Comparative life cycle assessment of Ni-based catalyst synthesis processes

Journal of Cleaner Production 162 (2017) 7e15

Contents lists available at ScienceDirect

Journal of Cleaner Production journal homepage: www.elsevier.com/locate/jclepro

Comparative life cycle assessment of Ni-based catalyst synthesis processes  Boris Agarski a, *, Vesna Nikoli c b, Zeljko Kamberovi c c, Zoran AnCi c d, Borut Kosec e, Igor Budak a Faculty of Technical Sciences, University of Novi Sad, 6 Trg Dositeja Obradovica Street, 21000, Novi Sad, Serbia Innovation Center of the Faculty of Technology and Metallurgy, University of Belgrade, 4 Karnegijeva Street, 11120, Belgrade, Serbia c Faculty of Technology and Metallurgy, University of Belgrade, 4 Karnegijeva Street, 11120, Belgrade, Serbia d Innovation Center of the Faculty of Chemistry, University of Belgrade, 12-16 Studentski Trg Street, 11000, Belgrade, Serbia e Faculty of Natural Sciences and Engineering, University of Ljubljana, 12 Askerceva Street, 1000, Ljubljana, Slovenia a

b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 8 November 2016 Received in revised form 24 May 2017 Accepted 3 June 2017 Available online 3 June 2017

Ni-based catalysts supported on ceramics are particularly suitable for industrial applications, for instance reforming of hydrocarbons to produce synthesis gas or hydrogen and production of carbon nanofibers. Conventional synthesis processes for all metal/ceramic catalysts are impregnation, precipitation, coprecipitation and others. The authors have previously developed a novel process for the synthesis of Ni-based catalysts supported on reticulated ceramic foams, including impregnation of foams with ultrasonically generated aerosols of dissolved metal chlorides. By using appropriate multi-criteria analysis methods, the authors concluded that the novel process for the synthesis of Ni-based catalysts was superior in terms of economic and technological aspects. The aim of this research was to compare the novel synthesis processes for a Ni-Pd/Al2O3 catalyst and for other Ni-based catalysts by performing life cycle assessment and evaluating the environmental impacts of each synthesis process. Characterisation results showed that the dominant environmental impact results from production of palladium (II) chloride for the Ni-Pd/Al2O3 catalyst synthesis process, while the other catalyst synthesis process had large environmental impacts associated with high energy consumption. The final outcome, obtained from comparison of normalisation results, indicates that the novel Ni-Pd/Al2O3 catalyst synthesis process had the smallest environmental impact. © 2017 Elsevier Ltd. All rights reserved.

Handling Editor: Yutao Wang Keywords: Novel synthesis process Life cycle assessment Ni-based catalyst

1. Introduction Catalysts based on metals supported on ceramic materials are used in a wide range of heterogeneous catalytic processes. Nickelbased catalysts are mostly used for the production of highly efficient energy sources such as H2 and synthesis gas (CO þ H2) (Calles et al., 2015; Tu et al., 2011) or advanced carbon materials (de Llobet et al., 2015; Takenaka et al., 2003). The Ni/Al2O3 system is the most common among those catalysts (Akande et al., 2005; Zhang et al., 2010) and suitable for dry methane reforming (Tu et al., 2011), oxidative methane reforming (Omata et al., 2008), ethanol steam reforming (Calles et al., 2015), production of carbon nanofibers (de Llobet et al., 2015; Takenaka et al., 2003), etc. Catalysts for the

* Corresponding author. E-mail address: [email protected] (B. Agarski). http://dx.doi.org/10.1016/j.jclepro.2017.06.012 0959-6526/© 2017 Elsevier Ltd. All rights reserved.

hydrocarbon reforming processes are usually modified by a small amount of a noble metal, for example 0.1 wt% of Pd, which prevents nickel deactivation (Fatsikostas et al., 2002; Profeti et al., 2009). Conventional thermochemical processes for catalyst synthesis include impregnation, precipitation, co-precipitation, sol-gel and others. In order to obtain initial precursors for catalysts, those processes require suspending of Al2O3 or other oxide powder in aqueous solution of metal salts, preparation of mixed salt solutions or forming of gels from compounds that contain metallic ions (Akande et al., 2005; Yurdakul et al., 2016). Initial precursors are calcined at high temperatures (from 773 K (Sengupta et al., 2014) to 823 K (Sharifi et al., 2014)) until oxide precursor mixtures are synthesised. Reduction is a final step to obtain catalysts where only active particles are transferred to metallic state while support remains as an oxide. It is also performed at high temperatures, most commonly from 823 K (Sengupta et al., 2014) to 973 K (Sharifi et al., 2014). In some cases, the calcination step is avoided and

