Power generation in dual chamber microbial fuel cells using dynamic membranes as separators

Energy Conversion and Management 165 (2018) 488–494

Contents lists available at ScienceDirect

Energy Conversion and Management journal homepage: www.elsevier.com/locate/enconman

Power generation in dual chamber microbial fuel cells using dynamic membranes as separators

T

⁎

Xinyang Lia,b, Guicheng Liuc, , Shaobin Suna,b, Fujun Maa,b, Siyu Zhoua,b, Joong Kee Leec, Hong Yaoa,b,1 a

School of Civil Engineering, Beijing Jiaotong University, 3 Shangyuancun, Beijing 100044, PR China Beijing Key Laboratory of Aqueous Typical Pollutants Control and Water Quality Safeguard, Beijing 100044, PR China c Center for Energy Convergence Research, Green City Research Institute, Korea Institute of Science and Technology (KIST), Hwarang-ro 14-gil 5, Seongbuk-gu, Seoul 02792, Republic of Korea b

A R T I C LE I N FO

A B S T R A C T

Keywords: Microbial fuel cells Dynamic membranes Separator Neylon mesh Low cost Scaling-up

Two dual-chamber microbial fuel cells (MFCs) that use dynamic membranes as separators were designed for power production. The performance of these dynamic membrane microbial fuel cells (DM-MFCs) was studied. Compared to an up-flow dual-chamber MFC (U-MFC), at the total volume of 1.1 L, DM-MFCs achieved a higher maximum power density (1923 mW m−3 versus 856 mW m−3). This is because the DM-MFCs have lower membrane resistance (0.6–5.4 Ω), oxygen diffusion coefficient (D0 = 1.8 × 10−7 cm2 s−1), and cost (0.3 USD m−2) than other reported separators; e.g., anion exchange membrane (ACM), cation exchange membrane (CEM), ultrafiltration membrane (UFM), and J-cloth. The dynamic membrane is primarily composed of filamentous bacteria and vorticellidae-like protozoa, which tightly attach to the nylon supporting layer. This microorganism layer consumes most of the dissolved oxygen and prevents oxygen transfer from the cathode chamber to the anode chamber, leading to the low D0 value of the dynamic membrane. Power production of DMMFCs was further optimized by increasing the NaCl concentration in the influent and the electrode area. The results show that DM-MFCs are feasible and suitable for scaling-up because of their sleeve-shaped configuration. These results indicate that dynamic membranes can be used to increase power production in MFCs relative to traditional separators and DM-MFCs are promising tools for practical applications.

1. Introduction

functions of separators is to prevent the transfer of dissolved oxygen from the cathode chamber to the anode chamber so that the cathode chamber can maintain an aerobic environment while the anode chamber maintains an anaerobic environment [8]. The use of a separator is associated with three problems: (1) the separator increases the internal resistance of microbial fuel cells [9]; (2) it is difficult to select an ideal separator with low resistance and low oxygen diffusion coefficient; (3) separators increase the cost of an MFC, because the cost of a separator is critical for the development and scaling up of MFCs [10]. These challenges must be overcome for further development of the practical applications of MFCs. In the past decade, various types of materials have been used as separators, including cation exchange membranes (CEMs) [11], anion exchange membranes (AEMs) [12], ultrafiltration membranes (UFMs) [9,13], and other coarse-pore filters materials; e.g., J-cloth [14] and glass fibers [15]. Although CEM and AEM have been widely used in MFCs because of their high cation or anion conductivity, they also show

The fuel cell, a kind of electrochemical device that converts the chemical energy from a fuel into electricity [1], has been considered as a promising green energy conversion technology [2]. Among these, microbial fuel cells (MFCs) are emerging biological tools that can sustainably generate electricity by oxidizing organic and inorganic matter in wastewater and other substrates [3]. By generating power from a renewable substrate that simultaneously treats wastewater, MFCs have been the subject of intense interest over the past decades and shown to have the potential to alleviate future energy crisis [4]. Although research on microbial fuel cells has shown significant development in recent years, there are still many limitations in their practical applications, such as their unstable system performance and high cost [5]. One of the most important constraints is the separators of microbial fuel cells [6]. In an MFC, a separator is used to physically separate anode and cathode chambers [7]. One of the important ⁎

1

Corresponding author. E-mail addresses: [email protected], [email protected] (G. Liu), [email protected] (H. Yao). Co-corresponding author.

https://doi.org/10.1016/j.enconman.2018.03.074 Received 28 February 2018; Received in revised form 19 March 2018; Accepted 24 March 2018 0196-8904/ © 2018 Elsevier Ltd. All rights reserved.

