MFC power extraction waveform.

Vth− Low determines when to stop the EH operation and is predefined at 50% of the initial measured The power extraction algorithm presented in Figure 6a is described below: VMFC . When Vth− Low is reached, the MFC is removed from the PMS until the voltage threshold high, 1.th− High The, is Power Ranking Algorithm subroutine Section executed. As aVreturn V reached. The recovery time is defined(described as the timein it takes for3.5) the is MFC to go from th− Low a look-up table is filled out with the individual MFCs’ power rankings to Vthvariable, . V and V are determined after selecting the voltage thresholds that yield the − High th− Low th− High 2. The look-up tableextraction is accessed to recovery interleavetime the for MFC power extractionwithout processover-depleting the most optimum power and energy harvesting, 3. The DC-DC converter is enabled start EH MFC from a seriesboost of threshold testing of an to adaptive algorithm. 4. The VMFCpower is compared to the low threshold voltage 𝑉 extraction algorithm presented in Figure 6a is described below: 5. When 𝑉 voltage is reached the DC-DC converter is disabled 1. The Power Ranking Algorithm subroutine (described in Section 3.5) is executed. As a return 6. If more MFCs in the array are waiting to be harvested, then the program jumps to instruction variable, a look-up table is filled out with the individual MFCs’ power rankings #2. Otherwise, the system stops its EH mode and waits until the recovery time of the highest 2. ranked The look-up is accessed to interleave the MFC power extraction process MFC table is reached 3. The DC-DC boost converter is enabledthe to algorithm start EH restarts. 7. After the recovery time is completed, 4. VMFC is compared to the low threshold voltage Vth− Low The proposed interleaved energy extraction allows the step-up DC-DC converter to dynamically 5. When Vth− Low voltage is reached the DC-DC converter is disabled find the MPP of each individual MFC, ensuring a maximum power output per MFC while at the same 6. If more MFCs in the array are waiting to be harvested, then the program jumps to instruction #2. time, the idle time for the remaining MFCs is used as an MFC’s recovery time that directly affects the Otherwise, the system stops its EH mode and waits until the recovery time of the highest ranked MFC’s health status. MFC is reached Another important feature of this approach is that if there is a failure at one or more MFCs, the 7. After the recovery time is completed, the algorithm restarts. system automatically adapts to it by assigning a rank of zero to that specific MFC, which allows the system neglect the damagedenergy cell. extraction allows the step-up DC-DC converter to dynamically Thetoproposed interleaved find the MPP of each individual MFC, ensuring a maximum power output per MFC while at the same time, the idle time for the remaining MFCs is used as an MFC’s recovery time that directly affects the MFC’s health status.

Appl. Sci. 2018, 8, 2404

8 of 16

Another important feature of this approach is that if there is a failure at one or more MFCs, the system automatically adapts to it by assigning a rank of zero to that specific MFC, which allows Appl.system Sci. 2018, x FOR PEER 8 of 16 the to8,neglect the REVIEW damaged cell.

(a)

(b)

Figure Figure 6.6. Flow diagram for MFC MFC power power extraction: extraction: (a) Main Program and and (b) (b) Power Power Ranking Ranking Algorithm Algorithmsubroutine. subroutine.

3.5. Power Ranking from Multiple MFCs 3.5. Power Ranking from Multiple MFCs Due to the bioelectrochemical nature of the MFC (parasitic reactive elements), its open circuit Due to the bioelectrochemical nature of the MFC (parasitic reactive elements), its open circuit output voltage does not hold a direct relationship to the total amount of energy. Measuring the MFC output voltage does not hold a direct relationship to the total amount of energy. Measuring the MFC open circuit output voltage it is not necessarily indicative of the amount of energy available from open circuit output voltage it is not necessarily indicative of the amount of energy available from a a specific device in the array because its internal resistance ( R MFC ) is not considered. Therefore, specific device in the array because its internal resistance (𝑅 ) is not considered. Therefore, the the followed approach was chosen to properly estimate the power available per MFC: followed approach was chosen to properly estimate the power available per MFC: The circuit shown in Figure 4 uses a 100 µF power measurement capacitor (CPM ) and a charging The circuit shown in Figure 4 uses a 100 μF power measurement capacitor (𝐶 ) and a charging switch (Cap Charge) in a series to an MFC under testing. A second switch (Cap Discharge) is in parallel switch (Cap Charge) in a series to an MFC under testing. A second switch (Cap Discharge) is in with CPM to reset its voltage conditions at the end of a power measurement by setting its initial voltage parallel with 𝐶 to reset its voltage conditions at the end of a power measurement by setting its to zero. By neglecting parasitic elements, the power in the circuit shown in Figure 4 can be calculated initial voltage to zero. By neglecting parasitic elements, the power in the circuit shown in Figure 4 using (2)using below: can beEquation calculated Equation (2) below: dV P = V · I = V (t) · CPM 𝑑𝑉. (2) dt (2) 𝑃 = 𝑉 ⋅ 𝐼 = 𝑉(𝑡) ⋅ 𝐶 . 𝑑𝑡 It can be noticed from Equation (2) that the voltage change in time dV dt of CPM is directly of 𝐶 is such directly It can betonoticed Equation (2) of that voltage in time proportional the totalfrom available power thethe MFC underchange test, and most importantly, as charging time,to as the estimated in Equation (3), intrinsically resistances R MFC and RSwitchsuch , where proportional total available power of the MFCconsiders under test, and most importantly, as the latter resistance is constant.inUsing this principle, the power ranking algorithm an accurate and 𝑅 , charging time, as estimated Equation (3), intrinsically considers resistancesregisters 𝑅 power measurement every 16.52 ms. where the latter resistance is constant. Using this principle, the power ranking algorithm registers an accurate power measurement every 16.52 ms. τDischarge = C MFC · Rout = C MFC · ( R MFC + RSwitch ). (3) 𝜏 =𝐶 ⋅𝑅 =𝐶 ⋅ (𝑅 + 𝑅 ). (3)

