H-type microbial fuel cells consisted of two 125 mL glass bottles connected through a flange on the sides where a cation exchange membrane (CMI-7000S, Membranes International Inc.) was placed. Two grams of finely ground pine tree (Avicel PH-101, Sigma-Aldrich) was weighed into each one of the bottles which serve as the anode, then 80 mL of culture medium and 20 mL strained rumen fluid described below were transferred into each anode. Under flushing of CO₂ gas, anode contents were well suspended using a magnetic bar and agitator. The magnetic bar is left in each anode to agitate anode suspension daily during MFC operation. An electrode (Graphite plate) connected with copper wire was placed in the middle of the anode. Copper wire was fixed to butyl rubber stopper which is also fixed with stopcock and Luer locks. To the other bottle which serves as a cathode, 100 mL of phosphate-buffered saline pH 7.4 (PBS) was transferred. Electrodes identical to anode electrode was placed in the anode chamber. A rubber stopper was placed on the cathode but left open to the air through tubing. A two-liter volume Mylar bag was attached to anode stopper Luer lock to prevent gas pressure on top in anode compartment during establishment and to collect biogas during the experimental period while cathode stopper was open to the air through Luer lock. Anode and cathode were connected externally through a copper wire with a resistor (300 ohm).

MFCs were placed in a water bath at 37˚C for operation. After 9d pre-experimental operation, treatments were added to anode chamber, and MFCs were operated for 10 d for each experiment after treatments were received.

For MFC anode compartment inoculum, strained rumen fluid collected from a cow was prepared through filtering through 4 layers of cheesecloth to remove solid debris and bubbling with CO₂ gas for 10 min. Strained rumen fluid was stored in a water bath at 37˚C until inoculated to MFCs. Culture media consisted of 0.048% KH₂PO₄, 0.048% K₂HPO₄, 0.048% (NH₄)₂SO₄, 0.096% NaCl, 0.1% Trypticase peptone, 1% yeast extract, 0.05% cysteine-HCl, 0.013% CaCl₂∙2H₂O, 0.02% MgSO₄∙7H₂O, 0.4% Na₂CO₃, 0.1% sodium fumarate, and 1 ppm of Resazurin and was prepared anaerobically by bubbling with CO₂ gas and autoclaved at 121˚C for 30 min. Culture media was stored in a water bath at 37˚C until inoculated to MFCs. PBS consisted of 137 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, and 2 mM KH₂PO₄ and was autoclaved at 121˚C for 30 min and stored at room temperature.

For treatment, freeze-dried red ginseng marc (RGM) was ground to pass through a 0.5 mm screen. Treatment MFCs received 1 g or 0.1 g of RGM was added to the anode for experiments 1 and 2, respectively. Control MFCs received 1 g or 0.1 g of celite (Celite^®, Celite Corp) as the inert placebo for experiments 1 and 2, respectively.

Using a digital multimeter, closed-circuit voltage (voltage across a resistor) end point potential (open circuit voltage) and current were measured daily for 10d after treatments addition. The following equations were used to calculate power density normalized to electrode surface area.

P = I V A

where I (A) is the current (V/R(ohm)); V (V) is voltage; R (ohm) is the external resistance; and A (m²) is the projected area of the anode.

The volume of biogas collected in a Mylar bag, produced from the anode, was measured with a glass syringe. Gas composition was analyzed with a gas chromatograph (Agilent 6890).

Experiment 1 (Exp. 1) and Experiment 2 (Exp. 2) were conducted independently and analyzed separately. Exp. 1 and Exp. 2 included 1% and 0.1% of RGM, respectively, but in the same MFC operational condition. Each dose of RGM was compared to its own control in the corresponding experiment.

Power generation measures, fermentation gas production, and gas composition were analyzed using the one-way ANOVA procedure of JPM 14.1.0 (SAS Institute Inc.). Where the significance found (P < 0.05), least square means of for treatments and days were separated using Student’s t-test (P < 0.05) and Tuckey HSD (P < 0.05), respectively.

Methane amount (Figure 1) produced in anode chamber for 10d operation did not change with red ginseng marc (RGM) addition in Exp. 1 (P = 0.0551) or Exp. 2 (P = 0.8640) at 1% or 0.1% of dose, respectively. However, carbon dioxide production (Figure 1) increased with RGM addition at both 1% (Exp. 1; P = 0.0010) and 0.1% (Exp. 2; P = 0.0005) of doses. Consequently, RGM addition increased total gas production (Figure 1) at both 1% (P = 0.0114) and 0.1% (P = 0.0082) of doses.

Microorganisms ferment carbohydrates for energy production required for growth. The anode in the current study contained cellulose, a polysaccharide, as experimental substrates and also soluble carbohydrates from an anolyte. Cellulose is degraded to cellobiose, a disaccharide of two glucose, then glucose, a monosaccharide by endo- and exo-cellulases as a function of cellulolytic bacteria. Most microorganisms can ferment glucose for their growth in anode in MFC, and the products of anaerobic respiration and fermentation are short-chain organic acids, carbon dioxide, and hydrogen. The organic acid can be further fermented and the ideal glucose decomposition in MFCs is into CO₂, proton, and electron in an anode chamber under anaerobic conditions (C₆H₁₂O₆ + 6H₂O → 6CO₂ + 24H⁺ + 24e⁻) [22] by symbiotic microorganisms. However, methane is another end product in most of the MFCs through acetoclastic (CH₃COO⁻ + H⁺ → CH₄ + CO₂) and hydrogenotrophic (4H₂ + CO₂ → CH₄ + 2H₂O) methanogenesis. In either way, the greater total gas production reflects the efficient cellulose fermentation in the current study.

