Comparison of Pre- and Post-treatments of Sugarcane Industry By-products to Increase Biomethane Production

(1)

Department of Thematic Studies Environmental Change

MSc Thesis (30 ECTS credits) Science for Sustainable development

Baitong Huang

Supervisor: Alex Enrich Prast and Rubens Perez Calegari Examiner: Francesco Ometto

Comparison of Pre- and Post-treatments of

Sugarcane Industry By-products

(2)

Copyright

The publishers will keep this document online on the Internet – or its possible replacement – for a period of 25 years starting from the date of publication barring exceptional circumstances. The online availability of the document implies permanent permission for anyone to read, to download, or to print out single copies for his/her own use and to use it unchanged for non-commercial research and educational purpose. Subsequent transfers of copyright cannot revoke this permission. All other uses of the document are conditional upon the consent of the copyright owner. The publisher has taken technical and administrative measures to assure authenticity, security and accessibility.

According to intellectual property law the author has the right to be mentioned when his/her work is accessed as described above and to be protected against infringement.

For additional information about the Linköping University Electronic Press and its procedures for publication and for assurance of document integrity, please refer to its www home page: http://www.ep.liu.se/.

(3)

Abstract

Even though the Brazilian ethanol and sugar production system (based on sugarcane industry) have been providing large amounts of bioenergy, the extensive amounts of organic wastes generated cannot be ignored when it comes to sustainability. Using these biomasses to produce biomethane through anaerobic digestion has been proven as a promising way to tackle this issue. This study investigated the biomethane potential of the co-digestion of these biomasses: SF (sugarcane straw : filter cake = 8:2), SFV (sugarcane straw : filter cake : vinasse = 1:4:45), and D (digestate separated from AD of SFV). Three treatments autoclaving (AU), alkaline (AL) treatment using 6% (w/w) NaOH and the combination of these two (AUAL) were then conducted on SF and SFV as pre-treatments; on D as post- treatments. In the biomethane potential tests of untreated material, the highest methane yield was achieved by SFV with 275.28 ± 11 N ml CH4/g VS, followed by SF with 223.25 ± 10 N ml CH4 g-1 VS, substrate D also

resulted in a methane potential of 144.69 ± 2 N ml CH4 g-1 VS. As pre-treatments, AL and AUAL

both showed increase in methane yield (between 36.0% and 49.1%) and methane production rate. As post-treatments, AU, AL and AUAL showed distinctive results in methane production, with 33.8%, 99.8% and 128.8% increase, respectively. In comparison with pre-treatment, post-treatment showed a better performance in increasing methane production. The following feeding experiments performed in continuous stirred-tank reactors showed that AL treatment led to an average of 248% increase in methane yield.

Keywords: Biomethane potential, sugarcane straw, filter cake, vinasse, co-digestion, autoclaving treatment, NaOH treatment, Brazil

(6)

2

List of abbreviations

AD – Anaerobic digestion AU – Autoclaving treatment AL – Alkaline treatment

AUAL – The combination of autoclaving treatment and alkaline treatment BMP – Biomethane potential (test)

C – Control sample

COD – Chemical oxygen demand CSTR – Continuous stirred-tank reactor D – Digestate from SFV after AD GHG – Greenhouse gas

HRT – Hydraulic retention time OLR – Organic loading rate

SF – Sugarcane straw and filter cake

SFV – Sugarcane straw, filter cake and vinasse TAN – Total ammonia nitrogen

TS – Total solids

VFAs – Volatile fatty acids VS – Volatile solids

(7)

3

1. Introduction

Brazil plays an important global role producing bioenergy as it has become the world’s second largest producer of bioethanol. Bioethanol is one of the most widely used biofuels in the world with a global production of 96 billion liters in 2015 (Bordonal et al., 2018). In Brazil, bioethanol is produced from one of its most important agricultural sectors- sugarcane industry. It is estimated that a total of 621 million tons of sugarcane were harvested during the 2018/2019 season, producing 29 million tons of sugar and 33 billion liters of ethanol (UNICA, 2020). According to Manochio et al. (2017), sugarcane as an energy crop is favorable for ethanol production, especially on its industrial phase. To address the fossil fuel replacement and GHG mitigation both domestically and globally, Brazil’s sugarcane production will continue to increase.

Even though the Brazilian ethanol and sugar production system is providing large amounts of bioenergy, the problem of massively generating organic wastes cannot be ignored when it comes to sustainability (Janke, 2014). The organic wastes produced by Brazilian sugarcane industry include mainly sugarcane straw, filter cake, vinasse and bagasse. Sugarcane straw is an important residue produced by sugarcane industry. Filter cake is the leftover from cane juice filtration. Bagasse is the fibrous residue remaining after crushing cane to obtain the juice. Vinasse is the main residue of the sugar-ethanol industry. It is estimated that during the 2018-2019 harvest, 155 million tons of bagasse, 124 million tons of straw, 18.6-24.8 million tons of filter cake and 272- 644 billion liters of vinasse were generated (Janke et al., 2015; UNICA,2020 ). These organic wastes are poorly disposed and mismanaged, for example, a part of the vinasse and filter cake is applied in the field as fertilizers on the sugarcane field; the bagasse is used in some low-efficiency cogeneration systems; straw used to be directly burned in the fields or just left to decay on the fields, which caused negative impacts on the environment, such as promoting GHG emission and increasing the salinity of the soil. (Regis et al., 2013; Janke, 2014; Janke et al., 2015; Bordonal et al., 2018). Obviously, a good strategy needs to be implemented in taking advantage of these organic residues.

A promising way of tackling this problem is to collect these organic wastes and convert them into biogas through anaerobic digestion. In this way, these residues will not be decayed on the fields or burned on the field thereby protecting the environment as a waste management method. Anaerobically digesting of these residues has a much higher efficiency of energy generation than simply combusting them. The biogas produced out of the residues is mostly composed of methane (50%- 75%), carbon dioxide (25%- 50%) and a small amount of other components, such as hydrogen, hydrogen sulfide, ammonia and water (Da Costa Gomez, 2013). Biogas can be used as fuels for heating and electricity generation, also, a purified biogas (reaches to approximately 98% of methane) can be implemented as an alternative of natural gas (Da Costa Gomez, 2013), which is a huge advantage of saving costs on equipment for implementation since natural gas has been widely used worldwide.

(8)

4

Sugarcane industry by-products straw, filter cake are rich in lignocellulose (cellulose, hemicellulose and lignin), which are difficult to be anaerobically degraded by the microorganisms (Mailin et al., 2014; Janke, Weinrich, Leite, Terzariol, et al., 2017; Suzana et

al., 2017). Therefore, to produce biogas more efficiently from these lignocellulose-rich

biomasses, researchers have conducted various treatment methods on them. These treatment methods can be generally classified as physical treatment, chemical treatment, biological treatment and the combination of them (Sayara and Antoni, 2019). In some studies, material that has been treated has yielded higher methane compared to untreated materials (Dai and Dong, 2018; Kim et al., 2018; Nair and Kabir, 2018).

In recent years, instead of conducting treatment on the raw material that are used in the anaerobic digestion (usually called ‘substrate’), research has been conducted on the treatment of product that is obtained after the anaerobic digestion (usually called ‘digestate’) (Lindner et

al., 2015; Tsapekos et al., 2016). This approach aims only at the non-degraded part in the

digestate. It is still a quite new topic to research on, but the effectiveness has been proven by the researches (Lindner et al., 2015; Tsapekos et al., 2016). This project conducts research on this approach and is called ‘post-treatment’ in this project report.

