Combined Pre-Treatment Technologies for Cleaning Biogas before Its Upgrading to Biomethane: An Italian Full-Scale Anaerobic Digester Case Study

Abstract

Biogas produced by anaerobic digestion contains different types of contaminants, and it is preferable to eliminate those contaminants before biogas’ energetic valorization or upgrading to biomethane as they are harmful to human health and detrimental to combustion engines. This study presents the biogas cleanup system optimized by an Italian full-scale anaerobic digester treating food waste (FW) and represented by micro-oxygenation, chemical scrubber, cooling, and activated carbon sections. The cleaned biogas is upgraded to biomethane using a membrane-based upgrading unit and injected into the natural gas network for transport sector use. H₂S and volatile organic compound (VOC) concentration in raw biogas was reduced from an annual average value of 1207 ppm_v and 895 mg/Nm³, respectively, to below 0.1 mg/Nm³ in the final biomethane. In the summer, the H₂S average content in raw biogas was 833 ppm_v due to a greater presence of low-sulfur-containing vegetables in FW, while in the winter it was an average of 1581 ppm_v due to a larger portion of protein-containing FW. On the other hand, raw biogas VOC content in the winter was an average of 1149 mg/Nm³, with respect to 661 mg/Nm³ in the summer, due to the greater consumption of citrus fruits containing high amount of terpene compounds. The concentration of other trace contaminants, such as HCl, NH_3, and siloxanes, was lowered from 17, 36, and 0.6 mg/Nm³ in raw biogas, respectively, to below 0.1 mg/Nm³ in the final biomethane. All the considerations and evaluations underlying the technological and plant engineering choices together with the individuation of the best operating conditions are discussed.

1. Introduction

To protect the climate and natural environment, world policies are increasingly pushing toward the transition from fossil fuels-based technologies to energy carriers obtained using biomasses and renewable sources. At the same time, the need to avoid high environmental impacts deriving from both resource depletion and pollution of waste landfilling is increasingly felt.

Technologies for the conversion of waste to energy and fuels can be applied to valorize and transform municipal solid waste [1] and, in particular, food waste (FW) [2] into steam, electricity, liquid, and gaseous fuels through thermochemical and biochemical processes. Although there are currently several technologies available for FW management [2], recycling FW through anaerobic digestion (AD) produces both an energy vector (biogas) [3] and an organic fertilizer (digestate) generating lower greenhouse gas emission than landfill and incineration [4]. Unlike thermal treatment methods, such as FW incineration where organic matter is lost as CO₂, FW AD is considered a better recycling process by both European [5] and the United States [6] standards.

Biogas produced by AD is composed of methane (CH₄, 45–70%), carbon dioxide (CO₂, 30–55%), and trace contaminants such as hydrogen sulfide (H₂S), volatile organic compounds (VOCs), ammonia (NH₃), nitrogen (N₂), oxygen (O₂), and water vapor. Biogas bulk composition and trace compound content are mainly linked to the type and origin of the organic matter used in the AD process [7,8]. Biogas can have multiple applications [9]: it can be directly applied in boilers and central heat and power (CHP) engines to generate heat, mechanical work, and electric energy or used in fuel cells. By moisture and trace contaminant removal (biogas cleaning pre-treatment) followed by CO₂ separation (biogas upgrading) [10,11], biogas can also be converted to biomethane (CH₄, 97–99%), which can be utilized as an alternative to natural gas, fed into distribution grid for domestic and industrial usage, or employed as a gaseous or liquid fuel for motor power vehicles. It has been reported that biomethane grid injection is the most profitable solution if connection costs are low [12] and upgrading biogas to biomethane for the transport sector has less impact on the environment than directly burning it in a CHP unit [13].

The need to remove trace contaminants from biogas [14] is linked to several factors: (i) high levels of compounds harmful to human health and the environment, such as non-burnt hydrocarbons, polycyclic aromatic hydrocarbons, VOCs, NOx, SOx, CO, and CO₂, are released in exhaust gases [15,16,17] during biogas combustion; (ii) corrosion of pipeline infrastructure and damage of internal combustion engine components occur due to the activity of NH₃, siloxanes, H₂S and its Sox, and H₂SO₄ oxidation products [18]; (iii) biogas impurities dilute CH₄ content lowering its lower calorific value [16]; (iv) biogas trace compounds can affect proper functioning of the fuel cells degrading significantly any catalytic process [7,18]; (v) trace contaminants have to comply with the limits set by the European standards for biomethane use in transport [19] and for injection in the natural gas network [20]; (vi) some biogas upgrading technologies, such as pressure swing adsorption and membrane separation, require the removal of compounds such as H₂S and VOCs, which could poison the material constituting the separation system [16].

