Physical Activation of Waste-Derived Materials for Biogas Cleaning

: Biogas produced from biomass is carbon neutral. In fact, the carbon feedstock of biomass is converted into gas phase. Biogas use in high efﬁcient energy systems, such as Solid Oxide Fuel Cells is a viable choice. One of the most important drawbacks for such systems is related to the interaction between trace compounds and anode section. Gas cleaning through physical removal mechanisms is the simplest and cheapest method adopted in the literature. Coupled with this solution, the recovery of waste materials is an efﬁcient application of the circular economy approach. In this work, a physical activation process was investigated experimentally for waste-derived materials at a temperature of 700 ◦ C. The removal of H 2 S was considered as the most abundant trace compound. Activated biochar showed an adsorption capacity comparable to commercial sorbents, while the performance of ashes are still too poor. An important parameter to be considered is the biogas humidity content that enters in competition with trace compounds that must be removed. ranged from 0.04 cm 3 /g to 0.26 cm 3 /g; and the total volume ranged from 0.06 cm 3 /g to 0.37 cm 3 /g. The ash sample was activated using the same process. The speciﬁc surface increased to 6 m 2 /g, while the micro pore and total pore volume also increased according to a similar ratio recorded for the biochar.


Introduction
Scientists believed that the Earth's climate change could be related to the high concentration of CO 2 and other greenhouse gases produced by human activities. How biogas can reduce or mitigate this problem is a scientific challenge. Biogas produced from biomass is considered to be carbon neutral, since the carbon feedstock of biomass is converted into gas phase. In other words, atmospheric carbon (CO 2 ) is captured in a relatively short time. Biogas is a secondary energy carrier and can be produced from many organic substrates, such as from agricultural sources, sewage sludge and animal waste [1][2][3][4][5][6]. In addition to the chemical process, a thermal process (gasification or pyrolysis) can exploit organic material for biogas production purposes. The European Biomass Organisation for 2020 estimated a production of biogas around 39.5 Mtoe, which corresponds to 10% of EU natural gas consumption [7]. Biogas can be adopted as a fuel for several energy generators, such as fuel cells, as well as external and internal combustion engines [8,9]. Alternatively, biogas can be upgraded with highly engineered systems into biomethane to be injected into the gas grid [10,11]. Among energy generators, high temperature solid oxide fuel cells have higher electrical efficiency values, high fuel feeding flexibility, low noise and pollutant emissions [8,[12][13][14]. The most important drawbacks for such generators is the low tolerability to trace compounds [3,[15][16][17][18][19][20][21][22]. Depending on the nature of the trace compound and its concentration in the gaseous matrix, a gas cleaning section has to be specifically designed. The most frequently detected trace compounds from Organic Fraction of Municipal Solid Waste (OFMSW) biogas are sulphurs, carbonyl compounds, carboxylic acids and terpenes [2,4]. Rasi et al. (2011) studied the trace compounds that affect biogas energy users [6]. They mainly

