Life cycle inventory data for ethyl levulinate production from Colombian rice straw

This data article is associated with the research article “Sustainable production of ethyl levulinate by levulinic acid esterification obtained from Colombian rice straw”. This paper shows the methodology to calculate the Life Cycle Inventory (LCI) of the foreground system to perform the Life Cycle Assessment (LCA) of the ethyl levulinate (EL) production from Colombian rice straw (RS). This process encompasses two main stages: (i) RS production (involving cultivation and harvesting) and (ii) EL production (involving acid hydrolysis, levulinic acid (LA) purification, and El production). On one hand, foreground data related to paddy rice cultivation was gathered from the literature review. Besides, emissions of the cultivation stage were calculated using the IPCC (Intergovernmental Panel on Climate Change) methodology. The SQCB (Sustainable Quick Check for Biofuels) methodology was used to calculate NH3, NOx, N2O and NO3 emissions, whereas the SALCA (Swiss Agricultural Life Cycle Assessment) model was used to calculate phosphorous emissions to water. The Turc method was employed to calculate the irrigation requirements based on the rainfall and agrological features of rice culture. On the other hand, foreground data related to RS conversion to EL within a biorefinery scheme was obtained from simulation using Aspen Plus v.12. Lastly, background data associated with raw materials, catalysts, and utilities were gathered from Ecoinvent database. All the inventories are meaningful to carry out future environmental assessments involving sustainable production processes using RS as raw material or biorefinery processes using dilute acid hydrolysis.


a b s t r a c t
This data article is associated with the research article "Sustainable production of ethyl levulinate by levulinic acid esterification obtained from Colombian rice straw". This paper shows the methodology to calculate the Life Cycle Inventory (LCI) of the foreground system to perform the Life Cycle Assessment (LCA) of the ethyl levulinate (EL) production from Colombian rice straw (RS). This process encompasses two main stages: (i) RS production (involving cultivation and harvesting) and (ii) EL production (involving acid hydrolysis, levulinic acid (LA) purification, and El production). On one hand, foreground data related to paddy rice cultivation was gathered from the literature review. Besides, emissions of the cultivation stage were calculated using the IPCC (Intergovernmental Panel on Climate Change) methodology. The SQCB (Sustainable Quick Check for Biofuels) methodology was used to calculate NH 3 , NO x , N 2 O and NO 3 emissions, whereas the SALCA (Swiss Agricultural Life Cycle Assessment) model was used to calculate phosphorous emissions to water. The Turc method was employed to calculate the irrigation requirements based on the rainfall and agrological features of rice culture. On the other hand, foreground data related to RS conversion to EL within a biorefinery scheme was obtained from simulation using Aspen Plus v.12. Lastly, background data associated with raw materials, catalysts, and utilities were gathered from Ecoinvent database. All the inventories are meaningful to carry out future environmental

Objective
The data presented in this data article is generated as a result of the techno-economic and environmental assessment performed to evaluate the preliminary feasibility for the production of ethyl levulinate from Colombian rice straw. This data is mainly focused on two aspects. First, Table 1 Data collection for the construction of the LCI in each stage of the product system. the description of the simulation process performed in Aspen Plus v12 (AspenTech, Bedford, MA, USA) as well as the mass and energy balance obtained for the overall process. Second, the life cycle inventory obtained from the simulation process and the literature review for the paddy rice cultivation in Colombian context and the subsequent biorefinery process. This Life Cycle Assessment was performed in OpenLCA v1.10 (GreenDelta, Berlin, Germany). Economic data for the process economic assessment is included in this data article. The detailed life cycle inventories presented in this article support the main article discussion about environmental impacts related to the production of ethyl levulinate from rice straw and the identification of improvement opportunities for the process under study. Likewise, this inventory data could be used for further research related to the valorization of agro-industrial residues in the Colombian context.