Thus, the final value of 𝐶 , holds a direct relationship to the MFC’s individual power. Both, the capacitor value and the constant charging time were selected to provide meaningful information regarding the electrical characteristics of all the MFCs connected to the system. The previously described process is repeated for each MFCs located in the MFC array. After all the MFCs are

Appl. Sci. 2018, 8, 2404

9 of 16

Thus, the final value of CPM , holds a direct relationship to the MFC’s individual power. Both, the capacitor constant charging time were selected to provide meaningful information Appl. Sci. 2018, 8,value x FOR and PEERthe REVIEW 9 of 16 regarding the electrical characteristics of all the MFCs connected to the system. The previously described is repeated forineach MFCs located in themandates MFC array. After all the MFCs are measured, measured,process the data is sorted a look-up table that the order in which each MFC is the data is sorted in a look-up table that mandates the order in which each MFC is harvested. harvested. The The description description of of the the power power ranking ranking sub-routine sub-routineshown shownin inFigure Figure6b 6bisispresented presentedbelow: below:

1. 1. 2. 2. 3. 3. 4. 4.

The power ranking measurement capacitor is initialized to 0 V by Cap Discharge switch switch The MFC under testing is connected to the ranking circuit using Cap Charge switch The MFC under testing is connected to the ranking circuit using Cap Charge switch ADC are stored stored in in the the look-up look-up table table ADC values values of of MFC MFC measured measured power power are The ADC values are sorted, assigning the highest rank to the MFC withwith maximum voltage, and The ADC values are sorted, assigning the highest rank to the MFC maximum voltage, lowest rank to the MFC with minimum voltage. MFCs that cannot provide significant power are and lowest rank to the MFC with minimum voltage. MFCs that cannot provide significant power ranked zerozero andand neglected in the subsequent EH EH process. are ranked neglected in the subsequent process.

3.6. Implementation 3.6. The proposed PMS was developed a developed on on aa two-layer two-layer PCB PCBshown shownininFigure Figure7.7.On Onthe thetop topside, side, microcontroller was placed along with the low-R switches and the power measurement capacitor. a microcontroller was placed along with the low-R switches and the power measurement The DC-DC DC-DC boost boost converter converter is is found found at at the the bottom bottom side side of of the the PCB. PCB. Detailed files files of this work can be The found in the Supplementary Materials section of this manuscript. found in the Supplementary Materials section of this manuscript.

(a)

(b)

Figure 7. Circuit Board of the multi-MFC power extraction, (a) top view; Figure 7. Printed Printed Circuit Board ofPMS the for PMS for multi-MFC power extraction, (a) (b) topbottom view; (b) bottom view. view.

The The collected collected power power is is stored stored in in an an off-board off-board output output supercapacitor supercapacitor that that powers powers the the IoT IoT sensor sensor node node used used to to test test the the system. system. The final final algorithm algorithm was was implemented implemented using MPLAB X IDE and and XC8 XC8 programmed programmed in the the PIC24F16KA102’s non-volatile internal flash memory. It is important to mention that even though the PIC24F16KA102’s non-volatile internal flash memory. It is important to mention that even though information retrieved by the ranking algorithm was stored in the volatile data memory, the use the information retrieved bypower the power ranking algorithm was stored in the volatile data memory, of a non-volatile Electrically Erasable Programmable Read-Only Memory (EEPROM) to store the use of a non-volatile Electrically Erasable Programmable Read-Only Memory (EEPROM) to such store power ranking valuesvalues is available in case off-line data analysis is required. The sensor data such power ranking is available infurther case further off-line data analysis is required. The sensor collected in thisinwork acquired with the purpose to be transmitted and stored a remote in data collected this is work is acquired with the purpose to be transmitted andinstored in aserver remote the cloud to enable further and more complex data analysis. server in the cloud to enable further and more complex data analysis. Even though though the proposed proposed system can natively support up to nine MFC MFC devices devices without any further modification only four MFCs were tested in this work. It is also important to notice that ifthat more further modification only four MFCs were tested in this work. It is also important to notice if

more MFCs are needed to increase the overall power available, the system can easily be adapted by adding more solid-state switches and digital multiplexers as needed.

