Process simulation–based scenario analysis of scaled-up bioethanol production from water hyacinth

Abstract

Water hyacinth (WH) is an aquatic weed with an experimentally proven potential as a feedstock for bioethanol production. Unlike other bioethanol feedstocks, water hyacinth has no requirement for land use and resource consumption for cultivation. This study evaluates scaled-up bioethanol production process routes, modelled using the Aspen Plus process simulator to analyse the process performance of water hyacinth as a bioethanol feedstock. Four process scenarios are developed by combining two different feedstock pretreatment methods (i.e., alkali pretreatment and diluted acid pretreatment) and bioethanol dehydration techniques (i.e., azeotropic distillation and extractive distillation). Mass and energy flows of the four scenarios are comparatively analysed. Results show that the alkali pretreatment method provides a higher bioethanol yield (i.e., 254 L/tonne-WH) compared with the dilute acid pretreatment method (i.e., 210 L/tonne-WH). In addition, the process route combining alkali pretreatment and extractive dehydration techniques indicates the least process energy consumption of 45,310 MJ/m³ of bioethanol. The process energy flow analysis evaluates two energy sustainability indicators, i.e., net energy gain and renewability factor, with further interpretation of variation effects of the key process parameters through a sensitivity analysis. The feasible ways of utilising water hyacinth as a fuel-grade bioethanol feedstock for industrial-scale production are discussed.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at 10.1007/s13399-023-03891-w.

Introduction

Global energy demand is increasing at a tremendous speed which is expected to have a further increase of 28% by the year 2040 [1]. The world energy supply is still dominated by fossil fuels; especially, the transportation sector heavily relies on petroleum-based fuels. However, excessive reliance on petroleum-based fuels leads to an impending energy crisis as their reserves are running low [2]. Rapid resource depletion is not the only consequence of consuming uncontrolled petroleum-based fuels; it also results in global economic instability and a plethora of environmental concerns, such as global warming, extreme weather conditions, polar caps melting, acid rains, etc. [3]. For example, the transportation sector contributes about 15% of global greenhouse gas emissions [4]. Therefore, alternative renewable fuels are urgently needed to reduce the dependency on petroleum-based fuels [5]. It is vital that such a renewable fuel is capable of maintaining a sufficient supply to fulfil the demand for an efficient and seamless switch from traditional or conventional fuels.

At present, the contribution of renewable fuels to global energy consumption is only 23.7% [6, 7]. Notably, the fraction of biofuel usage needs to be increased. For this mission, lignocellulosic biomass can serve as a reliable source to produce liquid biofuels like bioethanol [8]. Due to abundant availability, lignocellulosic biomass has the potential to cater for future bioethanol requirements by reducing petroleum-derived fuel usage. Bioethanol substitution can minimise carbon and other pollutant emissions from petroleum-derived fuels that can cause catastrophic environmental problems [9]. Further, local bioethanol production by a third-world country would mitigate its foreign currency outflow by reducing petroleum fuel imports and encouraging economic growth, thus lessening the influence during global energy crises [10, 11]. However, such countries are in search of the most appropriate feedstocks and process developments for commercial-scale bioethanol production. Thus, global bioethanol production is still 3% of the gasoline consumption in the world [12]. In 2019, the global bioethanol production was 115 billion litres, and an approximate increase of 28% is expected in the worldwide bioethanol demand by the year 2026 [13, 14]. Prior to the COVID-19 pandemic, production was forecasted to reach 130 billion litres by the year 2024 [15]. It is also predicted that India will be the third-largest producer in the bioethanol market by 2026.

The affinity for gasoline in the mass market and industry is because of its high calorific value of 35.7 MJ/L, whereas bioethanol has only a calorific value of 23.4 MJ/L [16]. Nevertheless, with other benefits like octane number enhancement and renewability approaches, anhydrous ethanol shows advantages blended with pure gasoline to produce gasohol mixtures named E5, E10, E20, or E85, depending on the percentage of ethanol in the blend [17]. Otherwise, the alternative option in terms of boosting the octane number is mixing methyl tert-butyl ether (MTBE) has adverse effects on the environment [18]. Most of the vehicles already available in the market can withstand biofuel blends up to 20% without any modification. Vehicles with modified engines that can use neat biofuels also are currently available, delivering an array of vehicle options to consumers.

At present, the majority of bioethanol is produced from First Generation (1G) bioethanol feedstocks, such as sugarcane, potato, corn, cassava, beetroot, etc. [19]. However, limitations due to the threats to food security by expanding the usage of arable land, freshwater utilisation, and intensive agricultural inputs prevent 1G biofuels from becoming a long-term solution [20]. Furthermore, the Ukraine-Russian political conflict has worsened the crops and grain shortage, putting strain on the diversion from 1G biomass to higher-generation biomass [21]. Over the last decade, Second Generation (2G) biomass which is known as lignocellulosic biomass, has been getting the attention of researchers [22]. For example, forestry residues, agricultural residues, and municipal solid waste can be mentioned as viable sources of 2G biomass with abundant availability, lesser raw material cost, and intrinsic renewable features [16, 23].

Water hyacinth (Eichhornia crassipes) can be considered as a 2G biomass which is a noxious, invasive aquatic weed that grows fast in eutrophic waters, currently having no efficient way of discarding [24]. The plant has a high content of cellulose and hemicellulose, which is an indication of the potential of fermentable sugar [25]. The low lignin content in water hyacinth (WH) has led scientists to study it more as a potential bioethanol feedstock [26, 27]. Other benefits of attracting attention towards WH include no competition for occupying arable land, not being utilised as a feedstock for other applications, and a high renewable growth rate of 220 kg/ha per day [27]. Recently published studies have focused on the investigation and optimisation of pretreatment methods and fermentation methods in the process of converting WH to ethanol [5, 28–32]. It has been found that WH is more suitable compared with metasequoia chips, sugarcane bagasse, miscanthus, and water peanut because of low hydrolysis time and low byproduct formation as a result of the unique cellulose properties [30].

