This is the pre-peer reviewed version of the following article: Catalán, E., Komilis, D. and Sánchez, A. A life cycle assessment on the dehairing of rawhides: chemical treatment versus enzymatic recovery through solid state fermentation in Journal of industrial ecology (Ed. Wiley), april 2018, which has been published in final form at : DOI 10.1111/jiec.12753. This article may be used for non-commercial purposes in accordance with Wiley Terms and Conditions for Use of Self-Archived Versions. © 2018. This manuscript version is made available under the “All rights reserved” license 1 A life cycle assessment on the dehairing of rawhides: Chemical 2 treatment versus enzymatic recovery through solid state 3 fermentation 4 5 Eva Catalán, Dimitrios Komilis*, Antoni Sánchez 6 7 Composting Research Group (GICOM) 8 Department of Chemical, Biological and Environmental Engineering (Escola d’Enginyeria) 9 Universitat Autònoma de Barcelona 10 Bellaterra, Cerdanyola del Vallès, 08193 Barcelona, Spain 11 12 *Corresponding author: 13 Tel/Fax.: +30-2541079391 14 E-mail address: dkomilis@env.duth.gr 15 16 17 18 19 20 21 1 22 Abstract 23 The leather industry needs to switch from the traditional chemically based dehairing process to an 24 environmentally friendly one so that the overall burdens to the environment are reduced. The primary 25 goal of the work was thus to compare the chemical leather dehairing process to an enzymatically based 26 one using the enzymes that are extracted after the application of solid state fermentation (SSF) on hair 27 wastes generated after dehairing. The environmental burdens of the dehairing stage were determined 28 using a life cycle assessment (LCA) approach by comparing the two aforementioned management 29 scenarios. The first scenario was the commonly used technology in which hair is removed via a 30 chemical process and then composted in open piles. This scenario included two sub-scenarios where 31 hair waste is either incinerated or landfilled. In the second scenario, the proteolytic enzymes extracted 32 during the solid-state fermentation of the residual hair are used to dehair the new rawhides instead of 33 chemicals. Industrial and laboratory data were combined with international databases using the 34 SimaPro 8.0 LCA software to make comparisons. The environmental impacts associated to the 35 enzymatic dehairing were significantly lower than the ones associated to the conventional chemical 36 dehairing process. This difference is attributed to the impacts associated to the original production of 37 the chemicals and to the electricity consumed in the conventional method. A sensitivity analysis 38 revealed that the results are affected by the amounts of chemicals used during dehairing. 39 Keywords: life cycle assessment; solid state fermentation; protease; dehairing; leather industry 40 41 42 43 44 45 46 47 2 48 Abbreviations 49 ALO: Agricultural land occupation 50 BOD: Biological oxygen demand (ppm) 51 CaO: Calcium oxide 52 CC: Climate Change 53 COD: Chemical oxygen demand (ppm) 54 DM: Dry matter (kg) 55 FD: Fossil fuel depletion 56 FE: Freshwater eutrophication 57 FET: Freshwater ecotoxicity 58 FU: Functional unit (kg) 59 HT: Human toxicity 60 IR: Ionizing radiation 61 LCA: Life cycle assessment 62 LCI: Life cycle inventory 63 LTUI: Leather Tanner’s Union in Igualada 64 ME: Marine eutrophication 65 MET: Marine ecotoxicity 66 MRD: Mineral resource depletion 67 Na2CO3: Sodium carbonate 68 NaHS: Sodium hydrosulfide 69 Na2S: Sodium sulfide 70 NLT: Natural land transformation 71 OD: Ozone depletion 72 PMF: Particulate matter formation 73 POF: Photochemical oxidant formation 74 SSF: Solid state fermentation 75 TA: Terrestrial acidification 76 TET: Terrestrial ecotoxicity 77 TS: Total solids 78 ULO: Urban land occupation 79 WD: Water depletion 80 3 81 1. Introduction 82 Large amounts of organic wastes are produced worldwide that can be treated by biological treatment 83 technologies under aerobic or anaerobic conditions. The principal aerobic treatment processes are 84 composting and solid state fermentation (SSF). The main aim of composting is to reduce the volume of 85 the wastes, to stabilize organic matter and to generate a compost for agricultural use. Solid state 86 fermentation is a process with the main objective of generating bioproducts (e.g. enzymes), after an 87 extraction procedure, and to generate compost with the fermented solid (Abraham et al. 2014; Abu 88 Yazid et al. 2016). SSF has proven to be a very promising technology in the development of several 89 bioprocesses and products, since it holds tremendous potential for the production of enzymes. It can be 90 of special interest in those processes where the crude fermented product may be used directly as an 91 enzyme source (Doelle et al. 2009). This technique has, therefore, become an attractive alternative for 92 specific applications. In recent years, for example, the production of enzymes from various organic 93 substrates using solid state fermentation has been evaluated. In particular, protease production by SSF 94 under different process conditions, microorganisms and substrates has been demonstrated in different 95 studies (Singhania et al. 2009). Recently, Abu Yazid et al. (2016) presented the protease production 96 and extraction using a pilot-batch mode operation using hair waste from the tannery industry. Also, the 97 yields and recovery achieved on an easily scalable low-cost downstream process have been presented 98 (Abraham et al. 2014), with an obtained recovery of 74%. 