Description of the Experimental Setup
The pyrolysis of wood waste produces a biochar that Gruppo RM Impianti srl (Arezzo, Italy) exploit internally. The reactor adopted is 200 kWe with operating temperature of around 450 • C (construction details reserved). Ashes from wood-chips from local forest were obtained from a boiler (3.3 MW, Viessman, Allendorf, Germany). The activation of biochar and wood ash was accomplished using an electric furnace of 500 We (C.I.T.T., Milano, Italy) regulated with a PID controller (Horst, GmbH Germany). Technical instructions are reported as follows: heat-up ramp fixed at 7 • C/min until 700 • C, with a CO 2 flow rate of 100 mL/min, see Figure 1a. The physical activation step was carried out for a period of 60 min under isothermal conditions. The operating temperature for the reactor was fixed using a furnace regulated by a K-type thermocouple placed in the centered bed. Two additional thermocouples (K-type, Tersid, Italy) were placed at the outlet and inlet of the bed. Experiments were accomplished with a Gas Hourly Space Velocity (GHSV) of 66 h −1 . Figure 1b epicts the sorbent material testing set-up. A simulated biogas mixture was accomplished (CH 4 /CO 2 = 1.5) using gas cylinders (Siad spa, Praxair with mass and liquid flow controllers (MFC, LFC) (Bronkhorst GmbH, Ruurlo, The Netherlands). A controlled evaporator mixer (CEM) was adopted to add water to the heated lines with a relative humidity of 30% (RH). A gas mixture prepared with CO 2 and H 2 S (190 ppm(v)) was used for the removal of trace compounds. The H 2 S concentration was fixed at 75.9 ppm(v) and obtained from the dilution with pure CO 2 , in order to simulate an average biogas mixture deriving from the anaerobic digestion of organic waste. This concentration is representative of a real biogas, as reported in literature [5,44,45]. Biochar and ashes were ground and sieved before and after the activation process, to obtain a grain size around 400 µm. These sorbent materials were placed in Teflon tubes with an internal diameter of 4 mm and an L/D ratio around 10. The removal test was accomplished three times, both for the activated and for the not activated sorbent materials. This concentration is representative of a real biogas, as reported in literature [5,44,45]. Biochar and ashes were ground and sieved before and after the activation process, to obtain a grain size around 400 μm. These sorbent materials were placed in Teflon tubes with an internal diameter of 4 mm and an L/D ratio around 10. The removal test was accomplished three times, both for the activated and for the not activated sorbent materials.

Methodology
In order to maintain the correct aspect ratio (~4) between the pellet and reactor diameter, the sorbents were ground and sieved [46]. This procedure was followed before the activation proces. The testing sample was selected using metallic sieves in the range 5.4-400 μm. In supplementary material, the pore distribution of activated carbons and waste-derived materials are reported. To remove residual gases inside the pores the samples were pre-treated using N2 flow (200 NmL/min) for 30 min at 120 °C. A gas flow of 200 NmL/min was used for the removal performance testing. Dynamic tests for the sulphur removal were accomplished for the selected materials at 25 °C. Below the adsorption capacity equation is reported, calculated in terms of mg/g for H2S removal: where:

Methodology
In order to maintain the correct aspect ratio (~4) between the pellet and reactor diameter, the sorbents were ground and sieved [46]. This procedure was followed before the activation proces. The testing sample was selected using metallic sieves in the range 5.4-400 µm. In supplementary material, the pore distribution of activated carbons and waste-derived materials are reported. To remove residual gases inside the pores the samples were pre-treated using N 2 flow (200 NmL/min) for 30 min at 120 • C. A gas flow of 200 NmL/min was used for the removal performance testing. Dynamic tests for the sulphur removal were accomplished for the selected materials at 25 • C. Below the adsorption capacity equation is reported, calculated in terms of mg/g for H 2 S removal: where: • t 0 , is the cleaning service time; • t 1 , is the last detection interval, corresponding to initial concentration for H 2 S of 1%, 10% and 100%; • Q tot = total gas flow rate (NL/h); • MW = molecular weight of the trace compound removed (g/mol); Considering the work of Barelli et al. (2016), Equation (1) was adjusted. In fact, the difference between the rectangle (t 1 (h) × C in (ppm(v))) and the triangle areas (0.5(t 1 − t 0 ) (h) × 1 (ppm(v)) approximates the area enclosed by the breakthrough curve and the saturation line (C out = C in ) [47]. An electrochemical gas sensor (ES, Meccos eTr-Leopold Siegrist GmbH, Lorsch, Germany) was adopted to measure the H 2 S concentration. This sensor is able to perform a continuous reading in the range of 0-200 ppm(v). The accuracy of the sensor reading is 0.5%, while a value of 1% is fixed for the electrochemical sensor. Flow meter controllers have an accuracy value of 0.1%, related to the full-scale value.

Results and Discussion
Results related to the adsorption capacity were reported below, starting from the characterization to the sorbent selection considering also the waste-derived materials.