1.1 Aim of the project

Based on the anaerobic co-digestion of sugarcane industry by-products (straw, filter cake and vinasse), the first aim of this project is to investigate the effectivity of the application of treatments in order to increase the lignocellulose- rich biomass degradation and consequently increase methane yield . A complementary aim was to investigate the effectiveness of post-treatment on the digestate (D) after the first substrate digestion.

1.2 Research questions

The following questions are the basis of this project:

⚫ What are the methane potentials of co- digestion of SF and SFV?

⚫ What is the methane potential of D (digestate of SFV after anaerobic digestion)?

⚫ As pre-treatments, what are the effects of AU (autoclaving) treatment, AL (alkaline) treatment and the combination on co-digestion substrate SF and SFV?

⚫ As post-treatments, what are the effects of AU treatment, AL treatment and the combination on substrate D?

⚫ Can autoclaving treatment maximize the positive effect of the alkaline treatment in the combination?

⚫ What is the performance of post-treated D in CSTR reactor?

(9)

5

2. Background

2.1 Sugarcane and sugarcane-based bioenergy industry in Brazil

Sugarcane has been a main source of sugar and alcohol (fermented from the plant juice) for almost 500 years (Souza et al., 2014). Nowadays, sugarcane biomass has become an important raw material for a variety of bio-based products, including food (sugar and related products), bioenergy (ethanol), bio-based bulk materials (bio-plastic) and bio-based high value products (e.g. flavors and fragrances) (Scheiterle et al., 2018). With the increasing demand of replacing fossil fuel use and developing renewable energy, sugarcane will always be important in providing bioenergy due to its exceptional ability to produce biomass (Souza et al., 2014). Sugarcane has many advantages of being used in bioenergy production system in terms of sustainability (Peng et al., 2014; Scheiterle et al., 2018): 1) it has high yields per unit land area in tropical environment (50-150 tons of cane per hectare, dry matter accounts for 30-40%), as well as high radiation use efficiency and high water use efficiency; 2) it can adapt to tropical environments, which leads to its perennial growth whereby the plants re-grow after harvesting for several successive years with only small annual decrease in yield; 3) the perennial growth also reduces the negative impacts on the environment (e.g. soil erosion), due to the less needs of inputting energy and costs on annual soil tillage and planting operations; 4) a high proportion of sugarcane biomass are naturally simple compounds (e.g. simple sugars, accounting for approximately 12-13% of fresh harvested weight basis, or 45-50% on a dry-weight basis), which are easy to ferment and available for high-technology processing.

The world’s leader in sugarcane production is Brazil, and it has already played a pioneering role in promoting the use of bioenergy derived from sugarcane (Scheiterle et al., 2018). The story can be earliest tracked back to the economic crisis of 1929, which pushed the federal government to invest in creating a sugar buffer stock, meanwhile the Institute of Sugar and Alcohol was created to help govern the regulation of sugarcane production (Souza et al., 2014; Scheiterle et al., 2018). Then a boost of sugarcane-based bioenergy industry happened in 1973 due to the first oil crisis. The government established the National Alcohol Program- ProÁlcool in 1975 to develop new sources of energy from sugarcane (Matsuoka, Ferro and Arruda, 2009; Souza et al., 2014; Scheiterle et al., 2018). The ProÁlcool program focused on producing ethanol as an alternative liquid fuel to substitute for imported oil (Arruda, 2011). The changes started from cars. The car engine was re-designed for using bioethanol, and the first light cars capable of using exclusively bioethanol were produced in 1978 (Souza et al., 2014). However, a remarkable change took place in 2002, as the flex-fuel technology was introduced. It allows consumers to run their cars with gasoline, hydrated ethanol, or any proportional mixture between the two fuels (Arruda, 2011). According to Matsuoka, Ferro and Arruda (2009), the flex-fuel technology completely changed the consumer’s attitude towards fuel ethanol, the sales of flex-fuel new cars grew rapidly and reached the level of approximately 86% of new car sales

(10)

6

around 2010. This increasing consumption of flex-fuel cars also boosted the Brazilian sugarcane industry, ethanol production increased by about 2.5 times from 2000 to 2009 (Arruda, 2011).

According to Bacovsky et al. (2016), Brazil has a very high share of renewable energy, representing around 43% of the total primary energy supply. Bioenergy plays a dominant role in renewable energy sources, accounting for 70.1%. Sugarcane as a superior crop for the production of bioenergy has been placed Brazil as a center of the renewable energy world, which also provides an good example for other countries, such as China and India (Arruda, 2011; Peng et al., 2014; Solomon, 2014).

2.2 Sugarcane industry by-products

The Brazilian ethanol and sugar production system has been providing large amounts of bioenergy, however, the large scale of sugarcane-based bioenergy industry also brings extensive amounts of agricultural residues, such as straw, filter cake, bagasse and vinasse (Janke et al., 2015), which cannot be ignored when it comes to sustainable development.

Sugarcane straw is also known as tops and trash, accounting for approximately one third of the total primary energy of the sugarcane (Janke, Weinrich, Leite, Schüch, et al., 2017). Filter cake is the leftover from cane juice filtration, it causes serious pollution and is considered as a waste in many sugar factories (Ochoa et al., 2010). Bagasse is the fibrous residue remaining after crushing cane to obtain the juice, it can be burnt in mill boilers to provide steam and electricity for self-consumption (Peng et al., 2014). Vinasse is the main residue of the sugar-ethanol industry and is produced on average of 13 L for each liter of ethanol (Bordonal et al., 2018). Vinasse is usually an acidic dark-brown slurry (pH = 3.5- 5) with a high organic content (Christofoletti et al., 2013).

During the 2013-2014 season, 91 million tons of straw (dry basis), 169 million tons of bagasse (wet basis), 22 million tons of filter cake (wet basis) and 286–678 billion liters of vinasse were generated from 653 million tons of sugarcane harvest (Janke et al., 2015). The bagasse is often used to generate electricity, while straw, filter cake and vinasse still remain mismanaged, for example, straw is directly burned or left decayed on the fields; vinasse and filter cake are directly applied as fertilizers without previous energetic utilization (Regis et al., 2013; Janke, 2014; Janke et al., 2015; Bordonal et al., 2018). These inappropriate waste managements are using these biomasses in unsustainable ways and can cause negative environmental impacts, for example, the non-controlled digestion of such wastes on the fields may cause the release of huge amounts of methane (Janke et al., 2015). A much more sustainable and promising way of handling these organic wastes is to produce biogas through anaerobically digesting these wastes, which is the focus in this project.

(11)

7 2.2.1 Methane potential and co-digestion

Many studies have proven that these sugarcane industry by-products have promising methane potentials (Janke et al., 2014, 2015; Sanchez-Herrera, 2018). Janke et al. (2014) obtained biogas yield of 395, 486 and 647 N ml/g VS for straw, filter cake and vinasse, respectively. And it is estimated that if all of the filter cake generation was converted to biomethane, it could substitute 10% of the total natural gas consumption in São Paulo State; if the entire vinasse could be used to produce biomethane, 17% of the natural gas consumption could be substituted in São Paulo state.

The practice of adding two or more substrates in the anaerobic digestion is called co-digestion. Researches have studied co-digestions of a variety of substrates, for example, sugarcane straw co-digested with sugarcane filter cake (Janke et al., 2017). Anaerobic co-digestion of sewage sludge and other organic matter has been proven to have advantages of increasing the biogas production and the degradation of organic matter (Martinez-jimenez et al., 2017). Co- digestion can also improve the buffering capacity of the mix substrates and balance the nutrients in the reactors (Giordano, 2012). In an anaerobic digestion reactor experiment, Moraes et al. (2015) found that the biogas production failed when only vinasse was fed to the reactor, due to the low content of nutrients and low C/N ratio of vinasse. Then straw is added for co-digestion, the biogas production reached close to BMP results of vinasse. Other advantages of the co-digestion, including increase in the biodegradable content and providing a wider diversity of microorganisms present in the process (Martinez-jimenez et al., 2017).

