Application of a CO₂-stripping system for calcium removal to upgrade organic matter removal and sludge granulation in a leachate-fed EGSB bioreactor

Abstract

The application of CO₂-stripping system for calcium removal to upgrade organic matter removal and sludge granulation in a leachate-fed EGSB bioreactor was evaluated. Three-dimensional excitation–emission matrix (3D-EEM) spectroscopy combined with parallel factor (PARAFAC) analysis was used to characterize the transformation of the effluent dissolved organic matter (DOM) during the operation. X-ray diffraction (XRD), Fourier-transform infrared (FT-IR) and scanning electronic microscopy (SEM) were used to assess the effects of a CO₂-stripping unit on the microstructure of the granules. The introduction of the CO₂-stripping system reduced the calcium concentration while upgrading methane evolution. The methane yield reached 0.33 L CH₄ per g COD_removed in the bioreactor with the CO₂-stripping unit compared with 0.31 L CH₄ per g COD_removed without the unit as the control. The combined system produced 80% and 50–60% chemical oxygen demand (COD) and total nitrogen (TN) removal under steady-state conditions, which were 6.3% and 41.0% higher than those of the control, respectively. With 3D-EEM-PARAFAC analysis, three fluorescence components, associated as tryptophan protein-like (component 1, E_x/E_m = 275–280/355–365 nm) and humic-like substances (component 2, E_x/E_m = 240(295, 340)/450 nm and component 3, E_x/E_m = 320/320 nm), were identified from the effluent samples. The componential characterizations confirmed the favorable influence of the CO₂-stripping unit on the transformation of DOM. Further analysis through XRD, FT-IR and SEM demonstrated that the use of the unit alleviated inactivation of the granules through removing calcium, which might be the core reason for the enhancement of the EGSB performance.

---

1. Introduction

Expanded granular sludge bed (EGSB) bioreactors, modified from up-flow anaerobic sludge blanket (UASB) reactors, have attracted much attention due to their low energy consumption, high treatment efficiency, low sludge production and renewable energy recovery in the form of methane.^1,2 In recent years, they has been universally used for treating various kinds of wastewater such as brewery wastewater, starch wastewater, molasses alcohol slops, slaughterhouse wastewater, and municipal solid waste (MSW) leachate.^3,4 Several parameters affect the long-term performance of EGSB bioreactors such as the physiochemical properties of anaerobic granules, organic loading rates (OLR), pH, hydraulic retention time (HRT), temperature and superficial velocity.^5–7 Anaerobic granules, which are regarded as the principle component of EGSB bioreactors, play a relatively critical role in the biological treatment process. Anaerobic granules support active biofilms but also provide the buoyancy and the settle ability necessary to enable very vigorous granule–liquid contact in the reactor.⁸

Many factors contribute to the sludge granulation process. One of the crucial elements is the calcium ion (Ca²⁺), which is found to exert a key role in the granulation process by neutralizing granules surface charges and/or binding extracellular polymeric substances (EPS).^9–11 However, insufficient or excessive calcium is often detrimental to granule formation. An insufficient supply of calcium would inhibit microbial activity in denitrifying granules, and weaken the three-dimensional structure of EPS–Ca²⁺–EPS and the microbial community of bio-granules.¹¹ An excessive dosage on the contrary seems to cause clogging in the bioreactor,¹² which substantially deteriorates the EGSB performance. Unstable COD removal efficiency and cementation of the sludge bed in a UASB reactor was observed by van Langerak et al.¹³ when they treated acidified wastewater with a calcium concentration of 1200 mg L⁻¹. Another study by Pevere et al.¹⁴ also found that calcium concentrations of 780–1560 mg L⁻¹ in the wastewater induced CaCO₃ precipitate formation, which led to scaling of granular sludge and decreased the diffusivity. Therefore, it is necessary to remove calcium from calcium-rich leachate to achieve successful performance of EGSB bioreactors. An alternative to alleviate calcium-related problems can be the carbonation pretreatment of wastewater through CO₂ stripping. Jo et al.¹⁵ also used CO₂ to remove Ca²⁺ ions from coal fly ash through a carbonation reaction of CO₂ with Ca²⁺ ions. Nir et al.¹⁶ utilized CO₂ in a seawater reverse osmosis membrane to minimize calcium carbonation to prevent fouling the membrane. Likewise, Kim et al.¹⁷ adopted biologically produced alkalinity in a UASB bioreactor to minimize the adverse effect of calcium hardness from industrial wastewaters with a respective process, and got a stable COD removal efficiency of over 90% after some calcium was removed. In our previous work,¹⁸ a scientific attempt on recirculating biogas (CO₂) from a leachate-fed EGSB bioreactor for the removal of calcium and simultaneous methane purification was also carried out; by optimizing the solution pH and imported biogas (CO₂), the calcium level suitable for microbial growth as well as and organic matter removal was achieved, which, as stated before, improved the sludge granulation and methane evolution to a large degree. However, the previous study mainly focused on the optimization of operational parameters (e.g. solution pH, imported biogas (CO₂), etc.) as well as the exploration of calcium precipitation mechanisms. More work in investigating the functions of biogas recirculation on improving the organic matter degradation and the micro-characteristics of granules are still required. Moreover, as the core component of the EGSB system, the characteristics of granular sludge affect, or even decide the overall performance of the process. Thus, characterizing the microstructure of granules would help to further investigate the stimulating effect of biogas recirculation on the overall EGSB performance.