First, cellulose and hemicellulose contained in lignocellulosic biomass are required to be recovered from lignin and converted into sugars prior to fermentation [33]. Because of the recalcitrant nature of lignin, a specific pretreatment step is essential to remove lignin and recover cellulose and hemicellulose before the hydrolysis step [23]. Hence, the 2G bioethanol production process is energy intensive compared to 1G bioethanol production. Therefore, it is important to identify energy-efficient bioethanol conversion process routes for lignocellulosic biomass for larger-scale biorefineries [34].

Multiple techniques are identified as successful methods to convert 2G biomass to bioethanol [7, 35]. Different pretreatment methods, such as dilute acid pretreatment, alkaline pretreatment, steam explosion, the use of organo-solvents, and microwave treatment, as well as different enzymes for hydrolysis and modern fermentation methods like simultaneous saccharification and fermentation, are several techniques applied in bioethanol conversion [7]. The bioethanol conversion route plays a crucial role in the effectiveness of the overall process system. The significance of identifying the most suitable techniques for the feedstock is not limited to gaining a high ethanol yield; it is equally important to have a thorough understanding of the mass and energy flows within the process, especially when adapting techniques for large-scale bioethanol processing plants [36]. Primarily, bioethanol yield depends on the feedstock and process conversion efficiencies, where the plant size, process conditions, and microorganisms used for fermentation would affect secondarily [37]. Published studies on scaled-up bioethanol production using WH as a feedstock are scarce in the existing literature, and a detailed process analysis is essential before scaling up such a bioethanol plant.

Thus, considering the maturity of the conversion method and the simplicity of the process, this study uses both alkaline pretreatment and dilute acid pretreatment as pretreatment techniques to further develop the process and virtually scale up. A noteworthy fact is that pretreatment is the most costly and complicated step in the conversion process of lignocellulosic biomass to bioethanol [38]. Furthermore, there is an unavoidable impact on the downstream stages by the pretreatment yield [39]. The conversion step of C₅/C₆ sugars to ethanol can either be done as Separate Hydrolysis and Fermentation (SHF) or Simultaneous Saccharification and Fermentation (SSF). The SSF process contains numerous benefits, including the use of a single reactor which gives economic benefits by reducing the cost and residence time [40]. Since the single reactor accommodates the saccharified products to be consumed by microorganisms instantly, the formation of inhibitors is prevented, unlike in SHF [41]. Hence, a new scaled-up process would include SSF instead of conventional SHF [42]. For application as fuel-grade bioethanol, the end product requires a minimum purity of 99.5 vol% [10]. Therefore, the converted bioethanol needs to be concentrated from the ethanol–water mixture. However, because of the low boiling azeotrope at the composition of 89.4 mol% of ethanol at 1 atm, the conventional direct distillation is not sufficient to achieve the targeted purity. Hence, the mixture needs to be dehydrated using an advanced distillation technique, such as azeotropic distillation, extractive distillation, molecular sieve adsorption, pressure-swing distillation, etc.

The use of process simulation offers the opportunity to study different conversion technology combinations using different feedstock options [43]. Process simulations can provide unbiased process data and minimum variations due to plant features and contingencies, such as plant design, age of the plant, location of the plant, advancements of the used technology, labour skills, etc., that facilitates a fair comparison to assess the suitability of different processing techniques for bioethanol conversion [44]. Previous studies have compared process simulation-based data of bioethanol production from a variety of feedstocks with actual bioethanol plant data to validate the reliability of this technique [23, 44, 45]. To the best of our knowledge, there are no reported industrial-scale process plants that use WH as a feedstock. Therefore, the process simulation technique is used in this study to scale up the bioethanol production process and perform a fair comparison of energy consumption among different scenarios of bioethanol conversion routes from WH.

This study focuses on conducting a fair comparison of mass and energy flow analysis in possible bioethanol production routes from WH with different combinations of pretreatment and bioethanol dehydration techniques using process simulation. This study would support decision-making for the establishment of new bioethanol production plants using WH as a new feedstock option and in designing/retrofitting scaled-up bioethanol plants from lignocellulosic feedstocks like WH. Further, this study provides an efficient and environmentally benign valorisation option that mitigates the burden for authorities to safely dispose of invasive WH growth in natural water bodies, municipal canals, etc.

Materials and methods

This study performs a comprehensive process analysis for a scaled-up production process of fuel-grade bioethanol from Water Hyacinth (WH) as the feedstock. Four process routes with combinations of two feedstock pretreatment methods (i.e., alkali pretreatment and diluted acid pretreatment) and two bioethanol dehydration techniques (i.e., azeotropic distillation and extractive distillation) are studied, separately, and compared with each other.

Process description

Table 1 lists the four process route combinations in the considered bioethanol production scenarios from WH. For clear identification, the four scenarios of the bioethanol production process from WH are abbreviated as WH1, WH2, WH3, and WH4 (i.e., WH1: Alkaline pretreated WH dehydrated using extractive dehydration technique, WH2: Alkaline pretreated WH dehydrated using azeotropic dehydration technique, WH3: Dilute acid pretreated WH dehydrated using extractive dehydration technique, WH4: dilute acid pretreated WH dehydrated using azeotropic dehydration technique).

Table 1.

Process route combinations in bioethanol production scenarios from WH

Process scenario	Bioethanol feedstock	Process route combination in each scenario
		Pretreatment technique	Dehydration technique
01	Water hyacinth (WH)	Alkaline pretreatment	Extractive distillation
02		Alkaline pretreatment	Azeotropic distillation
03		Dilute acid pretreatment	Extractive distillation
04		Dilute acid pretreatment	Azeotropic distillation

Figure 1 illustrates the fuel-grade bioethanol production process from WH. The considered bioethanol production scenarios differ from each other by the change of the feedstock pretreatment and bioethanol dehydration techniques. Hence, all four considered scenarios of bioethanol production from WH can be demonstrated via two diagrams of process routes, as shown in Fig. 1.