8

B. Agarski et al. / Journal of Cleaner Production 162 (2017) 7e15

catalytically active Pd and Ni can be obtained by low-temperature reduction of metal salts (at 573 K for 1 h) (Takenaka et al., 2003). In a previous research, Nikolic et al. (2014b) have developed a novel, unconventional synthesis process to prepare a monolithic Ni-Pd/Al2O3 catalyst, where reticulated a-Al2O3 based foam was impregnated with ultrasonically generated aerosol of dissolved metal chlorides. The foam was produced according to a previously described method (Nikolic et al., 2014a). The developed catalyst synthesis process enabled technological simplification and energy savings (Nikoli c et al., 2014b) because the calcination procedure was eliminated and the catalyst was reduced by hydrogen at a very low temperature, 533 K (Nikoli c et al., 2014b). In an earlier work of these authors, the synthesis processes of Ni-Pd/Al2O3 catalyst and other Ni-based catalysts were evaluated by using three multicriteria analysis methods (SAW, TOPSIS and PROMETHEE II). Research results indicated that the novel Ni-Pd/Al2O3 catalyst synthesis process ranked best in comparison with other processes (Nikoli c et al., 2016). Catalyst synthesis processes lead to formation of wastes that mainly comprise flue gases, liquid waste and a minor amount of solid waste. Environmental impacts are highly dependent on a synthesis process and starting materials. Most conventional thermochemical processes for catalyst synthesis require the preparation of a starting solution with a low concentration of dissolved chemicals. As an example, Abdedayem et al. (2015) have synthesised Cu/AC catalysts supported on activated carbon by impregnation process; the starting Cu(NO3)2 solution had a concentration of 0.001 mol L 1. Jung et al. (2012) have prepared Ni/ Al2O3 catalysts by using co-precipitation process; the concentration of Ni(NO3)2 and Al(NO3)3 in the starting solution was 0.085 mol L 1 and 0.098 mol L 1, respectively. Therefore, the synthesis of 1 kg of catalyst requires a significant amount of starting solution, which usually becomes waste after use. Oxidative calcination of catalyst supports impregnated with metal salts, such as nitrates, leads to emission of NO2 gas. Waste acid solutions are generated when mixtures containing metal nitrates and/or chlorides are reduced by hydrogen instead of calcined. Life cycle assessment (LCA) is often used for comprehensive

Background data

2. Materials and methods 2.1. Goal and scope definition The goal of this assessment is to perform comparative attributional LCA of five Ni-based catalyst synthesis processes and to assess their impacts on the environment. System boundaries include extraction of raw material from the environment, production and transport of semi-products and production of Ni-based catalysts (Figs. 1 and 2). Use phase and waste management of end

System boundaries Emissions to environment (elementary flows)

Foreground data

evaluation of process environmental impacts through the life cycle (Agarski et al., 2016; Curran and Young, 2014; Levasseura et al., 2016). To date, life cycle assessments of catalyst synthesis processes are scarce. Snowden-Swan et al. (2016) evaluated the greenhouse gas emissions of catalysts for hydrotreating of fast pyrolysis bio-oil with LCA. Yaseneva et al. (2014) performed a cradle-to-gate LCA study to compare potential water treatment processes based on two carbon nanofiber supported catalyst types. The environmental burden associated with the synthesis process of fine chemicals via TiO2 (solar) photocatalysis was analysed with LCA and compared with the same reactions under thermal conditions by Ravelli et al. (2010). Fernandez et al. (2016) examined enzyme-catalysed processes for biodiesel production and effects on the environment with LCA. Comparative LCA of pharmaceutical wastewater treatment with heterogeneous catalysts was performed by Rodríguez et al. (2016). Considering the fact that catalysts in general can have considerable different properties and applications, in this research, comparative LCA of Ni-based catalyst synthesis processes are performed. Based on previous studies on catalyst synthesis processes, the authors concluded that none of the research compared Ni-based catalyst synthesis processes by using LCA. In this study, comparative LCA was performed for five Ni-based catalyst synthesis processes. The novel process for synthesis of Ni-Pd/Al2O3 catalyst, developed by the authors, is compared with the other Ni-based catalyst synthesis processes; the environmental impacts of each process were evaluated and discussed.

Extraction of raw materials Transport Production of semi products Transport

Production of CSP5 (Ni-Pd/Al2O3 catalyst)

Emissions to environment (elementary flows)

Inputs from environment (elementary flows)

Transport Production of CSP1-4 (other Ni-based catalyst)

Use of Ni-based catalyst

Emissions to environment (elementary flows)

Waste management Fig. 1. System boundaries, foreground and background data for life cycles of CSP1-5.

B. Agarski et al. / Journal of Cleaner Production 162 (2017) 7e15

9

Filter paper Water, deionised Nickel

Hydrogen, liquid

Nitric acid, in water

Nickel II nitrate

Water, deionised

Aluminium oxide Titanium dioxide

Palladium Palladium II chloride Hydrochloric acid

CSP1 Catalyst synthesis process

Ni-based catalyst

CSP2 Catalyst synthesis process

Emissions to air (NO2)

CSP3 Catalyst synthesis process

Emissions to river (waste solutions)

CSP4 Catalyst synthesis process

Waste paper to landfill

Nitrogen, liquid Aluminium hydroxide

Electricity

Polyurethane

α-Al203

CSP5 Catalyst synthesis process

Refractry, fireclay Nickel

CSP1 flows CSP2 flows CSP3 flows CSP4 flows CSP5 flows Transport

Nickel II chloride Hydrochloric acid Fig. 2. System boundaries for life cycles of CSP1-5.