99 100 1.1 The leather processing industry 101 Leather processing involves a series of unit operations that can be classified into three groups: 102 I. Pre-tanning or “beamhouse” operations 103 II. Tanning 104 III. Post-tanning and finishing operations 105 Pre-tanning includes different steps such as soaking, fleshing, dehairing and liming, deliming, bating 106 and pickling. During pre-tanning, the previously dehydrated raw material (the rawhides) must be 107 carefully rehydrated before it can be subjected to mechanical action. The flesh layer (meat) is removed 108 to aid in the penetration of chemicals. Fleshing can be done after slaughter, after soaking or after 4 109 liming. Then, conventionally, the hides or skins are treated with different chemicals to destroy the 110 keratinous material of the epidermis and to remove hair. During this process, hair roots and pigments 111 are removed. Pickling increases the acidity of the hide to a pH of 3, and salts are added to prevent the 112 hide from swelling. For preservation purposes, fungicides and bactericides are applied to the dehaired 113 leather. The hair waste (solid residue) is produced at this stage (Thanikaivelan et al. 2004). 114 According to studies as Saran et al. (2013) or George et al. (2014), the pre-tanning operations uses 115 chemicals that may have hazardous effect to tannery workers and to the environment. Almost 70% of 116 the total pollution of the process is produced in these operations (Thanikaivelan et al. 2004). The use of 117 proteases in this step can be a viable and green alternative to the conventional chemical process that 118 can lead to substantial reduction of the amount of effluent and its toxicity (Kamini et al. 1999). 119 Daddi et al. (2016) and Laurenti et al. (2016) showed that the use of chemicals in the tanning process 120 contributes more than 60% to the environmental impact in the leather industry. Chemicals’ usage and 121 water resources depletion are the main environmental impacts of the whole tanning process. 122 Composting is traditionally used to treat hair residues (hair wastes) generated during the conventional 123 dehairing process in the leather industry. SSF, a process similar to composting, could be applied to 124 extract enzymes (mainly proteases) that can replace chemicals in the original dehairing process. The 125 use of extracted enzymes in the dehairing process can result in: a) shorter processing (dehairing) time, 126 b) practical elimination of chemicals and c) lower amount of wastewater generated (Valeika et al. 127 2009). In fact, the efficiency of proteases extracted during the SSF of hair waste to replace typical 128 chemicals used to dehair rawhides has been well demonstrated in recent studies (Abraham et al. 2014; 129 Abu Yazid et al. 2016). 130 The proteases can be excellent alternatives to decompose hair keratin. Around 40% of keratin has been 131 found to be decomposed through digestion with protease after 10 minutes (Park et al. 2004.) In recent 132 studies, it has been demonstrated that after 24h of incubation, between 90% and 95% of hair was 133 removed from raw hides following treatment with proteases, which is a value similar to that obtained 134 with the common chemicals (Abu Yazid et al. 2016). 135 136 5 137 1.2 Gaps in knowledge and scope of work 138 The use of SSF as an enzymatically sustainable process has rarely been studied. In particular, the use of 139 LCA to assess the environmental impact of this technology has never been reported. Hence, the main 140 objective of this study was to compare the two principal scenarios for dehairing in a typical leather 141 industry, namely: 142 • The existing chemical dehairing process in which hair waste is then composted to provide a 143 potential soil amendment 144 • An enzyme driven dehairing process, in which enzymes are extracted during the SSF of the hair 145 residues and used to fully replace the traditional chemicals used in dehairing. The resulting SSF 146 solid end product, after the enzyme extraction, is also directed to composting (or other treatment 147 techniques such as landfilling or incineration). 148 A LCA-based comparison was performed to quantify direct and upstream environmental burdens for 149 both scenarios (chemical vs enzymatic dehairing), including three sub-scenarios for the traditional 150 method, in order to investigate the optimum treatment (composting, landfilling, incineration) of hair 151 wastes. It is clarified that our study focused only on the dehairing stage and the composting of the 152 removed hair (hair waste). That is, we did not focus on the whole leather processing life cycle (Figure 153 1), since this was beyond the scope of our work. We chose the dehairing stage, however, since it 154 contains the two main stages in a leather industry that are responsible for the high usage of chemicals. 155 A sensitivity analysis was performed for both main scenarios to study the effect of key parameters on 156 the outputs. 157 158 2. Methodology and case study 159 To better study the process, we considered a tannery industry located in North Catalunya, Spain, as a 160 typical model industry. Operating data were collected from that industry that currently adopts the 161 conventional chemical based dehairing process to remove hair. Data related to the environmental 162 burdens of the SSF were based on pertinent laboratory experiments. 163 164 2.1. The conventional chemical dehairing process 6 165 The first scenario consists of the conventional chemical process as depicted in Fig. 1 and described 166 below. 167 168 Insert Figure 1 169 170 Conventionally, hair removal is carried out by chemical and mechanical means. The keratinous 171 material and fat are removed from the pelts mainly with sulphides (NaHS or Na2S) and lime. In the first 172 stage, soaking is performed to allow hides to re-absorb any water that may have been lost after flaying 173 to clean the hides, and to remove inter-fibrillary material. The data included in the first scenario 174 considered in this study and the system boundaries are from the Leather Tanner’s Union in Igualada, 175 Spain (LTUI 2015) for 100 kg of initial leather after the salting stage and before rehydration/soaking 176 (i.e. referred to as rawhide); these 100 kg of rawhide were defined as the functional unit (FU) in this 177 work. 178 The sequential chemical treatment of raw cow hides consists of the soaking, liming and dehairing 179 stages, namely: 180 1. In the soaking stage, the following reagents are added and stirred over 30h: 600 kg H2O at 181 25ºC, 0.2 kg surfactant, 0.5 kg Na2CO3, 0.25 kg NaHS and 0.1 kg NaOH. 182 2. In the liming and dehairing stages, the following reagents are added and mixed over a 5h 183 period: 700 kg H2O, 3.3 kg CaO, 0.3 kg NaOH, 0.8 kg NaHS and 0.5 kg Na2S. This step produces 13 184 kg of solid (mainly hair waste) and 703 kg of liquid residues. 