Energy Conversion and Management 165 (2018) 488–494

X. Li et al.

MFC(S), was used here. The detailed information of U-MFC is listed in supplementary material. The parameters of the U-MFC are listed in Table 1 and in the supporting information. To optimize the power production of DM-MFCs, NaCl salts were added in the influent at various NaCl concentrations (0.04 mol L−1, 0.08 mol L−1, 0.12 mol L−1) to increase electrolyte conductivity. Power production of DM-MFCs was further optimized by increasing the anode area followed by the increase in the cathode area (electrode area information is listed in Table S1).

relatively high Ohmic resistance and biofilm growth during long-time operations that would deteriorate the MFC performance. Compared to these ion exchange membranes (IEM), UFMs show a lower resistance, higher ability of proton transfer, and better MFC applicability owing to their non-ion selective properties and large pore sizes to facilitate overall charge transfer. Other types of separators, i.e., coarse-pore filters, such as J-cloth and glass fibers, have a lower cost than IEM and UFM and high potential for practical applications in MFCs [6]. These coarse-pore separators are favorable for ion transport and have low Ohmic resistance. Researchers found that the biofilm growth on a Jcloth coarse-pore filter did not reduce the performance of the filter but prevented oxygen diffusion from the cathode chamber to the anode chamber [15]. However, biodegradable characteristics of the cloth limited its practical application. Inspired by the biofilm growth on coarse-pore filters, we introduce the concept of dynamic membrane bioreactors in MFCs [16]. Dynamic membranes are also called secondary membranes and consist of cheap supporting material, e.g., a membrane, mesh, or a filter cloth [17]. The dynamic layer contains suspended solid particles, e.g., microbial cells and flocs. Organics and colloidal particles, which normally result in the fouling of the membrane, are entrapped in the biomass filtration layer, preventing the fouling of the support material. Dynamic membranes also consume dissolved oxygen and may prevent oxygen diffusion. In addition, because the dynamic layer is in situ formed on the supporting layer and can self-upgrade via the metabolism of microorganisms, it can be self-cleaned and is easy to operate [18]. In this study, a novel dynamic membrane MFC with the up-flow mode was designed. A cheap nylon mesh, which is nonbiodegradable, was used as the supporting layer of the dynamic membrane, while the sleeve-shape configuration was used to maintain a low electrode distance and a large separator area when the MFC was scaled up. The performance of the DM-MFC was compared to our previously studied up-flow MFC (U-MFC) [19], which does not have a sleeve-shape configuration or a dynamic membrane. Two DM-MFCs with different sizes were constructed to test the scaling-up property of the DM-MFCs, after the optimization of power production by increasing the NaCl concentration in the influent and the electrode area. Finally, the scalability of the DM-MFC was analyzed.

2.2. Inoculation and operation The anode of DM-MFCs and U-MFC were inoculated using granular anaerobic wastewater sludge from the Gaobeidian Domestic Wastewater Treatment Plant in Beijing, China. After being washed three times with distilled water, 100 mL of the sludge was added to 500 mL of a nutrient solution containing glucose (200 mg L−1 COD), NH4Cl, and NaH2PO4 such that the COD:N:P ratio was 100:5:1 by weight. The sludge was cultivated for one week at 35 °C and the resulting suspension was used to inoculate the anode material. The DM-MFCs and U-MFC were fed a synthetic medium containing 500 mg COD L−1 of glucose, 1000 mg L−1 of NaHCO3, 95.5 mg L−1 of NH4Cl, 19.25 mg L−1 of NaH2PO4, and 0.08 mol L−1 NaCl. Various NaCl concentrations (0.04 mol L−1, 0.08 mol L−1, and 0.12 mol L−1) were used to investigate DM-MFC optimization. If not otherwise specified, the NaCl concentration is 0.08 mol L−1. No trace metals or buffers were added. The substrate was pumped into the bottom of the anode chamber using a peristaltic pump (Lange Co., Hebei Province, China) (Fig. 1). The flow rate was adjusted to keep a constant hydraulic retention time of 1 d. The cathode chamber was continuously aerated using a circular aeration pipe placed at the bottom of the cathode chamber, producing a dissolved oxygen concentration of 6–7 mg L−1. All experiments were conducted in duplicate at room temperature. 2.3. Characterizations and calculations Cell voltage (V) across a resistor (R) was determined in relation to Ag/AgCl reference electrodes (Shanghai, China) and measured using a data acquisition system (AD8223h, Beijing, China). Data were recorded on a personal computer every 10 min. Polarization curves were constructed by varying the resistance in the circuit from 10 Ω to 100,000 Ω. For each change in external resistance, data were recorded at least 2 h after a stable voltage was achieved. Internal resistance was calculated by applying the polarization slope method to the linear portion of the current (I) versus voltage (V) plot [20]. Power was calculated using the equation, P = I × V, normalized to the reactor net volume for comparison with published power density values [21]. The total resistance Rt, anode resistance Ra, and the cathode resistance Rc of MFCs were calculated by the polarization slope method, using the linear portion of the plots of current (I) versus voltage (V), anode potential, and cathode potential, respectively. Membrane resistance Rm is calculated by subtracting Ra and Rc from Rt. Surface morphologies of dynamic membranes were observed using scanning electron microscope (SEM) (FEIQUANTA 200, Netherlands). Before scanning, dynamic membrane samples were pretreated as follows [22]: samples were immersed in 2.5% pentanediol solution for 4 h and cleaned three times with deionized water. They were then immersed in 1% osmic acid solution for 2–4 h and cleaned three times with phosphate buffer. Next, the samples were dehydrated using 30%, 50%, 70%, 85%, and 95% ethanol solution, successively. They were treated twice with isoamyl acetate and dried using a CO2 critical point drying machine (CPD030 Critical Point Dryer). Finally, after metal spraying treatment, the samples were scanned [23,24]. The oxygen transfer coefficient (k0) of dynamic membrane separators can be obtained using Eq. (1) [9], where Va is the liquid volume in the anode chamber, A is the membrane cross-sectional area, C1,0 is the saturated dissolved oxygen (DO) concentration in the cathode chamber,