The most important and studied biogas trace contaminant is H₂S. H₂S formation in FW AD is related to the biodegradation of sulfur-containing compounds, such as inorganic SO₄²⁻ and proteins containing methionine and cysteine, through known biological mechanisms [21]. Structures and functions of the microbial community responsible for H₂S formation from FW AD are also known [22]. H₂S is a toxic gas corrosive to metal equipment, poisonous for fuel cells [23], and harmful to human health even at concentrations of 10–100 ppm [8]. H₂S and its SOx combustion products are dangerous not only for humans but also for the environment as they are the main precursors of acid rain [23]. H₂S abatement and removal can be achieved using various technological strategies [21,24] that can be broadly divided into three groups: in situ desulfurization processes, ex situ biogas post-treatment, and feedstock pre-treatment.

In situ H₂S removal is performed through the precipitation of elemental sulfur localized within the digester using micro-aeration and iron dosing [25]. At the industrial level, micro-aeration showed higher desulfurization efficiencies (68–99%) than iron dosing (35–72%). Both techniques seem to be cheaper than most ex situ desulfurization methods [26] although the economic feasibility is essentially indicative due to a lack of data [25]. However, the main disadvantage of these techniques is finding the right operating conditions since an excessive dosage of iron and oxygen can lead to disturbances or inhibition of the AD process.

Ex situ biogas post-treatment is the most used desulfurization technique and is performed using (i) absorption technologies according to H₂S different solubility in water and organic solvents with respect to CH₄ (removal efficiency: 17–100%); (ii) adsorption-based system using H₂S adsorbents at high pressure and temperature and with high specific area and porosity, such as iron oxide-based material, activated carbon, and zeolites (breakthrough capacity: from several mg to about 10 g of adsorbed H₂S per each g of adsorbent material); (iii) biofiltration and biotrickling filtration performing biological H₂S oxidation by microorganisms grown on organic media or immobilized on chemically inert materials (removal efficiency: 60–100% and 85–100% for biofiltration and biotrickling filtration, respectively); (iv) polymeric membranes filtration (output H₂S > 50 ppm_v), (v) system based on the set of several ex situ techniques, such as physical, chemical, and biological scrubbing (removal efficiency: 80–99%).

The purpose of the feedstock pre-treatment is to remove, via liquid–solid separation processes, the sulfur contained in the organic substrate before it enters the digester in order to prevent H₂S formation during the AD process. By using inorganic iron salts, hydrogen peroxide, calcium oxide, and hydroxide, it is possible to precipitate soluble sulfur compounds. However, since the critical step of this technique is the separation of precipitated sulfur derivatives from the organic material before it enters the digester, feedstock pre-treatment is restricted to liquid substrates such as wastewater.

Aromatic and halogenated hydrocarbons, alkanes, terpenes, organosilicon derivatives, oxygenated and sulfur compounds represent most of the non-CH₄ VOCs produced by industrial FW AD that can cause serious harm to both the environment and human health [27]. Due to their different chemical-physical characteristics, biogas VOCs can also cause various problems to the equipment and pipeline. Halogen derivatives cause severe metal corrosion [16]; terpenes affect the performance of adsorption-based and membrane-based biogas upgrading [7]; siloxanes are harmful to fuel cells [23] and, after combustion, cause SiO₂ deposition on equipment lowering its life-time and performance [28]. The most used and cheaper industrial method to remove non-CH₄ VOCs is based on their adsorption on activated carbon [28], while other physical-chemical technologies include chemical oxidation in the liquid phase [29], absorption into organic solvents [30] or strong acids, and deep chilling [16]. Depending on the adopted technique, the VOC removal efficiency can reach up to 100%.

NH₃ is a toxic gas [8] present in biogas produced by FW AD. Moreover, the products of NH₃-containing biogas combustion (NOx) are among the most relevant compounds for air pollution. Due to its high solubility in water in the form of NH₄⁺, NH₃ can be easily removed by water and acidic-water washing or, alternatively, adsorbed onto an activated carbon system [16].