185 The water effluent generated by the whole process (from both steps combined) contains a high 186 inorganic and organic polluting load. The soaking stage is the most polluting stage of the tanning 187 process since it contributes up to 50-55% of the total polluting loading (Chowdhury et al. 2015). This is 188 due to the inorganic chemicals that are used for the treatment of the hides, such as lime, sodium 189 carbonate and sodium hydroxide. In the liming stage, hair, skin and emulsion fats (i.e. a mixture of 190 sodium soap and fat) are removed from the hides, are released to the effluent and increase its total 191 solids (TS) content (Chowdhury et al. 2015). Wastewater is normally treated by an on-site or central 192 wastewater treatment plant and the treated effluent is discharged to surface water. According to 193 Chowdhury et al. (2015), the principal pollutants in tannery wastewater is increase sulfates, chlorides 194 and sodium cations, as well as chemical oxygen demand (COD) (between approximately 4000 to 5000 7 195 mg/L for each of the above four parameters), whilst a typical BOD5 is around 900 mg/L (Chowdhury et 196 al. 2015). These literature values are in agreement with the corresponding field values provided after 197 the personal communication with LTUI (2015). The hair residues obtained after dehairing become 198 solid waste that is normally treated via composting or disposed of to a landfill. Composting takes place 199 either in turned windrows or in-vessel. At this point, the main environmental emissions of the process 200 are primarily ammonia due to the high nitrogen content of hair waste (Barrena et al. 2007). Alternative 201 techniques to treat hair waste is landfilling and incineration (sub-scenario 1: landfill; sub-scenario 2: 202 incineration) as illustrated in Figure 2. However, the dominant techniques are composting and 203 landfilling (LTUI 2015). 204 Μore detail about the conditions and the operations, during all stages, in a beamhouse that utilizes the 205 conventional dehairing process can be found in Thanikaivelan et al. (2004) and Ramasami et al. (1999). 206 207 Insert Figure 2 208 209 2.2. The enzymatic based dehairing process 210 The hair waste is a good source of protein with a content of 65-95% DM (Dawber 1996). Recent 211 studies have proven the efficiency of SSF to produce alkaline proteases from hair waste. These 212 enzymes could be used in the dehairing process of new rawhides instead of chemicals (Abu Yazid et al. 213 2016). This enzymatic dehairing was tried at laboratory scale (Abraham et al. 2014) and is briefly 214 described below: Fermented solid material was mixed thoroughly with buffer HCl-Tris 215 (tris(hydroxymethyl)aminomethane) according to Abu Yazid et al. (2016). The incubation for dehairing 216 performed is using the protease extract in the hides according to the method of Abraham et al. (2014). 217 SSF was performed out in pilot-scale reactors (10-50 L) that worked under near-adiabatic conditions 218 and a continuous aeration regime (Abraham et al. 2014; Santis-Navarro et al. 2011). Those experiments 219 provided all the necessary data (emissions, enzyme activity) to carry out the LCA study describe here. 220 Fermented solid material, after reaching the thermophilic phase, was mixed with buffer for 1 hr. The 221 enzyme extract was separated by centrifugation and filtration through a 0.45 mm filter. The complete 8 222 description of the SSF based enzymatic dehairing process, including operational conditions, material 223 conditioning, quality of final product, etc., can be found, in Abraham et al. (2014) and Abu Yazid et al. 224 (2016). It is noted however that the enzymatic dehairing process does not exist yet in the field scale. 225 Thus, all data necessary to perform our LCA were obtained from the aforementioned references 226 performed at laboratory scale. 227 To apply the enzymes, wet-salted cow hides were washed and cut in similar shapes (approximately 15 228 cm2), then they were incubated with enzymatic crude extract at 37ºC for 24 hours on a rotatory shaker. 229 After 24 hours of incubation, the hides treated with these specific proteases showed an easier removal 230 of hair when mechanically scraped compared to the chemical dehairing process (Abraham et al. 2014). 231 This modification in the process implies a likely less water consumption, decreased to negligible 232 chemical usage and likely less wastewater emissions. For example, BOD and COD loadings from the 233 effluent during enzymatic processing of buffalo hide skins is reduced by 82% and 85% respectively 234 when compared to conventional processing thought chemicals (Saran et al. 2013). A scheme of the 235 enzymatic dehairing process is presented in Figure 3, which also shows the internal loop of the enzyme 236 generation, during SSF, and the reuse of the enzymes during dehairing of new leather. It is this loop 237 that this work attempts to compare, on an LCA basis, with the conventional chemical dehairing 238 process. 239 240 Insert Figure 3 241 242 2.3. Functional unit (FU) 243 The FU in LCA provides a reference to related inputs and outputs and to allow comparisons among 244 systems (International Organization for Standardization, 2006). In this study, the functional unit was 245 determined to be the 100 kg of rewetted (soaked) hide, which have resulted from the salting stage of 95 246 kg of initial hide. That is, the 5kg gain is due to the rehydration (soaking) that follows the salting 247 performed during storage. Since several coefficients in the tannery industry are based on an area basis 248 (m2), the coefficient of 7.5 kg/m2 was used to convert units of kg to m2 and vice versa, where necessary 249 (LTUI, 2015). 