Prediction of hydrothermal carbonization with respect to the

Subscriber access provided by Nottingham Trent University

Biofuels and Biomass

Prediction of hydrothermal carbonization with respect to the biomass components and severity factor Mohammad Heidari, Omid Norouzi, Shakirudeen A. Salaudeen, Bishnu Acharya, and Animesh Dutta Energy Fuels, Just Accepted Manuscript • DOI: 10.1021/acs.energyfuels.9b02291 • Publication Date (Web): 30 Aug 2019 Downloaded from pubs.acs.org on August 30, 2019

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Prediction of hydrothermal carbonization with respect to the biomass components and severity factor Mohammad Heidaria, Omid Norouzia, Shakirudeen Salaudeena, Bishnu Acharyab, Animesh Dutta*,a a Mechanical

Engineering Program, School of Engineering, University of Guelph, Guelph,

Ontario N1G 2W1, Canada b

School of Sustainable Design Engineering, University of Prince Edward Island, 550

University Avenue, Charlottetown, PEI C1A 4P3, Canada KEYWORDS: HTC, Biomass components, Hydrochar prediction, Optimum condition

ABSTRACT: In recent years, hydrothermal carbonization (HTC) has been introduced as an attractive method for converting biomass into value-added products. The complex reaction chemistry and variable composition of biomass have, however, been barriers to find general equations for describing the HTC process. The goal of this study is to establish a link between the composition of biomass and the expected hydrochar from HTC. Based on the experimental design found from response surface methodology, the biomass components, namely pure cellulose, hemicellulose, and lignin, were submitted in different combinations into 39 HTC experiments with severity factors (SF) of 3.83, 5.01, and 6.19. Using the experimental data, an attempt was then made to predict the mass yield (MY), higher heating value (HHV), carbon content (C%), and energy recovery factor (ERF) of the hydrochars according to the biomass

*Corresponding author: Animesh Dutta, e-mail: [email protected], Tel: +1 519-824-4120 ext:52441 ACS Paragon Plus Environment

Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

composition and the process severity. The results revealed that the interactions between the biomass components do not have a major effect on the hydrochar characteristics whereas the interaction between cellulose and SF is the most significant. Moreover, it was shown that after the lignin content, the hemicellulose content has the highest positive impact on HHV and C% of the hydrochar. An optimization study showed that with a focus on minimizing the SF, while the HHV is maximized, biomass with a cellulose content of 40%, hemicellulose of 35%, and lignin of 25%, under the severity of 4.41 will be the most suitable case for HTC treatment. Finally, a comparison between the predictions and the experimental data in the literature suggests that the proposed equations can provide a good evaluation on the HTC of several biomass feedstocks especially when the amount of ash and extractives are low.



INTRODUCTION

It is widely accepted that the production of energy from fossil fuels contributes to harmful emissions. Hence, renewable sources should replace fossil fuels to generate environmentally friendly energy [1]. Lignocellulosic biomass such as corn cobs, rice husk, and grasses, are huge renewable sources that possess considerable carbon content. However, treatment is required to convert them from waste to energy. Hydrothermal carbonization (HTC) is a promising method to convert biomass to bioenergy. The main advantage of HTC over other thermochemical processes is that the high moisture content of biomass is in favor of the process [2] [3]. In this process, the raw biomass is mixed with water and introduced to a reactor. The reactor content is then pressurized and heated up to a certain temperature (180-350 oC) and is kept at that

2 ACS Paragon Plus Environment

Page 2 of 31

Page 3 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

temperature for a certain time (5-240 min) [4]. After the process, the main product called hydrochar is separated in a solid form from the liquid by-product (process water). Some of the ash content of the initial biomass leaves the biomass and dissolves in the process water which is another advantage of HTC [5]. The ash content is particularly important when the biomass is used as an energy resource. Bartochi et al. [6] compared the performance of other thermochemical processes in terms of ash removal. Their research indicated that pyrolysis at high temperatures (600 oC) releases a negligible amount of the initial ash, however, gasification and combustion are able to release 45%, and 55% of ash respectively. The presence of the water in the HTC process facilitates the dimerization of the inorganic compounds from the biomass to the process water. Hence HTC can achieve significant ash reduction (30-60%) of the biomass, even in low temperatures (around 200 oC) [7]. The advantages of HTC have led to a noticeable interest in the process among the researchers. To date, the effects of the process parameters (temperature [8], time [9], pressure [10], water to biomass ratio [11], catalysts [12], recycling of the process water [13], and particle size [14]) on the products as well as the product’s suitability for several applications (power generation [15], soil amendment [16], and separation [17]) have been assessed. However, all of these studies have been performed using lab-scale reactors and the major steps towards commercialization of HTC have not been taken yet [18]. It is known that the temperature of the process has the most significant role among process parameters in the carbonization of biomass [19]. The pressure of the reactor does not have significant effects after ensuring it is above the saturation pressure of water to prevent the 3 ACS Paragon Plus Environment

Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

formation of steam [10]. The residence time of the process can increase the intensity of the hydrothermal reactions. To indicate the impacts of reaction time and temperature together, severity factor (SF) has been introduced [20]. It is worthy to note that in an industrial scale HTC where continuous production is of interest, short residence times will probably be more favorable. The delay in commercialization of HTC is mainly due to unknown reaction orders, interactions, and intensities. In addition, the kinetics of the reactions are feed dependent and this hinders the development of a reliable model that can predict the amount and quality of the hydrochar after performing HTC at given conditions. In other words, the results found in each study are specific to the type of feedstock used in the study. Hence, there is no generic model to predict the behavior of different feedstock under HTC conditions. Biomass is generally categorized into different generations. The first generation usually sourced from crop plants and using them for non-food usages can threaten food security. The second and third generations of the biomass are lignocellulosic and algal types respectively and are non-food materials [21]. It is known that lignocellulosic biomass is mainly composed of hemicellulose (20-40%), cellulose (40-60%), and lignin (10-25%) [22]. Since the percentages of each of the main components vary in different biomass and as each component has its own specific characteristics, the behavior of each feedstock under HTC is specific. Cellulose is made of glucose subunits that are linked with via β-1,4-glycosidic bond while Hemicellulose is a branched heteropolysaccharide consisting of C5 sugars (usually xylose and arabinose) [23]. 4 ACS Paragon Plus Environment

Page 4 of 31

Page 5 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Cellulose has high crystallinity and consists of repeating units in linear chains with higher polymerization degree, while the structure of lignin and hemicellulose is amorphous [24]. Hemicellulose contains monomer units such as xylose in a branched, irregular and with a lower degree of polymerization. The structure of lignin is branched and has a threedimensional polymer shape [24], [25]. Hemicellulose starts degradation at as low temperatures as 180 oC, while the main degradations of cellulose and lignin starts at around 220 oC and 260 oC respectively. Regarding the carbon content (C%), lignin has the highest percentage (around 61%), then cellulose (around 42%) and then hemicellulose (around 40%) [26]. Other than these three main components, each biomass may also have small amounts of extractives and inorganic content (ash). The ash and extractive contents in lignocellulosic biomass are often less than 5%, however, in some cases, higher values have been reported [27]. The extractives such as resins, fats, waxes, fatty acids, alcohols, tannins, and flavonoids are organic substances which have low molecular weight and are soluble in neutral solvents [28]. They are highly reactive at low temperatures while the inorganic materials (ash) such as CaCO3, NaNO3, and SiO2 are chemically stable and require high temperatures to be fully removed [29]. As it was explained, of all the parameters that can affect the quality and quantity of the products, the severity factor (a combination of the temperature and time) and the type of the feedstock have the major roles. Therefore, studying the HTC of the main biomass components can help to find generic equations and rules for the prediction of the products based on their fiber composition and the severity of the process. 5 ACS Paragon Plus Environment

Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The studies on the main components of biomass have been mostly conducted for other thermochemical processes [30]–[32]. Regarding the HTC, Kim et al. [33]investigated the HTC of the biomass components at a temperature range of 150-280 °C and a residence time of 30 min and assessed their suitability to convert to biofuel. Borrero-López et al. [34] subjected the biomass components to HTC with a wide range of severity factor and assessed the mass yield (MY) of the hydrochar, and the pH, yield, and composition of the process water. Kang et al. [35] presented a similar study with a long residence time (20 h) resulting in proposed formation pathways of each of the components of the biomass. Although the aforementioned research discloses an insight of the hydrothermal reactions taking place for different components of the biomass, the interactions of the components and an effort to predict the hydrochar of real biomass under HTC have not been considered. Hence, finding acceptable predictions of the MY, HHV and C% with respect to the biomass composition and the SF of the HTC are the main goals of the current paper. The approach to achieve the goal of this research consists of first performing HTC on the synthesized biomass samples consisting one, two, and three biomass components at different severities and then finding the MY, HHV and C% of the obtained hydrochars. The next step is finding a correlation based on the obtained information using response surface methodology in Design Expert Version 11. The possible interactions by mixing the biomass components are also considered. Finally, the results are compared with hydrochars of real biomass and the possible errors are explained.