Appl. Sci. 2018, 8, 2404

10 of 16

MFCs are needed to increase the overall power available, the system can easily be adapted by adding more solid-state switches and digital multiplexers as needed. Appl. Sci. 2018, 8, x FOR PEER REVIEW

10 of 16

4. Experimental Results and Discussion 4. Experimental Results and Discussion

4.1. PMS for Multiple MFCs 4.1. PMS for Multiple MFCs

The operation of the power ranking subroutine operation is illustrated in Figure 8, which displays operation of the The power ranking subroutine operation when is illustrated in Figure 8, which the CPM The output voltage. label Sample 1–4, indicate the respective MFC 1–4 is displays the 𝐶 output voltage. The label Sample 1–4, indicate when the respective MFC 1–4 is power-measured by sampling its output power using CPM such value is held to perform an power-measured by sampling its output power using 𝐶 such value is held to perform an analoganalog-to-digital conversion. This sampling process is repeated for each MFC connected to the to-digital conversion. This sampling process is repeated for each MFC connected to the board boardperiodically periodically its ranking values are grouped in phases. Theobtained rankingatobtained phase is andand its ranking values are grouped in phases. The ranking the phaseatisthe shown shown in Table 3, phase the MFCs are sorted, theextraction power extraction can be performed. in Table 3, phase one.one. AfterAfter the MFCs are sorted, then thethen power can be performed. Phase Phasetwo two Table 3, represents a second power-ranking measurement recalibration. in in Table 3, represents a second power-ranking measurement recalibration.

Figure rankingwaveform. waveform. Figure8.8.MFC MFC power power ranking Table 3. MFC power rank. Table 3. MFC power rank.

Device Device MFC MFC 1 1 MFC MFC 2 2 MFC 3 3 MFC MFC 4 4 MFC

Ranking(Phase (PhaseOne) One) Ranking Ranking (Phase Two) Ranking (Phase Two) 1st 1st 1st 1st 4th 2nd2nd 4th 2nd 2nd 3rd3rd 3rd 3rd 4th 4th

In Figure 9, the time interleaved power extraction operation of the proposed PMS is presented

In Figure 9, the time interleaved power extraction operation of the proposed PMS is presented for for different extraction phases. The four channels representing the MFC output voltages are labeled different extraction phases. The four channels representing the MFC output voltages are labeled at the top at the top of the graph. Initially, before power extraction, the voltage at the MFC represents its open of thecircuit graph. Initially, before power extraction, the voltage at the MFC represents its open circuit voltage; voltage; then, when the MFC is harvested its voltage drops to its MPP voltage. Finally, when then,𝑉when the MFC is the harvested itsharvests voltagethe drops its MPP Finally, when Vth−Lowvoltage is reached, is reached, algorithm nextto MFC in thevoltage. look-up table only if its output the algorithm harvests is larger than 𝑉 the. next MFC in the look-up table only if its output voltage is larger than Vth− High . Figure 9a shows thethe energy harvesting rightafter afterperforming performing MFC power ranking Figure 9a shows energy harvestingoperation operation right thethe MFC power ranking presented in Figure 8. Thus, the pattern for power the power extraction, pointed with gray arrows, presented in Figure 8. Thus, the pattern for the extraction, pointed outout with gray arrows, follows followsdescribed the pattern Tableone; 3 forhowever, phase one; however, afterpower extracting fromfor the~20 h, the pattern in described Table 3 forinphase after extracting frompower the MFCs MFCspower for ~20ranking h, the MFCs’ power ranking look-up tablevalues recalibrates values shown in Table the MFCs’ look-up table recalibrates to the showntointhe Table 3, phase two. Such3,power phase two. Such power changes are illustrated in Figure 9b. It is important to highlight that, in phase changes are illustrated in Figure 9b. It is important to highlight that, in phase two, not only are the power two, not only are the power rankings changed for devices MFC 2, MFC 3, and MFC 4, but at this point rankings changed for devices MFC 2, MFC 3, and MFC 4, but at this point MFC 1 now takes more time to MFC 1 now takes more time to reach 𝑉 . This condition forces the algorithm to switch to the reachnext Vth−available condition forces the algorithm to switch to the next available MFC that complies with High . This MFC that complies with the 𝑉 specifications, which in this case is MFC 2. The the Vth specifications, which in this case is MFC 2. The algorithm continues andas MFC 1 − High algorithm continues its operation, and MFC 1 is only harvested again when 𝑉 its operation, is reached is only harvested again when V is reached as originally proposed. th− High originally proposed.