Based on the above-mentioned considerations, a CO₂-stripping unit, used as a pretreatment of fresh leachate with a high calcium concentration to eliminate the inhibitory effect of calcium ions, was conducted in an EGSB bioreactor in this study. Effects of biogas recirculation on calcium removal, methane production, COD, TN, and NH₄⁺–N were investigated and are discussed. Three-dimensional excitation–emission matrix (3D-EEM) spectroscopy was used to characterize the transformation of the effluent dissolved organic matter (DOM) during the operational period. Parallel factor (PARAFAC) analysis was employed to quantitatively analyze the 3D-EEM spectra and to identify the configuration and heterogeneity of the DOM fractions. In addition, X-ray diffraction (XRD), Fourier-transform infrared (FT-IR) and scanning electronic microscopy (SEM) were further performed to assess the possible effects of the CO₂-stripping unit on the properties and microstructure of granular sludge. This study would provide an in-depth and comprehensive insight into the biogas recirculation process (i.e. CO₂-stripping unit) while simultaneously advancing its practical applications in EGSB bioreactors as well as other kinds of bioreactor systems.

2. Materials and methods

2.1. Fresh leachate

The fresh leachate used as the feed in this study was obtained from Pudong Compost Factory in Shanghai, China. This factory has a daily handling capacity of 1300 tons of municipal solid waste (MSW), which generates about 100 tons of fresh leachate per day. The fundamental characteristics of the fresh leachate are shown in Table 1.

Table 1 Fundamental characteristics of the fresh leachate from the Shanghai Pudong Compost Factory (unit: mg L⁻¹, except for pH)

Item	Value	Item	Value
CODCr	37 156–73 865	Mg	351–387
BOD5	23 351–32 107	Fe	148–180
TOC	16 133–24 773	K	116–123
NH^+4–N	789–1332	Al	44.89–75.58
TN	1599–2507	Zn	14.18–26.35
TP	102–278	Li	0.33–0.38
pH	3.76–4.89	Sr	7.41–8.71
Ca	5073–5529	B	7.71–27.77

---

2.2. Source of seed sludge

The seed sludge was taken from a full-scale internal circulation bioreactor in a paper mill in Shanghai, China. The anaerobic granular sludge, consisting of well-settled black granules with about 95% showing sizes from 0.5 to 2 mm in diameter, had a mixed liquor volatile suspended solid (MLVSS) concentration of 75 972 mg L⁻¹ and a mixed liquor suspended solid (MLSS) concentration of 103 291 mg L⁻¹.

2.3. Experimental set-up and its operation

The continuous biodegradation of fresh leachate was carried out in parallel in two lab-scale EGSB bioreactors A and B. The detailed schematic diagram of both reactors was shown in our previous research.¹⁸ The reactors had a working effective volume of 6 L with an inside diameter of 8 cm and a height of 120 cm, and the granular sludge occupied 1/3 of the working volume of the EGSB bioreactor. A gas–liquid–solids separator was installed at the top of the reactors. Reactor B had a CO₂-stripping unit (φ 14 cm × 14 cm) with a working volume of 2 L. The biogas produced from the reactor was re-circulated to the CO₂-stripping unit without any external force, and then the biogas from the CO₂-stripping unit was collected by a gas sampling bag. The pH value in the carbonation unit was adjusted to 10.0 ± 0.5 through feeding a 2.0 N NaOH solution. In order to ensure sufficient mass transfer between the biogas and leachate as well as uniformity of the solution pH, the CO₂-stripping unit was constantly agitated by a magnetic stirrer during the carbonation process. Afterwards, the leachate from CO₂-stripping unit was used as the feed of the reactor. The temperature in the reactor was controlled with water recirculation using a heated water bath. The operational temperature was kept constant at 35 ± 1 °C and the liquid up-flow velocity (V_up) was maintained at 3.2 m h⁻¹ based on the results of our previous research.⁴ Two pumps were used to feed the influent and circulate the effluent. Reactor A without a CO₂-stripping unit operated under the same conditions and was also used as a control. The organic loading rate (OLR) for both reactors was set at 10–20 kg COD per m³ per d with a hydraulic retention time (HRT) of 72 h under steady-state conditions.

2.4. 3D-EEM fluorescence analysis

The leachate samples collected during the operation period were filtered through a 0.45 μm membrane and diluted by 100 times. 3D-EEM fluorescence spectra were measured using a luminescence spectrometer (FluoroMax-4, HORIBA Jobin Yvon Co., France).¹⁹ Leachate EEM spectra were recorded with scanning emission (E_m) spectra from 250 to 550 nm at 5 nm increments through varying the excitation (E_x) wavelength from 215 to 400 nm at 5 nm increments. Excitation and emission slits were both maintained at 5 nm, and the scanning speed was set at 4800 nm min⁻¹ for all the measurements. The software Origin 8.0 (Origin Lab Inc., USA) was used for handling the EEM data.

Parallel factor (PARAFAC) analysis was employed to quantitatively analyze the 3D-EEM spectra. The detailed approach of PARAFAC analysis can be found elsewhere.^20,21 PARAFAC is a generalization of bilinear principal component analysis (PCA) to higher order arrays. It can decompose the three-way array X of fluorescence EEM into three matrices, A (the score matrix), B and C (loading matrices) with the elements a_if, b_jf and c_kf using:

where, χ_ijk is the fluorescence intensity of the sample i at the emission wavelength j and excitation wavelength k; F is the number of components in the model; a_if is directly proportional to the concentration of the fluorophore f in the sample i (defined as scores); b_jf and c_kf are estimates of the emission and excitation spectra for the fluorophore component f (defined as loadings), respectively. ε_ijk is the residual term representing the invariability not accounted for by the model. PARAFAC analysis was carried out in Matlab 10.0 with the DOMFluor toolbox. The variance identified by the model, and the core consistency diagnostic as well as split-half analysis were used to find the suitable number of components. The relative concentration of fluorophore was estimated using the maximum fluorescence intensity (F_max) values (R.U., i.e. Raman units).