Page 6 of 31

Page 7 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

It should be noted that the model proposed in this paper, can be considered only for the second generation of biomass because instead of cellulose, hemicellulose, and lignin, the third generation of biomass consists of lipid, protein, and carbohydrates [36], [37]. In addition to this, the presence of ash and water extractives are not considered in this model. This is mainly due to their low content in most of the lignocellulosic biomass, their variable composition and to prevent the complexity of the model.



MATERIAL AND METHODS

Feed Pure hemicellulose (H2125), cellulose (435236), and alkali lignin (370959) were purchased from Sigma-Aldrich Canada Co. as the model components of the biomass. These materials were then mixed or used solely in several HTC experiments (Figure 1-a). The characteristics of the purchased materials are given in Table 1. The ultimate, proximate, and HHV analysis that are explained in the coming sections were utilized to find the data in table 1. Table 1. Ultimate, proximate, and HHV analysis of the biomass main components

Component

C

H

N

S

O

Ash

Volatile matter

Fixed carbon

(wt.%db)

HHV (MJ/kg)

Cellulose

42.1

5.8

0.3

0.1

51.7

0

93.6

6.4

15.80

Hemicellulose

41.9

5.2

0.3

0

51.1

1.5

76.7

21.8

15.02

Lignin

62.3

4.7

0.5

2

28.1

2.4

47.2

50.4

25.60

Hydrothermal carbonization experiments


Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 31

The HTC experiments were conducted using a Parr one-liter series 4571 high-pressure reactor equipped with a precise temperature controller and a cooling loop (Figure 1-c).

a

b

c

Figure 1. Experimental setup: a) the pure and synthesized feedstock, b) The three small reactors placed in the main reactor, c) The main reactor equipped with temperature controller and cooling loop

Considering the high number of the required experiments, and to save time without jeopardizing the consistency and accuracy of the results, three small vessels were constructed (Figure 1-b). The size of the vessels was considered based on the size of the main reactor, such that the three vessels would fit in the main reactor. The information about the construction and the size of these vessels can be found in another paper [14]. Each of the vessels in each run contains 2 g of pure or mixed sample. Hence, the weight of each component can vary in a range of 0-2 g in the vessels. 10 ml of water was added and stirred in each vessel (1:5 biomass to water ratio). The temperatures of the HTC experiments were set as 180, 220 and 250 °C and the residence time was 30 min for all experiments. Three replicates were performed for each experimental data point. The number of experiments and the percentage of each component is explained in Section 2.2. As the real biomass starts degradation from around 180 °C, and this paper tries to model the HTC of real biomass by mixing the biomass components, this


Page 9 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

temperature was selected as the minimum temperature of the experiments. Moreover, 250 °C was selected as the maximum temperature due to the reactor limitations. In addition, as residence time has a considerably lower impact than temperature, and its effects can be addressed by the severity factor, a relatively low time was considered (30 min). Based on the selected time and temperature of each experiment, severity factor (SF) is defined by equations (1) and (2) [20], [38]: [(T ― 100) 14.75]

R0 = t.e

(1)

SF = log R0

(2)

where R0 is the reaction ordinate (min), t is reaction time (min) and T is the reaction temperature (°C). To indicate the impacts of reaction time and temperature together, reaction severity (SF) given by equations (3) and (4) is defined based on Arrhenius equation [38]: [(T ― 100) 14.75]

R0 = t.e

(3) (4)

SF = log R0

where R0 is the reaction ordinate (min), t is reaction time (min) and T is the reaction temperature (°C). The severity factor then can be found as the logarithm of the reaction ordinate (log R0). Severity factor is widely used by the researchers for describing the hydrothermal processes. This equation assumes a first-order reaction with an Arrhenius temperature dependence. Abatzoglou et al. [39] proposed this equation using a reference temperature of 100 °C, however, 9 ACS Paragon Plus Environment

Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 31

the concept of severity was originally introduced by Genlesse and Reuter [40] to address the effects of time and temperature for cracking of the oils. They indicated that a temperature rise of 17 °C halves the required time for getting the same results. SI = 2t.[

(T ― 480) 17]

(5)