Appl. Sci. 2018, 8, 2404

11 of 16

Appl. Sci. 2018, 8, x FOR PEER REVIEW

11 of 16

Appl. Sci. 2018, 8, x FOR PEER REVIEW

11 of 16

(a)

(b)

Figure 9. MFC interleaved power extractionmeasurements: measurements: (a) one and (b) (b) phase two.two. Figure 9. MFC interleaved power extraction (a)Phase Phase one and phase (a) (b) As demonstrated, important characteristic characteristic of ranking feature previously As demonstrated, an animportant of the the(a)power power feature Figure 9. MFC interleaved power extraction measurements: Phase oneranking and (b) phase two. previously demonstrated, is that if an MFC does not provide proper power conditions or has an internal failure, demonstrated, is that if an MFC does not provide proper power conditions or has an internal failure, the system automatically isolates the damaged MFC and continues its operation without affecting demonstrated, an important characteristic the power ranking feature previously the systemAsautomatically isolates the damaged MFC andofcontinues its operation without affecting the the remaining MFCs interrupting process. demonstrated, is that or if an MFC doesthe notenergy provideharvesting proper power conditions or has an internal failure,

remaining MFCs or interrupting the energy harvesting process.

the system automatically isolates the damaged MFC and continues its operation without affecting 4.2. IoT Sensor Node Application

theSensor remaining or interrupting the energy harvesting process. 4.2. IoT NodeMFCs Application

An IoT application was implemented to test the capabilities of the proposed PMS for a power

An IoT application was implemented to test the capabilities of the proposed PMSand for built a power 4.2. IoT Sensor Node Application demanding extraction conditions. A Wi-Fi-enabled temperature smart node was designed demanding extractionThis conditions. A Wi-Fi-enabled temperature smart node was®designed and built for this sensor node was based on the platformof Photon (Particle , San Francisco, An purpose. IoT application was implemented to test the IoT capabilities the proposed PMS a power ®for for this purpose. This sensor node was based on the IoT platform Photonconnected (Particleto , aSan Francisco, CA, USA)[50]. As seen in Figure 10a, the IoT circuit board was wirelessly router to demanding extraction conditions. A Wi-Fi-enabled temperature smart node was designed and built send the collected data over the internet, and the information gathered was retrieved using CA, USA) [50]. As seen in Figure 10a, the IoT circuit board was wirelessly connected to a router for this purpose. This sensor node was based on the IoT platform Photon (Particle®, San Francisco,a to ® device. Figure 10b shows that the 5 F output supercapacitor was charged from 0 handheld iPhone send the collected data overinthe internet, andIoT thecircuit information gathered was connected retrieved using a handheld CA, USA)[50]. As seen Figure 10a, the board was wirelessly to a router to ® V to 3.3 V with an charging time 11.3 at 3.3 V the internet-enabled smart node iPhone Figureinitial 10b shows the(𝑡5) Fof output supercapacitor was charged from 0 Vwas toa3.3 V senddevice. the collected data over thethat internet, and theh; information gathered was retrieved using enabled, and the® data was remote supercapacitor server. After the packet was device. Figure shows 5 Fa output wasdata charged from 0 with handheld an initialiPhone charging time (ttransmitted 11.3 h;wirelessly atthat 3.3the V tothe internet-enabled smart node was enabled, i ) of10b received in the smartphone, the sensor was disconnected and the output capacitor voltage presented V to 3.3 V with an initial charging time (𝑡 ) of 11.3 h; at 3.3 V the internet-enabled smart node was and the data was transmitted wirelessly to a remote server. After the data packet was received in the an outputand voltage of 2.8was V. It transmitted took an average recharging time (𝑡 ) server. of threeAfter hoursthe to fully recharge the enabled, the data wirelessly to acapacitor remote data packet was smartphone, the sensor was disconnected and the output voltage presented an output voltage supercapacitor from 2.8 V to 3.3 V. The process was repeated six times in the span of 24 h. The received in the smartphone, the sensor was disconnected and the output capacitor voltage presented of 2.8collected V. It took an average recharging timein(tFigure hours to fully recharge the supercapacitor r ) of three areanshown 10c. time an outputtemperature voltage of 2.8readings V. It took average recharging (𝑡 ) of three hours to fully recharge the from 2.8 V to 3.3 V. The process was repeated six times in the span of 24 h. The collected temperature supercapacitor from 2.8 V to 3.3 V. The process was repeated six times in the span of 24 h. The readings are shown in Figure 10c. collected temperature readings are shown in Figure 10c.

(a) (a) Figure 10. Cont.

Appl. Sci. 2018, 8, x FOR PEER REVIEW

12 of 16

Appl. Sci. 2018, 8, 2404

12 of 16

Appl. Sci. 2018, 8, x FOR PEER REVIEW

12 of 16

(b) (b)

(c) (c) Figure 10. Internet of things sensornode. node. (a) (a) Hardware Hardware (b)(b) supercapacitor voltage waveform Figure 10. of sensor setup; supercapacitor voltage waveform Figure 10. Internet Internet of things things sensor node. (a) Hardwaresetup; setup; (b) supercapacitor voltage waveform and (c) temperature measurements. and and (c) (c) temperature temperature measurements. measurements.