2.5. Characterization of clogging materials and sludge granules

The clogging materials from the carbonation unit and sludge granules from both reactors were characterized by means of X-ray diffraction (XRD). The XRD patterns were collected over a 2θ from 5–90° in a scan step of 0.02° and measuring time of 0.1 s per step using graphite monochromated CuKα radiation (D8, Bruker AXS Inc., Germany). The accelerating voltage was 40 kV and the current was 40 mA. The diffractograms were attained with Diff-plus and analyzed using MDI Jade 5.0 software. The functional groups were analyzed using an ALPHA FT-IR spectrometer (Bruker Optics, Germany) in the range of 4000–400 cm⁻¹. For scanning electron microscopy (SEM) analysis,²² the samples were fixed for 2 h at 4 °C with 2.5% (v/v) glutaraldehyde in 0.1 M phosphate buffer (pH 6.8). After being washed three times with phosphate buffer solution, the samples were dehydrated via successive passages through 50%, 60%, 70%, 80% and 90% (v/v) ethanol followed by freeze drying, then sputter-coated with a thin layer of gold for SEM analysis (Hitachi S-520, Japan).

2.6. Other analytical methods

The chemical oxygen demand (COD), biological oxygen demand (BOD₅), total phosphorus (TP), NH⁺₄–N, pH, mixed liquor volatile suspended solids (MLVSS) and mixed liquor suspended solids (MLSS) were estimated following standard methods.²³ Total nitrogen (TN) and total organic carbon (TOC) were measured with a TOC-VCPN analyzer (SHIMADZU, Japan) using the combustion-infrared method. Metal concentrations were analyzed with ICP-AES (Optima 2100 DVICP-AES, PerkinElmer, USA). Biogas volume was recorded using a gas sampling bag connected to a gas meter, and the gas composition (CH₄ and CO₂) of the biogas was analyzed through gas chromatography (GC9800) with a stainless steel column (1 m × 3 mm) packed with TD-01 for the separation of CH₄ and CO₂ using a thermal conductivity detector (TCD) at room temperature.

3. Results and discussion

3.1. Methane production and purification

Calcium plays an important role in the EGSB reactor, and the presence of high calcium content in the leachate could damage the environment required for maintenance of the granular structure or the bacterial activity.²⁴ A CO₂-stripping unit with biogas recirculation as a pretreatment effectively enhanced the removal of calcium from the leachate, with the resulting removal efficiency being up to 92.8–96.5%.¹⁸ The noticeable decrease of calcium during the carbonation process resulted from its precipitation in the form of calcium carbonate in the presence of CO₂ from the biogas. In this sense, the methane content from reactor A varied between 41.5–57.9%. In comparison, it dramatically increased to around 87–91% when the CO₂-stripping unit was utilized,¹⁸ which is profoundly higher than the values (65–80%) reported by Ye et al.²⁵ The higher methane content in the biogas achieved from bioreactor B evidently indicate that most of the CO₂ was dissolved and consumed by generating CaCO₃ precipitates in the carbonation unit. A higher methane content would make the biogas easier to be utilized as an energy source. Detailed mechanisms for calcium precipitation were described in our previous publication.¹⁸

The production of methane during anaerobic digestion is strongly correlated with COD_removed.²⁶ The average methane yield was determined to be 0.31 L CH₄ per g COD_removed from reactor A, suggesting that approximately 88.6% of the removed COD was converted to methane and the remaining COD (11.4%) might be regarded as synthesis of biomass^25,26 as the theoretical methane production rate is 0.35 L CH₄ per g COD_removed.²⁷ Comparatively, the methane yield amounted to 0.33 L CH₄ per g COD_removed from reactor B, indicating that 94.3% of the removed COD was converted to methane whereas the remaining 5.7% was presumably changed into the form of biomass. The higher methane production rate in reactor B is presumably due to less inhibition from calcium. Mineral encrustation as inert suspended solids (ISS) easily accumulates onto the granules, which may exclude the active anaerobic methanogenic microorganisms by occupying the space of the granule, resulting in weakened activity of the biomass.^28,29 The presence of the CO₂-stripping unit efficiently reduced the calcium content in the fresh leachate and hampered its adverse interference, which accordingly enhanced methanogenic activity and the subsequent digestion processes. The abovementioned experimental results once again confirmed the beneficial influence of the application of a CO₂-stripping unit as an effective alternative to reinforce calcium removal and to improve the overall performance of an EGSB.

3.2. COD, TN and NH⁺₄–N removal

To examine the effect of carbonation pretreatment on the overall treatment performance of an EGSB bioreactor, the process parameters in terms of COD, TN and NH⁺₄–N removal were measured once steady-state conditions were achieved. The variation profiles of COD, TN, and NH⁺₄–N concentrations in the influent and effluent and the corresponding removal efficiencies from bioreactors A and B are given in Fig. 1. In reactor A (Fig. 1a), the effluent COD concentrations ranged from 5226 to 16 682 mg L⁻¹ at influent concentrations of 39 581–73 865 mg L⁻¹. The highest COD removal efficiency of 85–86% was observed during the first 3 days. With the progress of the experiment, the COD removal efficiency decreased stepwise and maintained at a relatively low and stable level of around 75%. This level of COD removal was similar to the previously reported removal of roughly 60–80% for an EGSB bioreactor.³⁰ The considerable decrease in COD removal efficiency after 4 days of operation was presumably ascribed to the methanogenesis inhibition from high concentrations of calcium in the fresh leachate. Severe calcium precipitation and accumulation on the surface of sludge granules, as hypothesized by Ye et al.,²⁵ lowered the mass transfer rate and decreased the specific methanogenesis activity of the granules, thereby greatly deteriorating the performance of the EGSB. Similar trends were also reported in a UASB treating paper water¹⁷ and an EGSB treating the fresh leachate.^4,18

Fig. 1 Performance of the EGSB bioreactors A and B: the concentrations and the removal efficiencies of (a) COD, (b) TN and (c) NH⁺₄–N.