of life catalyst are not included in the system boundaries. The functional unit in LCA is introduced for comparison of functionally equivalent products or processes. In this research, the functional unit is the production of 1 kg of Ni-based catalyst. All the catalysts have the same conversion efficiency. Catalysts synthesis processes were denoted as the following: CSP1 e for Ni-Pd/Al2O3 (Takenaka et al., 2003), CSP2 e for Ni-Pd/ TiO2 (Takenaka et al., 2003), CSP3 e for Ni/Al2O3 (Akande et al., 2005), CSP4 e for Ni/a-Al2O3 (Omata et al., 2008) and CSP5 e for Ni-Pd/Al2O3 catalyst, experimentally obtained in previous research (Nikoli c et al., 2014a,b). The CSP1-4 are the conventional impregnation processes described in the literature. System boundaries of CSP1-5 synthesis processes with material flows are presented in Fig. 2. Metal salts (Ni(NO3)2, NiCl2 and PdCl2) were used as a part of the starting materials for synthesis of catalysts. Raw materials required for production of metal salts were Ni, Pd, HNO3, HCl and H2O. Salts were obtained by dissolving of Ni and Pd in diluted HNO3 or HCl; Al(OH)3, polyurethane foam and refractory clay were required for the production of a-Al2O3-based ceramic foam. For the CSP1 and CSP2, the following materials were used: Ni(NO3)2, PdCl2 and oxide powders as catalytic supports, g-Al2O3 for CSP1 and TiO2 for CSP2. Oxide powders were immersed in aqueous solutions of metal salts in order to obtain initial precursor mixtures and dried after impregnation at 373 K for 12 h. The catalysts were reduced in hydrogen flow at 573 K for 1 h (Takenaka et al., 2003). Starting materials for the CSP3 process were Ni(NO3)2 and gAl2O3 powder. After impregnation of g-Al2O3 with aqueous solution of Ni(NO3)2, the initial precursor mixture was filtered, dried at 383 K for 10 h, calcined 873 K for 3 h and finally reduced at 873 K for 2 h (Akande et al., 2005). Materials used for the CSP4 were Ni(NO3)2 and a-Al2O3 powder,

which was prepared by calcination of g-Al2O3 powder at 1473 K for 2 h. After the impregnation procedure, water was evaporated and the obtained sample was calcined at 773 K for 3 h. The catalyst was reduced in a tubular reactor where hydrogen flow was supplied until the catalyst bed reached the temperature of 1073 K. After that, the catalyst bed was cooled down with flowing Ar (Omata et al., 2008). Materials required for the novel CSP5 process were NiCl2, PdCl2 and a-Al2O3 based foam. The foam was preheated at 473 K and then impregnated with ultrasonically generated aerosol of dissolved metal salts. After drying at 473 K for 1 h, the catalyst was reduced at very low temperature, 533 K, for 1.5 h (Nikolic et al., 2014a,b). During catalyst synthesis processes, different types of waste products are generated. The environmental impacts of catalyst production include formation of gaseous emissions and liquid or solid waste. The amounts of waste products per 1 kg of synthesised catalyst were estimated assuming that each of the catalysts obtained by CSP1-5 processes was prepared in our laboratory and that liquid waste is discharged into sewers, gaseous emissions into the atmosphere and solid waste into landfill. Considering the conventional impregnation process, the starting solution of metal salts usually becomes waste solution after the filtration procedure, unless the liquid phase is evaporated. Used filter paper represents solid waste in case of filtration. Precursor mixtures obtained by the impregnation process consist of metal salts on catalyst supports and tare usually calcined in the air atmosphere. After calcination, metal oxides are formed from salts and when metal nitrates are used as starting materials, the resulting flue gas mainly contains NO2 diluted in air. If precursor mixtures are directly reduced in hydrogen flow instead of calcined, salts are transferred to metals and waste acid solutions are formed. Depending on the metal salt used in the synthesis process, HNO3, HCl or mixtures of those acids

10

B. Agarski et al. / Journal of Cleaner Production 162 (2017) 7e15

assessment of Ni-based catalyst synthesis processes can be performed. The ReCiPe Midpoint method (Goedkoop et al., 2009) was used and the following 11 impact categories were selected: climate change, terrestrial acidification, freshwater eutrophication, human toxicity, photochemical oxidant formation, particulate matter formation, freshwater ecotoxicity, marine ecotoxicity, natural land transformation, metal depletion, fossil depletion. Other ReCiPe impact categories were excluded because they contributed less than 0.02 pt to normalisation results for all five analysed processes, which can be considered as insignificant compared to impacts of other midpoint categories. The excluded impact categories were ozone depletion, marine eutrophication, terrestrial ecotoxicity, ionising radiation, agricultural land occupation, urban land occupation and water depletion.

are formed. Those acid-containing gases are then let through a wash bottle where they are collected as diluted acids. Considering the novel synthesis process developed by the authors, the waste acid solution contains diluted HCl. The transport of semi products (chemicals) from the producers to the laboratory is assumed by a van < 3.5 t. The transporting distances for the substances are as follows: 1535 km for filter paper (considering the small amount of waste filter paper and short distance, transport for waste filter paper is not considered inside the system boundaries), 1070 km for Ni, HNO3, Pd, HCl, Al(OH)3 and TiO2, 200 km for polyurethane, flexible foam, 65 km for refractory clay and 1170 km for H2, Al2O3 and N2. 2.2. Life cycle inventory

3. Life cycle impact assessment results

Foreground data for the novel catalyst synthesis process (CSP5) were obtained on-site in laboratory conditions. The other four processes (CSP1-4) have also been carried out in laboratory conditions and they were described in the research of authors Takenaka et al., 2003, Akande et al., 2005 and Omata et al., 2008. Therefore, background data for the CSP1-4 processes were obtained from the literature. Foreground and background data within the system boundaries are shown in Fig. 1. The processes from ecoinvent 3.0 (Weidema et al., 2013) database were used for assembling the LCI of catalysts. Life cycle inventory per functional unit for CSP1-5 is presented in Tables 1e5.