Fig. 1.

Diagram of process system boundary: (a) Bioethanol conversion from WH via dilute acid pretreatment, (b) Bioethanol conversion from WH via alkali pretreatment

In this study, a uniform feedstock composition of WH (cellulose 19.2 wt%, hemicellulose 40 wt%, lignin 4.8 wt.%, and ash 36 wt% [46]) at the dry basis is considered for all the scenarios of bioethanol production. The pre-processed feedstock is fed into the pretreatment reactor to remove recalcitrant lignin available in WH and recover cellulose and hemicellulose. Removed lignin is separated along with the other solid residues, and the liquid phase with recovered cellulose/hemicellulose is fed to the enzymatic hydrolysis reactor. In the enzymatic hydrolysis reactor, the recovered cellulose and hemicellulose are converted into glucose, xylose, and C₅/C₆ sugars. The sugar solution is then neutralised and sent to the SSF reactor to convert sugars into ethanol. The separated lignin and solid residues are utilised as fuel for heat and power generation in the same process.

The purity of bioethanol is required to be higher than 99.5 vol% (> 99.5 vol%) for fuel-grade bioethanol. After fermentation, bioethanol in the water-rich solution can be separated only up to an azeotropic mixture (~ 96 vol%) using direct distillation [47]. Hence, a bioethanol dehydration technique is needed to shift the ethanol–water azeotrope and further purify bioethanol. In this study, extractive distillation and azeotropic distillation are used as the bioethanol dehydration techniques in the process scenarios.

In this study, the comparative process analysis comprised of Material Flow Analysis (MFA) and Energy Flow Analysis (EFA) is performed with a unit basis of 1 m³ (1000 L) of bioethanol production at 99.7 vol% as the final product. For the process stage-wise analysis, the process system boundary is divided into three stages, namely, (1) Feedstock pretreatment stage, (2) Bioethanol conversion stage, and (3) Bioethanol dehydration stage. Table S1 in the supplementary information lists relevant chemical reactions and process conditions of each process operation, from feedstock pretreatment to bioethanol dehydration.

Stage 1: Feedstock pretreatment stage

For scenarios WH1 and WH2, WH is pretreated with the alkaline pretreatment method. The mixing tank is loaded with WH having a solid loading of 10% (w/v) and 2% (w/v) NaOH as the alkaline agent. Alkaline pretreatment operation is performed at 121 C° for 40 min, supplied with saturated steam at 5 bar to maintain the required temperature. For scenarios WH3 and WH4, WH is pretreated with diluted H₂SO₄ under the same conditions [46]. Then the pH value of the pretreated mixture is adjusted to 4.8, either with HCl or Ca(OH)_2, for enzymatic activity in the hydrolysis process [46]. At this point, all the recoverable organic components are converted to reducing sugars in the liquid phase, and the residue solids are separated by a solid/liquid filtration unit before adding to the fermenter. At the fermentation unit process, the extracted sugars are converted to ethanol in the presence of yeast strains with nutrients.

Stage 2: Bioethanol conversion stage

In this process stage, the conversion of glucose, xylose, and C₅/C₆ sugars to ethanol via SSF reactor in the presence of yeast (S. cerevisiae) and nutrients is considered. Ethanol conversion efficiencies of 90% from glucose and 70–80% from xylose are possible with SSF in the presence of yeast and nutrients [48, 49]. The ethanol/water mixture obtained after SSF is separated near the azeotropic composition (up to 91 wt%) by sending through direct distillation. The initial input parameters, such as the reflux ratio, number of column stages, and recovery efficiency, are defined for simulating the distillation unit process, as indicated in Table S1 in the supplementary information.

Stage 3: Bioethanol dehydration stage

In this process stage, the distillate output at 91 wt% of bioethanol content is further purified via an advanced distillation technique, such as extractive distillation and azeotropic distillation, to dehydrate the bioethanol output up to the required fuel-grade purity level (> 99.5 vol%). Ethylene glycol and cyclohexane are utilised as the solvent/entrainer for extractive and azeotropic distillation techniques, respectively. The solvent/entrainer recovery efficiency is set at 99% for the dehydration column and solvent recovery column. The reflux ratio and the number of stages of the columns are varied to reach the set solvent recovery efficiency in each process scenario. Table S2 in the supplementary information summarises the column design specifications and required input parameters applied to develop the process simulation models for the four process scenarios of bioethanol production.

Process simulation methodology

Scaled-up bioethanol production plants that correspond to the considered four process scenarios of each feedstock are modelled and simulated using the Aspen Plus process simulation software. The process simulations adopt the ethanol–water binary properties, ethanol–water-ethylene glycol, and ethanol–water-cyclohexane ternary properties from the Aspen Plus property database.

The Non-Random Two Liquid (NRTL) activity coefficient model is applied as the thermodynamic property method in process simulations. The UNIQ-RK thermodynamic property method is applied to simulate the azeotropic and extractive distillation unit processes. The UNIQ-RK property method in the Aspen Plus properties database is a combination of the UNIQUAC activity coefficient model for the liquid phase, Redlich-Kwong equation of state for the vapour phase, Rackett model for liquid molar volume, and Henry's law for supercritical components that can accurately simulate the nonideality of the liquid phases and the vapour phases in a bio-ethanol mixture. The R-Stoic reactor model in the Aspen Plus equipment model library is used to simulate the reactors for feedstock pretreatment, neutralisation, enzymatic hydrolysis, and SSF unit processes. The distillation columns are modelled using the rigorous RadFrac distillation column model in the Aspen Plus equipment model library. The process simulation results are validated in comparison to the real bioethanol production plant performances of other lignocellulosic feedstocks reported in the published studies.