In comparison with reactor A, reactor B with a CO₂-stripping unit always achieved higher COD removal efficiencies. As presented in Fig. 1a, the COD removal efficiency rapidly increased to more than 85% after 5 days of operation at influent COD concentrations of 37 156–72 834 mg L⁻¹. Even though a slight reduction was observed after 8 days of operation, the COD removal still approached up to 80% with the effluent COD concentration of below 15 000 mg L⁻¹, 6.3% higher in contrast to reactor A. Similar results were reported by Kim et al.¹⁷ It is obvious that a higher COD removal efficiency can be achieved when biogas recirculation as a pretreatment is employed in an EGSB bioreactor.

Similar to the results of COD removal, the TN (NH⁺₄–N, NO⁻₂–N and NO⁻₃–N) removal efficiency was always higher in reactor B compared to that in reactor A (Fig. 1b). As illustrated in Fig. 1b, the average effluent TN of reactor A increased from 1040 to 1334 mg L⁻¹, and finally stabilized at 1579 mg L⁻¹; the average effluent TN of reactor B declined from 1518 to 986 mg L⁻¹, and then to 709 mg L⁻¹. The total removal efficiencies of TN in reactor A varied between 30% and 48% during the initial 11 days, and then it sharply declined to approximately 3% on day 13. This considerable reduction might result from the inhibition of calcium to the anaerobic ammonium oxidation (ANAMMOX) reaction of anaerobic granules.³¹ Different to reactor A, the TN removal efficiency of reactor B gradually increased at the beginning and attained around 50–60% after 7 days of continuous operation (Fig. 1b), roughly 41.0% higher than from reactor A. The increased removal efficiencies of COD and TN obtained in reactor B confirmed the profoundly positive role played by biogas recirculation in the promotion of the EGSB performance.

Fig. 1c shows that the NH⁺₄–N concentration in the influent was higher than that in the effluent, mainly attributable to the degradation of the nitrogenous organic substances in the leachate.²⁵ Very interestingly, the NH⁺₄–N concentrations in the effluent from bioreactors A and B were definitively different although the levels in the influent were relatively similar. The effluent NH⁺₄–N concentrations in reactor A varied between 1000 and 1600 mg L⁻¹ while the concentrations were almost less than 900 mg L⁻¹ in reactor B (Fig. 1c). This revealed that ammonia stripping probably took place in the carbonation unit owing to the high pH value (10.0 ± 0.5). It has been reported that high ammonium concentrations can cause elevated concentrations of free ammonium (FA) in the reactor, which inhibits the phase of methenogenesis.¹ Therefore, the reduced ammonium in the feed because of ammonia stripping might be another reason for the increase in the COD removal efficiency in reactor B.

3.3. 3D-EEM fluorescence analysis

3.3.1 EEM spectra. 3D-EEM spectra of the DOM fractions in leachate samples collected from the CO₂-stripping unit, influent and effluent are illustrated in Fig. 1S in the ESI.† Three main peaks could be identified from the fluorescence spectra at the excitation/emission wavelengths (E_x/E_m) of 275–280/355–365 nm (peak A), 230–235/350–370 nm (peak B) and 315–335/405–425 nm (peak C). The first peak was attributed to tryptophan protein-like substances^19,32 while the second peak was assigned to aromatic protein-like substances.³³ Compared to the fluorescence peak locations of proteins (E_x/E_m of 225/330–340 nm) in dissolved organic matter (DOM) for sludge samples described by Zhang et al.,³⁴ the locations of peak B for the leachate showed a blue shift in terms of emission wavelengths. The third peak was regarded as humic-like substances derived from the biodegradation of soluble organic matter.³⁵ Similar fluorescence signals have also been reported for DOM from landfill leachate³⁶ and a submerged membrane bioreactor (SMBR)²⁹ as well as extracellular substances of aerobic granules in a Plexiglas sequencing batch bioreactor (SBR).³³ Although all leachate samples had similar fluorescent features with typical fluorophores of protein-like and humic-like substances, the location and intensity of the fluorescence peaks showed significant differences in the absence or presence of a carbonation unit.

The fluorescence parameters including the peak location and maximum fluorescence intensity were analyzed from EEM fluorescence spectra and are summarized in Table 2. The peak locations of the effluent displayed slight shifts in comparison with those of the influent. For reactor A, peaks B and C were red or blue shifted by around 5–10 nm along the excitation/emission axis with the variety of the operational time. In the case of reactor B, the locations of peaks A, B and C after treatment were all red shifted by 5–20, 5–15 and 5–15 nm, respectively. Such a shift suggested changes in the conformations of the fluorescence components in the leachate after the EGSB process. A red shift was attributed to the increase of carbonyl, hydroxyl, alkoxyl, and amino groups in fluorophores while a blue shift is related to the elimination of particular functional groups (carbonyl, hydroxyl, amine and aromatic rings) or a reduction in the degree of π-electron systems.^34,37,38

Table 2 Fluorescence spectral parameters of organic substances present in the leachate from EGSB reactors A and B^a