Characterisation results from the ReCiPe method on midpoint level for CS1 and CSP2 catalyst synthesis processes for CSP3, CSP4 and CSP5 are shown in Figs. 3e5, respectively. Normalisation results from the ReCiPe method on midpoint level and comparison of five catalyst synthesis processes are shown in Fig. 6. 4. Interpretation and discussion Regarding the characterisation results in Figs. 3 and 4, it can be noted that CSP1 and CSP2 have very similar environmental impacts, and the same applies to CSP3 and CSP4. The CSP1 and CSP2 characterisation results (Fig. 3a and b) are almost identical, and this was expected since the only difference is that CSP1 uses Al2O3 while

2.3. Life cycle impact assessment After assemblage of all the data for the LCI, life cycle impact

Table 1 LCI of Ni(NO3)2 production per functional unit. Process (name from ecoinvent 3.0)

Unit

CSP1

CSP2

CSP3

CSP4

CSP5

Nickel, primary, from platinum group metal production/RU S Nitric acid, in water (60% HNO3) (NPK 13.2-0-0), at plant/RER mass Water, deionised, from tap water, at user {CH}| production | Alloc Def, S Transport (van < 3.5 t) for nickel and nitric acid

g g g kg km

275.00 980.00 28.39 1342.85

275.00 980.00 28.39 1342.85

150.00 540.00 15.49 783.30

100.00 360.00 10.32 492.20

e e e e

Table 2 LCI of PdCl2 production per functional unit. Process (name from ecoinvent 3.0)

Unit

CSP1

CSP2

CSP3

CSP4

CSP5

Palladium {GLO}| market for | Alloc Def, S Hydrochloric acid, without water, in 30% solution state {RoW}| market for | Alloc Def, S Transport (van < 3.5 t) for palladium and hydrochloric acid

g g kg km

25.00 17.13 45.08

25.00 17.13 45.08

e e e

e e e

0.20 0.14 0.36

Table 3 LCI of a-Al203 production per functional unit. Process (name from ecoinvent 3.0)

Unit

CSP1

CSP2

CSP3

CSP4

CSP5

Aluminium hydroxide {GLO}| market for | Alloc Def, S Polyurethane, flexible foam {GLO}| market for | Alloc Def, S Refractory, fireclay, packed {GLO}| market for | Alloc Def, S Transport (van < 3.5 t) for aluminium hydroxide, polyurethane, flexible foam, and refractory

g g g kg km

e e e e

e e e e

e e e e

e e e e

920.00 56.97 200.00 978.79

Table 4 LCI of NiCl2 production per functional unit. Process (name from ecoinvent 3.0)

Unit

CSP1

CSP2

CSP3

CSP4

CSP5

Nickel, primary, from platinum group metal production/RU S Hydrochloric acid, without water, in 30% solution state {RoW}| market for | Alloc Def, S Transport for nickel and hydrochloric acid

g g kg km

e e e

e e e

e e e

e e e

200.00 1500.00 1819.00

B. Agarski et al. / Journal of Cleaner Production 162 (2017) 7e15

11

Table 5 LCI of catalyst production CSP1-5 per functional unit. Process (name from ecoinvent 3.0)

Unit

CSP1

CSP2

CSP3

CSP4

CSP5

Paper, newsprint, 0% DIP, at plant/RER S (filter paper) Water, deionised, from tap water, at user {CH}| production | Alloc Def, S Hydrogen, liquid, at plant/RER S Ni(NO3)2 (Table 1) PdCl2 (Table 2) Aluminium oxide, at plant/RER S Titanium dioxide, production mix, at plant/RER S Nitrogen, liquid, at plant/RER S a-Al2O3 (Table 3) NiCl2 (Table 4) Electricity, low voltage, production CS, at grid/CS S Transport for filter paper Transport for hydrogen, liquid Transport for aluminium oxide Transport for titanium dioxide Transport for nitrogen, liquid Total transport Waste emissions Waste water (in waste solution) Waste water (in waste acid solution) Nitric acid (in waste acid solution) Hydrogen chloride (in waste acid solution) Disposal, paper, 11.2% water, to sanitary landfill/CH S (waste filter paper) Nitrogen dioxide (in flue gas)

g kg kg g g kg kg kg kg kg kWh kg km t km Kg km Kg km t km t km

17.80 1282.57 12.62 1362.48 41.66 0.70 e e e e 1579.45 27.32 14.77 819.00 e e 15.62