2. Experimental 2.1. DM-MFCs set-up A DM-MFC reactor was constructed in a cylindrical Plexiglas tube. In the center of the reactor, a hollowed smaller tube acted as an anode chamber, while the outside part of the anode chamber acted as a cathode chamber (Fig. 1). The anode and cathode compartments were separated by a nylon mesh (0.03 cm pore diameter, 0.0225 cm thick, Huawei Co., Beijing, China), used as a dynamic membrane supporting layer (Fig. 1). A tubular shape carbon felt (5 cm thick, 60.5 m2 g−1 specific surface area, Sanye Co. Beijing, China) and an activated carbon fiber felts (ACFF) (5 cm thick, 1000 m2 g−1 specific surface area, Senxin Co., Liaoning Province, China) were used as anode and cathode, respectively. These tubular shape electrodes were connected to the circuit by graphite rods and then placed vertically in the anode and cathode chambers (Fig. 1). Ag/AgCl reference electrodes (Luosu Co., Shanghai, China) were inserted into each chamber to test electrode potentials. The influent was pumped from the bottom of the anode chamber with an up-flow mode, passed through the dynamic membrane separator into the cathode chamber, and then flowed out from the reactor. To test the scalability of DM-MFC, a larger DM-MFC (denoted as DM-MFC(L)) with the same configuration was also constructed (the smaller DM-MFC is denoted hereafter as DM-MFC(S)). The basic parameters of these two DM-MFCs are listed in Table 1. To compare the power production of DM-MFC with other dual-chamber MFCs, our previously studied upflow MFC (U-MFC), whose empty volume is similar to that of DM489

Energy Conversion and Management 165 (2018) 488–494

X. Li et al.

Fig. 1. Schematic construction of the dynamic membrane microbial fuel cell (DM-MFC).

(135.7 Ω) resulted from its large Rm of 84.8 Ω, almost 62.5% of Rt. The Rm of U-MFC is 14.7 times greater than that of DM-MFC(S) (5.4 Ω). This great decrement in Rm is caused by employing the dynamic membrane as the separator. In DM-MFC(S) (240 cm2), a 0.02-cm thick dynamic membrane was used but in U-MFC, a PVC tube (12 cm2 cross-sectional area, 1.5 cm height) filled with granular porous sponges (0.5-cm diameter) was used as the separator. This sponge-made separator is incomparable with the dynamic membrane because of its low projected area and thickness, leading to a high Rm and Rt and low power production. Therefore, using dynamic membranes as separators is an effective way to increase power density.

and C2 is the DO in the anode chamber at time t. The diffusion coefficient (D0) of DO can be calculated using the thickness of the separator (LD) in Eq. (2).

k0 = −

Va ⎡ C1,0−C2 ⎤ In ⎥ A·t ⎢ ⎣ C1,0 ⎦

(1) (2)