Process	Samples	Peak A		Peak B		Peak C
		Ex/Em	Int. (×10^5)	Ex/Em	Int. (×10^5)	Ex/Em	Int. (×10^5)
a Inf.: influent; Eff.: effluent; Int.: intensity.
Reactor A	Inf.-1d	275/355	11.78	230/360	3.31	330/415	2.97
	Eff.-1d	275/355	3.68	230/370	1.3	335/425	1.45
	Inf.-3d	275/355	11.03	230/360	3.03	330/420	2.61
	Eff.-3d	275/355	4.06	230/355	1.42	330/420	1.52
	Inf.-5d	275/355	11.12	230/355	2.91	330/415	2.65
	Eff.-5d	275/355	4.41	230/355	1.59	325/425	1.67
	Inf.-7d	275/355	11.16	230/355	3.04	325/415	2.75
	Eff.-7d	275/355	4.31	230/360	1.57	330/420	1.66
	Inf.-9d	275/355	10.36	230/365	2.81	325/410	2.59
	Eff.-9d	275/355	3.31	230/360	1.31	325/420	1.64
	Inf.-15d	275/355	9.99	230/350	2.7	325/410	2.57
	Eff.-15d	275/355	9.99	230/350	2.7	325/405	2.58
Reactor B	Inf.-1d	275/355	7.15	230/355	2.1	325/405	2.24
	Eff.-1d	275/355	2.44	230/370	0.98	325/420	1.56
	Inf.-2d	275/355	7.74	230/365	2.16	325/405	2.29
	Eff.-2d	275/355	2.21	230/365	1.04	325/415	1.65
	Inf.-4d	275/355	7.58	230/355	2.19	325/410	2.31
	Eff.-4d	280/355	2.29	235/365	1.16	315/410	2.06
	Inf.-6d	275/345	8.5	230/360	2.27	330/410	2.13
	Eff.-6d	275/365	2.79	230/360	1.16	330/415	2.14
	Inf.-7d	275/345	8.2	230/355	2.22	330/415	2.25
	Eff.-7d	275/360	3.22	230/355	1.22	325/415	2.07
	Inf.-9d	275/355	8.78	230/360	2.27	325/415	2.19
	Eff.-9d	275/355	3.72	230/350	1.5	325/415	2.2
	Inf.-11d	275/355	10.3	230/365	2.77	325/415	2.45
	Eff.-11d	275/355	4.31	230/365	1.51	325/415	2.24
	Inf.-12d	275/360	10.28	230/360	2.47	325/415	2.58
	Eff.-12d	275/355	4.17	230/365	1.81	325/415	2.24
	Inf.-13d	275/355	10.23	230/360	2.34	325/405	2.63
	Eff.-13d	275/360	3.58	230/360	1.56	325/415	1.94
	Inf.-15d	275/355	10.96	230/365	2.84	325/415	2.79
	Eff.-15d	275/360	4.33	230/360	1.71	325/420	2.06

---

Generally, the change of the fluorescence peak intensities in the leachate before and after EGSB treatment is an indication of the biodegradation or transformation of the fluorescing materials. As given in Table 2, the fluorescence intensities of tryptophan protein-like and aromatic protein-like substances described respectively by peaks A and B for the effluent decreased significantly compared with the influent, however, the humic-like substances indicated by peak C maintained relatively stable regardless of the operational conditions. The experimental results clearly implied that protein-like substances in the leachate were more easily biodegraded for biogas production than humic substances. Moreover, the fluorescence intensities in the effluent from reactor B were weaker than those for reactor A, which obviously hinted that reactor B had higher efficiency in the biodegradation of organic matter and biogas recovery. The enhanced performance of an EGSB in this study was closely related to better sludge biological activity. The powerful sludge granules decomposed the high molecular fluorescing substances in the leachate, and therefore could enhance the following anaerobic digestion performance of the EGSB bioreactor.

3.3.2 PARAFAC analysis of EEM spectra. PARAFAC analysis was further employed to decompose fluorescence EEM spectra of the effluent samples from bioreactors A and B. Fig. 2 shows the contour plots of the three fluorescence components and the excitation and emission loadings of each component derived from the DOMFluor-ARAFAC model. The tryptophan protein-like fluorescence peak, which was located at the E_x/E_m of 275–280/355–365 nm, was detected in component 1, possibly originated from cell lysis. Li et al.³² and Ni et al.³⁹ observed a similar component in extracellular polymeric substances extracted from sludge. Component 2 consisted of three long emission wavelength humic-like fluorescence peaks, which were centered at 240/450, 295/450 and 340/450 nm. According to Coble,⁴⁰ the peak at 295/450 nm was related to marine humic-like substances; compared with the observation of Coble, the peak in this study showed a slight shift towards longer wavelengths (red-shifted). The peak at 340/450 nm, according to Cai et al.,⁴¹ represented continental humic-like substances. In addition, one fluorescence peak at 320/320 nm was identified in component 3. Based on the protocol of Chen et al.,⁴² however, the peak did not seem to belong to any excitation–emission region. Several pieces of work have been conducted in recent years regarding the characterization of leachate with the help of a three-dimensional excitation–emission matrix combined with PARAFAC analysis.^43–46 To date, very limited information about this component is available in the literature, with the exception of Yan et al.,⁴⁷ in analyzing marine chromophoric dissolved organic matter (CDOM) via PARAFAC analysis, and they attributed this component to humic-like substances.

Fig. 2 Contour plots of four components (left) identified using the DOMFluor-ARAFAC model and the split-half validation (right) of four components model (excitation to the left of emission spectra): (a) bioreactor A, and (b) bioreactor B.