By effectively poweringthe the IoT application application with the MFC array, the system robustness was was By By effectively effectively powering powering the IoT IoT application with with the the MFC MFC array, array, the the system system robustness robustness was confirmed after providing continuous power extraction in the span of 24 h, in such period the confirmed after providing continuous power extraction in the span of 24 h, in such period theMFC MFC did confirmed continuous power extraction in in thethis span of the 24 h, in suchuse period the MFC did notafter showproviding any abnormal stress or failure, validating way potential of MFC not show any abnormal stress or failure, validating in this way the potential use of MFC technology did technology not show for anypower-hungry abnormal stress or such failure, validating in this way the potential use of MFC systems, as IoT-applications.

for power-hungry systems, such as IoT-applications. technology for power-hungry systems, such as IoT-applications. 4.3. Total Power Consumption and Efficiency

4.3. Total Power Consumption and Efficiency 4.3. TotalThe Power Consumption and Efficiency PMS’ power efficiency (𝜂), as defined in (4), is presented in Figure 11. The peak end-to-end The PMS’ power efficiency , as defined in (4), is presented Figure 11. Thecompares peak end-to-end (η )for efficiency measured was ~50.7% 3.3 output voltage. Table 4 in summarizes the The PMS’ power efficiency (𝜂), asadefined in (4), is presented in Figure 11.and The peak end-to-end efficiency measured ~50.7% for a 3.3 output voltage. Table 4 summarizes and compares the presented work to was previously reported systems. efficiency measured was ~50.7% for a 3.3 output voltage. Table 4 summarizes and compares the presented work to previously )+𝑃 − [𝐼 (𝑅 +𝑅 ] − ∑𝑃 systems.𝑃 𝑃 𝑃reported presented work𝜂to = previously × 100 = reported systems. × 100 ≈ × 100. (4) 𝑃

𝑃

𝑃

) + )𝑃+ PLossConverter PMFC Iout ((𝑅R MFC+ + ] 𝑃 −−[[𝐼 𝑅 Rswitch ] − ∑𝑃 𝑃 P Pout −𝑃∑ Plosses = = ×MFC 100 = × 100 × 100. η= ×𝜂100 × 100 ≈ ≈ × 100. (4) (4) 𝑃PMFC 𝑃 Pin PMFC 𝑃

Figure 11. Measured Efficiency versus Output Voltage.

Figure Measured Efficiency Figure 11. 11. Measured Efficiency versus versus Output Output Voltage. Voltage.

Appl. Sci. 2018, 8, 2404

13 of 16

Table 4. Comparison of MFC power management units. Specification

[16]

[23]

[26]

[24]

[32]

[35]

This Work

1.4 V (DC-DC converter input)

300 mV

330 mV

Input voltage

300 mV

300 mV

300–720 mV

300–600 mV

Output voltage

1V

1.8 V

2.5 V

2.5 V

4.25 V

3.3 V

3.3 V

-

22 µH

Inductor

326.7 µH

2 µH

1.5 mH

1.5 mH

Transformer 31.8 mH (primary winding)

Output capacitor

8 µF

47 µF

0.1 F (supercapacitor)

0.1 F

68 mF

Multiple

5F

Maximum power extraction

-

-

Adaptable Maximum Power Extraction

Yes

MPPT

-

MPPT

Efficiency

73%

-

58%

30%

-

35.02%

50.7%

Custom integrated circuit

Discrete Components

Discrete Components

Discrete Components

Implementation approach

Discrete Components

Discrete Components

Custom integrated circuit

Multiple MFC power extraction

No

No

No

No

Yes

Multi anode

Yes (default support for 9 MFCs)

MFC health protection

-

-

-

-

No

No

Yes

5. Conclusions This paper presents a system design that can extract power from an array of MFCs using a time-interleave technique that not only enables a custom MPPT per MFC in the arrange but also allows a proper MFC charge-recovery time to avoid overly depleting the cells in the array, which can lead to an MFC failure or premature end-of-life. An optimal power harvesting process was measured when enabling a maximum power point tracking in the DC-DC power conversion block across four different MFCs with an output characteristic power at MPPT ranging from 6400 µW to 435 µW with an observed 50.7% peak efficiency. The results of the proposed system show a robust performance of the proposed PMS after more than 24-h of MFC EH. The system was able to run autonomously to charge an output supercapacitor of 5 F from 0 V to 3.3 V in 13.6 h and endure a 24-h power extraction with no signs of MFC over-depletion or failure. The system powered an internet-enabled smart node five times, and the information was correctly retrieved from a remote server using a smartphone device. By successfully completing this MFC-powered IoT-sensor node demonstration, EH based on MFC technology has been verified as a promising area that has the capabilities to run power intense systems, such as internet-enabled technologies, which can potentially enable the arrival of a new generation of self-powered smart nodes. Finally, future work in this area includes the development of efficient PMS startup circuits tailored to MFC technology, more efficient low-power data acquisition and processing techniques for remote sensing, in which the addition of new sensors and analytical models in the cloud can give accurate real-time and predictive information about the health state of the MFC. Supplementary Materials: The following are available online at www.mdpi.com/2076-3417/8/12/2404/s1, File S1: source code and circuit schematics of this work, applsci-386823. Author Contributions: A.C.-R. is the main author and thus responsible for writing the paper. C.E. has contributed to Section 2.1 and by providing MFC devices, equipment, laboratory space and materials, S.C.-B. contributed to Section 3.3 and industry insight in the area of power management systems. A.H. and E.S.-S. have been responsible for revising and supervising the paper in the MFC side and the electronics circuits aspects, respectively. Conceptualization: A.C.-R. and S.C.-B.; Data curation, A.C.-R.; Funding acquisition, A.H. and E.S.-S.; Investigation, A.C.-R.; Project administration, E.S.-S.; Resources, A.H. and E.S.-S.; Software, A.C.-R.; Supervision, A.H. and E.S.-S.; Validation, C.E.; Writing—review and editing, A.C.-R., C.E. and S.C.-B. Funding: This research was partially funded by the Qatar National Research Foundation (QNRF), grant number #NPRP 7-1634-2-604