Where SI is the time-temperature index, T is the temperature and t is the time. Similarly, in hydrothermal carbonization, as shown in Equation (3), 14.75 °C is the temperature increase which has an exponential relation with reaction order to produce the same changes in the hydrochar [38]. Characterizations of the hydrochars MY was calculated as the ratio of the weight of obtained hydrochar to the initial biomass weight. The proximate analysis for determining ash, volatile matter (VM), and fixed carbon (FC) content was performed on the raw material as per ASTM standards [41]. The ultimate analysis was conducted using a Flash 2000 Elemental Analyzer (Thermo Fisher Scientific, Waltham, MA, USA) on the hydrochar and raw material. HHV was measured via IKA-C200 bomb calorimeter (IKA Works, Wilmington, NC, USA). The energy recovery factor (ERF) was then calculated from equation (6):

ERF = MY ×

HHVhydrochar

(6)

HHVfeedstock

Where HHVhydrochar and HHVfeedstock are the HHV of the hydrochar and feedstock respectively in MJ/kg and are obtained from bomb calorimetry experiment.


Page 11 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Although ERF is defined by an equation from the mass yield and HHVs of the biomass and hydrochar, it is often used as an important measure to assess the feasibility of HTC processes. Usually, with an increase in SF, MY and ERF decrease, but HHV increases. However, some papers have reported that due to the type of the feedstock, or the process condition, by increasing SF of HTC, MY, and HHV have increased and resulted in an increase in ERF [2]. Response surface methodology Response surface methodology (RSM) explores the relationships between several explanatory variables and one or more response variables. The main idea of RSM is to use a set of designed experiments to obtain an optimal response [42]. In the current study, RSM was used to find the relations between the variables reported in Table 2, with the responses including MY, HHV, C%, and ERF (Table 3). Build information In this study, a flexible design structure (combined, and user-defined) was performed to accommodate a custom model for the HTC process. Cellulose, hemicellulose, lignin, temperature and time have been considered as variable factors. These factors are representative of lignocellulosic biomass and process condition. For simplicity, SF was replaced by both temperature and time. Table 2 presents the design factors and their defined ranges. Table 2. Range of the variables for the experimental design

Mixture Components and process Factor Component/Factor Name Units Type Minimum Maximum A Cellulose % Mixture 0 100 B Hemicellulose % Mixture 0 100


Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 Run 34 1 35 2 36 37 3 38 4 39 5 40 6 41 7 42 8 43 9 44 10 45 46 11 47 12 48 13 49 14 50 15 51 16 52 17 53 18 54 19 55 20 56 57 58 59 60

C D

Lignin Severity Factor

% -

Mixture Numeric

Page 12 of 31

0 3.83

100 6.19

HHV, MY, C%, and ERF were used as response factors and are given in Table 3. Table 3. Response factors and respective units

Response Name Units R1 HHV MJ/kg R2 MY % R3 C% % R4 ERF %

According to the variables of the study, a total of 39 runs were suggested by Design-expert 11 (Stat-Ease, Inc., Minneapolis, USA) in randomized order and the output responses were obtained via HTC experiments. Table 4 shows the I-optimal design matrix with 39 experimental runs. The standard deviation for HHV, MY, C%, and ERF were 0.48, 1.40, 2.34, and 1.18, respectively. Table 4. I-optimal design layout and experimental results Component 1 A: Cellulose % 16.67 16.67 41.67 50.00 0.00 16.67 33.33 0.00 100.00 0.00 41.67 50.00 16.67 50.00 16.67 0.00 0.00 16.67 41.67 0.00

Component 2 B: Hemicellulose % 16.67 66.67 41.67 50.00 0.00 41.67 33.33 50.00 0.00 50.00 41.67 50.00 41.67 50.00 66.67 50.00 0.00 41.67 41.67 100.00

Component 3 C: Lignin % 66.67 16.67 16.67 0.00 100.00 41.67 33.33 50.00 0.00 50.00 16.67 0.00 41.67 0.00 16.67 50.00 100.00 41.67 16.67 0.00

Factor 4 D:SF 6.19 3.83 6.19 3.83 6.19 6.19 6.19 3.83 3.83 6.19 5.01 5.01 3.83 6.19 5.01 5.01 5.01 5.01 3.83 6.19

Response 1 HHV MJ/Kg 28.90 24.15 27.71 20.47 30.41 28.33 28.30 25.82 18.26 28.70 27.15 25.76 23.55 26.06 27.09 27.75 28.70 27.37 22.77 27.57


Response 2 MY % 64.50 65.94 49.73 71.41 73.50 54.91 54.72 72.30 80.45 57.34 55.74 55.22 75.11 47.32 53.97 64.10 81.92 63.58 69.29 40.30