17.80 1282.57 12.62 1362.48 41.66 e 0.70 e e e 1579.45 27.32 14.77 e 749.00 e 15.55

17.80 666.18 6.56 743.17 e 0.85 e 126.05 e e 860.75 27.32 7.68 994.50 e 147.48 156.18

e 444.12 4.37 495.45 e 0.90 e e e e 625.84 e 5.11 1053.00 e e 6.16

e 22.36 17.70 e 0.33

l l g g g g

1282.54 9.60 590.48 17.13 17.80 e

1282.54 9.60 590.48 17.13 17.80 e

666.17 e e e 17.80 235.15

e e e e e 156.77

e 9.60 e 248.62 e e

a) CSP1 characterisation results

e e 0.80 0.81 22.50 e 20.71 e e e 20.71

b) CSP2 characterisation results

100%

100%

90%

90%

80%

80%

70%

70%

60%

60%

50%

50%

40%

40%

30%

30%

20%

20%

10%

10%

0%

0%

Aluminium oxide

Nickel II Nitrate

Palladium II chloride

Nickel II Nitrate

Palladium II chloride

Water, deionised

Water, deionised

Hydrogen, liquid

Transport

Hydrogen, liquid

Titanium dioxide

Transport

Filter paper

Electricity

Disposal of paper

Filter paper

Electricity

Disposal of paper

Fig. 3. CSP1 and CSP2 characterisation results.

CSP2 uses TiO2 in the same amount. For CSP1 and CSP2, the largest environmental impacts are based on the consumption of electricity and production of PdCl2. Largest environmental impacts from consumption of electricity apply to all of the considered impact categories, except freshwater eutrophication, freshwater ecotoxicity, marine ecotoxicity and metal depletion, where the environmental impact from PdCl2 production dominates. The environmental impacts of other processes for all impact categories are very small, except for natural land transformation category results where the environmental impact from transport is the second largest impact and more significant than the impact from PdCl2 production. Filter paper production, waste emissions and

disposal of waste paper to landfill have very small environmental impacts for CSP1 and CSP2. As previously noted, characterisation results from CSP3 and CSP4 are similar (Fig. 4a and b). The CSP3 consumes larger amounts of chemicals and electricity than CSP4 and uses additional liquid N2 (Table 5). The largest difference is in the environmental impact of the transport. The CSP3 has a large environmental impact from transport processes, since 126.05 kg of liquid N2 have to be transported over a distance of 1170 km. Consumption of electricity is the dominant environmental impact for all of the considered impact categories for CSP4 and almost for all the impact categories for CSP3 (exceptions are the natural land transformation and metal

12

B. Agarski et al. / Journal of Cleaner Production 162 (2017) 7e15

a) CSP3 characterisation results

b) CSP4 characterisation results

100%

100%

90%

90%

80%

80%

70%

70%

60%

60%

50%

50%

40% 40% 30% 30%

20%

20%

10%

10%

0%

0%

Aluminium oxide

Nickel II Nitrate

Water, deionised

Hydrogen, liquid

Nitrogen, liquid

Transport

Filter paper

Electricity

Disposal of paper

Waste emissions

Aluminium oxide

Nickel II Nitrate

Water, deionised

Transport

Electricity

Waste emissions

Hydrogen, liquid

Fig. 4. CSP3 and CSP4 characterisation results.

100% 90%

80% 70%

60% 50% 40%

30% 20% 10%

0%

α-Al2O3

Nickel II chloride

Palladium II chloride

Hydrogen, liquid

Transport

Electricity

Water, deionised

Fig. 5. CSP5 characterisation results.

depletion impact categories in CSP3). Compared to CSP1 and CSP2, nearly all of the environmental impact from CSP4 is based on electricity consumption, which has a more than ten times larger

impact than the other processes. Similar to CSP1 and CSP2, the environmental impact from transport in the natural land transformation impact category is the second largest impact for CSP4.

B. Agarski et al. / Journal of Cleaner Production 162 (2017) 7e15

13

1.6 1.4

1.2 1.0 0.8

0.6 0.4

0.2 0.0

CSP1

CSP2

CSP3

CSP4

CSP5

Fig. 6. CSP1-5 comparison of normalisation results.

The CSP3 transport process results in more than 70% of environmental impact in natural land transformation and more than 50% in the metal depletion impact category. Waste emissions from flue gas (NO2) provide less than 10% of environmental impact on particulate matter formation impact category for CSP3 and CSP4. Characterisation results for the novel (CSP5) synthesis process (Fig. 5) significantly stand out from the remaining results. In contrast to the previous catalyst syntheses processes CSP1-CSP4, where environmental impacts due to electricity consumption are dominant, for CSP5 processes, impacts are more or less evenly distributed among the impact categories. Overall, environmental impacts from CSP5 from the largest to the smallest impact are distributed in the following order: PdCl2, NiCl2, transport, electricity, H2, a-Al2O3. The PdCl2 has the largest environmental impact for all the considered impact categories, except for the terrestrial acidification, freshwater eutrophication and particulate matter formation impact categories where impact of NiCl2 production has the largest contribution to environmental loading. Normalisation enables the comparison of impact category results (Fig. 6). Normalisation of ReCiPe characterisation results was evaluated with normalisation values for Europe. Evidently, CSP1 and CSP2 have by far the largest environmental impact in all considered impact categories except for natural land transformation. After CSP1 and CSP2, third and fourth processes, considering a descending environmental impact order, are CSP3 and CSP4. The novel CSP5 synthesis process has the smallest environmental impact in all considered impact categories. Impact category marine ecotoxicity has the largest environmental impact for CSP1, CSP2 and CSP4, while the natural land transformation is the largest environmental impact for CSP3 and CSP5. 4.1. Sensitivity and scenario analysis Considering the fact that electricity consumption has a high environmental impact in all five catalyst synthesis processes, the impact of ±10% electricity consumption variation was investigated on the impact category normalisation results. The sensitivity analysis results are shown in Table 6. As can be expected, catalyst synthesis processes that consume