9 250 251 2.4. LCA software and inventory data per scenario 252 SimaPro® v.8 and the Ecoinvent v.3 database were used to perform the LCA based comparison. In 253 addition, the ReCiPe methodology (ReciPe 2016) was adopted to calculate the environmental impacts. 254 ReCiPe is a follow up of the Eco-Indicator 99 and the CML method and has two levels of indicators, 255 namely: i) Midpoint indicators ii) Endpoint indicators. At the midpoint level, which was adopted here, 256 there are 18 impact categories. The 18 midpoint impact categories of Recipe are presented in Table 1. 257 258 Insert Table 1 259 260 In the present study, the life cycle inventory (LCI) original data with regard to the conventional process 261 were obtained by the Leather Tanner’s Union in Igualada (Catalunya). Additional data were obtained 262 from the literature especially regarding the enzyme dehairing process that is still in an experimental 263 stage. The main inputs and outputs that are common in both processes are given in Table 2. Water 264 consumption is presented in supplementary Table S1. All data have been converted to correspond to 265 the functional unit used here (100 kg of rawhide prior to soaking). 266 267 Insert Table 2 268 269 a. Base scenario: conventional chemical dehairing 270 The flow diagram of the typical dehairing process that uses chemicals (herein referred to as chemical 271 process) was presented in Figure 1. This case is characterized, apart from the use of chemicals, by a 272 relatively high water consumption. According to the collected data, not all of the used water is fresh 273 water: 75% of the water usage is actually recycled water. After the process, wastewater is treated in a 274 wastewater plant. In the case of Igualada, which was the model tannery industry in our case (LTUI, 275 2015), wastewater treatment takes place in a central wastewater plant located in the industrial site. The 10 276 hair waste is stabilized via composting with bulking agent to produce compost with a high nitrogen 277 content (Barrena et al. 2007). The main inputs and outputs for the chemical process are shown in 278 supplementary Table S1. 279 280 281 b. Alternative scenario: Use of extracted enzymes to perform dehairing 282 The alternative enzymatic based dehairing scenario is shown in Figure 3 and the specific data used are 283 displayed in supplementary Table S2. It is noted that the water used to perform the enzyme extraction 284 was considered equal to the water necessary to perform the soaking step that is included in the 285 traditional process. For this reason, the soaking step does not appear in the enzymatic process (see 286 Figure 3). In addition, due to the lack of field data, we considered that the enzymatic process has the 287 same electricity consumption as the traditional process, during soaking/dehairing. 288 289 290 2.5. Technical assumptions and system boundaries 291 The characterization of each raw material, energy and water consumptions have been mainly obtained 292 from the Ecoinvent database v3. As the main goal of this paper is to analyze and compare the 293 environmental impacts of the dehairing process for the two selected dehairing scenarios (conventional 294 chemical, enzymatic based), within the boundaries of the systems, the following sub-systems have been 295 included: i) the production of the hair via the dehairing process, ii) the treatment of hair wastes: 296 composting or solid state fermentation and iii) the treatment of exhaust gases emitted during the 297 previous stage. In this study, the LCI includes the consumption of chemical products, energy, and water 298 during each process. In the evaluated scenarios, the treatment of wastewater is considered to lie outside 299 the limits of the system, while any transport of raw materials and products into or outside the system is 300 also not taken into account. 301 302 3. Results and discussion 303 3.1 Environmental impacts assessment of the conventional chemical and the enzymatic dehairing 304 processes 11 305 The environmental impacts associated to both processes are presented in Figure 4. For each process, 306 both direct and indirect emissions are presented. The environmental impacts due to direct emissions are 307 those associated to the dehairing process per se, while the environmental impacts associated to indirect 308 emissions are, for example, those that come from the production of wood chips, the production of the 309 chemicals, or the electricity production that highly depends on the local grid and the specific usage of 310 fuel. In our case, the (medium voltage) electrical grid of Spain was used as this is defined in the 311 Ecoinvent® database. To facilitate the comparison between the impact categories, all results are 312 normalized so that all category indicators have the same units (see Figures 4, 5 & 6). Normalized units 313 are calculated by the division of the actual impact emission, at each category by a reference emission. 314 A commonly used reference coefficient, for example, is the average yearly environmental (pollutant) 315 load in a country or continent, divided by the number of inhabitants. There are various methodologies 316 to perform normalization. The ReciPe methodology was utilized here. A detailed explanation of the 317 normalization procedure can be found in Sleeswijk et al. (2007). 318 Figure 4 (top) shows that the main direct environmental impacts (red color) in the conventional 319 chemical method are, in hierarchical order, freshwater ecotoxicity, marine ecotoxicity, human toxicity 320 and terrestrial ecotoxicity, which are mainly due to the wastewater discharge. 