D0 = k 0·LD 3. Results and discussion 3.1. Performance comparison between U-MFC and DM-MFCs

3.2. Oxygen transfer properties of DM-MFCs Electricity production from DM-MFCs is higher than from U-MFC (Fig. 2a). The maximum power densities of DM-MFC(S) and DMMFC(L) are 1923 mW m−3 and 1202 mW m−3, which are 2.2 times and 1.4 times that of U-MFC (856 mW m−3). Fig. 2a also shows that DMMFCs have a higher cell voltage than the U-MFC. For example, at the current density of 3.5 A m−3, the cell voltages of DM-MFC (S) and DMMFC (L) are approximately 310 mV, 150 mV and higher than the value of U-MFC. Interestingly, although having the same empty total volume (1.1 L), DM-MFC(S) performs much better than U-MFC because of its sleeve-shaped configuration, which allows for a low anode-cathode distance (1–3 cm). In sleeve-shaped reactors, the anode and the cathode are parallel to each other with a constant electrode distance. By contrast, in U-MFC, the anode chamber is placed under the cathode chamber (Fig. S1). The electrode distance varies from the top of the cathode to the bottom of the anode. The longest distance can be 33 cm and the shortest distance can be 6 cm. The resistance data (Fig. 2b) confirmed that DM-MFCs have a much lower Rt (35.0–52.2 Ω) than U-MFC (135.7 Ω). The large Rt of U-MFC

The characteristics of dynamic membranes were further investigated considering membrane resistance, oxygen transfer coefficient k0, diffusion coefficient D0, and cost. Dynamic membranes had a lower membrane resistance (5.4 Ω for DM-MFC(S) and 0.6 Ω for DM-MFC(L)) than the sponge separator in U-MFC (85 Ω), even less than various published separators such as CEM (42–78 Ω), AEM (about 9 Ω), UFM (3–4227 Ω), and other separators (2.2–23 Ω) [9], i.e., glass fiber, J-cloth (Table 2). This is because we used nylon mesh with a large pore size (about 0.03 cm) as the supporting material of the dynamic membrane. The large pore size enhances the charge transfer and the insulation characteristics of the nylon mesh can prevent a short circuit. Researchers also found that coarse pore materials have potential practical applications for MFCs [6,25,26]. However, the large pore size may also lead to a high oxygen transfer coefficient. Dynamic membranes showed a low oxygen transfer value in the present study, compared to the CEM, AEM, and UFM, whose pore sizes were less than 1 µm. For example, the dynamic membrane D0 of DM-MFC(S) (1.8 × 10−7cm2 s−1) was only 16.4% of that of the UFM-

Table 1 Basic parameters of DM-MFCs and U-MFC.

DM-MFC (S) DM-MFC (L) U-MFC

Inner diameter of anode chamber (cm)

Inner diameter of cathode chamber (cm)

Height of anode chamber (cm)

Height of cathode chamber (cm)

Electrode distance (cm)

Anode size (cm2) Cathode size (cm2)

Total volume (mL)

6.5 9.0 4.5

11.0 16.0 9.0

13.5 19.0 25.0

11.5 17.0 8.0

1.5 3 6

22.5 × 12.5 26.0 × 18.0 26.5 × 16.5

1100 3300 1100

490

33.0 × 11.0 41.0 × 16.0 52.5 × 8.5

Energy Conversion and Management 165 (2018) 488–494

X. Li et al.

Voltage (mV)

1800 1600 1400

700

1200

600

1000

500

800

400

600

300

400 200

200 100

-3

800

evenly distributed on the biofilm, promote the flocculation effect, make the biofilm more tightly attach on the nylon mesh, eat free bacteria in the penetrated substrate or filamentous bacteria on the dynamic membrane, update bacteria dynamically, and purify water. The biodiversity of the dynamic membrane resulted in a lower D0 than the D0 of other separators (Table 2). However, Fig. 3d shows that in the DM-MFC(L), some parts of the nylon mesh are not fully coated with bacteria, causing high flux of oxygen from the cathode chamber to the anode chamber, leading to a higher D0 (6.3 × 10−6 cm2 s−1) than the D0 of the membrane in the DMMFC(S) (1.8 × 10−7 cm2 s−1). This large D0 is comparable to the value of the other coarse-pore filters, e.g., J-cloth and glass fiber. Apart from the low membrane resistance and low D0, low cost is another advantage of dynamic membranes. Low-cost materials such as nylon mesh and non-woven fabrics are typically used as support layers of dynamic membranes. Nylon mesh is readily available and cheap in China; only 0.3 USD m−2, which is almost two to three orders of magnitude lower than the cost of CEM (Nafion) (1400 USD m−2), UFM (350 USD m−2), and J-cloth (400 USD m−2) [6]. Therefore, dynamic membranes have a great potential for practical applications of MFCs owing to their low cost, availability, low membrane resistance, and low D0.