2.3 Anaerobic digestion process

AD is the process of fermenting biogas from organic materials. The organic matter is degraded by microorganisms in an anaerobic environment, then a final product consisting of methane (50%- 75%), carbon dioxide (25%- 50%) and a small amount of other components, such as hydrogen, hydrogen sulfide, ammonia and water is produced (Da Costa Gomez, 2013). The biogas produced can be used as fuels in providing energy. AD is a very complex process that involves several groups of microorganisms (Wang, 2016). The degradation process can be generally divided into four steps: hydrolysis, acidogenesis, acetogenesis and methanogenesis (Wang, 2016; Deo and Bjerg, 2017; Sayara and Antoni, 2019). A schematic graph of the whole anaerobic degradation process is presented in Figure 1.

(12)

8

Figure 1. A schematic graph of the whole anaerobic degradation process (adapted from Angelidaki et al., 2011)

2.3.1 Hydrolysis

Hydrolysis is the first step of degrading organic matter in AD process. Macromolecules such as proteins, carbohydrates and lipids, are all broken down by extracellular hydrolytic enzymes into small molecules (e.g. amino acids, long chain fatty acids and sugars) (Angelidaki et al., 2011; Wang, 2016). In this process, the hydrolytic enzymes (e.g. cellulase, amylase, lipase and protease) are produced by hydrolytic microorganisms that work hydrolytically (Deo and Bjerg, 2017). Hydrolysis is usually considered as a rate-limiting step, as many materials consist of complex structures and thus protect the degradable component from being accessible to microorganisms (Angelidaki et al., 2011). A variety of treatments have been applied on such materials to help achieve a more efficient digestion in hydrolysis process (2.4).

(13)

9 2.3.2 Acidogenesis

Acidogenesis is the second step in AD. The hydrolyzed products from last step are converted into carbon dioxide, volatile fatty acids (VFAs), alcohols and hydrogen by acidogenic microorganisms (Wang, 2016). The most common VFAs formed are acetate and butyrate; whereas alcohols, lactate and propionate are formed when the process is overloaded with organic matter (Angelidaki et al., 2011). Acidogenesis is usually considered as the quickest step in AD. However, during this step, acidogenic microorganisms produce high concentrations of hydrogen, which may inhabit the production of acetate in the next step; by-products such as ammonia (NH3/ NH4+) and hydrogen sulfide (H2S) are also produced, which could also be toxic to the methanogens later (Kandylis et al., 2016; Sayara and Antoni, 2019).

2.3.3 Acetogenesis

In the step of acetogenesis, the VFAs and alcohols are converted into products such as acetic acid, hydrogen, as well as carbon dioxide by acetogenic microorganisms (Angelidaki et al., 2011). In this step, an important factor needs to be considered - the partial pressure of the

hydrogen, which needs to be kept in a low level (Sayara and Antoni, 2019). A high hydrogen

partial pressure will favor CO2 reduction rather than the formation of acetate from organic acids, which may inhibit the bioconversion process and result in accumulation of fatty acids (Treu et

al., 2016; Deo and Bjerg, 2017). The main bacteria performing in the acetogenesis are

hydrogen-utilizing bacteria and hydrogen-producing bacteria, where the former might compete with the methanogens for hydrogen, formate and methanol (Angelidaki et al., 2011).

2.3.4 Methanogenesis

In the last step of AD, the methanogenic bacteria and archaea consume the acetate, H2 and CO2

to produce methane, as well as CO2 (Angelidaki et al., 2011). The two paths for the methane

formation are: 1) splitting acetate to CH4 and CO2; 2) producing CH4 through H2 as an electron

donor and CO2 as a carbon acceptor (Dewil et al., 2008). However, approximately 70 % of the

methane is converted from acetate (Angelidaki et al., 2011). Methanogens are considered to be the most sensitive microorganisms in AD process, their activity can be easily inhibited by

factors such as pH and high concentrations of ammonia (NH3) (Wang, 2016).

Maintaining a suitable and stable state is significant throughout the whole AD process. A small fluctuation or decrease in the activity of one or several microbial groups could severely influence the process performance and efficiency, even lead to process failure (Sayara and Antoni, 2019).

(14)

10 2.3.5 Parameters that are important for AD process

Several parameters should be monitored/measured during AD process, including temperature, pH, VFAs, OLR and HRT. Other parameters, such as, TS&VS, are important for estimating the biogas potential for substrates.

Temperature

Temperature is an important factor in the AD process because it affects the activities of microorganisms (growth rate and metabolism) and the physiochemical properties of the compounds in the substrate (e.g. enzymes) (Dewil et al., 2008). Different microorganisms have different optimal temperatures, they can be active at four different temperature ranges: psychrophilic (5-25℃), mesophilic (30-35℃), thermophilic (50-60℃) and hyper-thermophilic (> 65℃) (Wang, 2016). According to Wang (2016), fluctuation in temperature in AD process should be minimized below 1℃ per day for thermophilic digestion, and 2-3℃ per day for mesophilic digestion.

Methanogens are one of the most sensitive groups to increasing temperatures, they are active in mesophilic and thermophilic conditions. Therefore, most of AD processes are run at these two conditions (Schnurer A, 2010). The final methane yield from substrate will not significantly change in the temperature interval of 30-60℃, but a higher temperature can lead to many advantages, including higher biological and chemical reaction rates and increase in the solubility of the organic compounds and destruction of pathogens (Dewil et al., 2008; Wang, 2016).

Generally, it is of great importance to keep the AD process in a suitable and stable temperature environment.

pH

pH is another important factor that affects the AD process. Different microorganism groups have various optimal pH range: methanogens are the most sensitive to pH with an optimal pH between 6.5 and 7.2; the acidogenic microorganisms are less sensitive and can function well in a wider range between 4 and 8.5; acetic and butyric acid are the main products when the pH is low; while acetic and propionic acid are mainly produced when the pH is around 8.0 (Dewil et

al., 2008). The maximal biogas yield can be achieved with an optimal pH range of 6.7-7.5

(Wang, 2016). VFAs

Volatile fatty acids (VFAs) are important intermediate products after the acidogenesis and are the main substrates for methanogens to use and produce methane (Angelidaki et al., 2011). The VFA levels are supposed to be monitored: 1) accumulation of VFAs can reduce the pH and

(15)

11

inhibit the methanogenesis process; 2) accumulation of VFAs can be related to an accumulation of free ammonia, which can inhibit the AD process; 3) accumulation of VFAs can also be related to variations in organic overloading and presence of toxic compounds. (Chen, 2008; Dewil et

al., 2008; Wang, 2016). Therefore, a high level of VFA can indicate that the AD process is not

working properly.

Total solids (TS) and volatile solids (VS)

TS and VS parameters are usually measured for estimating the biogas potential for a substrate. The value of total solids depends on the content of dry matters in the substrate with a subtraction of the moisture content. It is a common parameter for treatment design (3.1.3) and monitoring, and normally determined in an oven through a drying process (Peces and Astals, 2014). However, it is a variable value of the substrate that changes with time, storage or manipulation (Peces and Astals, 2014). The value of VS represents the content of organic matter in the substrate. VS can provide a first approach of the organic matter available to the biodegradation and its decrease is used as a process control parameter (by comparing the VS before and after AD) (Peces and Astals, 2014).