The ARAFAC analysis provides additional quantitative information describing the distribution of the individual components in each sample.⁴⁶ The F_max of each component related to the relative concentration of the corresponding component in each effluent EEM sample can be derived from the DOMFluor-ARAFAC model. The resulting F_max values are illustrated in Fig. 3. As can be seen, overall, the content of all of the components in the effluent samples from bioreactor B were much lower than those from bioreactor A. The componential characterizations clearly reveal that the introduction of the CO₂-stripping unit affected the transformation of dissolved organic matter present in the leachate and the overall performance of the EGSB bioreactor. Application of the CO₂-stripping unit alleviated the adverse impact of calcium, increased the bacterial activity, and accordingly stimulated the biodegradation and removal efficiency of the organic matter. As a result of this, the EGSB bioreactor with the CO₂-stripping unit could work more stably and efficiently.

Fig. 3 The maximum fluorescence intensities (F_max) in the three components collected from the effluent samples of bioreactors A (a) and B (b) after 15 days of operation.

3.4. Characterizations of the clogging materials and sludge granules

To better understand the mechanisms observed within the digestion experiments (especially reactor B) performed in this study, the calcium removal and microstructural properties of the clogging materials and sludge granules collected from the EGSB bioreactors were analyzed systematically.

3.4.1 Crystalline phase analysis. The clogging materials and sludge granules were characterized through XRD analysis (Fig. 4). The XRD patterns, as illustrated in Fig. 4a, depicted that the clogging materials were composed principally of CaCO₃, Mg_0.064Ca_0.936CO₃, and Mg_0.03Ca_0.97CO₃, which revealed that the dominant cause of clogging in the carbonation unit was calcium carbonate precipitates. It can be also found that the characteristic peaks for calcium carbonate drastically increased as the operation time elapsed, implying that the carbonation through biogas recirculation exerted a relatively positive role in the deposition of calcium carbonate. Obviously, biogas produced from the anaerobic digestion of the leachate can be recirculated for calcium precipitation and removal.

Fig. 4 XRD pattern of the clogging materials (a) from the CO₂-capturing system and sludge granules (b) from the EGSB bioreactors.

Regarding the sludge granules, the results of XRD analysis, presented in Fig. 4b, revealed that the granules from reactor A consisted of mostly calcium carbonate. The calcium content reached up to 305.96 mg g⁻¹ TS from an initial value of 86.06 mg g⁻¹ TS (seed sludge) on day 15 (Table 3), increasing by 255.5%, which is mainly attributed to the formation of calcium precipitates in the granular sludge. The precipitation of calcium carbonate in granular sludge was also observed by Ye et al.²⁵ and Yu et al.¹¹ in an UASB reactor, Lozecznik et al.²⁸ in an anaerobic sequencing batch reactor (ASBR), as well as Liu et al.⁴ in an EGSB bioreactor. In contrast, the level of calcium in the granular sludge taken from reactor B remained relatively stable at 87.63 mg g⁻¹ TS, similar to that in the seed sludge, confirming the positive role of carbonation on calcium removal. Therefore, to avoid calcium inhibition on the long-term stability of an EGSB bioreactor, leachate pretreatment, such as carbonation through biogas recirculation or chemical precipitation should be conducted.

Table 3 Metal content values of the sludge granules in the EGSB bioreactors (mg g⁻¹ TS)

Type of sludge	Ca	Mg
Seed sludge	86.06	5.77
Sludge from reactor A	305.96	4.30
Sludge from reactor B	87.63	6.02

---

In addition to calcium, magnesium might also form precipitates together with calcium in the granules. However, it is interesting to find that no marked increase of magnesium in granules from both reactors was observed compared with the seed sludge (5.77 mg g⁻¹ TS), and magnesium content values of around 4.30–6.02 mg g⁻¹ TS were measured in both types of granules (Table 3), indicating that the inhibitory effect of magnesium was negligible.

3.4.2 Functional group identification. Fig. 5 depicts the FTIR patterns of the clogging materials and granular sludge. As shown in Fig. 5a, the spectra of the clogging materials exhibited characteristic bands at about 3409 cm⁻¹, possibly due to the O–H stretching vibration (ν₁). The bands at 2925 cm⁻¹ corresponded to the asymmetric stretching vibration of the CH₂ of aliphatic structures. The typical bands located at 1654 and 1540 cm⁻¹ were attributed to the stretching and deformation vibration of the C=O, C–N and N–H peptidic bond of proteins.^48,49 The band present at 1429 cm⁻¹ was associated with the C=O stretching vibrations of carboxylates and OH deformation vibration of alcohols and phenols. The strong bands of ν₂ and ν₁-CO²⁻₃ were found at around 1473 and 871 cm⁻¹ respectively, mainly due to the presence of CaCO₃. The band at 1049 cm⁻¹ was linked to the C–O–C and C–O vibration of polysaccharides. These absorbance patterns at 1473 and 871 cm⁻¹ provided further evidence for the accumulation of CaCO₃ in the form of clogging materials as a result of the carbonation reaction. Moreover, the peak intensities of the functional groups, i.e., ν₂ and ν₁-CO²⁻₃ resulting from the presence of CaCO₃, gradually increased following the prolonged carbonation time, indicating more calcium was efficiently precipitated and removed during the carbonation process. It was also supported by the results observed in the X-ray diffraction analysis (Fig. 5a).

Fig. 5 FTIR spectra of the clogging materials (a) from the CO₂-capturing system and the sludge granules (b) from the EGSB bioreactors.