Water vapor contained in biogas can lower its heat value, thus reducing the energy performance of combustion engines. Moreover, the condensation of water and hydrophilic compounds, such as H₂S, inside the combustion engines can lead to corrosion problems and equipment malfunctions. Different methods can be used to dehydrate biogas water, but the most used are condensation by cooling and adsorption by dryers [16].

From an industrial point of view, the choice of the biogas pre-treatment system is not only related to the highest yield of impurity abatement efficiency but also to a series of considerations and needs for the full-scale AD plant. Of course, the first evaluation is linked to the level of contaminant concentration to be achieved as a function of biogas or biomethane end-use. Then, economic evaluation based on the initial costs of the equipment and the fixed management costs is another important factor. Management costs depend on the number of operators necessary for the equipment functioning, machinery maintenance, and days of plant downtime, as well as energy, water, and chemical consumption. Using automated technologies that do not require specialized operators and are easy to maintain is certainly advantageous. Likewise, using space-saving equipment is preferable. Consequently, the adoption of a specific biogas pre-treatment system is dictated by the balance of all the plant’s needs.

The raw biogas purification process has been extensively studied on a laboratory and pilot scale, but only a few data on an industrial scale are available. Furthermore, the published studies relating to industrial-scale plants have specifically analyzed only some steps of the entire biogas purification process [31]. The aim and novelty of this work is to describe the logical path followed for the choice of the technological solutions adopted by a currently working large-scale FW AD plant for the pre-treatment of biogas before its upgrading to biomethane. Compared with other published studies, a detailed analysis of each cleaning step of the biogas purification process at an industrial-scale plant is reported in this work. The reasons underlying the adopted technological choices and the optimization of the various steps making up the biogas pre-treatment system are discussed. Starting from the analytical characterization of raw biogas, the best operating conditions have been determined through the quantification of biogas components after each stage of the cleaning system.

2. Materials and Methods

A description of the characteristics of the FW AD process was reported elsewhere [32]. Briefly, the AD plant treats 40,000 tons/year of household FW produced and separately collected by citizens of the province of Cosenza (Italy). AD plant consists of two independent and horizontal reactors, each with an operating volume of 1100–1200 m³. A pre-treatment unit allows for the elimination of plastic bags and shredding FW down to a particle size of 1–30 mm. The pre-treated FW with dry matter of about 30% is then fed into the two anaerobic reactors. FW feeding was 51–54 tons/d and 55–58 tons/d in the summer and winter, respectively. AD process is conducted at 42–45 °C, and the hydraulic retention time of 21–22 days is ensured by a slow and continuous plug-flow movement. AD process produces about 180 Sm³ of raw biogas for each ton of pre-treated FW with a CH₄ content of about 59%. The produced raw biogas is cleaned according to the biogas pre-treatment scheme adopted by the plant, as reported in Figure 1. Analytical instrumentation, methods, and analyzed compounds are reported in Table S1. Biogas was automatically sampled and analyzed every eight hours, with the determination of CH₄ (%vol), CO₂ (%vol), O₂ (%vol), and H₂S (ppm_v). Quantification of VOCs, NH₃, N₂, inorganic contaminants, and water vapor was carried out approximately every 40 days over the course of the examined year.

3. Results and Discussion

The goal of the AD plant was constant production of about 450–500 Sm³/h of biomethane to be fed into the national distribution network for use in the transport sector provided that the limits imposed by Italian legislation [19,33] were respected.

The membrane-based technology was chosen for the biogas upgrading (CO₂ removal) system. Indeed, the advantages of using membranes for upgrading biogas to biomethane compared with other technologies are known [34] and make this technique the most used in Europe. The used membranes are selective for CO₂ permeation, leaving CH₄ in the retentate flow (>98% content) but unable to separate biogas contaminants that remain in the bio-CH₄ stream. Moreover, trace contaminant compounds in biogas can damage membranes and affect the performance of the CH₄/CO₂ separation process. For this reason, a pre-treatment system was optimized for the purification of biogas before it was sent to the membrane separation unit. The adopted biogas cleaning system consists of several purification technologies placed in series according to Figure 1 (details are discussed in Section 3.2, Section 3.3, Section 3.4 and Section 3.5). This choice was mainly based on the type of contaminants quantified in the biogas and their national regulatory limits to be respected in the final biomethane fed into the national distribution network. Details of the contaminants determined in the biogas produced by the industrial AD of food waste are discussed in Section 3.1.