Appl. Sci. 2018, 8, 2404

14 of 16

Acknowledgments: This work was supported by Consejo Nacional de Ciencia y Tecnología (Mexico City, Mexico), Texas Instruments (Dallas, TX, USA), Silicon Labs (Austin, TX, USA) and Intel (Hillsboro, OR, USA). Conflicts of Interest: The authors declare no conflict of interest.

References 1. 2. 3.

4. 5.

6. 7. 8. 9. 10.

11.

12.

13.

14. 15.

16.

17.

18.

Chong, C.Y.; Kumar, S.P. Sensor networks: Evolution, opportunities, and challenges. Proc. IEEE 2003, 91, 1247–1256. (In English) [CrossRef] Gèungèor, V.ð.; Hancke, G.P. Industrial wireless Sensor Networks: Applications, Protocols, And Standards; (Industrial Electronics Series), PP. 1 Online Resource (xxxii, 374 Pages); CRC Press: Boca Raton, FL, USA, 2017. Tiwari, A.; Lewis, F.L.; Ge, S.S. Wireless sensor network for machine condition based maintenance. In Proceedings of the 2004 8th Control, Automation, Robotics and Vision Conference, Kunming, China, 6–9 December 2004; pp. 461–467. (In English) Swartz, R.A.; Lynch, J.P.; Zerbst, S.; Sweetman, B.; Rolfes, R. Structural monitoring of wind turbines using wireless sensor networks. Smart Struct Syst. 2010, 6, 183–196. (In English) [CrossRef] Konstantopoulos, C.; Koutroulis, E.; Mitianoudis, N.; Bletsas, A. Converting a Plant to a Battery and Wireless Sensor with Scatter Radio and Ultra-Low Cost. IEEE Trans. Instrum. Meas. 2016, 65, 388–398. (In English) [CrossRef] Toh, W.Y.; Tan, Y.K.; Koh, W.S.; Siek, L. Autonomous Wearable Sensor Nodes with Flexible Energy Harvesting. IEEE Sens. J. 2014, 14, 2299–2306. (In English) [CrossRef] Wang, Y.; Lynch, J.P.; Law, K.H. Design of a low-power wireless structural monitoring system for collaborative computational algorithms. Proc. Soc. Photo-Opt. Instrum. 2005, 5768, 106–117. (In English) Atzori, L.; Iera, A.; Morabito, G. The Internet of Things: A survey. Comput. Netw. 2018, 54, 2787–2805. (In English) [CrossRef] Seah, W.K.; Tan, Y.; Chan, A.T. Research in energy harvesting wireless sensor networks and the challenges ahead. In Autonomous Sensor Networks; Springer: Berlin, Germany, 2012; pp. 73–93. Wan, Z.G.; Tan, Y.K.; Yuen, C. Review on energy harvesting and energy management for sustainable wireless sensor networks. In Proceedings of the 2011 IEEE 13th International Conference on Communication Technology, Jinan, China, 25–28 September 2011; pp. 362–367. Raghunathan, V.; Kansal, A.; Hsu, J.; Friedman, J.; Srivastava, M. Design considerations for solar energy harvesting wireless embedded systems. In Proceedings of the Fourth International Symposium on Information Processing in Sensor Networks, Boise, ID, USA, 15 April 2005; pp. 457–462. (In English) Zarate-Roldan, J.; Carreon-Bautista, S.; Costilla-Reyes, A.; Sanchez-Sinencio, E. A Power Management Unit With 40 dB Switching-Noise-Suppression for a Thermal Harvesting Array. IEEE Trans. Circuits Syst. I Reg. Pap. 2015, 62, 1918–1928. (In English) [CrossRef] Calhoun, B.H.; Daly, D.C.; Verma, N.; Finchelstein, D.F.; Wentzloff, D.D.; Wang, A.; Cho, S.-H.; Chandrakasan, A.P. Design considerations for ultra-low energy wireless microsensor nodes. IEEE Trans Comput. 2005, 54, 727–740. (In English) [CrossRef] Vullers, R.J.M.; Schaijk, R.v.; Visser, H.J.; Penders, J.; Hoof, C.V. Energy Harvesting for Autonomous Wireless Sensor Networks. IEEE Solid-State Circuits Mag. 2010, 2, 29–38. [CrossRef] Moghe, R.; Yang, Y.; Lambert, F.; Divan, D. A Scoping Study of Electric and Magnetic Field Energy Harvesting for Wireless Sensor Networks in Power System Applications. In Proceedings of the 2009 IEEE Energy Conversion Congress and Exposition, San Jose, CA, USA, 20–24 September 2009; Volumes 1–6, pp. 3423–3430. (In English) Wang, H.M.; Park, J.D.; Ren, Z.Y. Active Energy Harvesting from Microbial Fuel Cells at the Maximum Power Point without Using Resistors. Environ. Sci. Technol. 2012, 46, 5247–5252. (In English) [CrossRef] [PubMed] Logan, B.E.; Hamelers, B.; Rozendal, R.; Schröder, U.; Keller, J.; Freguia, S.; Aelterman, P.; Verstraete, W.; Rabaey, K. Microbial fuel cells: Methodology and technology. Environ. Sci. Technol. 2006, 40, 5181–5192. (In English) [CrossRef] [PubMed] Lovley, D.R. The microbe electric: Conversion of organic matter to electricity. Curr. Opin. Biotech. 2008, 19, 564–571. (In English) [CrossRef] [PubMed]