2000

(a)

DM-MFC(S) DM-MFC(L) U-MFC

900

Power density (mW m )

1000

0 0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

-3

Current density (A m ) 160

(b)

Resistance (Ohm)

140 120 100

Rt Rc

80 60

Ra Rm

3.3. Optimization of DM-MFCs performance

40

To further improve power production by DM-MFCs, the effect of NaCl concentration in the substrate on the performance of DM-MFCs was investigated. Increasing the NaCl concentration from 0.04 mol L−1 to 0.08 mol L−1 drastically improved the power production by both DM-MFCs (Fig. 4a and b). In DM-MFC(S), the maximum power density doubled from 957 mW m−3 to 1923 mW m−3, while in DM-MFC(L), the value increased by 51% from 795 mW m−3 to 1202 mW m−3. However, a further increase in NaCl concentration to 0.12 mol L−1 did not improve power production, instead, caused a decrease in power production by both DM-MFCs. The same trend was also observed in the resistance value of the DM-MFCs (Fig. 4c and d). At NaCl concentration of 0.08 mol L−1, the optimal power density and the lowest Rt were obtained for both DM-MFCs (Fig. 4c and d). NaCl concentrations lower than 0.8 mol L−1 may significantly increase electrolyte conductivity, promote charge transfer, and reduce the resistance of the MFCs, resulting in an increase in power production. However, a further increase in NaCl concentrations to 0.12 mol L−1 may be beyond the salt tolerance level of the bacteria on the anode, cathode, and dynamic membranes, inhibiting bacterial growth and leading to a reduction in power generation [28,29]. Therefore, 0.08 mol L−1 was selected as the optimal NaCl concentration. Subsequently, power production of DM-MFCs was further improved by increasing the area of the carbon felt anode and ACFF cathode (Tables S1 and S2). When the cathode area was first increased without changing the anode area, the maximum power density of DM-MFC(S)

20 0

UP-MFC

DM-MFC(S)

DM-MFC(L)

Fig. 2. Polarization (dashed line), power density (solid line) curves (a) and resistance distribution (b) of DM-MFC (S), DM-MFC (L) and U-MFC. NaCl concentration of 0.08 mol L−1.

3 K membrane (1.1 × 10−6 cm2 s−1), and 4.2% of that of AEM and CEM (CMI-7000) membranes (4.3 × 10−6 cm2 s−1) [6,9]. D0 was also much lower than 8.69 × 10−5 cm2 s−1 and 5.0 × 10−6 cm2 s−1 for other coarse pore filters, i.e., J-cloth and glass fiber, respectively [6]. The success of the dynamic membranes relies largely on the biofilm growth on nylon mesh surfaces. Under steady conditions of DM-MFC, a biofilm growth was observed on a nylon mesh (Fig. 3a) after three months of cultivation. This biofilm is mainly composed of filamentous bacteria (Fig. 3b), tightly attached on the nylon support layer, and thus can consume dissolved oxygen and prevent oxygen transfer from the cathode chamber to the anode chamber, leading to a low D0 of the dynamic membrane. The biofilm also contained some vorticellidae-like protozoa (Fig. 3c), which often exist in biofilter reactors or aeration tanks in wastewater treatment plants [27]. The presence of vorticellidae maintains the ecosystem balance of dynamic membranes. The vorticellidae,

Table 2 Summary of oxygen transfer parameters of different separator materials. Separator

Thickness (cm)

Area (cm2)

k0 (×10−4 cm/s)

D0 (×10−6 cm2/s)

Resistance Rm (Ω)

Cost (USD/m2)

Reference

AEM CEM (Nafion) CEM (CMI-7000) UFM-0.5K UFM-1K UFM-3K J-Cloth Glass fiber Sponge in U-MFC Dynamic membrane in (DM-MFC (S)) Dynamic membrane in (DM-MFC (L))

0.046 0.019 0.046 0.0265 0.0265 0.0265 0.03 0.1 1.5 0.0225 0.0225

3.5 3.5 3.5 3.5 3.5 3.5 7 7 12 240 480

0.94 1.3 0.94 0.19 0.41 0.42 29 0.5

4.3 2.4 4.3 0.51 1.1 1.1 86.9 5.0

80 1400 200 350 350 350 400

a

a

0.08 2.80

0.18 6.3

9 42 78 4779 9 3 23 2.2 85 5.4 0.6

[6,9] [6,9] [6,9] [6,9] [6,9] [6,9] [6,14] [6,15] This study This study This study

a b

Not measured. Value not reported by the authors and could not be calculated from other data. 491

b a

0.3 0.3

Energy Conversion and Management 165 (2018) 488–494

X. Li et al.

Fig. 3. Surface morphologies of the nylon supporting layer (a), the dynamic membrane in DM-MFC (S) (b, c) and the dynamic membrane in DM-MFC (L) (d).