321 Regarding the indirect impacts associated to the conventional chemical method, the highest ones are 322 related to electricity consumption, and the production of sodium sulphite and sodium hydroxisulfide 323 that are used in chemical dehairing. The categories with the highest impact were marine ecotoxicity, 324 freshwater ecotoxicity and natural land transformation. It is noted that the electricity consumption 325 shown (yellow color) is the sum of the electricity consumption at the dehairing stage (LTUI 2015) and 326 in the composting facility (Colón et al. 2011). 327 The negative values observed are due to the beneficial use of the compost that can partly replace some 328 chemical fertilizer, due to its nitrogen content. The negative values imply environmental benefits rather 329 than net burdens since environmental burdens associate to the production of the chemical fertilizer are 330 now reduced due to compost usage. 331 According to the results shown in Figure 4 (bottom) for the enzymatic scenario, the impacts related to 332 fresh water ecotoxicity, marine ecotoxicity and fresh water eutrophication, had positive values, which 333 means high environmental loads. These burdens are due to the high electricity consumption, mainly 12 334 due to the usage of an in-vessel system in both the composting and SSF systems (the same reactor is 335 assumed to be used in both technologies). 336 The direct impacts associated to the enzymatic based dehairing process are relatively low and are 337 basically caused by the ammonia emissions during the SSF. In this case, obviously, no indirect 338 emissions are observed from the usage of chemicals, since they are not used anymore. As observed in 339 Figure 4 (bottom), most of the environmental impacts are negative, due to the benefits gained by the 340 compost application that partly replaces some chemical fertilizer (Sánchez et al. 2016). Another way to 341 analyze the environmental impacts is to study at which stage the main impacts are created in both 342 scenarios. It is revealed from Figure 4 that the highest contribution to each impact category is due to 343 the direct emissions (mainly as wastewater effluent) during the chemical dehairing (red bars) and the 344 electrical energy consumption (yellow bars), with regard to the indirect emissions. In the case of the 345 enzymatic process (Figure 4 bottom), the electrical energy consumption justifies most of the emissions. 346 Since in the typical chemical dehairing method, there were much more available wastewater effluent 347 data than in the enzymatically based dehairing process, the LCA simulation was run by using COD as 348 the only effluent parameter which is the common available wastewater parameter from both processes. 349 The chemical dehairing process was proven to be still the one with the highest environmental impacts. 350 This indicates that it is the production of chemicals and the electrical energy consumed in the chemical 351 dehairing process that are mostly responsible for the overall emissions of the (base) chemical dehairing 352 scenario. The wastewater does contribute in the environmental impacts, but not as much as the 353 chemical production and the electrical energy. 354 355 3.2 Comparison of three hair waste traditional treatment techniques in the conventional chemical 356 dehairing process. 357 Hair waste generated in the leather industry can be treated via composting, although landfilling and 358 incineration are two viable treatments options too. Figure 5 shows the comparison to the possible three 359 sub-scenarios. According to Figure 5, composting has the lowest environmental impacts for almost all 360 categories. Incineration, for example, is the most environmentally polluting with regard to freshwater 361 and marine ecotoxicity, likely due to the generation of the acidic off-gases CO2, NOx and SOx, 13 362 (Assamoi et al. 2012) that lower the pH of aquatic systems. Landfilling follows as the second most 363 polluting method. In conclusion, in-vessel composting is the method that generates the lowest 364 environmental impacts. For this reason, it was considered in this work as the optimal hair waste 365 treatment method. 366 367 Insert Figure 5 368 369 3.3 Comparison of the conventional chemical dehairing process with the enzymatically driven 370 dehairing process. 371 Figure 6 shows the comparison of the conventional chemical dehairing process with the enzymatic 372 dehairing process via the use of SSF. This figure shows the total emissions, namely the sum of direct 373 and indirect environmental impacts per category and per scenario. No categorization per type of 374 parameter is done as was done in Figure 4. According to Figure 6, it can be clearly observed that the 375 conventional chemical process has significantly higher environmental impacts than the enzymatic 376 dehairing process via the application of SSF. This difference is mainly attributed to the avoidance of 377 the chemicals during dehairing, when SSF is applied, and to the reduction of the electrical energy 378 consumption. Table 3 clearly shows that there is significant reduction in all impact categories when the 379 dehairing is changed from chemically induced to enzymatically induced, proving the benefit of 380 adopting this technology in the dehairing stage. The range of reductions is from 20% (Ionizing 381 radiation) to up to 1942% (Metal depletion). 382 383 Insert Figure 6 384 Insert Table 3 385 These results agree with the study of Nielsen (2013) who performed an LCA to compare the 386 environmental impact of conventional soaking and liming processes against enzyme assisted processes. 387 Their work showed that environmental impacts of producing the enzyme were much smaller compared 14 388 to the impacts of the traditional method (Nielsen 2013). It is noted that although the functional unit and 389 all calculations used in this paper is per FU of 100 kg of rawhide, the environmental impacts are 390 expressed on a per kg basis due to SimaPro® limitations (see Figures 5 and 6). 