Appl. Sci. 2018, 8, 2404

19. 20. 21. 22. 23. 24.

25. 26. 27. 28.

29.

30.

31. 32.

33. 34. 35. 36.

37.

38.

39. 40.

15 of 16

He, Z.; Minteer, S.D.; Angenent, L.T. Electricity generation from artificial wastewater using an upflow microbial fuel cell. Environ. Sci. Technol. 2005, 39, 5262–5267. (In English) [CrossRef] [PubMed] Logan, B.E. Exoelectrogenic bacteria that power microbial fuel cells. Nat. Rev. Micro. 2009, 7, 375–381. [CrossRef] [PubMed] Pant, D.; van Bogaert, G.; Diels, L.; Vanbroekhoven, K. A review of the substrates used in microbial fuel cells (MFCs) for sustainable energy production. Bioresour. Technol. 2010, 101, 1533–1543. [CrossRef] [PubMed] Ivars-Barcelo, F.; Zuliani, A.; Fallah, M.; Mashkour, M.; Rahimnejad, M.; Luque, R. Novel Applications of Microbial Fuel Cells in Sensors and Biosensors. Appl. Sci. 2018, 8, 1184. [CrossRef] Donovan, C.; Dewan, A.; Heo, D.; Beyenal, H. Batteryless, Wireless Sensor Powered by a Sediment Microbial Fuel Cell. Environ. Sci. Technol. 2008, 42, 8591–8596. (In English) [CrossRef] [PubMed] Tender, L.M.; Gray, S.A.; Groveman, E.; Lowy, D.A.; Kauffman, P.; Melhado, J.; Tyce, R.C.; Flynn, D.; Petrecca, R.; Dobarro, J. The first demonstration of a microbial fuel cell as a viable power supply: Powering a meteorological buoy. J. Power Sources 2008, 179, 571–575. (In English) [CrossRef] Bond, D.R.; Holmes, D.E.; Tender, L.M.; Lovley, D.R. Electrode-reducing microorganisms that harvest energy from marine sediments. Science 2002, 295, 483–485. (In English) [CrossRef] [PubMed] Park, J.D.; Ren, Z. Hysteresis-Controller-Based Energy Harvesting Scheme for Microbial Fuel Cells with Parallel Operation Capability. IEEE Trans. Energy Convers. 2012, 27, 715–724. [CrossRef] Meehan, A.; Gao, H.; Lewandowski, Z. Energy Harvesting with Microbial Fuel Cell and Power Management System. IEEE Trans. Power Electron. 2011, 26, 176–181. [CrossRef] Erbay, C.; Carreon-Bautista, S.; Sanchez-Sinencio, E.; Han, A. High Performance Monolithic Power Management System with Dynamic Maximum Power Point Tracking for Microbial Fuel Cells. Environ. Sci. Technol. 2014, 48, 13992–13999. (In English) [CrossRef] [PubMed] Carreon-Bautista, S.; Erbay, C.; Han, A.; Sanchez-Sinencio, E. Power Management System with Integrated Maximum Power Extraction Algorithm for Microbial Fuel Cells. IEEE Trans. Energy Convers. 2015, 30, 262–272. (In English) [CrossRef] Carreon-Bautista, S.; Erbay, C.; Han, A.; Sanchez-Sinencio, E. An Inductorless DC-DC Converter for an Energy Aware Power Management Unit Aimed at Microbial Fuel Cell Arrays. IEEE J. Emerg. Sel. Top. Power Electron. 2015, 3, 1109–1121. (In English) [CrossRef] Zanella, A.; Bui, N.; Castellani, A.; Vangelista, L.; Zorzi, M. Internet of Things for Smart Cities. IEEE Int. Things 2014, 1, 22–32. (In English) [CrossRef] Lee, I.; Kim, G.; Bang, S.; Wolfe, A.; Bell, R.; Jeong, S.; Kim, Y.; Kagan, J.; Arias-Thode, M.; Chadwick, B.; et al. System-On-Mud: Ultra-Low Power Oceanic Sensing Platform Powered by Small-Scale Benthic Microbial Fuel Cells. IEEE Trans. Circuits Syst. I Reg. Pap. 2015, 62, 1126–1135. (In English) [CrossRef] Khaled, F.; Ondel, O.; Allard, B. Optimal Energy Harvesting from Serially Connected Microbial Fuel Cells. IEEE Trans. Ind. Electron. 2015, 62, 3508–3515. (In English) [CrossRef] Shantaram, A.; Beyenal, H.; Raajan, R.; Veluchamy, A.; Lewandowski, Z. Wireless sensors powered by microbial fuel cells. Environ. Sci. Technol. 2005, 39, 5037–5042. (In English) [CrossRef] [PubMed] Dallago, E.; Barnabei, A.L.; Liberale, A.; Torelli, G.; Venchi, G. A 300-mV Low-Power Management System for Energy Harvesting Applications. IEEE Trans. Power Electron. 2016, 31, 2273–2281. (In English) [CrossRef] Plekhanova, Y.; Tarasov, S.; Kolesov, V.; Kuznetsova, I.; Signore, M.; Quaranta, F.; Reshetilov, A. Effects of Polymer Matrices and Carbon Nanotubes on the Generation of Electric Energy in a Microbial Fuel Cell. Membranes (Basel) 2018, 8, 99. [CrossRef] [PubMed] Yang, G.; Erbay, C.; Yi, S.; de Figueiredo, P.; Sadr, R.; Han, A.; Yu, C. Bifunctional nano-sponges serving as non-precious metal catalysts and self-standing cathodes for high performance fuel cell applications. Nano Energy 2016, 22, 607–614. [CrossRef] Erbay, C.; Pu, X.; Choi, W.; Choi, M.; Ryu, Y.; Hou, H.; Lin, F.; de Figueiredo, P.; Yu, C.; Han, A. Control of geometrical properties of carbon nanotube electrodes towards high-performance microbial fuel cells. J. Power Sources 2015, 280, 347–354. [CrossRef] Erbay, C.; Yang, G.; de Figueiredo, P.; Sadr, R.; Yu, C.; Han, A. Three-dimensional porous carbon nanotube sponges for high-performance anodes of microbial fuel cells. J. Power Sources 2015, 298, 177–183. [CrossRef] You, J.; Greenman, J.; Ieropoulos, I. Novel Analytical Microbial Fuel Cell Design for Rapid in Situ Optimisation of Dilution Rate and Substrate Supply Rate, by Flow, Volume Control and Anode Placement. Energies 2018, 11, 2377. [CrossRef]