Mass and energy flow analysis

Mass and energy flows of the bioethanol production process scenarios are evaluated based on the obtained process simulation results, and MFA and EFA are performed. Process energy supply is catered by in situ heat and power cogeneration via the combustion of biogas from the effluent/wastewater treatment plant and lignin-rich solid residue from the feedstock pretreatment step. When the in situ fuel sources are insufficient to supply the required process heat energy in each process scenario, biomass (wood chips, wood residues, etc.) from external resources is utilised as additional fuel for heat energy generation. The electricity requirement for the process is fulfilled by the national power grid. The electrical power generation from in situ combined heat and power (CHP) cogeneration using lignin/solid residues/biogas is credited to the national power grid.

Table S3 in the supplementary information indicates the efficiencies of boilers and CHP units along with the energy values of the fuel sources. For biogas generation with wastewater treatment, up-flow anaerobic sludge blanket (UASB) reactors are selected as per the industrial practice for production plants of similar nature. The amount of biogas generation with 65% methane is determined using Eq. (1), where 0.25 is the maximum methane generation potential, and 0.8 is the methane correction factor [44, 45]. Solids and biogas are considered to burn in two separate boilers.

$$Generatedmethaneamount\left(m^3\right)=Wastewatervolume\times COD\times 0.25\times 0.8$$ 1

The study evaluates the process net energy ratio and the renewability factor of the process to understand the sustainability of the selected process scenarios. These two energy indicators are calculated using Eqs. (2) and (3) under the process net energy analysis.

Process energy ratio

$$NetEnergyRatio\left(N,E,R\right)=\frac{Totalenergyoutput}{Totalenergyinput}$$ 2

Renewability indicator

$$Renewabilityindicator\left(R,n\right)=\frac{Totalbioenergyoutput}{Totalfossilenergyinput}$$ 3

Sensitivity analysis

A sensitivity analysis for the bioethanol yield and the total process energy consumption is performed in this study by varying five sensitivity parameters, i.e., lignin content in WH, cellulose content in WH, the ratio for solid loading, pretreatment efficiency (cellulose/hemicellulose recovery efficiency), and fermentation efficiency. The variation ranges of the sensitivity parameters are determined based on lab-scale data for bioethanol synthesis from WH, reported in the published literature. Table 2 indicates the five sensitivity parameters and their variation ranges considered in the study.

Table 2.

Key parameter variations for sensitivity analysis

Sensitivity parameter	Initial value used in this study	Variation range	Reference
Lignin percentage in WH (wt.%)	4.8%	3–14%	[50, 51]
Cellulose percentage in WH (wt.%)	19.2%	18–55%	[51, 52]
Solid loading (biomass: water)	1:10	1:3–1:15	[53]
Pretreatment process efficiency (%)	Cellulose: 99 Hemicellulose: 57	Cellulose: ± 5% Hemicellulose: ± 5%	[54]
Fermentation efficiency (%)	Glucose: 90 Xylose: 80	Glucose: ± 5% Xylose: 5%	[54]

Results and discussion

The selected four process route scenarios were modelled and simulated in the Aspen Plus process simulation software to embody the respective scaled-up bioethanol plants to obtain final bioethanol output at 99.7 vol% purity. The process simulation results were calculated for a unit basis of 1000 L (1 m³) of bioethanol production and analysed the four scenarios.

Process simulation results

Figure S1, S2, S3, S4 in the supplementary document represent the simulated process flowsheet illustrating detailed material and energy flows for scenarios WH1 to WH4. According to simulation results, the bioethanol yield significantly depends on the pretreatment technique. As such, bioethanol yield is 254 L/tonne-WH (dry basis) for WH1, WH2 and 210 L/tonne-WH (dry basis) for WH3, and WH4 scenarios, respectively, which suggest a requirement of 3.96 and 4.82 tonnes of dry WH to produce 1 m³ of bioethanol at 99.7 vol% depending on the pretreatment method considering alkaline pretreatment and dilute acid pretreatment. As the literature reports, bioethanol yield varies from 239 to 330 L/tonne-biomass (dry basis) for different lignocellulosic feedstocks, such as rice straw, wheat straw, switch grass, etc. Table 3 shows the bioethanol yield with 99.7 vol% purity from different lignocellulosic feedstocks. Furthermore, the bioethanol yield obtained from process simulations in this study is in close agreement with the laboratory-scale bioethanol yield parameters reported in published studies for WH that is in the range of 0.12–0.35 kg-EtOH/kg-biomass (dry basis) [46, 55–57]. The comparison of the present findings can also be extended to other aquatic weeds; duckweed (Lemna minor and Lemna gibba) 0.218 and 0.197, water lettuce (Pistia stratiotes) 0.215, common reed (Phragmites australis) 0.165 and mosquito fern (Azolla microphylla) 0.167, all in kg bioethanol per kg of biomass [58–60].

Table 3.

Bioethanol yield for other lignocellulosic feedstocks

Feedstock	Yield (L/tonne biomass)	Reference
Rice straw	239	[48]
Corn stover	300	[61]
Wheat straw	330	[62]
Switchgrass	318	[63]
Softwood forest residues	248	[34]
Hardwood chips	250	[64]
Water hyacinth (WH1)	254	This study
Water hyacinth (WH2)	254	This study
Water hyacinth (WH3)	209	This study
Water hyacinth (WH4)	209	This study

Table 4 summarises the process simulation-based energy consumption results for WH1, WH2, WH3, and WH4 process scenarios. The results clearly indicate the segregation of steam and electricity requirements to function the scaled-up model process plant. Total energy consumption values for the four scenarios are 45,310 MJ, 52,000 MJ, 53,040 MJ, and 61,210 MJ per 1 m³ of fuel-grade bioethanol, respectively.

Table 4.