OLR

The organic loading rate is the amount of dry organic solids loaded per unit time, per unit volume of a digestion process (Siddique and Wahid, 2018). A lower OLR can lead to inefficiency in AD process, however, when the OLR is beyond a specific range, it can cause accumulation of VFA and ethanol, poor heat transfer, and uneven distribution during stirring (Siddique and Wahid, 2018).

HRT

The hydraulic retention time is the amount of time needed for any microorganism to consume and synthesize the substrate (Siddique and Wahid, 2018). Uncontrolled HRT has inhibitory effects on the metabolic activity of microorganisms, and long HRTs can cause the death of microorganisms due to shortage of nutrients (Siddique and Wahid, 2018).

COD

The chemical oxygen demand (COD) the amount of oxygen that is needed to completely oxidize the organic matter in the sample, in lab experiment it is determined by measuring the amount of a chemical oxidizing agent needed to fully oxidize a sample (Angelidaki et al., 2011).

Ammonia

Ammonia is one important nutrient for bacterial growth, however, a high concentration during anaerobic digestion process can inhibit methanogenesis and lead to digester failure when it

(16)

12

exceeds the inhibition threshold levels (Yenigün and Demirel, 2013; Sayara and Antoni, 2019). Total ammonia nitrogen (TAN) concentration is usually measured to monitor the concentration of ammonia in anaerobic digestion, which includes both free ammonia nitrogen (FAN) and

ammonium (NH4+). A range of TAN concentration between 1500–7000 mg N/L was found to

be inhibitory for anaerobic digestion (Sayara and Antoni, 2019). However, it should be noted that FAN was found to be the major cause of ammonia inhibition as it is membrane-permeable, causing proton imbalance and/or potassium deficiency inside cells (Yang et al., 2015).

2.4 Difficulties of degradation of lignocellulosic materials

As mentioned above (2.3.1), in hydrolysis process, complex structures of some materials can protect the degradable component from being accessible to microorganisms, therefore, reduces the methane yield. Lignocellulosic biomasses such as sugarcane straw and filter cake are materials with such structure. These lignocellulosic biomasses have a rigid crystalline structure that make them resistant to microbial degradation (Dai and Dong, 2018). The cellulose that is easy to be degraded is tightly wrapped by lignin and hemicellulose, preventing the contacts between microorganisms and enzymes, and thus significantly influences the biomethane production.

A suitable treatment can help to open the structure, make the degradable component more accessible and greatly increase methane yield. In general, these treatment methods are divided into physical treatment, chemical treatment, biological treatment and the combination of them (Sayara and Antoni, 2019).

2.4.1 Pre-treatment

Pre-treatment is the treatments that are conducted directly on the raw materials before AD. There are many articles exploring different pre-treatment methods on lignocellulosic materials. The most effective parameters in the biological conversion of lignocelluloses are believed to be cellulose crystallinity, accessible surface area, lignin and hemicellulose protection, cellulose degree of polymerization, degree of hemicelluloses acetylation, cellulase adsorption and desorption, and the biomass swelling capacity (Reddy, 2015).

Physical pretreatments can change the structure of biomass, typically increasing the enzyme accessible surface area, and reducing the degrees of polymerization of biomass (e.g. milling, irradiations) (Reddy, 2015). Thermal and autoclaving can change the crystalline structure of cellulose by disrupting hydrogen bonds (Kim et al., 2018). Physical pretreatments have the advantage of being eco-friendly due to no additional chemicals, which avoids the formation of toxic compounds (Bolado-rodríguez et al., 2016).

Chemical pretreatment can modify the crystalline structure of cellulose and remove and/or modify hemicelluloses and lignin by using chemicals (Reddy, 2015), and are widely used on

(17)

13

lignocellulosic biomasses. Alkaline processes are among the most efficient pretreatment processes, especially for hardwoods and agricultural residues (Reddy, 2015). Alkali pretreatment can remove the crosslinks between hemicellulose and other components, and reduces cellulose crystallinity (Kim et al., 2018). Acid pretreatment is one of the most applied methods for cleavage of glycosidic bonds in hemicellulose (Reddy, 2015).

2.4.2 Post-treatment

A new strategy of treating lignocellulosic biomasses has been investigated – conducting treatments on the digestate (Lindner et al., 2015; Tsapekos et al., 2016). Then the treated digestate can be re-used by feeding back to the AD reactor or feeding to another sequencing batch reactor. After AD, only the non-degraded part is left in the digestate, this approach thus aims only at these ‘stubborn’ matters. Lindner et al., (2015) applied mechanical treatment on three different digestates, they obtained methane yield from untreated digestate between 290- 320 ml CH4/g VS; as well as 9-17% increase achieved by the post- treatments.

2.4.3 Treatment method selection

The method selecting with personal information from T Mendes (2019, personal communication) that is conducting a review on the most efficient pre- treatment methods for the biogas sector. Based on the sort of substrate and demands of equipment, three articles were selected from the study, which contain feasible treatment methods, and with highlighted effects on increasing methane production on lignocellulosic biomass (Bolado- rodríguez et al., 2016; Dai and Dong, 2018; Kim et al., 2018). Then a further study on autoclaving treatment and alkaline treatment was conducted.

Bolado-rodríguez et al. (2016) and Kim et al. (2018) presented that physical treatment- autoclaving has a positive effect on increasing methane production of rice straw. Kim et al. (2018) obtained 114% increase on methane yield by treating rice straw with autoclaving in 121℃ for 60min. Considering the easy accessibility of heat and high pressure in the sugarcane and biogas industries, autoclaving treatment undoubtedly has the advantage of being convenient and cheap. Therefore, autoclaving is chosen as one of the treatments that will be conducted in this study.

Alkaline treatment using NaOH as a reagent has been proved to have remarkable effects on increasing methane yield for lignocellulosic biomasses (Reddy, 2015). Janke et al. (2017b) pointed out that using NaOH to pre-treat sugarcane straw can lead to more releasing of COD in comparison to the control sample, which is considered to be the reason for the high performance in methane yield. In another study Janke et al. (2016) demonstrated that the alkaline pre-treatment using 6% (w/w) NaOH on sugarcane filter cake improved both the COD solubilization and the VFA yield by 37%, consisted mainly by n-butyric and acetic acids. Morais

(18)

14

et al. (2016) reported that using 5% and 10% (w/w) NaOH to pre-treat sugarcane straw achieved

the lignin removal to 60% and 94%, respectively. The removal of lignin can make the degradable components more accessible for microorganisms, as mentioned above (2.4). Dai and Dong (2018) achieved 207% of increase on methane yield by using 6% (w/w) NaOH to treat rice straw. They also presented that 6% (w/w) has the highest impacts among other NaOH concentrations. Tsapekos et al. (2016) used thermal alkaline treatment to treat lignocellulose-rich fibers, the best result was obtained by using 6% (w/w) NaOH. However, Janke et al. (2017b) obtained a slightly higher methane yield by pre-treating sugarcane straw with a high concentration of 12% (w/w) NaOH. Mcintosh and Vancov (2011) investigated the combination of NaOH alkaline treatment and 121℃ autoclaving treatment, the combination of the treatments caused an increase in total sugar yields from the wheat straw. In this study, the alkaline treatment is chosen as another treatment and the concentration of NaOH is determined as 6% (w/w). The combination of autoclaving and alkaline treatment was also applied on the substrates to investigate if autoclaving treatment can improve alkaline treatment in the combination.

(19)

15

3. Materials and methods

The aims are accomplished by first conducing treatments on different substrates. Then Biomethane Potential (BMP) tests were conducted on the treated materials to test the methane yield and the effectiveness of the treatment methods. Finally, one treatment method was selected and conducted on D. The treated material was then applied on a continuous stirred-tank reactor, and was compared with a reactor fed with untreated D.