(a)

DM-MFC(S)

1800

0.04 mol L NaCl

1600

0.08 mol L NaCl

1400

0.12 mol L NaCl

1200

-1

-3

-1

1200 1000 800 600 400 200

1000 800 600 400 -1

200

0.04 mol L NaCl -1

0.08 mol L NaCl 0

0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5

-1

0.12 mol L NaCl

-0.5 0.0

0.5

-3

DM-MFC(S)

80

(c)

Rt

Ra

70

60

Rc

Rm

60

Resistance (Ohm)

Resistance (Ohm)

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

Current density (A m )

70

50 40 30 20

DM-MFC(L)

50

(d)

Rt

Ra

Rc

Rm

40 30 20 10

10 0

1.0

-3

Current density (A m ) 80

(b)

DM-MFC(L)

-1

Power density (mW m )

-3

Power density (mW m )

2000

0.04

0

0.12

0.08

0.04

0.12

0.08 -1

-1

NaCl concentration (mol L )

NaCl concentration (mol L )

Fig. 4. Power density curves of DM-MFC (S) (a) and DM-MFC (L) (b), resistance distribution of DM-MFC (S) (c) and DM-MFC (L) (d) at various NaCl concentration (0.04 mol L−1, 0.08 mol L−1, 0.12 mol L−1).

492

Energy Conversion and Management 165 (2018) 488–494

X. Li et al. 2400

(a)

DM-MFC(S)

2000

-3

2500 2000 1500 1000

No change Increase cathode area Increase anode area

500

1600 1200 800

No change Increase cathode area Increase anode area

400 0

0 0

1

2

3

4

5

6

7

8

9

10

11

0

-3

(c)

40

Rt

Ra

Rc

Rm

30 20

3

4

5

DM-MFC(L)

(d)

40 30

Rt

Ra

Rc

Rm

20 10

10 0

2

Current density (A m )

50 Resistance (Ohm)

Resistance (Ohm)

60

DM-MFC(S)

50

1

-3

Current density (A m ) 60

(b)

DM-MFC(L)

3000

Power density (mW m )

-3

Power density (mW m )

3500

No change

Increase cathode Increase anode area area

0

No change

Increase cathode Increase anode area area

Fig. 5. Power density curves of DM-MFC (S) (a) and DM-MFC (L) (b), resistance distribution of DM-MFC (S) (c) and DM-MFC (L) (d) under different conditions: electrodes area didn’t change; anode area is increased; on this basis, cathode area is increased. NaCl concentration of 0.08 mol L−1.

largely increased from 1923 mW m−3 to 2724 mW m−3 and the value of DM-MFC(L) was largely improved from 1202 mW m−3 to 1438 mW m−3 (Fig. 5a and b). Increasing the anode area significantly improved the power density from 2724 mW m−3 to 3377 mW m−3 for DM-MFC(S) and from 1438 to 2123 mW m−3 for DM-MFC(L) (Fig. 5a and b) because the increment of the cathode and anode area provides more space for charge transfer or anaerobic anode bacterial growth, promoting substrate degradation and power production. The resistance data also confirmed that the increment of the anode and cathode area resulted in a significant decrement of Rt for both DM-MFCs (Fig. 5c and d).

even if V was not changed, power production by the DM-MFC largely improved owing to a much higher S0/Vt of 21.8 and lower d/Vt of 13.6 (Table 3). The threefold increase in V significantly decreased d/Vt from 13.6 (m m−3) to 9.1 (m m−3), while d slightly increased from 1.5 cm to 3.0 cm, which is much lower than the d value of the U-MFC (6 cm). Likewise, for DM-MFC, the increase in Vt caused a twofold increase in S0, indicating that the unique sleeve-shaped design of DM-MFC is feasible and suitable for enlargement because, with the expansion of the reactor, the sleeve-shaped design can provide sufficient separator area and maintain a low electrode distance enabling charge transfer from the anode chamber to the cathode chamber. Other studies regarding MFC scale-up have also demonstrated that tubular reactors with sandwiched membrane/electrode structures are more suitable for scaling-up than flat-plate reactors [30], because the electrode distance, d is basically unchanged and the d/Vt remains in a low value when the tubular MFC is scaled up. However, the dual-chamber DM-MFCs in present study should also be further optimized by reducing air pump aeration to achieve simplified configurations and improved power production. Thus, further studies to examine the performance of DM-MFCs in improved MFC designs, such as using air-breathing cathode MFCs, are in progress.