Methane to carbon dioxide ratio (Figure 1(c)) is less in RGM treatments in both 1% (Exp. 1; P = 0.0033) and 0.1% (Exp. 2; P = 0.0052) of doses and the lower methane to carbon dioxide ratio implies the less (P < 0.01) production of methane per substrates. Although the methane production in volume was similar between control and RGM, methane took out 67.9% and 67.3% of total gas in control MFCs while 58.6% and 59.1% in RGM MFCs in Exp. 1 and Exp. 2, respectively. Ideally, electrons and the protons should move to the cathode via the external electrical circuit and the cation exchange membrane, respectively, then reduce oxygen and produce water (24H⁺ + 24e^– + 6O₂ → 12H₂O) in the cathode, and generate the current, the flow of electrons. As methane acts as an electron acceptor in an anode, the reduced methane formation in the anode is expected

to improve power generation efficiency in MFCs.

End point potential in control MFCs in Exp. 1 (Figure 2(a)) was constant (P = 0.9892) during 10d operation at the average of 631 ± 9.1 mV. It changed with operation time in 1% RGM MFCs and the average was 765.9 ± 40.5 mV ranged from 574 to 849 mV. It was higher (P < 0.05) on d2 through d8 comparing to d0 in 1% RGM MFCs. End point potential was also greater (P < 0.05) in 1% RGM MFCs than control MFCs on d5, d8 and d10 of operation. In Exp. 2 (Figure 2(b)), end point potential in control MFCs was also constant (P = 0.1389) during 10d and the average was 570 ± 7.9 mV, however it changed (P = 0.0006) with operation time in 0.1% RGM MFCs from 573 mV to 695 mV with the average of

638 ± 15.1 mV. Comparing to d0, end point potential increased (P 0.05) on d3 - 5 and d8 - 9 within 0.1% RGM MFCs. In Exp. 2 0.1% RGM MFCs had greater (P < 0.05) end point potential than control MFCs on d2 - 4 and d8 - 10.

Power density in control MFCs in Exp. 1 (Figure 2(a)) increased numerically (P = 0.0616) with the average of 61.5 ± 6.29 mW/m² ranged from 45.2 to 76.3 mW/m². Power density in 1% RGM MFCs changed with operation time and it was greater (P < 0.05) on d 2 - 10 than d 0 - 1, the average was 83.8 ± 2.99 mW/m² with range between 44.0 and 97.2 mW/m². 1% RGM MFCs (Exp. 1) had greater (P < 0.05) power density than control MFCs on d2, 3, 7, 9 and 10. In Exp. 2 (Figure 2(b)), power density was constant (P = 0.7985) over 10d of incubation in control MFCs. The average of power density in control MFCs was 49.7 ± 3.16 mW/m² and the range was between 45.1 and 54.1 mW/m². 0.1% RGM MFCs had a sort of spike of power density on d3 and was greater (P < 0.05) than d0 - 1. The average of power density in 0.1% RGM MFCs was 60.9 ± 3.98 mW/m² and ranged from 44.8 to 75.6 mW/m². The power density in 0.1% RGM MFCs was greater (P < 0.05) than control MFCs on d2, 3, 7 and 9 (Figure 2(b)).

To determine the bioelectrochemical performances of MFCs, the power generated by the MFC must be normalized to a relevant geometric characteristic of the MFC reactor such as anode surface area or anolyte volume [23]. The power density, normalized to electrode surface area in the current study, the observed values were between 44 and 97 mW/m² in Exp. 1 and 2. When sole rumen fluid was used as an anolyte like the current study, the maximum power densities were 55 mW/m² from microcrystalline cellulose [24], 100 mW/m² from carboxymethyl cellulose [13], and 405 mW/m³ from Canna indica [14]. Therefore, the performance of MFCs in the current study served as control or treatment can be considered as acceptable or better than previous reports. Furthermore, potassium ferricyanide solution (50 mM K₃Fe(CN)₆) was used as the catholyte to enhance oxygen reduction in the cathode in the previous reports while only PBS was used in the current study to maintain the environmental friendly conditions.

Power generation was improved (P < 0.0001) by red ginseng marc at both 1% or 0.1% dose and the average of power density and end point potential were shown on Figure 3(a) and Figure 3(b), respectively. For power density (Figure 3(a)), 36% and 22% is increased (P < 0.0001) by 1% or 0.1% RGM supplementation, respectively. A 21% or 12% of higher end point potential was generated by 1% or 0.1% RGM supplementation (Figure 3(b)), respectively. These improvements in means for 10d represent the daily measurement point while the power generation in MFCs was a continuous reaction during the operation, thus the measured improvement also reflects the total power generation improvements even with different values and units. Two experiments, Exp. 1 and Exp. 2, were conducted independently, therefore two different doses of red ginseng marc could not be compared. In each experiment, comparing to its own control MFCs, 1% RGM had shown bigger improvements for both power density and end point potential than 0.1% RGM, however total gas production improvement

was 7.7% in 1% RGM and 14.4% in 0.1% RGM.

In conclusion, methane production inhibited, and power generation increased by the red ginseng marc addition at 1% or 0.1% dose to the anode of MFC generating electricity from cellulose using rumen microorganisms.