3.1 Treatments

In this study, three treatment methods and control are applied on three substrates (Table 1). ⚫ Three substrates: SF (straw and filter cake), SFV (straw, filter cake and vinasse) and D

(digestate from the co-digestion of straw, filter cake and vinasse, see 3.1.1.2)

⚫ Three treatments: AU (Autoclaving) treatment, AL (Alkaline) treatment and AUAL (the combination of autoclaving and Alkaline) treatment.

Table 1. The list of treatments that were applied

Control AU AL AUAL

SF SFC SFAU SFAL SFAUAL

SFV SFVC SFVAU SFVAL SFVAUAL

D DC DAU DAL DAUAL

3.1.1 Raw material collection

SF and SFV substrate

The raw materials of straw, filter cake and vinasse were collected in Brazil and transported to Sweden, they were kept at -20 ˚C until the experiments were carried out. The vinasse came from Ethanol and sugar industry Iracema, from São Martinho Group, located in the city of Iracemapolis, São Paulo State, Brazil. The straw was collected from a cane field near the same industry that the vinasse was collected. The filter cake was from the same industry that the vinasse was collected. After being collected, straw and filter cake were dried in an oven with 70 ˚C with air circulation for 24 hours in order to remove moisture. Then, they were grounded in a 1mm smash mill, packed in vacuum bags and transported.

Before preparing samples for treatments, water was added to each raw material to offset the water loss caused by transportation and storage. The content of SF and SFV is determined according to the proportion of the productions of straw, filter cake and vinasse in the industry, which is 8% of straw, 2% of filter cake and 90% of vinasse according to investigation. For example, in order to make 100g SF sample, 80g of straw and 20g of filter cake should be added,

(20)

16

while 100g SFV sample is mixed with 8g of straw, 2g of filter cake and 90g of vinasse. D substrate

A study of co-digestion of sugarcane straw, filter cake and vinasse in a 5L CSTR reactor is being performed at Linköping University, Sweden. The reactor was incubated at 37℃, HRT 20 days, the OLR was at a proper level which can keep the AD process inside of the reactor stable. The reactors were continuously stirred (100 rpm). Every day 250 ml of the reactor sludge was

exchanged. No special treatment has been employed on the substrates fed in the reactor. TheD

substrate in this study is directly obtained from the outgoing digestate of the co-digestion reactor.

3.1.2 TS and VS analysis

The total solids (TS) and volatile solids (VS) were measured followed by the standard method (APHA, 2017). The process was achieved by a two-step heating procedure. Three crucibles were prepared for each sample, marked with a pencil and weighed (weight; empty). Approximately 8 g of the substance was measured and transferred to each crucible and weighed (weight; wet). The crucibles were then kept in 105 °C for at least 20 hours using an oven

(Termaks®_{). The crucibles were then cooled down and weighed (weight; dried at 105℃) for TS}

calculation (Equation 1). After TS measurement, the crucibles were heated to 550 °C for 2 hours in a muffle furnace (Naber Industrieofenbau®, Bremen). Later they were cooled, weighed again (weight; combusted at 505℃) for the VS calculation (Equation 2).

𝑇𝑆 (%) = 𝑤𝑒𝑖𝑔ℎ𝑡; 𝑑𝑟𝑖𝑒𝑑 𝑎𝑡 105℃−𝑤𝑒𝑖𝑔ℎ𝑡;𝑒𝑚𝑝𝑡𝑦

𝑤𝑒𝑖𝑔ℎ𝑡;𝑤𝑒𝑡 − 𝑤𝑒𝑖𝑔ℎ𝑡,𝑒𝑚𝑝𝑡𝑦 × 100 (Equation 1)

𝑉𝑆 (% 𝑜𝑓 𝑇𝑆) = 𝑤𝑒𝑖𝑔ℎ𝑡; 𝑑𝑟𝑖𝑒𝑑 𝑎𝑡 105℃ − 𝑤𝑒𝑖𝑔ℎ𝑡;𝑐𝑜𝑚𝑏𝑢𝑠𝑡𝑒𝑑 𝑎𝑡 550℃

𝑤𝑒𝑖𝑔ℎ𝑡;𝑑𝑟𝑖𝑒𝑑 𝑎𝑡 105℃ − 𝑤𝑒𝑖𝑔ℎ𝑡;𝑒𝑚𝑝𝑡𝑦 × 100 (Equation 2)

3.1.3 TS adjustment

There is a large difference in TS between the three substrates. SF consists of only straw and filter cake, which has the largest TS of 73.7%; followed by SFV of 10.7% TS due to the liquid fraction of vinasse; D has a much smaller TS of 1.8% compared to the other two substrates (Table 2). Therefore, in order to compare the effects caused by treatment methods on SF, SFV and D, the TS of these three substrates should be adjusted to the same, which was decided as 10% based on Dai and Dong's (2018) work.

For TS adjustment of D substrate, centrifuging was performed. The centrifuging was performed

on a centrifuging machine (Beckman Coulter®_{, Ja-10 rotor) in a condition of 10,000g for 20}

min. After centrifuging, a certain amount of supernatant has been removed from the sample to increase the TS. However, for samples of AL and AUAL treatment, the addition of NaOH needs to be taken into consideration when determining the amount of supernatant that has to be

(21)

17

removed. It means that the TS for D is supposed to be more than 10% after centrifuging, then it can be achieved to 10% by adding NaOH and water (for AL and AUAL treatment). It is worth mentioning that even if the centrifugation has been very thorough, it is still inevitable that a little amount of solid will remain in the supernatant, which will cause the TS of D to be slightly less than calculated value. For SF and SFV substrate, ultrapure water (milli-Q®_{) is used as liquid} agent to maintain 10% TS. Table 2. TS, VS of SF, SFV and D Substrate TS (%) VS (%) SF 73.7 ± 0.3 90.3 ± 0.1 SFV 10.7 ± 0.1 83.6 ± 0.1 D 1.8 ± 0.1 66.0 ± 1.6

Average ± standard deviation

3.1.4 Conducting treatments

Each substrate (SF, SFV and centrifuged D substrate) was weighed in four borosilicate bottles of 250 ml to perform the following treatments. In the example of SF substrate, four bottles were marked as SFC, SFAU, SFAL, SFAUAL. The amount should be sufficient for BMP tests and the follow-up experiments.

Control

Control samples were samples without treatments, they were made in order to be compared with the treated samples. For control, ultrapure water (milli-Q®_{) was used as a liquid agent to} maintain 10% total solids concentration for all the substrates. The samples were well-mixed and kept in room temperature until the following BMP tests.

AU treatment

For AU treatment, ultrapure water (milli-Q®_{) was used as a liquid agent to maintain 10% total} solids concentration. Then, the bottles were covered with autoclaving bags and placed in the autoclaving machine. The bottles were left slightly open due to the high pressure will be caused by autoclaving. The substrates were autoclaved at 121℃, 1 atm for 60min. After autoclaving, the machine took approximately 1 hour to cool down. Later, the samples were taken out from the machine with autoclaving bags. The bottles were closed with the bags on, then the bags were removed. After AU treatment, the samples were kept in room temperature until BMP tests. Due to vinasse itself is acidic, therefore, in SFVAU treatment, substrate SFV is autoclaved in an acidic environment. This can be seen as a combination of acid treatment and autoclaving treatment.

(22)

18

AL treatment

For AL treatment, powdered 6% (w/w) NaOH (99.0% purity, EMSURE®_{) was added into}

substrates, followed by addition of ultrapure water (milli-Q®_{) to maintain 10.0% total solids} concentration. The samples were stirred to achieve a good mixture for chemical reactions. Then the bottles were closed and kept in room temperature.