Energy consumption results for individual plant equipment (MJ/m³-EtOH)

Unit	WH1		WH2		WH3		WH4
	Steam	Electricity	Steam	Electricity	Steam	Electricity	Steam	Electricity
Unit
Pretreatment reactor	17,2400	108	17,240	108	20,540	108	20,650
Nutch filter		12		12
Neutralising tank		1098		1098		1250		1250
Hydrolysis rector	1098		1098		960		960
Residual solids removal (filter press)		3.00		3.00		4.00		4.00
Gypsum removal (filter press)						3.00		3.00
SSF		197		197		218		218
CO2 scrubber		1.87		1.87		45.00		45.00
Distillation column	16,500	828.6	16,500	828.6	20,230	1071	20,230	1071
Dehydration column	1178	174.6	4866	981.4	1218	257.3	5570	1181
Recovery column	410	104	2199	457	293	86	2700	500
Decanter				135				163
Other Utilities		6351		6269	5.00	6744	5.00	6664
Total	36,430	8879	41,910	10,100	43,250	9787	50,110	11,100
Stage-wise energy distribution
Conversion	23,440		23,440		27,160		27,160
Distillation	17,330		17,330		21,300		21,300
Dehydration	4537		11,230		4572		12,750
Total energy	45,310		52,000		53,040		61,210

Total process energy consumption by each scenario

Herein, total process energy consumption by WH as a bioethanol feedstock is higher than most of the other lignocellulosic feedstocks in the reviewed publications. Table 5 shows energy consumption results from past studies for a variety of feedstocks to produce 1 m³ of bioethanol. The existing studies have elucidated 25,520 MJ/m³-EtOH for wood [65], 25,500 MJ/m³-EtOH for switchgrass [65], 22,050 MJ/m³-EtOH for tall fescue [20] and 29,800 MJ/m³-EtOH for rice straw [66] for a bioethanol production facility. However, a study conducted on corn stover concluded an energy consumption of 41,280 MJ/m³-EtOH, which is comparable to WH of this study. The possible reasons for the energy difference in many other lignocellulosic feedstocks can be discussed through sensitivity analysis.

Table 5.

Energy consumption for bioethanol production from other feedstocks

Feedstock	Energy consumption (MJ/m3)	Reference
Tall fescue	22,050	[20]
Corn stover	41,280	[61]
Hard wood	52,500	[64]
Wood	25,520	[65]
Switch grass	25,500	[65]
Rice straw	29,800	[65]
Willow chips	56,095	[67]

Mass flow analysis

Table 6 summarises process mass flow results, including all the material requirements to produce 1 m³ of bioethanol. Based on the obtained bioethanol yield, an additional amount of 0.86 tonnes of dry WH is required to produce 1 m³ of fuel-grade bioethanol in a biorefinery that valorises WH via dilute acid pretreatment. On account of the high moisture content in WH (93.35% [56]), consequently being a water body, 60 tonnes and 73 tonnes are established as the harvesting demand once pretreated with an alkaline solution or a dilute acid.

Table 6.

Material flow results for the four scenarios

Inventory	Mass flow (kg/m3-EtOH)
	WH1	WH2	WH3	WH4
Feedstock preparation
Harvested WH	59,540	59,540	72,520	72,520
Bioethanol conversion
WH (dry basis)	3960	3960	4823	4823
Water	39,600	39,600	48,230	48,230
NaOH	792	792
H2SO4			965	965
Enzyme	23		23
HCl	723	723
Lime			729	729
Yeast	9.75	9.75	9.75	9.75
(NH4)2HPO4	9.48	9.40	10	9.88
CaCl2	210	210	242	239
Bioethanol purification
Cyclohexane		2.55		2.57
Ethylene glycol	0.84		0.77

When WH is pretreated at 121 °C with NaOH, delignification was observed by giving access to 99% cellulose and 57% hemicellulose [46]. The process utilises 791.9 kg/m³-EtOH of NaOH to alkaline the medium for the expected bioethanol output. The delignification happens when an alkaline solution reacts with ester links in-between hemicellulose and lignin. As a result, lignin gets solubilised, creating black liquor [68]. This black liquor can be separated via a filtering process, and the rest of the lignin will be removed later, with the rest of the solid residues after the enzymatic hydrolysis for maximum cellulose and hemicellulose recovery. Delignification of WH opens the possibility of obtaining convertible sugars approximately at 0.470 kg/kg-WH. On the other hand, dilute acid pretreatment can only recover 97% cellulose and 44% hemicellulose in WH under similar conditions [46].

According to the mass flow results, sulfuric acid consumption for dilute acid pretreatment is observed as 964.53 kg/m³-EtOH giving a reduced sugar yield of 0.407 kg/kg-WH, compared with the alkaline pretreated process. This is close when compared with the sugar yield under similar conditions using duckweed as the feedstock, which is 0.426 kg/kg-duckweed (dry basis) [69]. After adapting lab scale data, it is apparent that alkaline pretreatment is more suitable for WH since it gives a better conversion. By using NaOH to make the substrate alkaline, it is possible to gain 21.8% more recovery of cellulose and hemicellulose.

In both pretreatment methods, cellulose recovery was higher compared with hemicellulose recovery. As a percentage, 99% of cellulose was recovered and converted to glucose from the initial cellulose content, while only 57% of hemicellulose was able to convert to xylose after being treated with NaOH, whilst 97% of cellulose and 44% of hemicellulose conversions are observed once pretreated with dilute sulphuric acid as per mass flow results. During these two pretreatment methods, while NaOH converts into NaCl, H₂SO₄ converts into gypsum. This gypsum is a usable byproduct that can be utilised as a fertiliser or in construction applications. The recoverable gypsum mass as a byproduct of bioethanol production is 1693 kg/m³-WH.