To reveal the influence of calcium on the functional groups of the sludge granules, the granules collected from both reactors were characterized using FTIR analysis. The results from Fig. 5b showed that the spectral bands at 1458 cm⁻¹ and 873 cm⁻¹ corresponding to ν₂ and ν₁-CO²⁻₃ of CaCO₃ from the granules in reactor B were relatively weaker compared to those in reactor A, which clearly revealed that the adherence of calcium inside or onto the surface of granular sludge was effectively inhibited because of the efficient calcium removal during the carbonation process. These results further demonstrated the crucial role played by biogas recirculation in controlling the negative effect of calcium ions. In addition, the strong bands located at 1653, 1540 and 1047 cm⁻¹, representing proteins and polysaccharides respectively, confirmed the presence of a high content of extracellular polymeric substances (EPS). EPS, present outside bacterial cells and in the interior of microbial aggregates, play a key role in protecting the microorganisms in granules against toxic compounds through sorption and reaction.⁵⁰ The functional groups of EPS such as hydroxyl, carboxyl, and amine can generate a negative surface charge. The negatively charged groups can act as calcium binding sites to bind calcium ions in the influent, thereby partially reducing the toxicity to anaerobic granular sludge in the EGSB bioreactor.

3.4.3 SEM observations. To locate the CaCO₃ crystals in the granular sludge in more detail, SEM analysis of the granules collected from reactors A and B during the operation was performed (Fig. 6). As shown in Fig. 6a, large numbers of hexagonal-like CaCO₃ phases were found to extensively form within the granules from reactor A. Apart from the CaCO₃ encapsulated in the granules, CaCO₃ crystals could also be identified between granules. The surface topography of the granules was densely packed, rigid and rough, suggesting serious accumulation and precipitation of calcium. When in the presence of biogas recirculation, the granular sludge seemed more regular in shape with a smooth exterior surface, and there were nearly no bulk hexagonal crystals detected (Fig. 6b), agreeing well with the XRD and FTIR observations (Fig. 4 and 5). A previous study by Ye et al.,²⁵ reported that Methanosaeta-like species, known to exert a crucial role in the initial granulation, were predominant in the core of developed granules. Therefore, the higher CaCO₃ deposition would form a hard crust in the anaerobic granules, which inevitably reduced the activity of the methanogens in the interior layer. However, the adverse effect of calcium seemed to be effectively controlled when biogas recirculation was applied. Biogas recirculation through a carbonation reaction induced the reduced calcium level in the influent, which promoted Methanosaeta-like species survival and faster granule formation.¹⁸ As a consequence, a stable and improved performance of the EGSB bioreactor was achieved.

Fig. 6 SEM spectra of the sludge granules from bioreactor A (a) and B (b).

Based on the abovementioned results, it is therefore very reasonable to state that biogas recirculation is a cost-effective alternative for the removal of calcium and the maintenance of bacterial activity and it will be a method to eliminate the calcium inhibition of granular sludge activity and subsequent anaerobic digestion during the EGSB process of fresh leachate with high calcium concentrations.

4. Conclusion

A CO₂-stripping system was developed in a leachate-fed EGSB bioreactor, with the purpose of stimulating organic matter removal and sludge granulation. The continuous performance of the bioreactor was greatly improved when the CO₂-stripping unit was employed as the pretreatment to carbonize calcium-rich leachate. This strategy induced a considerable increase in methane yield, as well as COD and TN removal under steady-state conditions. Further analysis using 3D-EEM combined with PARAFAC, XRD, FT-IR and SEM evidenced that the introduction of the CO₂-stripping unit promoted calcium precipitation and reduced the adverse effects on clogging and granular activity, which subsequently enhanced the methanogenic efficiency and overall process stability. The general principle described here can also be transferred into related calcium-rich wastewater treatment systems for upgrading calcium removal, process stabilization, and the sustainable production of bioenergy. This study provides an in-depth insight into the improvement of applying a CO₂-stripping unit in organic matter degradation and sludge granulation improvement, advancing its practical applications.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (21377079, 51578329), the key research project from the Science and Technology Commission of Shanghai Municipality (13DZ0511600) and the Program for Innovative Research Team in University (No. IRT13078).