Data Description
This article shows the Life Cycle Inventory (LCI) of the foreground system needed to perform a Life Cycle Assessment (LCA) of the production of ethyl levulinate (EL) from Colombian rice straw (RS). These data give transparency to the main results shown in the reference article [1] . LCI was gathered from process simulation using Aspen Plus v.12 (AspenTech, Bedford, USA), Ecoinvent database v.3.4, scientific, academic reports, and websites. Table 1 shows the data collection sources for the LCI construction. Table 2 shows the proximate analysis of RS coming from the Orinoquia Region (Colombia). Table 3 depicts the LCI of the paddy rice cultivation stage. Fig. 1 presents the main flowsheet associated with the conversion of RS into EL. Herein, five hierarchy blocks were employed. Fig. 2 shows the flowsheet for the ACID-HYD block that represents the hydrolysis of RS. Fig. 3 shows the detailed purification (HYD-SEP, in Fig. 1 ) of LA and furfural (FFR). Fig. 4 shows the detailed purification of FFR (FFR-SEP, in Fig. 1 ). Fig. 5 depicts the ESTERIF block that simulated the esterification of LA with ethanol to produce EL and its subsequent purification through a separation train. Lastly, Fig. 6 portrays the combustion of solid waste to produce low pressure steam (LPS) and medium pressure steam (MPS). Detailed information of subroutines for the aforecited process is described in Tables 4 and 5 . Table 6 presents the list of all reactions used for the RStoics units. Table 7 shows the operating conditions for the    Two scenarios were assessed. On one hand, a base scenario without the combustion of solid hydrolysed residue corresponding to the hierarchy block HEAT-GEN ( Fig. 6 ) and using a paddy rice yield of 4.95 t/ha. The mass and energy balance for base scenario are presented in Table 8 . LCI of the base scenario is presented in Table 9 . The contribution analysis performed for the base scenario is presented in Fig. 12 . On the other hand, the alternative scenario includes the combustion of solid waste along with an increment of the paddy rice yield to 5.7 t/ha. The mass and energy balance for the alternative scenario is presented Table 10 . LCI for alternative scenario             is presented in Table 11 . The contribution analysis for the alternative scenario is presented in Fig. 13 . Finally, an economic assessment for the alternative scenario was performed. Table 12 shows the cost of raw materials, utilities, and the selling price of all products (i.e., FFR, LA, EL, sodium sulfate). Table 13 presents the cost distribution for all subroutines listed in Table 5 . Table 14 and  Table 15 portray the CAPEX and OPEX distribution.

Experimental Design, Materials and Methods
The purpose of this document is to gather all the relevant information to calculate the LCI to carry out the LCA to produce EL from RS, as shown in the main manuscript. Table 1 shows the methodology to calculate the LCI of each stage involved in the conversion of RS into EL. Detailed information is shown in the upcoming sections.

Rice straw production
LCI for the production of RS was obtained for Colombia conditions, in the Orinoquia region [2] . RS production encompasses both cultivation and harvest stages. An average paddy rice yield of 4.97 t/ha was used for inventory calculation [3] . Diesel, fertilizers, land resources, and energy were the main inputs. Emissions to air and water were also considered. Paddy rice and RS were considered as main outputs. Diesel, land, and fertilizer requirements for rice production was obtained from the literature review [2] . Emissions of the agricultural stage was determined according to the IPCC methodology. Emissions associated with the use of nitrogen fertilizers included NH 3, NO x , N 2 O, and NO 3 . The latter was calculated according to the SQCB (Sustainable Quick Check for Biofuels) methodology. Phosphorous emissions to water were calculated using the SALCA (Swiss Agricultural Life Cycle Assessment) model [4] . Emissions associated with diesel combustion in the agroindustry were calculated based on the emissions factors reported by Martinez-Gonzales et al. [5] . CH 4 emission associated with rice cultivation was also included. Carbon sequestered by the crop was calculated in terms of the proximate analysis rice and the carbon content of each fraction as shown in Table 2 reported in literature [6] . Rainfall and irrigation needs were calculated using the TURC methodology. The LCI for paddy rice culture in Orinoquia region in Colombia is present in Table 3 .  Fig. 1 shows a complete process to produce EL from RS was developed in Aspen Plus v.12 (Aspen Tech, MA, USA) by considering four main stages: (i) RS acid hydrolysis; (ii) purification of LA; (iii) purification of FFR; and (iv) production and purification of EL. The Non-Random Two-Liquid with Redlich-Kwon (NRTL-RK) was used as the main thermodynamic package for phase equilibrium and thermodynamic estimations. However, due to the scarcity of some binary parameters for modelling the equilibrium phase with 3,5-hydroxymethylfurfural (HMF) and other lignocellulosic by-products (e.g., FFR, LA, formic acid (FA) and acetic acid (AA)), the Dortmund modified UNIFAC (UNIFAC-DMD) group contribution was employed to calculate the activity coefficients. The UNIFAC-DMD provides more reliable behavior of the phase equilibria of compounds than the traditional UNIFAC method. RS was modelled in terms of cellulose (38.3 wt.%), hemicellulose (28 wt.%), lignin (14.9 wt.%), and ashes (18.8 wt.%) based on data presented in Table 2 .