By the end of fermentation, the beer feed entering the distillation column has 1.92 wt% for WH1 and WH2 and 1.64 wt% for WH3 and WH4 of ethanol. Fermented ethanol is purified to 91% mass purity with 99% recovery using distillation columns. Subsequently, the aqueous ethanol mixture is dehydrated using either extractive distillation or azeotropic distillation methods to achieve anhydrous bioethanol at > 99.5 vol% purity for commercial fuel purposes. The volume purity of the final product is 99.7%. In the use of both of the dehydration techniques, the efficiency of recovery of columns is considered as 99% in this study. Since the recovery is considered as 99% with given specifications, dehydration can be ruled out as an influencing factor for the bioethanol yield. Hence, as per the mass flow results, bioethanol yield solely depends on stage 1 of the process. Spent wash generation was observed as 41,200 L/m³-WH, where this carbon-enriched effluent stream is directed to anaerobic digestion tanks to produce in-house biogas.

Energy flow analysis

Once process simulation results are analysed, it is evident that the conversion stage consumes most of the energy regardless of the pretreatment method in this study. Figure 2 shows the breakdown of energy consumption for the defined scenarios with the respective energy distribution as a percentage for each stage. As per the simulation results, the alkaline pretreatment process consumes less energy (23,440 MJ/m³-EtOH) compared with the dilute acid pretreatment process (27,160 MJ/m³-EtOH). The said energy consumptions reflect 51.73% of the total energy consumptions for the plant for WH1, 45.08% for WH2, 51.21% for WH3 and 44.37% for WH4. The reason for WH2 and WH4 to have a relatively low distribution for pretreatment is that the azeotropic dehydration technique consumes more energy compared with extractive distillation to give a similar output. In this study, the two simulated scaled-up plants that use azeotropic distillation consumed 11,220 MJ/m³-EtOH in WH2 and 12,750 MJ/m³-EtOH in WH4, whereas for WH1 and WH3, which uses extractive distillation as the dehydration technique shows consumption of 4537 MJ/m³-EtOH and 4572 MJ/m³-EtOH reflecting a comparative 148% difference in average. Hence, it is possible to suggest that the alkaline pretreatment technique is better for pretreating WH from the energy perspective. The study also proves that extractive distillation has better performance in terms of energy consumption when compared with the azeotropic dehydration technique.

Fig. 2.

Stage-wise energy breakdown for scenarios WH1–WH4

Despite the scenario, the designed bioethanol plant consumes steam the most and percentage-wise, the steam demand remains at 81 ± 1% of the total energy requirement. Figure 3 depicts the heat and steam flows within the three stages considering WH1 of the process. As shown in the chart, 36,430 MJ/m³-EtOH of steam input gets distributed among the three stages as 40%, 36% and 4%, indicating stage 1 with the highest steam demand; and it is consistent with electricity intake as well. However, stage 3 requires 5% more electricity than stage 2, unlike the steam demand.

Fig. 3.

Energy flow diagram for WH1 scenario

Proper waste management can reduce total energy input from external sources. The lignin-rich solid residue is utilised as a fuel to burn in the boiler for heat and power generation in a CHP unit. Lignin-rich solid residue contributes more to the energy requirement by providing 30,640 MJ/m³-EtOH in scenarios WH1 and WH2. For WH3 and WH4, the energy generation is 40,910 MJ/m³-EtOH. As a result of low conversion efficiency in WH3 and WH4, the mass of combustible solids elevation can be noted. Furthermore, the produced methane by treating the spent wash via anaerobic digestion contributes to the energy input. For the four scenarios: WH1 to WH4, 3,588 MJ/m³-EtOH, 3,435 MJ/m³-EtOH, 4,436 MJ/m³-EtOH, and 4,315 MJ/m³-EtOH worth biogas quantities are produced, respectively. Table S2 in the supplementary document lists the efficiencies for heat and power generation. Waste woodchips are used as fuel to supply the remaining heat requirement that cannot be catered from the in-house fuel sources, such as lignin and solid residue, generated biogas, etc. After factoring in the efficiencies of the boilers and CHP unit, the energy flows for steam and electricity generation are shown in Table 7.

Table 7.

Energy flow analysis results

	WH1		WH2		WH3		WH4
	Steam	Electricity	Steam	Electricity	Steam	Electricity	Steam	Electricity
Heat and electricity generation from the CHP unit
Biogas	1614	1256	1546	1202	1996	1553	1942	1510
Lignin and solid residues	24,950	5691	24,950	5691	33,310	7597	33,310	7597
Wood chips	9864	2250	15,410	3514	7942	1811	14,860	3389
Total generated energy	36,430	9197	41,910	10,410	43,250	10,960	50,110	12,500
Electricity contributed to the grid		9197		10,410		10,960		12,500
Electricity consumed from the grid		8879		10,090		9788		11,100
Fossil energy input (60% of electricity)		5327		6055		5873		6659
Process net energy input	45,310		52,000		53,040		61,210
Process net energy output	30,400		31,610		32,160		33,700
Net energy ratio	0.67		0.61		0.61		0.55
Renewability	5.71		5.22		5.48		5.06

Bioethanol would have to provide a net energy gain during production to be a viable alternative to a non-renewable energy source. Despite the attractiveness of biomass to produce bioethanol, the process cannot be considered sustainable or economically feasible if there is a net energy loss. Figure 4 shows how the process net energy ratio and the renewability factor change depending on the process method. Amongst the scenarios, all four in the present study had a less than one energy ratio. Comparatively, WH1 showcases better performance with a 0.67 energy ratio and a 5.71 renewability factor. The scenario with the worst performance is WH4, with ratios of 0.55 and 5.06 NER and Rn. An interesting fact is that the renewability element of the plant is positive in all cases. One of the reasons for the renewability factor to be highly positive is the usage of biomass for the entire requirement of heat/steam generation instead of fossil-based fuels. The use of woody biomass in the burner benefits the environment by carbon neutralisation, stabilisation of heavy metals in ash etc., providing positive environmental aspects. However, the thermodynamic properties of biomass can change depending on the type of biomass. Moreover, inorganic substances absorbed into biomass in their growth phase can vary based on geographical location and soil quality. Thus, an accurate representation of biomass property data can be obtained through analytical techniques, such as Fourier transform infrared spectroscopy, gas chromatography, thermogravimetric analysis, mass spectrometry, etc. [70].