Sorbent Characterisation-EDS/SEM
The sorbent materials adopted were characterised in terms of their surface area, composition, micropore volume and morphology structure. The elemental composition measurements were performed by scanning electron microscopy (SEM) (FEI Inspect, Philips 525 M) coupled with EDS analysis (SW9100 EDAX). The results are reported in Table 1, nitrogen and hydrogen fractions are not reported due to the limits of the instrument.
The most abundant compound for biochar samples is carbon followed by oxygen, potassium, calcium, and magnesium. The activation process does not substantially modify the elemental composition. A decrease in potassium content with a corresponding increase in calcium content was recorded. The most abundant compound for the ash samples is oxygen, followed by silicon, carbon, potassium, calcium, magnesium and aluminium. The activation process does not substantially change the elemental composition.
Tepper and Richardson et al. (2002) demonstrated how the concentrations of K and Ca are important for the sulphur compound removal [48,49]. Transition metals are also important for salt formation with trace compounds. The adsorption isotherms were determined using a Quantachrome Autosorb 1 (Boynton Beach, FL, USA) using for N 2 at 77 K. The specific surface areas were calculated by Langmuir equation in the relative pressure range of 0.04-0.1. The micropore volumes were determined by means of t-plot method in the relative pressure range of 0.15-0.3. For the carbon-based materials, the pore size was evaluated through the DFT method (Density Functional Theory), using the NLDFT equilibrium model for slit/cylindrical pores, see Table 2. The isotherm graph recorded for the biochar, according to IUPAC classification, is of type I. Such isotherm is typical of microporous materials. In Figure 2 the sharp increase in adsorbed volume at low relative pressures is depicted. This is due to the capillary condensation of the adsorbate inside the micropore. The presence of a hysteresis loop suggests the presence of slit porosities, due to the particles aggregation.  The isotherm graph recorded for the biochar, according to IUPAC classification, is of type I. Such isotherm is typical of microporous materials. In Figure 2 the sharp increase in adsorbed volume at low relative pressures is depicted. This is due to the capillary condensation of the adsorbate inside the micropore. The presence of a hysteresis loop suggests the presence of slit porosities, due to the particles aggregation. In the following Figure 3, an analysis to SEM images is necessary for the sorbent materials adopted, on the left side before activation and on the right side after the activation process.
The waste-derived biochar and ash were physically activated at a fixed temperature and CO2 flow rate. The structure before this physical process was denser and the homogeneity of the structure cannot be guaranteed by the un-optimized production phase. At the end of the activation process, especially for the biochar sample, the increase of porosity and consequently the specific surface area value was evident. This value after the activation process was around 600 m 2 /g, while the micro pore volume ranged from 0.04 cm 3 /g to 0.26 cm 3 /g; and the total volume ranged from 0.06 cm 3 /g to 0.37 cm 3 /g. The ash sample was activated using the same process. The specific surface increased to 6 m 2 /g, while the micro pore and total pore volume also increased according to a similar ratio recorded for the biochar. In the following Figure 3, an analysis to SEM images is necessary for the sorbent materials adopted, on the left side before activation and on the right side after the activation process.
The waste-derived biochar and ash were physically activated at a fixed temperature and CO 2 flow rate. The structure before this physical process was denser and the homogeneity of the structure cannot be guaranteed by the un-optimized production phase. At the end of the activation process, especially for the biochar sample, the increase of porosity and consequently the specific surface area value was evident. This value after the activation process was around 600 m 2 /g, while the micro pore volume ranged from 0.04 cm 3 /g to 0.26 cm 3 /g; and the total volume ranged from 0.06 cm 3 /g to 0.37 cm 3 /g. The ash sample was activated using the same process. The specific surface increased to 6 m 2 /g, while the micro pore and total pore volume also increased according to a similar ratio recorded for the biochar.