391 Based on the above results, the impacts of four categories under each scenario are specifically 392 discussed below: 393 Human toxicity: Figure 6 demonstrates that the enzymatic process has a negative impact on human 394 health indicating environmental benefits. A detailed analysis of the results confirms that the impact of 395 all stages equals 6.12∙10-5 kg of 1,4-DCB eq/kg rawhide for the chemical process. 396 Freshwater aquatic ecotoxicity: Both analyzed systems show positive values here; yet, the impact of 397 the chemical process is again much higher. This is mainly attributed to the high amounts of chloride 398 and sulfate contained in the effluent from the chemically based dehairing process. In the enzymatic 399 process, energy production emission is the most important contributor associated at indirect production. 400 The contribution of this impact in the conventional chemical process is 43% more than that in the 401 enzymatic process. 402 Marine ecotoxicity: In the conventional chemical process, the largest positive contributions to the 403 marine ecotoxicity impact category are from effluent emissions, mostly due to emission into water; the 404 consumption of electrical energy (indirect emissions) and to the chemicals production (indirect 405 emissions). In the case of the enzymatic process, agricultural compost application is responsible for 406 avoided marine ecotoxicity impacts due to compost application that replaces chemical fertilizer (which 407 explains the negative impact in this category). 408 Natural land transformation: This is the natural land transformed and occupied for a certain time. The 409 unit is m2∙year (PRé Consultants 2015). As a reference, in the case of the conventional chemical 410 method, the land transformation (a positive impact) is associated to the energy production and the 411 sodium sulfite production used in the chemical dehairing. 412 413 3.4 Sensitivity analysis 15 414 Given the uncertainties that characterize the LCA phases, the final outcomes of a LCA should be tested 415 via sensitivity and uncertainty analysis to improve its robustness and transparency (Guo et al. 2012). A 416 sensitivity analysis was therefore performed to determine the effects of changing selected model or 417 assessment parameters on the results. In the sensitivity analysis, three alternative scenarios were 418 formed, as described below: 419 • Alternative 1: Doubling of the amount of chemicals used in the conventional chemical 420 process. 421 • Alternative 2: Removing the fertilizing ability of the produced compost in both processes 422 (chemical and enzymatic). 423 • Alternative 3: Use of the turned windrow system instead of the in-vessel system during 424 composting. The energy requirements of the turned windrow system were based on Colón et 425 al. (2011). 426 Alternative 1 observes the effect of the use of the chemicals used in the conventional dehairing step. 427 Specifically, under this alternative scenario, we doubled the quantity of the chemical substances 428 compared to the base scenario. Results are depicted in supplementary Figure S1 (top). The first thing 429 that can be observed is a marked increase in the final environmental impacts under all categories. This 430 is expected, since major impacts associated to the chemical process involve the production of the 431 specific chemicals. A notable increase (more than double compared to the base scenario) is observed 432 in the human toxicity scenario. By noticing the other impact categories too, it is observed that an 433 increase does occur when doubling the use of chemicals, but this does not necessarily lead to a 434 corresponding doubling of the environmental impacts in all categories. 435 In the case of the 2nd alternative, Figure S1 (bottom) demonstrates that the application of 436 compost to replace a chemical fertilizer is also important. When the compost "looses" its fertilizing 437 abilities, all impact categories have higher values compared to those of the base scenario. In particular, 438 the impacts under the freshwater and marine ecotoxicity are now mostly increased and have the higher 439 values among all 3 scenarios. Figure S1 (bottom) also confirms the importance of the land application 440 of compost even with the enzymatic dehairing process. That is, the environmental impacts in all 441 categories highly increase when compost looses its fertilizing ability in the enzymatically based 16 442 dehairing system as well. Figure S1 (bottom) also reveals that certain categories are actually not 443 influenced by the application of compost. For example, categories such as ozone depletion, ionizing 444 radiation and agriculture and urban land occupation are not affected at all. Another important aspect to 445 take into account is the wastewater obtained. In the enzymatic process, this wastewater implies a 446 significant reduction of the environmental impact in different categories. But it should be kept in mind 447 that in the enzymatic process, there were available values in the literature only for BOD and COD 448 loadings in the wastewater and not for the other parameters as occurs for the chemical dehairing 449 process. It should be also pointed out that the highest reduction is observed in the fresh water 450 ecotoxicity follow the marine water ecotoxicity. 451 Regarding alternative 3, Figure S2 reveals that by changing the composting system to turned 452 windrow slightly reduces the environmental impacts. This is because the turned windrow system 453 consumes less electricity than the in-vessel system (which traditionally operates on electricity) despite 454 its higher usage of diesel. Actually, Colón et al. (2011) have shown that the total energy consumption 455 in the turned windrow system is 16% less than that in-vessel system, which accounts for the combined 456 use of electricity and diesel. That is, although the turned windrow system has a higher diesel 457 consumption than the in-vessel, the overall impacts of the turned windrow technology are lower than 458 those of the in-vessel system. 