Rice straw valorization to EL
The three former was modelled according to the properties shown by Wooley and Putsche [7] . Whereas ashes were modelled as dioxide silicon due to this is the main constitutive element in this biomass fraction. Humins were modelled as cellulose and hemicellulose since those are decomposition products during the acid hydrolysis of biomass. Auxiliary units such as heat exchangers, pumps, compressors, valves, mixers, and splitters were considered within the simulation of the overall process. Kinetics models for reactions were not considered due to the main objective with the simulation is to obtain mass and energy balances for life cycle inventory purposes and the sizing or design of the equipment is not part of the study scope. Table 4 shows a description of main subroutines using in the process simulated with Aspen Plus v12. A brief description, as well as the operating conditions and assumptions used for each unit are presented. And Table 5 present the purpose of each subroutine employed among the simulation in Aspen Plus v12. Mixers and Flow Splitters were not considered for data presented in Table 4 and Table 5 . Fig. 2 presents the detailed diluted acid hydrolysis flowsheet, which corresponds to the Hierarchy block name ACID-HYD shown in Fig. 1 . Herein, steam explosion and diluted acid hydrolysis were employed together as pretreatment method to pretreat RS. Steam explosion was used as alternative to remove the hemicellulose fibers and ease the hydrolysis of hemicellulose and cellulose. In this first reaction stage, sulfuric acid was employed based on Biofine process at 210 °C and 20 bar. A second diluted acid hydrolysis with sulfuric acid was employed since its widely used at industrial level to pretreat lignocellulosic biomass at 190 °C and 18 bar. Table 6 shows the conversion rates of main reactions during the acid hydrolysis of RS and LA esterification based on the literature review. Aside from the acid hydrolysis stage, the pretreatment also includes the neutralization of sulfuric acid with sodium hydroxide, the recovery of unreacted cellulose, hemicellulose, lignin, and humins, hereafter names as HYD-SOL. Also includes the recovery of sodium sulfate as by-product for sale. All reactors employed to pretreat the biomass were RStoic subroutine. The heat generated in the first reaction stage was used to preheat the process water used for dilute acid hydrolysis. The liquid to solid ratio employed for dilute acid hydrolysis was 8:1, using a water recycle equivalent to 77% of the water used as solvent in the hydrolysis reaction. Two solid separation units were employed to separate the unreacted sugars and the sodium sulfate produced in the neutralization reactor (RS-103).   Fig. 3 (LA production) present the detailed flowsheet for LA purification. Distillation column was modelled using a RadFrac module with Strongly non-ideal liquid convergence method. The purification of LA was designed based on the literature review, where a distillation train formed by two flash separators and a distillation column was used. Flash separators were modelled using Flash-2 module, and the heat remanent from the vapor phase of the first flash unit was used to reheat the liquid phase feed to the second flash unit. The flash units were modeled as adiabatic units and no pressure drop inside de vessel. Temperature and pressure conditions required for each flash were set using a heat exchanger and a valve before each flash feed.  Fig. 4 (FFR purification) presents the detailed flowsheet for FFR purification. The FFR purification was carried out using azeotropic distillation based on Zeitsch [8] . Azeotropic distillation columns were also modelled with RadFrac module but using the Azeotropic convergence method. Efficiency of distillation columns were adjusted to 65% according to heuristics rules. A decanter was used to separate the distillate from the first distillation tower into a furfural-rich phase, which was subsequently distilled to obtain a high purity FFR stream and the removal of volatiles such as formic acid (FA) and the remaining water in the distillate. The aqueous rich phase from decanter DC-201 was recirculated to the azeotropic column to maximize FFR recovery. The bottom stream from RF-201 (azeotropic column) was recycled to the dilute acid hydrolysis stage to reduce the total water consumption of the overall process and to minimize the FFR lost in the two-step separation train.

Ethyl levulinate production
Fig. 5 presents the detailed flowsheet for EL production. This was done through the esterification of LA with ethanol using a desilicated DH-ZSM-5 zeolite as catalyst with a catalyst load of 13wt%. Esterification was modelled with a RStoic unit at 120 °C and 4.5 bar, using ethanol as solvent (8:1 ratio) with a conversion of LA of 85% as shown in Table 6 . Kinetic model for this reaction was not considered due to the simulation purpose is not the equipment design or sizing instead the mass and energy balances calculation for LCI. Purification of EL was performed by using a distillation train where RadFrac modules with Strongly non-ideal liquid convergence method were employed to module the two distillation columns. Efficiency of distillation columns were adjusted to 65% according to heuristics rules. In the first column, a distillate rich in ethanol and water is obtained, which is subjected to a separation process with molecular sieves for the recovery and recirculation of 95% of the remaining ethanol using a SEP-2 unit. The bottom stream, containing LA and EL is fed to a second distillation tower, where two high purity streams are obtained, one of EL and the other of LA.