Response 3 C% % 66.70 69.21 67.38 50.90 71.31 66.43 68.11 62.21 43.96 65.42 64.72 60.84 57.88 66.10 64.44 63.13 63.90 62.94 54.98 65.89

Response 4 ERF % 84.28 93.82 81.29 94.25 86.99 79.68 82.96 91.26 93.02 80.98 88.81 92.53 89.64 80.25 86.52 87.36 91.13 88.97 91.88 73.95

Page 13 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39

Energy & Fuels

16.67 0.00 66.67 50.00 0.00 33.33 0.00 16.67 66.67 100.00 66.67 41.67 41.67 50.00 16.67 33.33 50.00 100.00 41.67

66.67 100.00 16.67 0.00 0.00 33.33 100.00 16.67 16.67 0.00 16.67 16.67 16.67 0.00 16.67 33.33 0.00 0.00 16.67

16.67 0.00 16.67 50.00 100.00 33.33 0.00 66.67 16.67 0.00 16.67 41.67 41.67 50.00 66.67 33.33 50.00 0.00 41.67

6.19 3.83 3.83 5.01 3.83 3.83 5.01 5.01 5.01 5.01 6.19 5.01 6.19 3.83 3.83 5.01 6.19 6.19 3.83

standard deviation



27.87 23.37 20.66 27.37 27.40 23.11 26.77 27.62 26.39 25.86 27.18 25.78 27.51 22.75 25.18 27.52 28.34 26.26 22.33 0.48

47.25 60.68 78.33 69.93 88.52 76.11 47.76 73.29 61.51 58.40 52.31 68.26 60.09 84.31 82.38 61.51 61.80 50.10 83.41 1.40

66.12 60.09 49.22 63.11 63.10 55.99 63.29 61.88 60.41 61.96 67.71 60.12 65.14 53.82 59.61 64.82 68.38 68.56 52.33 2.34

RESULTS AND DISCUSSION

Experimental results The results of the 39 HTC experiments and characterization of the obtained hydrochars are given in Table 4. These results were then entered into the Design Expert software as inputs. As expected, the highest HHV (30.41 MJ/kg) and C% (71.31%) occurs when the sample is solely made of lignin and the severity factor is at its highest (6.19). Whereas, the lowest HHV (18.26 MJ/kg) occurs at the lowest SF and when the sample is only made of cellulose. This is mainly due to the fact that at low SFs, the hemicellulose undergoes significant carbonization and increases the HHV, but cellulose starts the main carbonization at medium SFs. Other than these easily observable results from Table 4, the relations and interactions of the parameters can be of interest which is described in the following sections. Data Analysis and statistical significance of each model 13 ACS Paragon Plus Environment

78.02 93.62 93.19 92.80 93.68 92.95 84.98 90.95 94.36 96.42 82.79 89.10 83.97 92.10 92.64 90.54 85.04 84.02 93.57 1.18

Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

An adequate functional relationship between the four output responses of the HTC and input variables is developed using response surface study. This relationship between the design factors and output responses can be modeled using a quadratic polynomial equation due to its higher accordance with the experimental results of the current study in comparison to other equations. Analysis of variance (ANOVA) was used to determine the statistical significance of each model and is given in the supporting information. Considering the level of significance as 5%, The Model F-value implies that all models are significant. The predicted R2 in the models is in reasonable agreement with the Adjusted R2 of 0.99 (i.e. the difference is less than 0.2). In the below sections, the effects and the interactions of the design parameters on HHV, C%, ERF, and MY are investigated. Higher Heating Value (HHV) Response surface methodology was employed to analyze the effects of each parameter on HHV. According to the methodology described in Section 3.1, the mathematical relationship between the dependent and independent variables, and the value of the regression coefficient of the respective equation were calculated. The final equation in terms of actual factors is as follows: HHV -54.74158 -3.83619 +26.63995 -0.594304 -0.688557 +28.77532 +0.413725 +10.68084 -0.493077

= C H L C×H C×L C × SF H×L H × SF L × SF

Impact High Low Medium Low Low Medium Low Medium Low 14 ACS Paragon Plus Environment

Page 14 of 31

Page 15 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

-2.53471 -0.909431 +0.175012

(7)