Fig. 4.

Results for process net energy ratio and renewability indicator for WH1, WH2, WH3 and WH4

Sensitivity analysis

Due to possible variations in the key process parameters, the mass and energy flows can change, influencing the interpretation of the results. Therefore, after evaluating fundamental process aspects, the key mass and energy flow indicators, such as bioethanol yield, total energy consumption, process net energy ratio and renewability, were analysed against highly varying process parameters to comprehensively understand the behaviour of the process. In this assessment, five process parameters were identified with a high possibility for variations. They are the lignin composition of WH, cellulose composition of WH, solid loading of the biomass, pretreatment efficiency, and fermentation efficiency. This sensitivity analysis provides an idea of the parameters which can influence the final result the most and the mass/energy flow indicators that are more sensitive to parameter variations. Based on the lowest energy consumption and the best overall performance, the WH1 scenario is selected and considered as the base case for the sensitivity analysis. Table 2 shows the selected threshold for key parameter fluctuations backed by the literature references. Table 8 lists the positive and negative deviations of the mass/energy flow indicators depending on sensitivity parameters change. Figure 5a to d illustrates results from the sensitivity analysis compared with WH1 as the base case. According to the sensitivity results, the parameter solid loading shows the highest sensitivity for key mass-energy flow indicators except for bioethanol yield.

Table 8.

Sensitivity analysis results

	99.7 Yield (kg/tonne)	Energy (MJ)	NER	Rn
Base case (WH1)	202	45,310	0.67	5.71
Lignin percentage in WH (wt.%)	205–184	44,910–49,020	0.67–0.63	5.06–5.74
Cellulose percentage in WH (wt.%)	197–355	46,480–27,210	0.66–0.99	5.60–8.33
Solid loading (biomass: water)	202–202	20,430–63,760	1.84–0.54	11.82–5.00
Pretreatment process efficiency (%)	187–212	48,990–43,370	0.63–0.69	5.39–5.89
Fermentation efficiency (%)	190–214	47,980–42,920	0.65–0.69	5.52–5.86

Fig. 5.

Sensitivity analysis results indicating; (a) Yield, (b) Process energy consumption, (c) Net energy ratio, (d) Renewability indicator

The composition of WH changes globally depending on the maturity of the plant and the climate [25]. For lignin and cellulose, there is a variation of 3–14% for lignin and 18–55% for cellulose mentioned in previous studies [50–52]. In the selected previous study in the laboratory scale, the composition of the used sample for pretreatment conversions reported 4.8% for lignin and 19.2% for cellulose, where both of the said components are towards the lower end of the variation. As shown in Fig. 5a, lignin percentage has a low sensitivity towards the yield and the rest of the indicators, whilst cellulose percentage variation has a more significant impact within the possible variation range. Since lignin is used as a fuel source to generate steam for the plant, a low lignin content affects the energy supply, reflecting on process energy gain and the renewability factor. Results of this study revealed that when the lignin percentage is 14, NER is 0.63, and Rn is 5.74. Even though it can be anticipated that when the lignin amount rises, NER and Rn simultaneously increase, because the percentage increase of lignin affects the percentage of the rest of the components, the sugar supply reduces, directly affecting the yield. Based on this behaviour of the system, it is evident that, ultimately, all the observed indicators in the sensitivity analysis give better values when the bioethanol yield is high. In reality, it is partial to expect the composition of every biomass stock arriving at the plant to be consistent. Nevertheless, the yield reduction due to composition variations can be minimised with improved conversion efficiencies aiding in a consistent supply of sugars.

Reviewed studies indicate a wide range of pretreatment and fermentation efficiencies [54, 71]. Hydrolysis efficiency can vary due to the structure of the plant, composition as well as the most present part of the plant in the biomass inflow, such as roots, stems and leaves. On the other hand, fermentation efficiencies also can give a degree of uncertainty because it changes, for example, based on the yeast strain used for the fermentation. A 5% decrease in pretreatment efficiency reduces the yield by 10.21 kg/tonne-WH and an increase of 5% makes the yield reduced by 14.79 kg/tonne-WH. Nevertheless, a 5% fermentation fluctuation gives a 12.21 kg/tonne-WH positive difference and an 11.79 kg/tonne-WH negative difference. At this point, the correlation between the yield and the process energy uptake to produce 1 m³ of fuel-grade bioethanol is noteworthy which is, the better the bioethanol yield, the lesser the process energy consumption. Being consistent with all of the scenarios discussed in this paper, even with the pretreatment variation the energy requirement is lesser where the yield is high and wise versa. Moreover, results indicate not only + 0.02 and − 0.04 and, + 0.32 and − 0.19 differences in energy ratio and the renewability compared with the base case with pretreatment efficiency variation, but also + 0.02, − 0.02 and + 0.18, − 0.16 difference with ± 5 fermentation variation.

Within the range of variation, the initial solid loading parameter shows the highest sensitivity on the key indicators. Bioethanol yield shows no sensitivity, and process energy ratio shows the heist sensitivity towards solid loading parameter. With the level of diluteness in the stream, there is high energy demand during distillation to separate excess water. In the laboratory scale studies, 1 kg of dry WH was added to a water mass of 10 kg before pretreatment to attain cellulose and hemicellulose from WH. Well-mixing is important within the reactor to achieve optimum output. There are studies that use 1 kg of biomass in a water mass of 15 kg for homogeneous mixing [53]. Therein suggest, an appropriate reactor design with minimum stirring limitations is crucial to have satisfactory mixing [72]. However, for industrial applications, high solid loading is desirable for energy efficiency. The majority of industrial and pilot-scale published studies with competitive results are with high solid loading representing more than 20% (w/w). Therefore, it is essential to thoroughly understand the industrial implications of how the critical factors, such as yield, energy consumption and environmental impacts, change along with the liquid volume in a biorefinery.