AUAL treatment

The same procedures for AL treatment were conducted. Then, the bottles were covered with autoclaving bags and placed together with AU treatment samples in the autoclaving machine. The subsequent operations are the same as AU treatment.

3.2 BMP tests

3.2.1 Inoculum collection

The inoculum was collected from outgoing digestate at TekniskaVerken®_{wastewater treatment}

plant in Linköping, Sweden. It receives wastewater from inhabitants and industries in Linköping. The treatment plant is working under the condition of 38℃ temperature, the organic loading rate (OLR) is around 2 kg VS/m3_{d, the hydraulic retention time (HRT) is 20 days and} the total reactor volume is 6000 m3_{. After collecting, the inoculum is stored for 5 days of} degassing before being used in BMP tests, according to the method recommendations.

3.2.2 BMP tests setup

According to the method reported by Ekstrand et al. (2013), the BMP tests were performed in triplicates for substrates and inoculum. In the experiments, 20 ml inoculum and substrate (ratio of inoculum: substrate = 2:1) were added in 330 ml serum bottles. For substrates treated by AL and AUAL treatment, pH was neutralized by adding 37% hydrochloric acid solution (VWR chemicals®_{). Ultrapure water (milli-Q}®_{) was used to maintain a total volume of 100 ml. The} bottles were flushed with nitrogen gas to obtain an anerobic environment, then sealed with rubber stoppers. Needles were used to release the overpressure caused by flushing nitrogen. Finally, the bottles were placed and incubated in two water bath sets with a temperature of 37 ℃. The bottles were incubated for 42 days in total (more incubation time cannot be achieved due to the limitation caused by covid-19).

(23)

19 3.2.3 Gas measurements

Gas pressure measurement

The gas in BMP bottles was measured every day in the first 4 days, then it was measured approximately 2-3 times a week, and once in the last two weeks. It was done by measuring the overpressure in the bottles, using a pressure meter (Testo® 312-3, 0 ~ 6000 hPa). A needle(BD Microlance®) was first connected with the pressure meter, when measuring, the needle was inserted in the bottles through the rubber stopper. One needle is used for one triplicate. After every measurement, the pressure gauge was re-calibrated. The gas pressure data was used to calculate the gas volume.

Gas sampling

After measuring the gas pressure, the gas was sampled for methane concentration determination. An 1ml syringe was used to inject into BMP bottles and pumped five times to get a well-mixed gas sample. Then the sample was transferred into a small bottle with a volume of 31 ml, which was filled with air and sealed with a small rubber stopper. The diluted air samples will be used in following step to determine the methane concentration.

Methane concentration measurement

The methane concentration of the gas samples was measured with a GC-FID (HP5880A series gas chromatograph equipped with poraplotT column), a 0.5 ml of the diluted gas sample was needed for the measurement, according to (Deo and Bjerg, 2017). Before measuring the gas sample, three standards were run on the machine for calibration and standardizing. The standards were with 0.07 %, 0.63 % and 1.71% of concentrations of methane. The error of three times measurements for one standard sample should be within approximate 5 % for accuracy. The gas samples were then injected in the machine. The data of methane content was calculated with Microsoft Excel® using a calibration curve from the known standards.

3.2.4 Analyzed parameters

The following parameters were evaluated (and further described): pH, Volatile Fatty Acids (VFA), Total Solids (TS), Volatile Solids (VS). Besides the triplicate used in BMP test, another set was also prepared in the meantime, in order to taking samples for the analysis of VFA and pH before BMP test.

pH analysis

After the bottles had been prepared, the pH was measured for each solution of substrate, control

and inoculum. The pH measurements were done with a pH meter (WTW® inoLab pH 7310)

(24)

20

together, and the pH was measured again. The pH was measured three times for each sample to verify the results. Before performing the pH measurements, the pH meter was calibrated and a pH reference with pH of 8 was used to check the accuracy.

VFA analysis

For comparison, the VFA analysis was conducted twice on samples ‘before BMP test and ‘after BMP test’. The measurement was done by following the methods reported by Jonsson and Boren (2002). The sample was centrifuged first, then the measurements were performed on a

gas chromatograph (6890 Series, Hewlett Packard®_{, USA), which quantifies the concentrations}

of VFAs including acetic acid, propionic acid, isobutyric acid, butyric acid, isovaleric acid, n-valeric acid, isocapronic acid, n-capronic acid and heptonic acid.

TS and VS analysis

The analysis was made on each substrate before BMP tests, after BMP tests (on the remaining batch liquids) and inoculum (followed by 3.1.2).

COD and ammonia analysis

For COD and ammonia, measuring kits from Hach-Lange (LCK 014, LCK 302 for COD and ammonia, respectively) were used following the manufacturer’s instructions.

3.2.5 Data analysis

All the data were summarized and saved in Excel files. An average value was calculated for each triplicate and used in further data analysis. The graphs in this study were made by using Origin 2017®_{program, designed by Origin Lab.}

Conversion of gas pressure to gas production

The total gas production in BMP bottles was calculated by converting the measured pressure to atmospheric pressure. First, the total gas volume in the headspace is obtained by multiplying headspace volume and atmospheric pressure. Then the volume of produced gas can be calculated by subtracting the headspace volume.

The calculation was done for each pressure measurement and the values were summarized to obtain the accumulated gas. The volume of the gas was converted to standard pressure by applying equation (Wang, 2016) .

𝑉𝑆𝑇𝑃 =𝑝𝑔𝑎𝑠

𝑝𝑆𝑇𝑃×

𝑇𝑆𝑇𝑃

(25)

21

Analysis by using Gompertz model

Many studies have been using Gompertz model on processing gas produciton data (Yekta et al., 2017), to obtain statistical parameters such as standard error, lag phase, rate and total methane production. In this study, the data were analyzed in Gompertz model with statistical software

SPSS®, designed by IBM, USA.

The Gompertz model that is inserted (Equation 5):

𝑀 = 𝑃 × exp (− exp (𝑅𝑚×𝑒×(𝐿−𝑡)

𝑃 + 1)) (Equation 5)

where M is total methane production, P is total methane potential, Rm is rate and L is lag phase.

Statistical analysis

The statistical analysis for significant difference between results was conducted by using SPSS®_. The analysis includes paired-samples T test and one-way ANOVA test.

Comparison between pre-treatment and post-treatment

In order to compare the effects of pre-treatment and post-treatment, the values below are compared.

⚫ For pre-treatment, the increase in methane yield caused by conducting pre-treatment is calculated: Methane pre-AU = Methane SFVAU - Methane SFVC, representing the increase caused by pre-treatment AU Methane pre-AL = Methane SFVAL - Methane SFVC, representing the increase caused by pre-treatment AL Methane pre-AUAL = Methane SFVAUAL - Methane SFVC, representing the increase caused by pre-treatment AUAL

⚫ For post-treatment, the increase in methane yield caused by conducting pre-treatment is calculated: Methane post-AU = Methane DAU – Methane DC, representing the increase caused by post-treatment AU Methane post-AL = Methane DAL – Methane DC, representing the increase caused by post-treatment AL Methane post-AUAL = Methane DAUAL – Methane DC, representing the increase caused by post-treatment AUAL

3.3 Reactor experiment

Two lab-scale continuous stirred-tank reactors (CSTR) with 5 L working volume were employed in the experiments.

(26)

22 3.3.1 Preparation and substrate

Before starting the experiments, the two reactors R1 and R2 were gradually filled with the same digestate from the reactors of an ongoing study to 5L (D substrate, 3.1.1). R1 is the control reactor and is fed with the substrate D without treatment. R2 is fed with the treated substrate (AL treatment, 4.1.2).