Heat integration
The integration consisted of using the remaining heat from the first acid hydrolysis reactor and the output stream of the second acid hydrolysis reactor to reduce the energy consumption associated with the generation of the steam required for hydrolysis, according to Fig. 2 . Likewise, the two steam streams obtained from the flash units associated with the separation of the FFR  6 presents the detailed flowsheet for LPS and MPS generation. RStoic unit was used to model combustion by stoichiometric reactions of solid waste combustion assuming complete combustion and disregarding the generation of methane, NOx, and SOx. The combustion temperature was set in 904 °C at a pressure of 10 bar. Excess air was used to control the reactor temperature.
The condenser was modeled using a Heater unit cooled by air and the boiler was modeled using a non-rigorous MHeatX unit. Ashes from the combustion unit was separated from the reactor outlet stream using an SEP-2 unit. Two isentropic turbines were used to generate the MPS (saturated steam at 175 °C) and the LPS (125 °C). A total of 5,600 kg/h of water was used to generate 3,100 kg/h of LPS (100% of the requirement of the process) and 2,500 kg/h of MPS (55% of the total process requirement). Both LPS and MPS generated was recirculated to the generation cycle, reducing the water consumption to a make-up stream of 280 kg/h of water.

Sensitivity analysis for distillation columns design
A sensitivity analysis was performed to determine the best operation conditions for the five distillation columns used among the process. The number of equilibrium stages, optimal feed stage, reflux molar ratio, and distillation or bottom to feed ratio were assessed to minimize the energy requirements in the distillation column and maximize the recovery of the interest product. Table 7 presents the operation conditions selected for each distillation column. Detailed sensitivity analysis is shown in Fig. 7 (RF-110), Fig. 8 (RF-201), Fig. 9 (RF-202), Fig. 10 (RF-301), and Fig. 11 (RF-302). The operation pressures were selected seeking the reduction of reboiler and condenser duty requirement.

Mass and energy balance, Life Cycle Inventory, and Contribution Analysis
Mass and energy balance were retrieved from the simulation results in Aspen Plus. The Life Cycle Inventory (LCI) was estimated using the Aspen Plus balances and the Ecoinvent Database for background data. Contribution analysis was performed in Open LCA v 1.10 software using the ILCD 2011 midpoint + baseline method with eight impact categories assessed: (1) Acidification (molc H + -eq); (2) Climate Change (kg CO 2 -eq); (3) Freshwater Eutrophication (kg P-eq); (4) Marine Eutrophication (kg N-eq); (5) Ozone Depletion (kg CFC-11-eq); (6) Photochemical Ozone Formation (kg NMVOC-eq); (7) Terrestrial Eutrophication (molc N-eq); and (8) Water Resource Depletion (m 3 water-eq). Table 8 presents the mass and energy balance to produce EL from RS in the base scenario. Fig. 12 presents the contribution analysis of all impact categories described above in the base scenario. Finally, Table 9 presents the LCI of the production of EL from RS in the base scenario. Table 10 presents the mass and energy balance to produce EL from RS in the alternative scenario and Table 11 presents the LCI of the production of EL from RS in the alternative scenario. Finally, Fig. 13 shows the LCI of the production of EL from RS using a paddy rice yield of 5.7 t/ha.

Economic assessment
An economic assessment for the alternative scenario was performed to give insights into the feasibility of the process in an integrated way. Table 12 presents the cost of raw materials, products, by-products, and co-products retrieved from the literature review and used for calculation of economic indicators. Table 13 presents the equipment cost and installation cost, obtained from Aspen Plus Economic Analyzer and used for CAPEX calculation. Table 14 depicts the CAPEX distribution estimated using Lang-Factors and Table 15 present the OPEX distribution using Lang-Factors.

Ethics Statement
This work did not involve human subjects or laboratory animal, therefore did not meet any ethical issues.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships which have or could be perceived to have influenced the work reported in this article.

Data Availability
Ethyl _ Levulinate _ from _ Colombian _ Rice _ Straw (Original data) (Mendeley Data) Life Cycle Inventory data for ethyl levulinate production from Colombian rice straw (Original data) (Data in brief).