As per the results of this study, leaving cellulose and hemicellulose recovery the same as the base case (WH1), changing the dry biomass input against the water input does not affect the yield. Hence, the bioethanol yield remains the same as 202 kg/tonne of dry WH. In contrast, in all of the scenarios considered in this study based on available data from past studies, the only parameter that has the potential to vary the process energy consumption enough to achieve an energy gain is the solid loading parameter. The highest energy ratio given is 1.84, with an energy consumption of 20,430 MJ/m³-EtOH at a solid loading ratio of 1:3. This high solid loading value also gives high renewability with an 11.82 factor, while low solid loading only gave a renewability factor of 5. Results of the sensitivity analysis suggest to maintain a process energy gain (NER > 1), the threshold of the water inflow that has to be maintained is 5.2 kg or less per 1 kg of dry biomass in a plant that uses WH as the lignocellulose source to produce fuel-grade bioethanol.

Considering all the objectives of the process, which are high bioethanol yield, low energy consumption, NER > 1, and Rn > 1, the ideal scenario for a bioethanol production plant utilising WH is when lignin percentage is the lowest as available with a high cellulose percentage, and high conversion efficiencies. Therefore, for optimal performance, starting from the WH composition: cellulose - 55%, hemicellulose - 30%, lignin - 3%, and ash - 12%, cellulose recovery: 100%, hemicellulose recovery: 62%, glucose conversion: 95% and xylose conversion: 85% can be noted as the benchmark based on this study. These optimum parameters have the potential to achieve a 380 kg/tonne-WH bioethanol yield with energy consumption as low as 12,800 MJ/m³ of bioethanol.

The key findings from this study, including the sensitivity analysis results, can be used to identify favourable process conditions for better bioethanol production performance using WH as the feedstock. The findings would also support any stakeholder to design, develop, and implement a future biorefinery using WH for the purpose of fuel-grade bioethanol production.

Limitations and future work

Process simulations in this study were conducted using the existing models in the Aspen Plus process simulator. Results obtained by using these models can be different and deviate from a realistic representation of an actual plant. Moreover, there is no accessible real plant that converts WH into bioethanol as of now to practically pursue a trial run.

With the pressing demand to shift away from gasoline, economically and environmentally friendly process pathways to produce bioethanol is mandatory. Future research can be directed to study the possibilities of incorporating fungal strains, such as Aspergillus, Fusarium, Paecilomyces and Trichoderma that are capable of delignification, scarification, and fermentation. This allows for combining steps in the bioethanol process, making it possible to reduce the number of reactors [73]. Also, the co-production of bioethanol and biodiesel in the same biorefinery is a consideration for future studies [74]. Further, GHG and other emissions in a similar plant can be investigated along with health effects and location-centric social matrix. By means of such strategies, the bioethanol production cost can be reduced, which would encourage third-world countries to establish bioethanol plants.

Energy analysis is commonly used in the industry to aid decision-making to assess resource utilisation and to optimise the process designs to reduce energy consumption where possible. However, theoretical mass and energy analysis on its own can hinder actuality in real-world commercial applications. For a better understanding of the process, especially environmental and financial implications, it is better to couple with other sustainability assessment tools, such as Life Cycle Assessment (LCA), techno-economic analysis and/or exergy analysis [75]. Further, exergy analysis provides results based on not only the quantity of the material and energy flows but also considering the quality of the process streams. In other words, exergy analysis gives insight to decision-makers to pinpoint cost losses by identifying thermodynamic losses and environmental impacts [76]. Hence, for a better portrayal of the suitability of WH as a feedstock to produce bioethanol, future research can expand the scope for other sustainability assessment tools based on the results presented in this study.

Conclusion

In this study, four scaled-up bioethanol production process routes with combinations of two pretreatment and dehydration techniques using WH as the feedstock were analysed and compared. The process simulation-based analysis provided important findings on the key mass and energy flow indicators of scaled-up bioethanol production from WH as the feedstock. The findings reveal that there is a significant improvement in bioethanol yield when WH is processed via alkaline pretreatment compared with dilute acid pretreatment under the same conditions. Approximately 0.86 tonnes of an additional dry WH quantity are required to produce 1 m³ of bioethanol when dilute acid pretreatment is adopted as the pretreatment method. Further, the process route with the combination of alkaline pretreatment along with extractive dehydration indicated less energy consumption in comparison to other studied process routes. All considered process route scenarios did not show positive energy gains under the selected process conditions. Nevertheless, the findings from sensitivity analysis results pointed out that the solid loading ratio (dry biomass: water) plays a key role in the process performance of fuel-grade bioethanol production using WH, where it is advisable to maintain a ratio of 5.2 kg of water threshold per 1 kg of dry WH for positive energy gain from the overall bioethanol production process. The results from this study show the feasible ways of utilising WH as a fuel-grade bioethanol feedstock for industrial-scale production, and the findings support future LCA and implementations of new bioethanol production plants using WH as a feedstock.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contribution

All authors of this manuscript have directly contributed to the concept, design, execution, and analysis of this study. The first draft of the manuscript was written by Dulanji Imalsha Abeysuriya and all authors commented on previous versions of the manuscript and approved the final manuscript.

Funding

This research study was supported by the Senate Research Committee Grant (Grant no. SRC/LT/2021/18) at the University of Moratuwa, Sri Lanka.

Data availability

Additional data is electronically provided as supplementary information.

Declarations

Consent for publication

The authors give their consent for publication upon acceptance of the article.

Conflict of interest

The authors declare no competing interests.