3.4. Scalability properties of DM-MFCs The scalability of MFC is necessary for its practical applications. Although the total volume of DM-MFC increased three times, the power density of MFC did not decrease significantly (Table 3). The maximum power density (2123 mW m−3) of DM-MFC(L) is still 62.8% of that of DM-MFC(S) (3377 mW m−3). Generally, the parameters S0/Vt and d/Vt are important indicators used to test the scalability of MFC, where S0 is the separator cross-section area, d is the distance between the anode and cathode, and Vt is the total empty volume of MFC. When U-MFC configuration was changed to the sleeve-shaped design of DM-MFCs,

4. Conclusions The dynamic membrane microbial fuel cells (DM-MFCs) were constructed for energy generation. Using nylon mesh as a supporting layer, microorganisms could form stable dynamic membranes on this layer. Dynamic membranes exhibited low oxygen transfer properties and low membrane resistance. They are readily available at lower costs than the other separators, e.g., AEM, CEM, UFM, and J-cloth. Owing to these unique features, the optimal power density of DM-MFC(S) (3377 mW m−3) was much higher than that of the previously studied UMFC without dynamic membranes (856 mW m−3). The scalability

Table 3 Parameters of scalability for DM-MFCs and U-MFC.

DM-MFC (S) DM-MFC (L) U-MFC

S0 (cm2) Vt (mL) S0/Vt (m2 m−3)

d (cm) d/Vt (m m−3)

Maximum power density (mW m−3)

240 480 12

1.5 3 6

3377 2123 856

1100 3300 1100

21.8 14.5 1.1

13.6 9.1 54.5

493

Energy Conversion and Management 165 (2018) 488–494

X. Li et al.

properties also indicate that DM-MFCs are feasible and suitable for scaling-up because of their sleeve-shaped configurations. Therefore, by combining dynamic membrane reactor with MFC, the developed DMMFCs make full use of the advantages of both sides and achieve simultaneously wastewater treatment and power production. DM-MFCs thus shows a great potential for MFC practical applications.

[12]

[13] [14]

Acknowledgments [15]

Thanks for the support from the Fundamental Research Funds for the Central Universities, China (2017JBM342 and 2016JBZ008).

[16]

Appendix A. Supplementary material [17]

Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.enconman.2018.03.074.

[18]

References

[19]

[1] Liu G, Zhou H, Ding X, Li X, Zou D, Li X, et al. Effect of fabrication and operating parameters on electrochemical property of anode and cathode for direct methanol fuel cells. Energy Convers Manage 2016;123:366–71. [2] Ye F, Xu C, Liu G, Li J, Wang X, Du X, et al. A novel PtRuIr nanoclusters synthesized by selectively electrodepositing Ir on PtRu as highly active bifunctional electrocatalysts for oxygen evolution and reduction. Energy Convers Manage 2018;155:182–7. [3] Huang D, Song BY, He YL, Ren Q, Yao S. Cations diffusion in Nafion117 membrane of microbial fuel cells. Electrochim Acta 2017;245:654–63. [4] Angelov A, Bratkova S, Loukanov A. Microbial fuel cell based on electroactive sulfate-reducing biofilm. Energy Convers Manage 2013;67:283–6. [5] Liang P, Zhang C, Jiang Y, Bian Y, Zhang H, Sun X, et al. Performance enhancement of microbial fuel cell by applying transient-state regulation. Appl Energy 2017;185:582–8. [6] Li WW, Sheng GP, Liu XW, Yu HQ. Recent advances in the separators for microbial fuel cells. Bioresour Technol 2011;102:244–52. [7] Asensio Y, Mansilla E, Fernandez-Marchante CM, Lobato J, Canizares P, Rodrigo MA. Towards the scale-up of bioelectrogenic technology: stacking microbial fuel cells to produce larger amounts of electricity. J Appl Electrochem 2017;47:1115–25. [8] Wang M, Liu G, Tian Z, Shao Y, Wang L, Ye F, et al. Microstructure-modified proton exchange membranes for high-performance direct methanol fuel cells. Energy Convers Manage 2017;148:753–8. [9] Kim JR, Cheng S, Oh SE, Logan BE. Power generation using different cation, anion, and ultrafiltration membranes in microbial fuel cells. Environ Sci Technol 2007;41:1004–9. [10] Trapero JR, Horcajada L, Linares JJ, Lobato J. Is microbial fuel cell technology ready? An economic answer towards industrial commercialization. Appl Energy 2017;185:698–707. [11] Venkatesan PN, Dharmalingam S. Development of cation exchange resin-polymer