3.1. Trace Compounds in Biogas

The results related to the analyses carried out on raw biogas are reported in

Table 1. Raw biogas composition.

	Mean a	SD b	Min	Max
CH4 (%vol)	59.4	1.0	58.1	60.5
CO2 (%vol)	40.1	0.9	39.2	40.8
H2S (ppmv)	1207	421	613	1740
VOCs (mg/Nm3)	895	304	432	1425
H2O (%vol)	0.3	0.2	0.1	0.6
N2 (%vol)	0.2	0.1	0.1	0.3
HCl (mg/Nm3)	17	4	12	20
NH3 (mg/Nm3)	36	7	30	46
Siloxanes (mg Si/Nm3)	0.6	0.2	0.2	0.8

^a annual mean; ^b SD = standard deviation.

. As biogas production and composition depend on many factors, such as operational parameters of the AD process and seasonal variability of FW characteristics, the mean, minimum, and maximum values of the parameters measured over a year are reported in Table 1.

As demonstrated [32], the characteristics of the FW used by the plant to feed the anaerobic digester were different in the summer and winter. In particular, in the warmer months, total volatile solid (TVS) content in FW was higher (TVS = 23.1%) than in the colder months (TVS = 21.0%), while pH value was higher in the winter (pH = 5.7) than in the summer (pH = 4.5). Also, the temperature of the AD process reached values of 44–45 °C in the summer and 42–43 °C in the winter. The variation in both FW composition and AD operational parameters led to different trends of trace contaminant formation in biogas.

The varying levels of H₂S concentration measured in biogas during the examined year were linked to the variable amount of sulfur-containing FW constituents. It is known that AD of FW made up of vegetables such as potatoes, lettuces, and tomatoes produces no measurable quantity of H₂S [35]. In the months from April to November, the H₂S content in the biogas was lower (mean H₂S concentration: 833 ppm_v) than in the other months (December–March, mean H₂S concentration: 1581 ppm_v) due to greater consumption of low-sulfur-containing vegetables, such as lettuce, tomatoes, zucchini, cucumbers, eggplants, carrots, potatoes, and fruits, including peaches and watermelons, widely produced in this area. In the colder months (December–March), on the other hand, the higher H₂S content in biogas was due to the AD of a larger amount of protein-containing waste, mainly composed of meat remains. Mean, minimum, and maximum values of H₂S measured in biogas were comparable to those observed by other similar full-scale AD plants treating FW [7].

Total VOC content in biogas also strongly depended on FW characteristics. From December to May (

Table 2. VOCs speciation (raw biogas).

Compound (mg/Nm3)	December–May			June–November			Annual
	Min	Max	Mean	Min	Max	Mean	Mean
Total VOCs	582	1425	1149	432	746	661	895
limonene	199	670	449	58	128	94	272
α-pinene	113	237	173	26	71	42	108
β-pinene	118	218	152	37	80	44	98
p-cymene	89	262	178	14	85	38	108
γ-terpinene	0	1	0	0	15	4	2
myrcene	0	0	0	0	4	1	<1
Δ-3-carene	0	1	0	0	10	3	1
propane	0	8	2	6	40	27	15
butane	0	10	3	3	45	28	15
pentane	0	2	1	0	5	2	1
hexane	0	6	1	8	25	17	9
heptane	0	7	2	10	32	24	13
2-methyl-pentane	0	1	0	0	3	1	<1
benzene	2	32	23	26	112	42	33
toluene	1	24	17	22	34	25	21
ethyl-benzene	1	26	22	22	39	26	24
xylene a	7	41	26	37	82	51	39
mesitylene	0	5	1	0	1	0	1
cumene	3	40	28	34	72	43	36
acetone	5	14	9	32	64	35	22
methyl-ethyl-ketone	18	38	32	79	147	122	77
hexanal	1	12	8	0	3	1	3
heptanal	1	10	4	0	1	0	1
octamethylcyclotetrasiloxane	0	1	0	0	2	1	<1
decamethylcyclopentasiloxane	0	2	0	0	3	1	<1

^a sum of isomers.