459 The sensitivity analysis of the LCA methods shows that water ecotoxicity (fresh water and 460 marine), and terrestrial acidification are the impact categories that are mostly affected by the chemical 461 process. The enzymatic process leads, for all impact categories, to lower environmental burdens 462 rendering it as an environmental friendly alternative that needs to be tried and practiced in full scale by 463 tannery industries. 464 465 3.5 Preliminary economical assessment 466 In this section is an account of the initial exploration in ongoing work on the economic impact of 467 conventional chemical dehairing process into account the cost of chemicals, hair waste treatment and 468 wastewater treatment. The objective of this preliminary estimation is to calculate the potential 469 economic savings after using the proteases obtained through SSF that replace chemicals. by hair waste 17 470 thanks to save these costs. The preliminary economic assessment has shown that nearly 7 million € per 471 year can be saved due to avoidance of chemicals and due to reduced cost of wastewater treatment 472 compared to conventional dehairing. The unit typical costs encountered in conventional treatment are: 473 chemicals: 0.168 €/m2 leather, hair waste treatment: 0.043 €/m2 leather, wastewater treatment: 1.07 474 €/m2 leather (LTUI 2015). No actual data exist for the enzymatic dehairing process, since this technique 475 has not been applied to the field yet. 476 477 4. Conclusions 478 There is already enough knowledge on the environmental impacts associated to the production of using 479 the conventional chemical dehairing process. Resource consumption (chemical or water usage mainly) 480 and residue treatment are the principal contributors to the overall environmental impacts of the process. 481 In this work we attempted to integrate this knowledge and the broader perspective offered by LCA into 482 waste management so that to compare conventional dehairing techniques with greener alternatives that 483 abide to the principles of circular economy. Thus, we analyzed an alternative of how to valorize a 484 residue produced by the process (hair waste) so that to recover a compound (protease) that can replace 485 the chemicals used typically in dehairing. 486 The conclusions from this study are: 487 • The LCA results show that the substitution of chemicals by enzymes obtained from SSF of hair 488 wastes leads to substantially lower environmental impacts compared to the conventional chemical 489 method during leather dehairing as revealed in Figure 6 and Table 3. The highest reduction was 490 observed in the metals depletion category. 491 • The categories with the highest impact in the conventional chemical dehairing process were the 492 water ecotoxicity (freshwater, marine), water eutrophication and human toxicity followed by 493 natural land transformation and terrestrial acidification. 494 • Enzymatic process has a contribution only in five impact categories, namely freshwater 495 eutrophication, marine eutrophication, freshwater ecotoxicity, ionizing radiation and fossil fuel 496 depletion, whilst there is minimal influence on the other thirteen categories. 18 497 • Based on the sensitivity analysis, it can be concluded that the overall burdens associated to the 498 enzymatic process are lower compared to those of the conventional process. Therefore, the results 499 demonstrate the eco-efficiency of the hair waste management by the enzymatically induced 500 dehairing, since the enzyme produced in SSF can successfully substitute the chemicals used in 501 dehairing. 502 • Composting proved to result in the least environmental burdens when treating hair wastes 503 compared to landfilling and incineration. 504 • Enzymatically induced dehairing can lead to a sustainable production of leather and to the 505 reduction of the overall environmental impacts compared to the traditional chemical dehairing 506 process. 507 508 Acknowledgements 509 E. Catalán appreciates the financial support provided by UAB with a pre-doctoral grant by Mineco 510 CTM2015-69513R. All authors thank Igualada tanneries, and particularly Mr. Miquel Vila, for their 511 collaboration in providing detailed raw data from the leather industry. Dr. Dimitrios Komilis thanks 512 Techniospring-Generalitat de Catalunya for the financial support provided during years 2014-2016 in 513 UAB (TECNIOspring project TECSPR13-1-0006). 514 19 515 References 516 Abraham, J., T.Gea, and A.Sánchez. 2014. Substitution of chemical dehairing by proteases from solid- 517 state fermentation of hair waste. J Cleaner Prod 74: 191-198. 518 Assamoi, B. and Y. Lawrysshyn. 2012. The environmental comparison of landfill vs. incineration of 519 MSW accounting for waste diversion. Waste Manage 32:1019-1030. 520 Abu Yazid, N., Barrena R., and Sánchez A., 2016. Assessment of protease activity in hydrolysed 521 extracts from SSF of hair waste by an indigenous consortium of microorganisms. Waste Manage 49: 522 420-426. 523 Banar, M., Z.Cokaygil, and A.Ozkan. 2009. Life cycle assessment of solid waste management options 524 for Eskisehir, Turkey. Waste manage 29: 54-62. 525 Barrena, R., E.Pagans, A.Artola, F.Vázques, and A. Sánchez. 2007. Co-composting of hair waste from 526 the tanning industry with de-inking and municipal wastewater sludges. Biodegradation 18: 257-268. 527 Barton, J.R, D.Dalley, and V.S.Patel. 1996. Life cycle assessment for waste management. Waste 528 Manage 16: 35-50. 529 Chowdhury, M., M.G. Mostafa, T.K. Biswas and A.K. Saha. 2013. Treatment of leather industrial 530 effluents by filtration and coagulation processes. Water resources industry 3:11-22. 531 Chowdhury, M., M.G.Mostafa, T. Kumar, A. Mandal, and A. Kumar. 2015. Characterization of the 532 effluents from leather processing industries. Environmental processes 2: 173-187. 533 Colón, J., E.Cadena, M.Pognani, R.Barrena, A.Sánchez, and X.Font. 2012. Determination of the 534 energy and environmental burdens associated with the biological treatment of source-separated 535 municipal solid wastes. Energy Environmental Science 5:5731-5741. 