Appl. Sci. 2018, 8, 2404

41. 42. 43. 44.

45. 46. 47. 48.

49. 50.

16 of 16

Lin, H.; Wu, S.; Zhu, J. Modeling Power Generation and Energy Efficiencies in Air-Cathode Microbial Fuel Cells Based on Freter Equations. Appl. Sci. 2018, 8, 1983. [CrossRef] Zhang, X.; Ren, H.; Pyo, S.; Lee, J.I.; Kim, J.; Chae, J. A High-Efficiency DC-DC Boost Converter for a Miniaturized Microbial Fuel Cell. IEEE Trans. Power Electron. 2015, 30, 2041–2049. [CrossRef] Ren, Z.Y.; Ward, T.E.; Regan, J.M. Electricity production from cellulose in a microbial fuel cell using a defined binary culture. Environ. Sci. Technol. 2007, 41, 4781–4786. (In English) [CrossRef] [PubMed] Aelterman, P.; Rabaey, K.; Pham, H.T.; Boon, N.; Verstraete, W. Continuous electricity generation at high voltages and currents using stacked microbial fuel cells. Environ. Sci. Technol. 2006, 40, 3388–3394. (In English) [CrossRef] [PubMed] Dewan, A.; Beyenal, H.; Lewandowski, Z. Scaling up Microbial Fuel Cells. Environ. Sci. Technol. 2008, 42, 7643–7648. (In English) [CrossRef] [PubMed] Microchip Technology Inc. PIC24F16KA102 Datasheet. Available online: www.microchip.com (accessed on 9 October 2011). Texas Instruments. BQ25505 datasheet. Available online: www.ti.com (accessed on 28 August 2013). Zhang, Y.; Zhang, F.; Shakhsheer, Y.; Silver, J.D.; Klinefelter, A.; Nagaraju, M.; Boley, J.; Pandey, J.; Shrivastava, A.; Carlson, E.J.; et al. A Batteryless 19µW MICS/ISM-Band Energy Harvesting Body Sensor Node SOC for EXG Applications. IEEE J. Solid-State Circuits 2013, 48, 199–213. [CrossRef] Ramadass, Y.K.; Chandrakasan, A.P. A battery-less thermoelectric energy harvesting interface circuit with 35 mV startup voltage. IEEE J. Solid-State Circuits 2011, 46, 333–341. [CrossRef] Photon Particle. Available online: https://www.particle.io/ (accessed on 11 November 2017). © 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).