References

1. J. Y. Liu, J. H. Luo, J. Z. Zhou, Q. Liu, G. R. Qian and Z. P. Xu, Bioresour. Technol., 2012, 113, 239–243.
2. S. M. Tauseef, T. Abbasi and S. A. Abbasi, Renewable Sustainable Energy Rev., 2013, 19, 704–741.
3. Y. J. Zhang, L. Yan, L. Chi, X. H. Long, Z. J. Mei and Z. J. Zhang, J. Environ. Sci., 2008, 20, 658–663.
4. J. Y. Liu, J. Hu, J. P. Zhong, J. H. Luo, A. H. Zhao, F. Liu, R. J. Hong, G. R. Qian and Z. P. Xu, Bioresour. Technol., 2011, 102, 5466–5472.
5. A. Ahmad, R. Ghufran and Z. Abd Wahid, J. Hazard. Mater., 2011, 198, 40–48.
6. D. Puyol, A. F. Mohedano, J. L. Sanz and J. J. Rodriguez, Chemosphere, 2009, 76, 1192–1198.
7. M. X. Zheng, K. J. Wang, J. E. Zuo, Z. Yan, H. Fang and J. W. Yu, Bioresour. Technol., 2012, 107, 33–40.
8. T. Abbasi and S. A. Abbasi, Renewable Sustainable Energy Rev., 2012, 16, 1696–1708.
9. X. Lu, G. Zhen, M. Chen, K. Kubota and Y. Y. Li, Bioresour. Technol., 2015, 198, 691–700.
10. Y. J. Liu and D. D. Sun, Process Biochem., 2011, 46, 987–992.
11. H. Q. Yu, J. H. Tay and H. H. P. Fang, Water Res., 2001, 35, 1052–1060.
12. R. P. Beaven, A. P. Hudson, K. Knox, W. Powrie and J. P. Robinson, Waste Manage., 2013, 33, 431–444.
13. E. P. A. van Langerak, G. Gonzalez-Gil, A. van Aelst, J. B. van Lier, H. V. M. Hamelers and G. Lettinga, Water Res., 1998, 32, 1255–1263.
14. A. Pevere, G. Guibaud, E. D. van Hullebusch, W. Boughzala and P. N. L. Lens, Colloids Surf., A, 2007, 306, 142–149.
15. H. Y. Jo, J. H. Ahn and H. Jo, J. Hazard. Mater., 2012, 241, 127–136.
16. O. Nir, M. Herzberg, A. Sweity, L. Birnhack and O. Lahav, Chem. Eng. J., 2012, 187, 275–282.
17. Y. H. Kim, S. H. Yeom, J. Y. Ryu and B. K. Song, Process Biochem., 2004, 39, 1393–1399.
18. J. Luo, X. Lu, J. Liu, G. Qian and Y. Lu, Bioresour. Technol., 2014, 173, 317–323.
19. G. Zhen, X. Lu, Y. Y. Li, Y. Zhao, B. Wang, Y. Song, X. Chai, D. Niu and X. Cao, Bioresour. Technol., 2012, 119, 7–14.
20. G. H. Yu, P. Jing He and L. M. Shao, Water Res., 2010, 44, 797–806.
21. X. Guo, X. He, H. Zhang, Y. Deng, L. Chen and J. Jiang, Microchem. J., 2012, 102, 115–122.
22. G. Zhen, X. Lu, Y. Y. Li and Y. Zhao, Appl. Energy, 2014, 128, 93–102.
23. APHA, American Public Health Association, Washington, D.C., 1998.
24. H. Q. Yu, H. H. P. Fang and J. H. Tay, Chemosphere, 2001, 44, 31–36.
25. J. X. Ye, Y. J. Mu, X. Cheng and D. Z. Sun, Bioresour. Technol., 2011, 102, 5498–5503.
26. Y. Chen, J. J. Cheng and K. S. Creamer, Bioresour. Technol., 2008, 99, 4044–4064.
27. H. Timur and I. Ozturk, Water Res., 1999, 33, 3225–3230.
28. S. Lozecznik, R. Sparling, S. P. Clark, J. F. vanGulck and J. A. Oleszkiewicz, Bioresour. Technol., 2012, 104, 37–43.
29. T. Liu, Z. L. Chen, W. Z. Yu and S. J. You, Water Res., 2011, 45, 2111–2121.
30. R. H. Dai, Y. Liu, X. Liu, X. D. Zhang, C. Y. Zeng and L. Li, J. Hazard. Mater., 2011, 192, 1161–1170.
31. J. L. Wang and K. Jing, Process Biochem., 2005, 40, 1973–1978.
32. W. H. Li, G. P. Sheng, X. W. Liu and H. Q. Yu, Water Res., 2008, 42, 3173–3181.
33. L. Zhu, H. Y. Qi, M. L. Lv, Y. Kong, Y. W. Yu and X. Y. Xu, Bioresour. Technol., 2012, 124, 455–459.
34. Y. X. Zhang, P. Y. Zhang, J. B. Guo, W. F. Ma and L. P. Xiao, Bioresour. Technol., 2013, 135, 616–621.
35. S. S. Adav and D. J. Lee, J. Taiwan Inst. Chem. Eng., 2011, 42, 645–651.
36. L. Bu, K. Wang, Q. L. Zhao, L. L. Wei, J. Zhang and J. C. Yang, J. Hazard. Mater., 2010, 179, 1096–1105.
37. W. Chen, P. Westerhoff, J. A. Leenheer and K. Booksh, Environ. Sci. Technol., 2003, 37, 5701–5710.
38. Z. W. Wang, Z. C. Wu and S. J. Tang, Water Res., 2009, 43, 1533–1540.
39. B. J. Ni, W. M. Xie, S. G. Liu, H. Q. Yu, Y. Z. Wang, G. Wang and X. L. Dai, Water Res., 2009, 43, 751–761.
40. P. G. Coble, Mar. Chem., 1996, 51, 325–346.
41. W. Cai, X. Xu, X. Du, H. Zhu and G. Luo, Res. J. Environ. Sci., 2012, 25, 276–281 , in Chinese.
42. W. Chen, P. Westerhoff, J. A. Leenheer and K. Booksh, Environ. Sci. Technol., 2003, 37, 5701–5710.
43. H. Wu, Z. Zhou, Y. Zhang, T. Chen, H. Wang and W. Lu, Bioresour. Technol., 2012, 110, 174–183.
44. J. Wu, H. Zhang, L.-M. Shao and P.-J. He, Environ. Pollut., 2012, 162, 63–71.
45. F. Lu, C.-H. Chang, D.-J. Lee, P.-J. He, L.-M. Shao and A. Su, Chemosphere, 2009, 74, 575–582.
46. X. S. He, B. D. Xi, X. Li, H. W. Pan, D. An, S. G. Bai, D. Li and D. Y. Cui, Chemosphere, 2013, 93, 2208–2215.
47. L. Yan, The degree of Master, Ocean University of China, 2012.
48. G. Zhen, X. Lu, B. Y. Wang, Y. C. Zhao, X. L. Chai, D. J. Niu, A. H. Zhao, Y. Y. Li, Y. Song and X. Y. Cao, Bioresour. Technol., 2012, 124, 29–36.
49. J. Laurent, M. Pierra, M. Casellas and C. Dagot, Chemosphere, 2009, 77, 771–777.
50. I. D. S. Henriques and N. G. Love, Water Res., 2007, 41, 4177–4185.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra26444h

---