larger amounts of electricity had larger deviations in impact category results. This applies for CSP1-CSP4, where sensitivity analysis revealed that impact category results have maximum variations of ±9.4% for CSP1 and CSP2, ±9.6% for CSP3 and ±9.7% for CSP4. Once again, CSP1 and CSP2 had almost the same results with largest variation for the category climate change (±9.4%) and smallest for metal depletion (±1.2%). Almost all of the considered impact category results varied more than ±9% for CSP3 and CSP4, except for natural land transformation and metal depletion that varied below ±9%. The CSP5 consumed significantly smaller amounts of electricity than the CSP1-4 and therefore, results for CSP5 impact category have maximum variation of ±1.9%. The novel CSP5 catalyst synthesis process had largest variation for particulate matter formation (±1.9%) and smallest for natural land transformation (±0.3%). Overall, CSP5 had the smallest deviation in impact category results, followed by CSP1 and CSP2, while CSP3 and CSP4 had the largest deviations in impact category results. Furthermore, scenarios with different country mix of electricity from ecoinvent 3.0 (Weidema et al., 2013) LCI database have also been considered. Default scenario with Serbian electricity mix (CS) has been compared with the ones from France, Switzerland, United Kingdom, Germany, and Poland. Serbian electricity mix consists mainly from burning of lignite in power plants (66%) and hydropower (32%). France has 78% of electricity that originate from nuclear power plants, and 11% from hydropower. On the other side, Switzerland’s electricity mainly comes from hydropower (55%) and nuclear power plants (40%). United Kingdom electricity mix originates mainly from burning of natural gas (41%) and hard coal (33%) in power plants, as well as nuclear power plants (20%). Germany has 23% of electricity that originate from hard coal, 25% from lignite, 10% from natural gas, and 27% from nuclear power plants. Polish electricity mix consists mainly from burning of hard coal (55%) and lignite (36%) in power plants. The scenario analysis results (ReCiPe endpoint LCIA method with average weighting set) are shown in Fig. 7. The total environmental impact from three endpoint categories (human health, ecosystems, and resources) shows that the Serbian electricity (Fig. 7a) has the largest environmental impact, and is closely followed by Polish electricity (Fig. 7f). Use of electricity mix

14

B. Agarski et al. / Journal of Cleaner Production 162 (2017) 7e15

Table 6 Sensitivity analysis of normalisation results (±10% variation of electricity consumption).

Impact category

CSP1

CSP2

CSP3

CSP4

CSP5

Climate change

± 9.4%

± 9.4%

± 9.2%

± 9.7%

± 1.1%

Terrestrial acidification

± 5.4%

± 5.4%

± 9.1%

± 9.3%

± 1.5%

Freshwater eutrophication

± 1.7%

± 1.7%

± 9.3%

± 9.5%

± 1.0%

Human toxicity

± 5.5%

± 5.5%

± 9.6%

± 9.8%

± 1.6%

Photochemical oxidant formation

± 7.1%

± 7.1%

± 9.0%

± 9.4%

± 0.9%

Particulate matter formation

± 6.7%

± 6.7%

± 9.4%

± 9.6%

± 1.9%

Freshwater ecotoxicity

± 4.5%

± 4.5%

± 9.3%

± 9.5%

± 1.6%

Marine ecotoxicity

± 4.8%

± 4.8%

± 9.3%

± 9.5%

± 1.4%

Natural land transformation

± 7.8%

± 7.7%

± 7.4%

± 8.5%

± 0.3%

Metal depletion

± 1.2%

± 1.2%

± 8.2%

± 8.6%

± 0.4%

Fossil depletion

± 9.2%

± 9.2%

± 9.0%

± 9.5%

± 0.8%

a) Scenario with Serbian electricity (default)

b) Scenario with French electricity

c) Scenario with Swiss electricity

250

250

250

200

200

200

150

150

150

100

100

100

50

50

50

0

0 CSP1

CSP2

CSP3

CSP4

0 CSP1

CSP5

d) Scenario with United Kingdom electricity

CSP2

CSP3

CSP4

CSP5

CSP1

250

250

200

200

200

150

150

150

100

100

100

50

50

50

0

CSP1

CSP2

CSP3

CSP4

CSP5

CSP3

CSP4

CSP5

f) Scenario with Polish electricity

e) Scenario with German electricity

250

0

CSP2

0

CSP1

Human health

CSP2

CSP3

Ecosistems

CSP4

CSP5

CSP1

CSP2

CSP3

CSP4

CSP5

Resources

Fig. 7. Scenario analysis of endpoint results.