), a higher mean concentration of VOCs (1149 mg/Nm³) was measured compared with the one detected in the other months (June–November, mean VOCs concentration: 661 mg/Nm³).

The higher VOC concentration corresponded to the period of production and consumption of citrus fruits, such as oranges and tangerines, of which the area is an important producer. In fact, from speciation analysis of the VOCs (Table 2), it was observed that the amount of terpenes in biogas and, in particular, of limonene, α-pinene, β-pinene, and p-cymene, which citrus peel is rich in, considerably increased in the period between December and May, while the concentrations of aromatic, carbonyl, and hydrocarbon compounds barely varied throughout the year.

A similar trend was observed in the study of a full-size FW AD [27], in which the greatest presence of terpenes in biogas, mainly represented by limonene, was measured in the spring period. The average concentration of limonene and p-cymene was lower than that reported by Esposito et al. [31], while that of β-pinene was similar. Instead, the average annual concentration of several terpenes, such as limonene, α-pinene, β-pinene, and p-cymene, was higher compared with that estimated in other FW AD plants. Calbry-Muzyka et al. [7] measured an average terpene concentration of about 187–298 mg/Nm³ (limonene: 158–187 mg/Nm³), while the value obtained by Zheng et al. [27] was 1.5 mg/m³. However, in the case studied by Zheng et al. [27], most of the terpenes were eliminated in the hydrothermal pre-treatment of FW before they underwent the AD process. As in the study reported by Esposito et al. [31], carbonyl compounds were mainly represented by acetone and methyl-ethyl-ketone, with an average concentration of 22 mg/Nm³ and 77 mg/Nm³, respectively, which was also in line with the values reported by Calbry-Muzyka et al. [7]. The amounts of some other VOCs in biogas were different from those reported in other studies. The concentrations of aliphatic and aromatic hydrocarbon compounds were 28 mg/Nm³ and 112 mg/Nm³, respectively. These values were found to be higher than those measured in other full-scale FW digesters [17]. Unlike other studies carried out on biogas produced by FW full-scale digesters [7,17,27], no halocarbon compounds were quantified.

Gaseous ammonia annual average concentration, which was 36 mg/Nm³, was higher than that recorded in similar FW AD plants (0.2–6.5 mg/m³) [7]. The mean concentration of N₂ (0.2%vol), water vapor (0.3%vol), HCl (17 mg/Nm³), and siloxanes (0.6 mg Si/Nm³) also fell within the range of typical FW AD biogas values.

3.2. Micro-Oxygenation

The first biogas purification step was carried out by insufflating into the digester headspace a low quantity of oxygen with a purity > 95%, produced by a PSA-based O₂ generator. It is known that micro-aeration and micro-oxygenation allow for the bio-oxidation of biogas H₂S to elemental sulfur using the biological reactions of sulfide-oxidizing bacteria in the digester headspace [36]. The choice not to insufflate air was mainly attributable to the high quantity of N₂, which would have been introduced in the reactor without any possibility of being separated by the membrane-based upgrading system. Consequently, N₂ would have diluted biogas, decreasing CH₄ concentration and lowering the higher heating value (HHV) and the Wobbe index (WI), bringing them in the final biomethane below the minimum limits established by the national legislation (HHV: 34.95–45.28 MJ/Sm³; WI: 47.31–52.33 MJ/Sm³). It is known that introducing O₂ into the headspace and not into the liquid phase of the digester minimizes biogas dilution, limits the need for O₂ overdose due to the co-oxidation of organic compounds, avoids methanogenic inhibition, and reduces operating costs [25]. It was also demonstrated [37] that using 95% O₂ produced by a PSA generator, despite the higher fixed costs of the initial equipment, was more profitable than using pure O₂, air, or FeCl₃ thanks to both lower operating costs and long-term profitability.

In order to evaluate the efficiency of H₂S oxidation and to optimize O₂ volume to be blown into the headspace of an anaerobic digester, some tests were performed. Since O₂ cannot be removed by membrane-based upgrading systems, a test range between 0.1 and 0.6% (vol) of O₂ was chosen based on the maximum O₂ concentration allowed in the final biomethane according to national legislation (O₂ ≤ 0.6%). The tests were performed over 5 weeks, in which biogas production and H₂S concentration remained almost constant, equal to 468 ± 19 Sm³/h and 860 ± 24 ppm_v, respectively.