3.3.2 Reactor experiment

The reactors were incubated at 37℃, HRT 20 days, the OLR was changed depending on the digestate obtained (Table 3). The reactors were continuously stirred (100 rpm). The experiment was run for 44 days, every day 250 ml of the reactor sludge was exchanged.

Table 3. OLR of substrate for reactors

time OLR g VS/L day 1- day 19 0.558 day 20- day 30 0.455 day 31- day 44 0.592 3.3.3 Measured parameters Gas production

The measurement was performed with a gas meter that was connected to the reactor. The gas production was recorded every day when feeding the reactor.

Methane content

The methane content was measured with BIOGAS 5000 (Scantec Nordic®_{) once a week. The}

biogas was stored in a balloon and the stored gas was used to measure methane content. pH

The pH measurement of fresh digestate from R1 and R2 was carried out 1-2 times a week. The measurement process is the same as 3.2.4.

VFA

VFA analysis on fresh digestate was performed once a week. The VFA concentration was measured in duplicates following 3.2.4.

(27)

23

TS and VS

TS and VS analysis of fresh digestate was also conducted once a week, following the method 3.1.2.

3.3.4 Data analysis

All the data were summarized in Excel. The methane yield of the reactors was calculated using

OLR, gas production and methane content. The graphs were made by Origin 2017®_program,

(28)

24

4. Results and discussion

Figure 10. (A)& (B) (see appendix, p50) were made from the original data that measured in the experiments. The graphs are very fluctuating and hard to follow. In order to have clearer presentations and better interpretations (parameters, e.g. lag phase), the data were later fitted in the Gompertz model, therefore, Figure 2. (A)& (B), as well as Figure 5. (A)& (B), Figure 6. (A)& (B) and Figure 7. (A)& (B) are made.

4.1 Biogas production and methane yield

In Figure 2. (A), under the same treatment, substrate SFV has the biggest methane yield, followed by SF, substrate D generates the least amount of methane. AL treatment and AUAL treatment showed the most effectiveness on increasing methane production for all the substrates - SF, SFV and D, whereas AU treatments were less effective. The highest methane production was verified by the treatment SFVAL with a production of 378.22 ± 21 N ml CH4 g-1 VS,

followed by SFVAUAL with a production of 374.46 ± 26 N ml CH4 g-1 VS. The high methane

yield obtained from SFV substrate is probably due to the high soluble sugar content in the vinasse (Janke et al., 2014).

In Figure 2. (B), the biogas production of these three substrates follows the order SFV > SF> D (under the same treatment condition). AL and AUAL treatment also showed the biggest improvements on biogas production, followed by AU treatment. The highest biogas production was observed at the treatment SFVAUAL of 804.05 ± 9 ml g-1_{VS, followed by SFVAL of}

792.56 ± 24 ml g-1_{VS. Compared with Figure 1. (A), the biogas production rate was much}

larger than that only of methane, most of biogas has been produced within 15-20 days. In Figure 3, under AL and AUAL treatments, all three substrates showed similar results in methane concentration. However, for control and AU treatment, under the same treatment condition, the methane concentration follows the order of SFV > SF > D. The low concentrations that were measure in the 27th day can be caused by the low production of substrates during this period or the instrument error of the GC instrument.

(29)

25 A

B

Figure 2. Results from the BMP experiments for all the substrates (after adjusted by gompertz model). (A) shows the accumulated methane production under normal conditions (273. K, 1 atm), (B) shows the accumulated biogas. Methane and biogas produced by inoculum has been subtracted.

(30)

26

Figure3. The methane concentration in percentage, methane concentrations are only based on the substrates, not the inoculum.

Figure 4 shows the methane production of inoculum and the predicted value is obtained by fitting in Gompertz model. The methane production of inoculum shows a nearly linear correlation with time. It means that the inoculum is still active by the time is was used in the BMP test, even though there has been 5 days of degassing. There are studies attesting that sludge continues to produce biogas for long periods without feeding (Gioelli, Dinuccio and Balsari, 2011). However, the inoculum showed a low and steady result regarding the methane production and the R2_{results in 99.2% in the Gompertz model, which proved that this inoculum} is advantageous to be used in the BMP tests.

Figure 4. Results from the BMP experiments for inoculum. (A) shows the original data, (B) shows the data after being fitted in Gompertz model

(31)

27

Table 4 shows the overview of the methane yield, accumulated biogas, methane concentration, as well as the lag phase and R-squared obtained from being fitted in the Gompertz model for all the substrates investigated.

The methane yield of SFC resulted in 223.25 ± 10 N ml CH4 g-1 VS a similar value to the BMP tests carried out by Janke et al. (2017) with 231.1 ± 1 N ml CH4 g-1 VS for sugarcane straw, 231.3 ± 11 N ml CH4 g-1 VS, and Janke et al. (2015) with 228.3 N ml CH4 g-1 VS for straw. It is lower than the methane yield of filter cake of 245-281 N ml CH4 g-1 VS and 260 N ml CH4 g-1_{VS (Janke et al. (2015) and Janke et al., 2017b).Co-digestion of sugarcane straw and filter} cake did not result in a higher methane production than no co-digestion by the end of the

experiments. However, the biogas production of SFC resulted in 512.12 ± 15 ml g-1_{VS, which}

is higher than both of the results obtained from straw and filter cake in the study carried out by Janke et al. (2014), with 394 N ml g-1_{VS and 486 N ml g}-1_{VS biogas yield, respectively. It} should be mentioned that the values in Janke et al. (2014) are based on normal condition (273. K, 1 atm), but they do not differ significantly from the values that are not based on normal condition, so do the data in this project.

The methane yield of SFVC obtained in this study is 275.28 ± 11 N ml CH4 g-1 VS, which is

slightly higher than the methane yield of 267.40 ± 5 N ml CH4 g-1 VS in Moraes et al. (2015); similar as the methane yield result of vinasse in Janke et al. (2015) between 246 - 302 N ml CH4 g-1 VS. It is higher than the methane yield of sugarcane straw and filter cake in Janke et al. (2015), which is probably caused by the higher soluble sugar content in vinasse (Janke et al., 2014). The biogas yield of SFVC of 666.96 ± 21 ml g-1_{VS is slightly higher than vinasse biogas} yield in Janke et al. (2014) with 647 N ml g-1_{VS. The result of SFV is close to vinasse, which} indicates that the straw and filter cake produced a limited amount of methane, therefore, treatments must be conducted on these lignocellulosic biomasses to obtain a promising methane production.

The result of statistical analysis showed that co-digestion of straw, filter cake and vinasse (275.28 ± 11 N ml CH4 g-1 VS) had a significantly higher methane production than co-digestion of straw and filter cake (223.25 ± 10 N ml CH4 g-1 VS) (p<0.01, T test). One reason could be that vinasse has a higher methane yield than the other two materials (Janke et al., 2014, 2015), therefore, the high content of vinasse in SFV increases the total methane yield of the co-digestion. Other reasons could be the mixture of vinasse, straw and filter cake balances the nutrients in the reactor and increase the buffering capacity of the mix substrates or a wider diversity of microorganisms (Giordano, 2012; Moraes et al., 2015; Martinez-jimenez et al., 2017).

High values of R2_{.can be overserved in the results (Table. 4). For D substrate, DC, DAU and}

DAL have lag phases between 3-4 days; lag phase was also shown in SFC and SFAU around 3 days. Substrate SFV do not have lag phase; AUAL treatment did not result in lag phase either. This is beneficial for working in a short HRT of 20 days.

(32)

28

Table 5 shows the increase on methane yield on the 18th_{, 21}st_{, 27}th_{, 42}nd_{(final) day, compared} with control substrate. The highest improvements on methane production were achieved on the 21st_{day. The increases caused by treatments are reducing due time. The data from table 5 are} later used in this chapter.