C × SF ² Low H × SF ² Low L × SF ² Low

This equation demonstrates the variation of HHV by changing the design factors. The equation for HHV in terms of the design factors and Figure 2-a illustrates that cellulose exhibits the highest negative effect on HHV with a coefficient of -54.74 while lignin has a linear positive effect with a coefficient of 26.64. This is due to the fact that the C% and HHV of lignin are already high without carbonization. Moreover, the results reveal that the effect of hemicellulose on HHV is negligible. Considering hemicellulose and cellulose, HHV and C% of hemicellulose is slightly lower than that of the cellulose, however, hemicellulose starts to carbonize from lower severity factors which finally will have a better contribution in increasing the HHV and C% of the final hydrochar. The proposed equation shows that the interaction between cellulose and severity factor is the most significant interaction whereas it does not detect considerable interactions among components. Among the biomass components, only interaction of lignin and hemicellulose increases the HHV. 15 ACS Paragon Plus Environment

Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

As shown in Figure 2-b, the effect of a change in cellulose, hemicellulose and lignin values on HHV are noticeable when the severity factor is more than 5.4. This is in accordance with the findings in other experimental studies [14]. Another interesting observation from Figure 2-b is that the effect of the biomass components' percentage on HHV is more significant in lower SFs than the higher ones.


Page 16 of 31

Page 17 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Figure 2. 3D response surface plot of HHV considering the interactions of the biomass components and severity factor

Mass yield (MY) Equation (8) for the MY in terms of design factors indicates that cellulose and hemicellulose with a coefficient of 243.4 and 132.85 respectively have the highest positive effects, whereas the interaction between cellulose and severity with a coefficient of -61 was the highest negative linear effect on MY. The interaction plot (Figure 3) shows that when the severity factor is less than 4.42, the effect of biomass components changes on MY is more dominant as compared to when the severity factor is more than 4.42. MY +243.40517 +132.84854 +96.71417 +1.40643 +1.68739 -61.00096 -4.88037 -26.21369 +0.543887 +4.81376 +1.83544 -0.682525

= C H L C×H C×L C × SF H×L H × SF L × SF C× SF² H × SF² L × SF²

(8)

Impact High High High Low Low Medium Low Medium Low Low Low Low


Energy & Fuels

Design-Expert® Software Component Coding: Actual

Mass Yield (%)

6.188

Factor Coding: Actual Mass Yield (%) Design Points 40.3

88.52

5.599

X1 = B: Hemicellulose

60

X2 = A: Cellulose X3 = D: SF Actual Component

D: SF

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 31

5.01

C: Lignin = 33.3333

4.421

70

80

3.832 B:

0

A: 66.6667

16.6667

33.3333

50

66.6667

50

33.3333

16.6667

0

B: Hemicellulose (%) A: Cellulose (%)

Figure 3. Contour plot of MY considering the interactions of biomass components and severity factor

Carbon content (C%) C% is the main indicator of carbonization level, thus its value directly relates to the hydrochar or solid fuel quality. The final equations for C% in terms of design factors and mixture components indicate that cellulose has the highest negative effect on the carbonization with a coefficient of -91.42 while the most effective factors on the carbonization of biomass were related to the lignin and hemicellulose (with a coefficient of 84.86 and 70.18 respectively). On the other hand, the most significant interaction between the parameters was observed between cellulose and severity factor with a coefficient of +50.33. Both 3D surface and


Page 19 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

counterplot (Figure 4) depicts that when the severity factor is more than 5.01, the effect of changes in mixture components on C% is less obvious as compared to when the severity factor is less than 5.01. C% -91.42253 +70.17892 +84.85954 -0.309030 -1.90956 +50.33209 +0.429702 -3.38168 -10.47161 -3.95784 +0.417360 +1.23801

= C H L C× H C×L C × SF H×L H × SF L × SF C × SF² H × SF² L× SF²

(9)

Impact High High High Low Low Medium Low Low Low Low Low Low

Figure 4. 3D response surface and contour plot of the C% considering the interactions of biomass components and severity factor

Energy recovery factor (ERF)


Energy & Fuels 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 31

ERF is usually an indicator of the feasibility of the proposed system. The final equation for ERF in terms of SF and mixture components is given as equation 10. In this case, cellulose, hemicellulose, lignin and their first order interaction with severity factor have been considered as significant model terms. According to the coefficient values, hemicellulose has a significant positive impact on ERF with a coefficient of 105.99. On the other hand, the only noticeable interaction between the parameters was observed between cellulose and severity factor with a coefficient of 50.07. As seen in Figure 5 in a lower amount of cellulose (

Prediction of hydrothermal carbonization with respect to the

Recommend Documents