Adsorption Capacity
The recovery of waste materials is an interesting application of the circular economy approach. In fact, literature studies consider a wide range of wastes, such as from agricultural sector for the production of activated carbons [50]. In this work two waste-derived materials have been selected, such as char from the pyrolysis of wood waste, and ashes from the combustion of wood chips in a boiler. Virgin chars, as reported also in this article have lower adsorption capacity for the removal of trace compounds, especially sulphur compounds [37]. The activation methods are divided into chemical and physical origin. Due to technical simplicity and economical aspects, physical activation was adopted as the activation method for the sorbent materials removal of sulphur compounds [51][52][53]. The pore size distribution and pore fraction of the activated biochar vary according to the quantity and type of biomass, operating temperature and activating gas [54]. According to Nabais et al. (2007), the CO2 activation leads to samples with higher BET surface areas and pore volumes compared with samples produced by steam activation and with similar burn-off values [55]. Zhang et al. (2004) studied the duration of CO2 activation and temperature on the characteristics of the char produced from the wood fast pyrolysis [56]. Chars type selected showed an increasing of specific surface area and micropore volume with increasing activation temperature. The increasing of pore size and surface area was verified up to 900 °C, above this value the carbon characteristics decrease [57]. Table 3 shows the adsorption capacity for commercial carbons and waste-derived materials for the removal of H2S. Commercial carbons have an adsorption capacity of around 2 mg/g while commercial metal oxides, such as ZnO has an average removal capacity of around 1 mg/g at t1 = 1% of initial concentration. The two commercial carbons selected have a similar adsorption capacity also at the saturation condition, t1 = 100%. Carbox compared to RST3 shows a steeper profile for the removal of H2S. Waste derived materials not physically activated show an adsorption capacity lower than commercial materials. Here, the surface area and microporous volume weight strongly on the