The results of O₂ insufflation tests are reported in

Table 3. Results of O₂ insufflation test.

Entry	O2 (%vol ± 0.01)	H2S (ppmv)	H2S Removal (%)
1	0.0	860	//
2	0.1	506	41.2
3	0.2	319	62.9
4	0.3	264	69.3
5	0.4	268	68.8
6	0.5	262	69.5
7	0.6	268	68.8
8 a	0.3	161	81.3
9 a	0.4	137	84.1
10 a	0.5	135	84.3

^a Increased digester headspace volume by 20%.

.

The extent of biogas H₂S oxidation rose with an increasing O₂ concentration up to 0.3 %vol (entries 1–4), after which it remained constant reaching a plateau level (entries 4–7). In fact, the maximum oxidation yield of H₂S was about 69% and did not vary significantly with percentages of insufflated O₂ volume between 0.3% and 0.6% (entries 4–7). A direct comparison with the desulfurization efficiency obtained by similar FW AD industrial plants was not possible due to a lack of published data. The H₂S removal efficiency varies from 68.2% up to 99.5% if full-scale AD plants treating manure and sewage sludge are taken into account [21,25]. The abatement efficiency of 69.3% achieved in this study is comparable to the minimum value reached by other full-scale AD plants treating manure as reported by Azizi et al. [25].

Further tests (entries 8–10) were carried out by lowering the operational volume of the digester in order to increase its headspace volume by 20%, maintaining the same operating conditions of the AD process. Insufflating an O₂ volume of 0.3%, a higher percentage of H₂S oxidation was measured (81.3%, entry 8) than that observed with a smaller headspace volume (69.3%, entry 4). By increasing the amount of insufflated O₂ to 0.4–0.5% vol (entries 9–10), the plateau level was reached with an H₂S removal efficiency of about 84%. In these cases, a higher digester headspace volume allowed for a greater contact time between O₂ and biogas H₂S for the sulfide oxidation process and a larger surface area colonized by sulfur-oxidizing bacteria for the H₂S biological oxidation. Indeed, it is known [25] that sulfur-oxidizing bacteria grow on the walls of the digester headspace where elemental sulfur is biologically produced and deposited. Consequently, the greater colonizable surface area resulted in higher bacterial activity and H₂S removal efficiency. On the other hand, increasing the digester headspace implied a smaller useful volume for the AD process with consequent lower biogas production and economic income.

In all of the tests carried out, an O₂ concentration equal to 50–55% of that introduced into the digester headspace was measured in the outgoing biogas flow. Therefore, it can be deduced that just 45–50% of the O₂ supplied to the digester headspace was used for the H₂S oxidation process. These data were in agreement with those reported by the study of Díaz et al. [38] according to which just 30–40% of the total O₂ supplied to the digester serves H₂S oxidation.

3.3. Chemical Scrubber

After the micro-oxygenation process, the biogas coming out from the digester was conveyed to a wet chemical scrubber. The biogas passed through the scrubber in a counter-current manner and H₂S was oxidized by a nebulized 40% FeCl₃ solution at pH = 8.6. pH was maintained at a constant value by automatically dosing a 30% NaOH solution.

A series of tests were carried out to determine the proper quantity of FeCl₃ to use and evaluate the H₂S abatement efficiency. The tests were performed with a constant biogas flow of 935 ± 27 Sm³/h entering the chemical scrubber. The relative results are reported in

Table 4. Results of biogas H₂S removal by chemical scrubber.

Entry	40% FeCl3 (L/h)	H2Sin (ppmv)	H2Sout (ppmv)	H2S Removal (%)	pH
1	0.5	260	96	63.1	8.6
2	0.8	255	65	74.5	8.6
3	1.0	257	52	79.8	8.6
4	1.5	252	33	86.9	8.6
5	2.0	263	34	87.1	8.6
6	3.0	261	32	87.7	8.6
7	1.5	270	52	80.7	8.3
8	1.5	259	25	90.3	9.0

.