Table 4. Overview of the methane yield, accumulated biogas, methane concentration, lag phase and R-squared for the substrates investigated.

Substrate Methane yield (N ml CH4 g-1 VS) Accumulated biogas (ml /g VS) Methane concentration (% v/v) Lag phase (days) R2 (%) SFC 223.25 ± 10 512.12 ± 15 43 3.5 99.6 SFAU 205.71 ± 10 540.65 ± 32 38 2.6 99.3 SFAL 329.61 ± 29 692.70 ± 55 43 0.0 97.8 SFAUAL 332.96 ± 8 696.53 ± 5 44 0.0 96.1 SFVC 275.28 ± 11 666.96 ± 21 40 0.0 99.5 SFVAU 299.44 ± 25 710.45 ± 33 40 0.0 99.3 SFVAL 378.22 ± 21 792.56 ± 24 45 0.0 99.1 SFVAUAL 374.46 ± 26 804.05 ± 9 43 0.0 97.3 DC 144.69 ± 2 420.96 ± 13 34 4.3 99.7 DAU 193.58 ± 18 547.61 ± 68 35 3.1 99.6 DAL 289.06 ± 7 558.22 ± 4 50 3.1 99.7 DAUAL 331.00 ± 12 617.20 ± 20 51 0.0 98.2

* methane and biogas produced by inoculum has been subtracted.

Table 5. Increases of methane production compared with control

Substrate 18 days 21 days 27 days 42 days (Final)

SFAU 4.7% 2.9% -4.5% -7.9% SFAL 174.1% 146.9% 99.2% 47.6% SFAUAL 163.9% 142.7% 98.3% 49.1% SFVAU 7.6% 9.1% 10.8% 8.8% SFVAL 69.7% 62.0% 54.9% 37.4% SFVAUAL 70.2% 61.2% 53.9% 36.0% DAU 45.6% 39.2% 39.3% 33.8% DAL 234.6% 195.0% 165.5% 99.8% DAUAL 275.4% 232.4% 199.1% 128.8%

(33)

29

4.1.1 Effects of pre-treatment (substrate SF and SFV)

BMP results of substrate SF and SFV are investigations on methane yield of co-digestion (SFC and SFVC), as well as the effects of pre-treatments on these two substrates.

In Figure 5. (A), AL and AUAL treatments have almost the same effects on methane production (statistically no significant difference, p>0.05, ANOVA test), the curve of AU treatment and control are also similar. Therefore, AL treatment has a dominant effect on substrate SF. AU treatment has nearly no effects on increasing methane production of SF and cannot help with improving the AL treatment in the combination of AUAL either.

In Figure 5. (B), Al and AUAL treatments showed nearly the same effects on increasing biogas production (statistically no significant difference, p>0.05, ANOVA test), which is similar as that shown in methane production. However, AU treatments shows a slightly higher biogas production than control.

Figure 6. (A) and (B) showed the similar curves on both methane production and biogas production. AL treatment has a clear effect on increasing methane and biogas production. As mentioned above, AU treatment on SFV is regarded as a combination of acid and autoclaving treatment due to the low pH of vinasse. Many studies have proven that acidic treatments can degrade hemicellulose and cellulose, which makes the degradable part in the biomass more accessible to hydrolysis (Sun and Cheng, 2002; Kim et al., 2018). As shown in the picture, AU treatment slightly increases the rate of methane and biogas production in the first 15 days and increases the methane and production during the second half of the experiment, which can be caused by the acidic environment provided by vinasse in AU treatment. However, there is no significant difference between the BMP results of SFVAU and SFVC samples, statistically (p>0.05, ANOVA test).

A positive effect of alkaline treatment on methane concentration can be seen both in Figure 5. (C) and Figure 6. (C).

Alkaline treatment using 6% (w/w) NaOH achieved 47.6% and 37.4% increase of methane yield on co-digestion substrates SF and SFV, respectively. In comparison with the study carried out by Janke et al. (2017b), resulted in an increase of 5.8% in methane yield by using 6% (w/w) NaOH to pre-treat sugarcane straw, the increases obtained in this study are much higher. However, it is not strong enough to say that NaOH treatment has better effects on co-digestion substrates than on one substrate, further studies need to be conducted.

Besides, in this study, the AL treatment is obviously playing a dominant role in affecting the methane yield. AU treatment only slightly increases the methane yield regarding SFV. But there is no evidence to say that AUAL combination pretreatment improves the AL treatment (p>0.05, no significant difference, ANOVA test). Therefore, AL treatment can be considered as a better pre-treatment method for SF and SFV than AUAL treatment.

(34)

30 A

B C

Figure 5. Results from the BMP experiments for substrate SF (after adjusted by gompertz model). (A) shows the accumulated methane production under standard conditions (273. K, 1 atm), (B) shows the accumulated biogas and (C) shows the methane concentration in percentage, methane concentrations are only based on the substrates, not the inoculum. In (A) and (B), methane and biogas produced by inoculum has been subtracted.

(35)

31 A

B C

Figure 6. Results from the BMP experiments for substrate SFV (after adjusted by gompertz model). (A) shows the accumulated methane production under standard conditions (273. K, 1 atm), (B) shows the accumulated biogas and (C) shows the methane concentration in percentage, methane concentrations are only based on the substrates, not the inoculum. In (A) and (B), methane and biogas produced by inoculum has been subtracted.

(36)

32 4.1.2 Effects of post- treatment (substrate D)

In Figure 7. (A), three treatments showed distinctive improvements on methane production of substrate D, the largest increase was caused by AUAL treatment, followed by AL and AU with a relatively smaller increase. Different from the results of AU treatment on substrate SF and SFV, it showed improvements on both production and rate on substrate D. The reason of AU treatment is more effective on D can be that D is digestate from SFV co-digestion, the structure of D has been broken down into very small pieces, while straw and filter cake in SF and SFV are more complete in terms of the structure. Therefore, water in D has much larger contact areas with biomass cells than it in SF and SFV. According to Mailin et al. (2014), water can penetrate into the biomass cell structure, hydrating cellulose, solubilizing hemicellulose, and slightly removing lignin. In the autoclaving process, the water fraction penetrates the biomass cells and results in more solubilized products. Therefore, AU treatment is more effective for D than SF and SFV.

In Figure 7. (B), AUAL treatment showed the largest improvement on biogas production, as well as a highest rate of producing biogas. Al and AU treatments resulted in similar biogas production (p>0.05, ANOVA test), however, AL treatment has a much higher rate than AU treatment, most of biogas has produced in the first 15 days, whereas biogas was gradually produced through the whole experiment for AU treatment and showed a flatter trend by the end (42 days).

Substrate D is the out-flow digestate from anaerobic digestion of SFV, which results in methane yield of 144.69 ± 2 N ml CH4 g-1 VS without any treatment (DC). In comparison with the BMP result

of raw material SFV (SFVC) of 275.28 ± 11 N ml CH4 g-1 VS, DC achieved 52.6% of this value.

However, the result of DAL (289.06 ± 7 N ml CH4 g-1 VS) reaches 76.4% of the result of SFVAL

(378.22 ± 21); the result of DAUAL (331.00 ± 12 N ml CH4 g-1 VS) reaches even 87.5% and 88.4%

of the result of SFVAU and SFVAUAL (374.46 ± 26), respectively. Therefore, the digestate still has

a big potential of methane production, especially when the AL and AUAL treatment are applied on it. Despite AUAL treatment lead to a highest result in increasing methane yield, statistically, there is no significant difference between AUAL and AL in both methane yield and biogas yield (p>0.05, ANOVA test).

(37)

33 A