536 Daddi T., Nucci B., Fabio I., Testa F., 2016. Enhancing the adoption of life cycle assessment by small 537 and medium enterprises grouped in an industrial cluster. A case study of the tanning cluster in Tuscany 538 (Italy). J Ind. Ecol. 20: 1199-1211. 539 Doelle, H.W., J.S.Rokem, and M.Berovic. 2009. Encyclopedia of life support systems. Biotechnology, 540 Biotechnology- Part II, Volume VI. United Nations Educational, Scientific and Cultural Organization 20 541 George, N., P. Singh Chauhan, V.Kumar, N.Puri, and N.Gupta. 2014. Approach to ecofriendly leather: 542 characterization and application of an alkaline protease for chemical free dehairing of skins and hides 543 at pilot scale. J Cleaner Prod 79: 249-257. 544 Goedkoop M, M. Oele, J. Leitjing, T. Ponsioen, and E. Meijer. 2013. SimaPro 8. PRé-Sustainability, 545 The Netherlands. 546 International Organisation for Standardisation. 2006. ISO 14040-14044. Environmental Management, 547 Life Cycle Assessment. International Standard, Geneva, Switzerland. 548 Kamini N.R, Hemachander C., Geraldine J., Puvankrishnan R., 1999. Microbial enzyme technology as 549 an alternative to conventional chemicals in leather industry. Curr. Sci. 77: 80-86. 550 Laurenti R., Redwood M., Puig R., Frostell B., 2016. Measuring the environmental footprint of leather 551 processing technologies. J Ind. Ecol. In press. DOI: 10.1111/jiec.12504. 552 Leather Tanner's Union of Igualada (LTUI). 2015. Personal communication with technical staff (Sr. 553 Miquel Vila). Igualada, Catalunya, Spain (October) 554 Nielsen, P.H. 2013. Environmental assessment of enzyme application in industrial production. Journal 555 Cleaner Prod 42: 228-240. 556 Maulini-Duran, C., J.Abraham, S. Rodríguez-Pérez, A.Cerda, P.Jiménez-Peñalver, T.Gea, R.Barrena, 557 A.Artola, X.Font, and A.Sánchez. 2015. Gaseous emissions during the solid-state fermentation of 558 different wastes for enzyme production at pilot scale. Bioresource Technol. 179: 211-218. 559 PRé Consultants. 2015. SimaPro Database Manual. Methods library. PRé-Sustainability, The 560 Netherlands. 561 Ramasami T., Rao J.R., Screeram K.J., 1999. Beamhouse and tanning operations: process chemistry 562 revisited. J Soc. Leath. Tech. Ch 83:39-45 563 Recipe Methodology. 2016. http://www.lcia-recipe.net/project-definition. Accessed 15 February 2016 21 564 Sánchez, A., X.Gabarrell, A.Artola, R.Barrena, J.Colón, X.Font, and D.Komilis. 2016. Composting of 565 wastes. In: Taherzadeh MJ and Richards T (ed) Resource recovery to approach zero municipal waste, 566 CRC Press, Taylor and Francis, Boca Raton, FL, pp. 77-106. 567 Saran S, R.Mahajan, R.Kaushik, J.Isar, and R.K.Saxena. 2013. Enzyme mediated beam house 568 operations of leather industry: a needed step towards greener technology. J Cleaner Prod 54: 315-322. 569 Singhania, RR, A.K.Patel, C.R.Soccol, and A.Pandey. 2009. Recent advances in solid-state 570 fermentation, Biochemical Engineer 44:13-18. 571 Sleeswijk A.W., van Oers L.F., Guinée J.B., Huijbregts M.A.J., 2007. Normalization in product life 572 cycle assessment: An LCA of the global and European economic systems in the year 2000. Sci Total 573 Environ 390:227-240. 574 Thanikaivelan, P., J.R.Rao, B.U.Nair, and T.Ramasami. 2004. Progress and recent trends in 575 biotechnological methods for leather processing. Trends in Biotechnology 22:4, 181-188. 576 Park H.Y., Son K.H., Kwon Y.K., Shin D., Min S.G., 2004.. Method for preparing leather using 577 protease and method for treating wastes derived from leather processing. US 20040214309 A1. 578 Valeika, V., K. Beleska, V. Valeikiene, and V. Kolodzeiskis. 2009. An approach to cleaner production: 579 from hair burning to hair saving using a lime-free unhairing system. J Cleaner Prod 17: 214-221. 580 22 581 Legends to figures 582 Fig 1 Flow diagram to conventional chemical dehairing process it included the composting process to 583 treat hair waste. The flows of the inputs and outputs are shown with dashed lines. 584 Fig 2 Flow diagram of the conventional chemical dehairing process with landfilling or incineration to 585 treat hair waste instead of composting. The flows of inputs and outputs are shown with dashed lines. 586 Fig 3 Flow diagram of the enzymatic dehairing process. The flows of inputs and outputs into the 587 system are shown with a dashed line. The enzyme production is not taken into account. 588 Fig 4 LCA results for the dehairing of 1 kg of rawhides via the conventional chemical process (top) 589 and the enzymatically based process (bottom). Category impacts are based on the ReCipe midpoint 590 method. Units per impact are normalized (i.e. calculated by the division of the impact emission at each 591 category by a reference emission coefficient, according to the Recipe methodology). 592 Fig 5 LCA results for three hair waste treatment sub-scenarios (composting, landfilling and 593 incineration) based on 1 kg of rawhide dehaired via the conventional chemical process (it is noted that 594 SimaPro® uses 1 kg as the unit basis to express all normalized results). Units per impact are 595 normalized (i.e. calculated by the division of the impact emission at each category by a reference 596 emission coefficient, according to the Recipe methodology) 597 Fig 6 Overall LCA environmental impacts (presented as a summation) to dehair 1 kg of rawhide via the 598 conventional chemical process and the enzymatic process (it is noted that SimaPro® uses 1 kg as the 599 unit basis to express all normalized results). Units per impact are normalized (i.e. calculated by the 600 division of the impact emission at each category by a reference emission coefficient, according to the 601 Recipe methodology). 602 23 603 Supplementary Figures 604 Fig S1 Comparison of alternative scenarios 1 and 2 with the base scenario. 605 Fig S2 Comparison of alternative scenario 3 with the base scenario 24 25