The H₂S removal yield rose to about 87% with an increasing amount of FeCl₃ till 1.5 L/h (entries 1–4). A further increase in the quantity of used FeCl₃ did not generate a significant rise in removal efficiency (entries 5–6). This trend could be attributed to the fixed contact time and surface between biogas and FeCl₃ solution, determined by scrubber dimensions and biogas pressure. Thus, raising the amount of FeCl₃ did not increase time and surface contact between the substrate and the oxidizing agent and, consequently, did not enhance H₂S elimination efficiency.

Keeping the amount of used FeCl₃ constant (1.5 L/h), the correlation between H₂S abatement yield and pH variation was also studied (entries 7–8). At a pH of 8.3 and 9.0, H₂S removal of 80.7% and 90.3%, respectively, was estimated. Varying the pH from 8.6 (entry 4) to 9.0 (entry 8) led to a higher H₂S removal efficiency but also to an excessive consumption of NaOH solution, which was economically disadvantageous. The enhanced removal efficacy is due to the ability of alkaline solutions to further absorb traces of biogas H₂S [39]. The spent Fe²⁺ solution discharged from the scrubber was sent to an oxidation tank. Here, an air flow was bubbled through the liquid and the Fe³⁺ solution was restored in order to be sent back to the chemical scrubber. The oxidation tank was automated and synchronized with the scrubber to supply it with the quantity of the re-oxidized Fe³⁺ solution necessary for the abatement of biogas H₂S. In this manner, a lower consumption of oxidant agents and a considerable reduction of costs were obtained.

Biogas VOCs were also subjected to partial abatement in the chemical scrubber (

Table 5. VOC speciation (biogas exiting chemical scrubber).

Compound (mg/Nm3)	December–May			June–November			Annual
	Min	Max	Mean	Min	Max	Mean	Mean
Total VOCs	532	1246	1022	306	584	422	722
limonene	185	648	442	50	96	72	251
α-pinene	101	213	159	17	62	33	96
β-pinene	117	215	137	34	77	40	89
p-cymene	84	241	162	14	74	28	95
γ-terpinene	0	1	0	0	6	2	1
myrcene	0	0	0	0	3	1	<1
Δ-3-carene	0	0	0	0	5	2	1
propane	0	8	1	2	33	23	12
butane	0	11	2	1	35	24	12
pentane	0	1	0	0	1	0	<1
hexane	0	6	1	6	24	15	8
heptane	0	6	2	7	26	20	11
2-methyl-pentane	0	0	0	0	2	1	<1
benzene	2	32	21	24	110	38	34
toluene	0	21	16	22	31	24	20
ethyl-benzene	0	23	21	20	34	24	22
xylene a	5	37	24	28	69	36	34
mesitylene	0	3	1	0	0	0	<1
cumene	2	35	26	30	68	33	31
acetone	0	1	0	0	4	2	1
methyl-ethyl-ketone	0	2	1	0	6	2	2
hexanal	0	2	0	0	0	0	0
heptanal	0	1	0	0	0	0	0
octamethylcyclotetrasiloxane	0	1	0	0	1	0	<1
decamethylcyclopentasiloxane	0	1	0	0	2	0	<1

^a sum of isomers.

) due to both the oxidation process and solubility in water.

In particular, total VOCs contained in biogas flow in the period between December and May and mainly represented by hydrophobic terpenes suffered a concentration reduction of 9.5%. During the rest of the year, total VOCs were characterized by a higher amount of hydrophilic compounds, so they underwent a concentration lowering of 36.2%. Likewise, a partial decrease in HCl and NH₃ concentration down to 8 mg/Nm³ and 4 mg/Nm³, respectively, was observed.

In the chemical scrubber, the water solution acted as an absorbent not only for H₂S but also for CO₂ [40]. Consequently, the average CO₂ concentration in biogas was reduced by about 2% when going out from the scrubber, thus increasing the CH₄ percentage. CH₄ and CO₂ average concentrations in biogas respectively passed from 59.4% and 40.1% before the scrubber to 60.3% and 39.2% after the scrubber.

3.4. Cooling Section

As water vapor drastically reduces the HHV of the final biomethane, it was necessary to eliminate the water fraction contained in biogas exiting the scrubber. For this purpose, a biogas dehumidification system consisting of a heat exchanger, chiller, and droplet separator was used.