[20]

[21]

[22] [23] [24]

[25] [26] [27]

[28] [29]

[30]

494

electrolyte membranes for microbial fuel cell application. J Mater Sci 2015;50:6302–12. Adeniran JA, Huberts R, De-Koker JJ, Arotiba OA, Olorundare OF, Van-Zyl E, et al. Energy generation from domestic wastewater using sandwich dual-chamber microbial fuel cell with mesh current collector cathode. Int J Environ Sci Technol 2016;13:2209–18. Zuo Y, Cheng S, Call D, Logan BE. Tubular membrane cathodes for scalable power generation in microbial fuel cells. Environ Sci Technol 2007;41:3347–53. Fan YZ, Hu HQ, Liu H. Enhanced Coulombic efficiency and power density of aircathode microbial fuel cells with an improved cell configuration. J Power Sources 2007;171:348–54. Zhang XY, Cheng SA, Wang X, Huang X, Logan BE. Separator characteristics for increasing performance of microbial fuel cells. Environ Sci Technol 2009;43:8456–61. Hu YS, Wang XCC, Tian WR, Ngo HH, Chen R. Towards stable operation of a dynamic membrane bioreactor (DMBR): operational process, behavior and retention effect of dynamic membrane. J Membr Sci 2016;498:20–9. Hu YS, Wang XCC, Ngo HH, Sun QY, Yang Y. Anaerobic dynamic membrane bioreactor (AnDMBR) for wastewater treatment: a review. Bioresour Technol 2018;247:1107–18. Ersahin ME, Ozgun H, Dereli RK, Ozturk I, Roest K, van Lier JB. A review on dynamic membrane filtration: materials, applications and future perspectives. Bioresour Technol 2012;122:196–206. Deng Q, Li XY, Zuo JE, Ling A, Logan BE. Power generation using an activated carbon fiber felt cathode in an upflow microbial fuel cell. J Power Sources 2009;195:1130–5. Liu G, Wang M, Wang Y, Ye F, Wang T, Tian Z, et al. Anode catalyst layer with novel microstructure for a direct methanol fuel cell. Int J Hydrogen Energy 2012;37:8659–63. Liu G, Li X, Wang M, Wang M, Lee JK. A study on anode diffusion layer for performance enhancement of a direct methanol fuel cell. Energy Convers Manage 2016;126:697–703. Little B, Wagner P, Ray R, Pope R, Scheetz R. Biofilms – an ESEM evaluation of artifacts introduced during SEM preparation. J Ind Microbiol 1991;8:213–21. Liu G, Yang Z, Halim M, Li X, Wang M, Kim JY, et al. A gradient activation method for direct methanol fuel cells. Energy Convers Manage 2017;138:54–60. Liu G, Li X, Wang H, Chen M, Woo JY, Kim JY, et al. Design of 3-electrode system for in situ monitoring direct methanol fuel cells during long-time running test at high temperature. Appl Energy 2017;197:163–8. Zhou M, Chi M, Luo J, He H, Jin T. An overview of electrode materials in microbial fuel cells. J Power Sources 2011;196:4427–35. Wei J, Liang P, Huang X. Recent progress in electrodes for microbial fuel cells. Bioresour Technol 2011;102:9335–44. Wang YK, Sheng GP, Li WW, Yu HQ. A pilot investigation into membrane bioreactor using mesh filter for treating low-strength municipal wastewater. Bioresour Technol 2012;122:17–21. Lefebvre O, Tan Z, Kharkwal S, Ng HY. Effect of increasing anodic NaCl concentration on microbial fuel cell performance. Bioresour Technol 2012;112:336–40. Adelaja O, Keshavarz T, Kyazze G. The effect of salinity, redox mediators and temperature on anaerobic biodegradation of petroleum hydrocarbons in microbial fuel cells. J Hazard Mater 2015;283:211–7. Janicek A, Fan Y, Liu H. Design of microbial fuel cells for practical application: a review and analysis of scale-up studies. Biofuels 2014;5:79–92.