Adsorption Capacity
The recovery of waste materials is an interesting application of the circular economy approach. In fact, literature studies consider a wide range of wastes, such as from agricultural sector for the production of activated carbons [50]. In this work two waste-derived materials have been selected, such as char from the pyrolysis of wood waste, and ashes from the combustion of wood chips in a boiler. Virgin chars, as reported also in this article have lower adsorption capacity for the removal of trace compounds, especially sulphur compounds [37]. The activation methods are divided into chemical and physical origin. Due to technical simplicity and economical aspects, physical activation was adopted as the activation method for the sorbent materials removal of sulphur compounds [51][52][53]. The pore size distribution and pore fraction of the activated biochar vary according to the quantity and type of biomass, operating temperature and activating gas [54]. According to Nabais et al. (2007), the CO 2 activation leads to samples with higher BET surface areas and pore volumes compared with samples produced by steam activation and with similar burn-off values [55]. Zhang et al. (2004) studied the duration of CO 2 activation and temperature on the characteristics of the char produced from the wood fast pyrolysis [56]. Chars type selected showed an increasing of specific surface area and micropore volume with increasing activation temperature. The increasing of pore size and surface area was verified up to 900 • C, above this value the carbon characteristics decrease [57]. Table 3 shows the adsorption capacity for commercial carbons and waste-derived materials for the removal of H 2 S. Commercial carbons have an adsorption capacity of around 2 mg/g while commercial metal oxides, such as ZnO has an average removal capacity of around 1 mg/g at t 1 = 1% of initial concentration. The two commercial carbons selected have a similar adsorption capacity also at the saturation condition, t 1 = 100%. Carbox compared to RST3 shows a steeper profile for the removal of H 2 S. Waste derived materials not physically activated show an adsorption capacity lower than commercial materials. Here, the surface area and microporous volume weight strongly on the removal performance, as reported [58]. In fact, as reported in the following Figure 4, the adsorption capacity is directly correlated to the specific surface area and microporus volume. Higher micropore volume along with smaller pores results in enhanced capacity of H 2 S removal. As reported by Adib et al. (2000), the mesoporous volume seems less important for the adsorption capacity performance [59].
Energies 2018, 11, x 7 of 12 removal performance, as reported [58]. In fact, as reported in the following Figure 4, the adsorption capacity is directly correlated to the specific surface area and microporus volume. Higher micropore volume along with smaller pores results in enhanced capacity of H2S removal. As reported by Adib et al. (2000), the mesoporous volume seems less important for the adsorption capacity performance [59]. In Figure 4, the coefficient of determination is higher than 0.95 for surface area and adsorption capacity, while this coefficient is higher than 0.98 for microporous volume and adsorption capacity. ZnO sample, as reported in the characterization section is defined by an adsorption isotherm related to chemical mechanisms, demonstrated by not observing the linear correlation reported in Figure 4. This removal mechanism is stronger than physisorption, both for removal and regeneration of material. To explain these poor performances, Aravind et al. (2012) demonstrated how the optimal metal oxides performances are shown at a higher temperature [60].  In Figure 4, the coefficient of determination is higher than 0.95 for surface area and adsorption capacity, while this coefficient is higher than 0.98 for microporous volume and adsorption capacity. ZnO sample, as reported in the characterization section is defined by an adsorption isotherm related to chemical mechanisms, demonstrated by not observing the linear correlation reported in Figure 4. This removal mechanism is stronger than physisorption, both for removal and regeneration of material. To explain these poor performances, Aravind et al. (2012) demonstrated how the optimal metal oxides performances are shown at a higher temperature [60]. The physical activation with CO 2 and higher temperature allows us to improve the average adsorption capacity from 0.18 mg/g to 1.38 mg/g for biochar, and from 0.05 mg/g to 0.38 mg/g for ashes.
As reported by Nabais et al. (2007), physical activation with CO 2 promotes pore size and surface area increasing, with better performance than steam activation [55]. Activated biochar shows an adsorption capacity comparable to commercial sorbents, as reported also from [37,61]. Increasing the saturation zone of the filter, from 1% to 10% up to 100%, the adsorption capacity increases up to the saturation of pores, see Table 3. The latter value is a limit of the filter bed filled by sorbent material. An average value of 10 mg/g was registered for activated biochar. This value is not comparable with those obtained with activated ash, while appears similar with those obtained with activated carbons [39]. Metal oxides are able to promote the dissociation of H 2 S and the HS − reaction to form sulphides. These reaction are reported in literature studies, as studied from Shang et al. (2016) [37]. This was due to the chemical mechanisms developed between trace compounds to be removed and metals [62]. Table 4 depicts the adsorption capacity of H 2 S in a wet biogas using commercial activated carbons, waste-derived materials and activated waste-derived materials. Biogas water vapor (30% RH) enters into competition with hydrogen sulphide that has to be removed. This competition decreases the adsorption capacity [31]. In fact, as reported elsewhere [31], when the RH value is above 50%, the adsorption capacity tended to be zero. At t 1 = 1% of the initial H 2 S concentration, the average adsorption capacity value registered for Carbox decrease from 1.9 mg/g up to 1.39 mg/g. A stronger decrease is obtained using ZnO. In fact, the adsorption capacity in the same conditions decreases of 55%.
Considering waste-derived materials, the effect of water content on the removal performance has the same effect registered for commercial sorbents [63]. Because this material is not activated chemically with alkali metals, water content competes with sulphur for their removal [30]. The adsorption capacities decrease more in the case of not possessing physically activated material. In fact, pores and surface area being lower suffer more from the competition of water content. Biochar shows a decrease of 80%, while with the physical activation, this decrease stands at around 30%.

Conclusions
Hydrogen sulphide removal performance of waste-derived material, such as biochar show lower adsorption capacity compared to optimized commercial sorbents. These results, although preliminary are important for the future optimization of waste-derived materials. In fact, such types of materials can be reinserted in the gas production cycle as fertilizer or structuring material for the anaerobic digestion process.
Physical activation with CO 2 and temperature promotes for activated biochar an adsorption capacity that increases with values comparable to commercial sorbents, while ashes performance are still too poor.
Humid biogas should be dried, because water enters into competition with hydrogen sulphide that has to be removed. A more pronounced decrease of adsorption capacities is recorded for not physically activated material. This is due to the higher competition between water content and trace compounds to be removed with low surface area and microporous volume.
Author Contributions: S.F. for the chemical analysis and D.P. and M.S. for the research activity and for the supervision work.