from United Kingdom (Fig. 7d) and Germany (Fig. 7e) show smaller total environmental impact and same ranking of CSPs as in scenario with Serbian electricity. Use of electricity mix from France (Fig. 7b) and Switzerland (Fig. 7c) show about three times less total environmental impact than the default scenario with Serbian electricity

mix. The most important difference in use of electricity mix from France and Switzerland instead of Serbian electricity is the rank change of total environmental impact between the CSP4 and CSP5. Since Switzerland and France have cleaner production of electric energy, and CSP4 has majority of environmental impacts that

B. Agarski et al. / Journal of Cleaner Production 162 (2017) 7e15

originate from electricity consumption, CSP4 becomes the process with the smallest total environmental impact. Total environmental impact from CSP5 is quite solid and insensitive to change of electricity in these six scenarios. 5. Conclusions In this research, LCA of the novel synthesis process for Ni-Pd/ Al2O3 catalyst was performed and compared with the other Nibased catalyst synthesis processes. The novel Ni-Pd/Al2O3 catalyst synthesis process had previously been developed by the authors of this paper. It involved impregnation of a-Al2O3 based foam with an ultrasonically generated aerosol of dissolved metal chlorides at 473 K, drying at that temperature for 1 h and hydrogen reduction at very low temperature, 533 K, for 1 h. Characterisation results revealed that consumption of electricity and production of PdCl2 mainly contribute to environmental impacts of observed impact categories for CSP1 and CSP2. The CSP3 and CSP4 had the largest impacts on the environment due to the consumption of electricity. Production of PdCl2 is a dominant environmental impact of CSP5 synthesis process, while other significant impacts come from NiCl2 production, electricity consumption and H2 production. The obtained LCA normalisation results showed that, in comparison with other processes for the synthesis of Ni-based catalysts, the novel Ni-Pd/Al2O3 catalyst synthesis process (CSP5) has the lowest environmental impacts. Sensitivity analysis was performed by changing the electricity consumption by ± 10%. The catalyst synthesis processes that consume larger amounts of electricity had larger deviations in impact category results, while the novel Ni-Pd/Al2O3 catalyst synthesis process had a small deviation of ±1.9%. Scenario analysis revealed that the use of cleaner resources for electricity (hydropower and nuclear power plants for production of electricity country mix in Switzerland and France) can significantly reduce the total environmental impacts for CSP1-4, and that CSP4 becomes the catalyst synthesis process with the lowest environmental impact. Future research should investigate the expansion of system boundaries and include use and end of life phases. This research was conducted under experimental laboratory conditions; however, studies of catalyst production processes under large scale industrial conditions would be beneficial. Industrial production of Ni-based catalysts would use different process apparatus and have different environmental impacts. Acknowledgement This research was financially supported by the Ministry of Education, Science and Technological Development of the Republic of Serbia and is a result of the projects No. 34033 and 35020. References Abdedayem, A., Guiza, M., Ouederni, A., 2015. Copper supported on porous activated carbon obtained by wetness impregnation: effect of preparation conditions on the ozonation catalyst’s characteristics. Comptes Rendus Chim. 18, 100e109. Agarski, B., Budak, I., Vukelic, Dj, Hodolic, J., 2016. Fuzzy multi-criteria-based impact category weighting in life cycle assessment. J. Clean. Prod. 112, 3256e3266. Akande, A.J., Idem, R.O., Dalai, A.K., 2005. Synthesis, characterization and performance evaluation of Ni/Al2O3 catalysts for reforming of crude ethanol for hydrogen production. Appl. Catal. A General 287, 159e175. Calles, J.A., Carrero, A., Vizcaíno, A.J., Lindo, M., 2015. Effect of Ce and Zr addition to Ni/SiO2 catalysts for hydrogen production through ethanol steam reforming.

