Nutrient Removal Process and Cathodic Microbial Community Composition in Integrated Vertical-Flow Constructed Wetland – Microbial Fuel Cells Filled With Different Substrates

Tensions the paper itself points to. Each side quotes the passage it comes from.

• Within the paper Power generation collapses while nutrient removal remains stable as HRT decreases

  One side

  Voltage output drops from 120.3 mW/m³ at 7.6 d HRT to below 0.02 V at 2.8 d HRT in CM-A

  “With an HRT of 7.6 d (H1), the average voltage outputs in both the up-flow and down-flow chambers in CM-A were higher than those in CM-B and CM-C... However, when the HRT was decreased to 4.0 d (H2) and 2.8 d (H3), low voltage outputs (below ”

  Other side

  NH₄⁺-N removal efficiency in CM-A remains high at 93.8% at 2.8 d HRT, comparable to 95.8% at 7.6 d HRT

  “With an HRT of 2.8 d, the ammonium (NH4+-N) and orthophosphate (PO43−-P) removal efficiencies in CM-A were as high as 93.8 and 99.6%, respectively.”

  Authors’ reading. Authors do not explicitly reconcile this; the decoupling suggests that nutrient removal is driven by bioelectrochemical and biological processes (SND, DSR) that are less dependent on the redox gradient needed for power generation, or that the electrode configuration (anode in down-flow, cathode in up-flow) is optimized for treatment rather than power.

• Within the paper Quartz substrate shows lowest N removal but highest COD removal at short HRTs

  One side

  CM-B (quartz) achieves only 0.4% NH₄⁺-N removal at 2.8 d HRT but 91.8% COD removal

  “The lowest effluent COD concentrations in CM-B were observed among the three CW-MFCs under HRTs of 4.0 and 2.8 d.”

  Other side

  CM-A (ceramsite) achieves 93.8% NH₄⁺-N removal at 2.8 d HRT but only 47.3% COD removal

  “With an HRT of 2.8 d, the ammonium (NH4+-N) and orthophosphate (PO43−-P) removal efficiencies in CM-A were as high as 93.8 and 99.6%, respectively.”

  Authors’ reading. Authors attribute this to substrate-dependent microbial communities: quartz enriches sulfate-reducing bacteria (Desulfovibrio, Desulfomonile 2.2% in CM-B vs. 0.5% in CM-A) that consume organic carbon but do not support nitrification/denitrification; ceramsite enriches denitrifiers and iron-oxidizers that prioritize N removal over COD oxidation.

Specific gaps in this paper, with an experiment that would close each.

• High N₂O emissions not measured; complete denitrification pathway (to N₂ vs. N₂O) unknown despite low NOₓ⁻-N effluent

  Effluent NOₓ⁻-N consistently <0.3 mg/L across all conditions, but paper does not report N₂O or N₂ production; SND and DSR bacteria can produce N₂O as a side product, especially under variable DO and pH

  Try. Measure N₂O and N₂ in headspace or dissolved phase using gas chromatography or isotope tracing (¹⁵N) across HRT range 2.8–7.6 d and DO gradient (2–4 mg/L observed); quantify N recovery balance

• High Cathode potential and current density not reported; electrochemical driving force for denitrification unknown

  Paper reports voltage output and power density but not cathode potential (vs. Ag/AgCl or SHE), current density, or polarization curves; these are critical to understanding whether denitrification is bioelectrochemical (electron transfer from cathode) or purely biological (organic carbon oxidation)

  Try. Measure cathode potential, anode potential, and current density at each HRT using potentiostat; correlate with denitrifier abundance and NOₓ⁻-N removal rate to quantify bioelectrochemical contribution

• Medium Influent COD/N ratio not varied; low organic carbon dependence of N removal not tested

  Synthetic wastewater has fixed COD/N ratio (~14 mg COD / mg N based on 243–262 mg/L COD and 16–18 mg/L NH₄⁺-N); paper claims autotrophic denitrification via Fe cycling but does not test whether N removal persists at lower COD/N ratios (e.g., <5)

  Try. Operate CM-A at HRT 2.8 d with COD/N ratios of 2, 5, 10, 14 (current); measure NH₄⁺-N removal, NOₓ⁻-N production, and denitrifier abundance to confirm autotrophic pathway

• Medium Ferrous/ferric iron concentrations not measured; Fe cycling mechanism inferred but not quantified

  Paper proposes that Sphaerotilus and Azospira/Zoogloea/Dechloromonas drive autotrophic denitrification via Fe²⁺ oxidation, but ferrous and ferric ion concentrations are not reported; SO₄²⁻-S increase in effluent is noted but not linked to Fe cycling

  Try. Measure dissolved Fe²⁺ and Fe³⁺ at sampling ports U4, U2, D1, D2 at HRT 2.8 d and 7.6 d using ICP-OES or colorimetric assay; correlate with NOₓ⁻-N removal and Sphaerotilus abundance

• Medium Catholyte pH range (8.0+) not explored; ammonia volatilization in up-flow chamber not quantified

  Effluent pH in CM-A up-flow chamber >8.0 (Supplementary Table S1); authors note this 'could contribute to the ammonia loss via volatilization' but do not measure NH₃ gas or quantify volatilization loss vs. biological removal

  Try. Operate CM-A at controlled pH (6.5, 7.0, 7.5, 8.0, 8.5) in up-flow chamber using buffer or acid/base dosing; measure NH₄⁺-N removal, effluent NH₃ (headspace), and nitrifier/denitrifier abundance to decouple volatilization from biological pathways

• Medium Substrate adsorption capacity for NH₄⁺ and PO₄³⁻ not quantified; contribution to removal efficiency unclear

  Paper states zeolite (CM-C) can adsorb NH₄⁺ at long HRT (7.6 d) and ceramsite shows 'substrate adsorption and denitrifying phosphorus removal' for PO₄³⁻, but adsorption isotherms, saturation time, and regeneration are not measured

  Try. Conduct batch adsorption tests on fresh and used ceramsite, quartz, zeolite granules with NH₄⁺ and PO₄³⁻ at pH 6–8; measure Langmuir/Freundlich parameters and compare to biological removal rates in columns