Using a heat exchanger and chiller, the biogas temperature was lowered to 4 °C. Then, condensate was removed in the droplet separator. After this step, a percentage of water vapor below 0.05%vol was measured. The condensed liquid was composed not only of an aqueous phase but also of an organic layer. In the aqueous fraction, an average concentration of NH₄⁺, Cl⁻, 2⁻, and VOCs of 1.2 g/L, 660 mg/L, 135 mg/L, and 980 mg/L, respectively, was quantified. The organic layer was composed of terpenes (84%vol), hydrocarbons (12%vol), siloxanes (0.8%vol), water (0.2%vol), and unidentified compounds (3%vol). Consequently, in addition to water condensation, partial abatement of NH₃, HCl, H₂S, siloxanes, and VOCs occurred at the biogas dehumidification stage.

Leaving the droplet separator, biogas contained on average 2–4 mg/Nm³ of NH₃, 0.0–0.1 mg/Nm³ of HCl, 5–30 mg/Nm³ of H₂S, 0.0–0.1 mg/Nm³ of siloxanes, and 100–300 mg/Nm³ of VOCs.

3.5. Activated Carbon and Biogas Upgrading

Activated carbon was used to eliminate the last traces of biogas contaminants before biogas entered the membrane-based upgrading unit. Two types of commercial activated carbon were placed in series: the first one represented by the classic nonpolar activated carbon for the removal of organic and inorganic compounds and the second one impregnated with 10% sodium carbonate for specific removal of H₂S. It is, indeed, known that activated carbon impregnated with alkali compounds is effective for H₂S elimination [41]. At the outlet of the activated carbon treatment unit, the concentration of contaminants such as H₂S, silicon derivatives, and VOCs was below 0.1 mg/Nm³.

Contaminant-free biogas was then compressed to around 10 bar and passed through the polymer polyimide membranes system. Selective CO₂ permeation through the membrane surface allowed us to obtain biomethane with a purity > 98%.

Upon leaving the upgrading plant, biomethane had the characteristics (Table S2) such as to satisfy the requirements established by the national legislation and was introduced into the natural gas distribution network. Both plant and natural gas distribution network managers had continuous monitoring systems at the point of delivery of the biomethane to verify its quality and compliance. The economic value of the biomethane injected into the natural distribution network varies daily based on the market price. The average price for the thermal year October 2022–September 2023 was 56,789 EUR/MWh while in the thermal year October 2021–September 2022, the price was higher (121.727 EUR/MWh) due to the Russia–Ukraine conflict.

4. Conclusions

An optimized pre-treatment system able to remove trace contaminants found in biogas produced by a full-scale FW anaerobic digester was reported in this study. The trace compounds detected in biogas were H₂S, VOCs, NH₃, HCl, H₂O, and N₂ with an annual average concentration of 1207 (ppm_v), 895 (mg/Nm³), 36 (mg/Nm³), 17 (mg/Nm³), 0.3 (%vol) and 0.2 (%vol), respectively. The goal of the AD plant was to upgrade biogas to biomethane using a membrane-based system and to inject it into the natural gas network for the transport sector. For this purpose, a pre-treatment system composed of micro-oxygenation, chemical scrubber, cooling, and activated carbon sections was optimized based on the type and concentration of the traces contaminants found in biogas and, of course, on the desired final biomethane fuel characteristics. Insufflating an O₂ volume of 0.3% in the digester headspace allowed for a reduction of H₂S concentration by 69.3%. Biogas treatment in a chemical scrubber with FeCl₃ and NaOH solutions led to further H₂S elimination and the partial abatement of VOCs, NH₃, HCl, and CO₂. The selected operating values of FeCl₃ and pH were the result of the balance between the H₂S removal efficiency and the operating costs of the chemical scrubber. The subsequent cooling section was able not only to remove water vapor from the biogas but also to further reduce the concentration of H₂S, VOCs, NH_3, and HCl. After the cooling unit, the H₂S concentration in biogas had values lower than 30 mg/Nm³. The activated carbon section allowed for the elimination of the last traces of biogas contaminants before biogas entered the upgrading system. At last, the membrane-based upgrading unit was able to separate the CO₂ stream and release biomethane, meeting national regulatory limits for trace contaminants. The biogas plant plans to improve some downstream processes of the entire purification system, such as the recovery and purification of elemental sulfur at the exit of the chemical scrubber and of CO₂ at the exit of the membrane system for subsequent industrial uses.