15

Catalysts 5, 58e76. Curran, M.A., Young, S.B., 2014. Critical review: a summary of the current state-ofpractice. Int. J. Life Cycle Assess. 19, 1667e1673. de Llobet, S., Pinilla, J.L., Moliner, R., Suelves, I., 2015. Effect of the synthesis conditions of Ni/Al2O3 catalysts on the biogas decomposition to produce H2-rich gas and carbon nanofibers. Appl. Catal. B Environ. 165, 457e465. Fatsikostas, A.N., Kondarides, D.I., Verykios, X.E., 2002. Production of hydrogen for fuel cells by reformation of biomass-derived ethanol. Catal. Today 75, 145e155. Fernandez, I.A.P., Liua, D.H., Zhaoa, J., 2016. LCA studies comparing alkaline and immobilized enzyme catalyst processes for biodiesel production under Brazilian conditions. Resour. Conserv. Recycl. http://dx.doi.org/10.1016/ j.resconrec.2016.05.009. Goedkoop, M.J., Heijungs, R., Huijbregts, M., De Schryver, A., Struijs, J., Van Zelm, R., 2009. ReCiPe 2008, A life cycle impact assessment method which comprises harmonised category indicators at the midpoint and the endpoint level. In: Report I: Characterisation, first ed. 6 January 2009. http://www.lcia-recipe.net Jung, Y., Yoon, W., Seo, Y., Rhee, Y., 2012. The effect of precipitants on Ni-Al2O3 catalysts prepared by a co-precipitation method for internal reforming in molten carbonate fuel cells. Catal. Commun. 26, 103e111. Levasseura, A., Cavalett, O., Fuglestvedt, J.S., Gasserd, T., Johansson, D.J.A., Jørgenseng, S.V., Raugeih, M., Reisingeri, A., Schivleyj, G., Strømmank, A., Tanakal, K., Cherubinik, F., 2016. Enhancing life cycle impact assessment from climate science: review of recent findings and recommendations for application to LCA. Ecol. Indic. 71, 163e174.  AnCi Nikoli c, V., Agarski, B., Kamberovi c, Z., c, Z., Budak, I., Kosec, B., 2016. Multicriteria analysis of Ni-based catalysts synthesis methods. Mater. Technol. 50, 553e558.  AnCi Nikoli c, V., Kamberovi c, Z., c, Z., Kora c, M., Soki c, M., 2014a. Synthesis of a-Al2O3 based foams with improved properties as catalyst carriers. Mater. Technol. 48, 45e50.  AnCi Nikoli c, V., Kamberovi c, Z., c, Z., Kora c, M., Soki c, M., Maksimovi c, V., 2014b. Influences of synthesis methods and modifier addition on the properties of Nibased catalysts supported on reticulated ceramic foams. Int. J. Miner. Metall. Mater. 21, 806e812. Omata, K., Endo, Y., Ishii, H., Masuda, A., Yamada, M., 2008. Effective additives of Ni/ a-Al2O3 catalyst at low methane conversion of oxidative reforming for syngas formation. Appl. Catal. A General 351, 54e58. Profeti, L.P.R., Dias, J.A.C., Assaf, J.M., Assaf, E.M., 2009. Hydrogen production by steam reforming of ethanol over Ni-based catalysts promoted with noble metals. J. Power Sources 190, 525e533. Ravelli, D., Dondi, D., Fagnoni, M., Albini, A., 2010. Titanium dioxide photocatalysis: an assessment of the environmental compatibility for the case of the functionalization of heterocyclics. Appl. Catal. B Environ. 99, 442e447. Rodríguez, R., Espada, J.J., Pariente, M.I., Melero, J.A., Martínez, F., Molina, R., 2016. Comparative life cycle assessment (LCA) study of heterogeneous and homogenous Fenton processes for the treatment of pharmaceutical wastewater. J. Clean. Prod. 124, 21e29. Sengupta, S., Ray, K., Deo, G., 2014. Effects of modifying Ni/Al2O3 catalyst with cobalt on the reforming of CH4 with CO2 and cracking of CH4 reactions. Int. J. Hydrogen Energy 39, 11462e11472. Sharifi, M., Haghighi, M., Rahmani, F., Karimipour, S., 2014. Syngas production via dry reforming of CH4 over Co- and Cu-promoted Ni/Al2O3eZrO2 nanocatalysts synthesized via sequential impregnation and solegel methods. J. Nat. Gas Sci. Eng. 21, 993e1004. Snowden-Swan, L.J., Spies, K.A., Lee, G.J., Zhu, Y., 2016. Life cycle greenhouse gas emissions analysis of catalysts for hydrotreating of fast pyrolysis bio-oil. Biomass Bioenergy 86, 136e145. Takenaka, S., Shigeta, Y., Tanabe, E., Otsuka, K., 2003. Methane decomposition into hydrogen and carbon nanofibers over supported PdeNi catalysts. J. Catal. 220, 468e477. Tu, X., Gallon, H.J., Twigg, M.V., Gorry, P.A., Christopher Whitehead, J., 2011. Dry reforming of methane over a Ni/Al2O3 catalyst in a coaxial dielectric barrier discharge reactor. J. Phys. D Appl. Phys. 44, 1e10. Weidema, B.P., Bauer, Ch., Hischier, R., Mutel, Ch., Nemecek, T., Reinhard, J., Vadenbo, C.O., Wernet, G., 2013. The Ecoinvent Database: Overview and Methodology, Data Quality Guideline for the Ecoinvent Database Version 3. www.ecoinvent.org. Yaseneva, P., Marti, C.F., Palomares, E., Fan, X., Morgan, T., Saz Perez, P., Ronning, M., Huang, F., Yuranova, T., Kiwi-Minsker, L., Derrouiche, S., Lapkin, A.A., 2014. Efficient reduction of bromates using carbon nanofibre supported catalysts: experimental and a comparative life cycle assessment study. Chem. Eng. J. 248, 230e241. Yurdakul, M., Ayas, N., Bizkarra, K., El Doukkali, M., Cambra, J.F., 2016. Preparation of Ni-based catalysts to produce hydrogen from glycerol by steam reforming process. Int. J. Hydrogen Energy 41, 8084e8091. Zhang, A.J., Zhu, A.M., Guo, J., Xu, Y., Shi, C., 2010. Conversion of greenhouse gases into syngas via combined effects of discharge activation and catalysis. Chem. Eng. J. 156, 601e606.