Você está na página 1de 103

Design of a Small-scale Double Auger Pyrolysis Reactor

João Lobo Navarro Gonçalves

Thesis to obtain the Master of Science Degree in

Mechanical Engineering

Supervisors: Dr. Ana Isabel Marques Ferreiro


Dr. Raquel Inês Segurado Correia Lopes da Silva

Examination Committee

Chairperson: Prof. José Manuel da Silva Chaves Ribeiro Pereira


Supervisor: Dr. Ana Isabel Marques Ferreiro
Member of the Committee: Prof. Maria Margarida Boavida Pontes Gonçalves

July 2022
Declaração

Declaro que o presente documento é um trabalho original da minha autoria e que cumpre todos os
requisitos do Código de Conduta e Boas Práticas da Universidade de Lisboa

i
Declaration

I declare that this document is an original work of my own authorship and that it fulfills all the require-
ments of the Code of Conduct and Good Practices of the Universidade de Lisboa.

ii
Acknowledgments

Gostaria de agradecer ao Professor Miguel Mendes, à Doutora Raquel Segurado e à Doutora Ana
Isabel Marques Ferreiro pelo conhecimento e orientação que me proporcionaram durante todo este
processo que é o fim do meu percurso no Instituto Superior Técnico. Em particular, gostava que fazer
um agradecimento especial à Isabel pela sua incessante disposição para ajudar e sempre com um
positivismo contagiante.
Gostaria ainda de agradecer aos meus amigos, aos que já conhecia e aos que conheci durante a
minha experiência no IST. Obrigado por me ajudarem a criar momentos incrı́veis que espero nunca me
esquecer.
Em último lugar, devo um agradecimento muito especial a toda a minha famı́lia pelo apoio e confiança
que sempre demonstraram e projetaram em mim. Um particular e sincero obrigado aos meus pais, aos
meus irmãos, à minha avó e à Adriana pela orientação e compreensão durante este percurso.

iii
Resumo

O presente trabalho foca-se no desenvolvimento e dimensionamento de um reator de parafuso, à es-


cala laboratorial, para realizar pirólise rápida de biomassa lenhosa (usando como amostra de referência
madeira de pinho), em que as condições de funcionamento (e.g. temperatura, tempo de residência)
sejam controláveis. Estudos termogravimétricos foram utilizados como dados de entrada numa ferra-
menta de otimização para, em conjunto com métodos experimentais de determinação de composição
orgânica, definir a distribuição orgânica completa da amostra de pinho. Posteriormente, estes dados
foram implementados num modelo cinético de maneira a realizar um estudo paramétrico e determinar
as condições ideais para a produção de bio-óleo. O reator de parafuso foi desenvolvido e dimensionado
seguindo as diretrizes de maximização da quantidade produzida de bio-óleo. Em seguida, os dados
provinientes da geometria final do reactor e da composição orgânica da amostra de biomassa foram
utilizados como dados de entrada no modelo cı́netico com o objetivo de prever os rendimentos dos
subprodutos da pirólise. As principais conclusões foram as seguintes: 1) a distribuição da composição
orgânica da amostra de pinho é: 37.72 %(m/m) celulose, 25.91%(m/m) hemicelulose, 26.73 %(m/m)
lenhina e 9.64 %(m/m) extrativos; 2) da percentagem total de lenhina, a distribuição dos seus compo-
nentes de referência utilizados para o modelo cinético é: 0.22 %(m/m) LIG-C, 19.01 %(m/m) LIG-H e
7.5 %(m/m) LIG-O; 3) o tempo de residência e valor de temperatura ótimos para a produção de bio-óleo
é 65 s e 500 ◦ C, dado que a eficiência de produção de bio-óleo não cresce significativamente acima
destes valores e descresce drasticamente abaixo de 65 s e 500 ◦ C de tempo de residência e temper-
atura, respetivamente; 4) a melhor eficiência prevista na produção de bio-óleo foi de 56.85 %(m/m),
e foi obtida a 550 ◦ C de temperatura de controlo do reator e a 10 rpm de velocidade de rotação dos
parafusos. 5) as quantidades de produção de bio-óleo previstas à temperatura de controlo do reator de

v
550 ◦ C, foram compreendidas entre 53.69 %(m/m) e 56.85 %(m/m).

Palavras Chave

pirólise rápida, design de reactor de parafuso, modelação cinética, tecnologias de conversão termoquı́mica
de biomassa, biomassa lenhosa, determinação de composição orgânica

vi
Abstract

The present work focuses on the develop and design of a rotating screw system in a laboratory scale
reactor to perform fast pyrolysis of woody biomass (with pinewood as a reference sample), having con-
trollable operating conditions(e.g. temperature, residence time). Fitting and optimization tools were used
with thermogravimetric studies and alongside experimental organic composition methods to determine
the distribution of the organic composition of pinewood. Implementing a kinetic tool, these results were
used as inputs to perform a parametric study and determine the optimal conditions for bio-oil produc-
tion. The guidelines to maximize bio-oil yield were followed throughout the reactor design. Afterwards,
the kinetic model was used with inputs from reactor geometry and organic composition of the biomass
to predict the yields of the pyrolysis byproducts. The main conclusions were as follows:1) the organic
composition of the pinewood sample is: 37.7 wt.% of CELL, 25.9 wt.% of HCE, 26.7 wt.% of LIG and
9.6 wt.% of extractives; 2) the weighted distribution of the LIG reference components is: 0.2 wt.% of
LIG-C, 19.0 wt.% of LIG-H and 7.5 wt.% of LIG-O; 3) optimal residence time and temperature for liquid
production is equal to 65 s and 500◦ C since the yield of bio-oil does not increase significantly above
these values, and decreases drastically below 65 s and 500◦ C of residence time and temperature, re-
spectively; 4) the best predicted yield of bio-oil was 56.9 wt.%, at 550 ◦ C of reactor control temperature
and at 10 rpm of auger rotation speed; 5) at 550 ◦ C of reactor control temperature, the predicted yields
of bio-oil were comprised between 53.7 wt.% and 56.9 wt.%.

Keywords

Fast pyrolysis, auger reactor design, kinetic modelling, biomass thermochemical conversion technolo-
gies, woody biomass, organic composition measurement

vii
Contents

1 Introduction 1
1.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Previous Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.2.1 Fast Pyrolysis Fundamentals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.2.2 Fast Pyrolysis Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.2.3 Bio-Oil Properties And Applications . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.3 Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

2 Theoretical Foundations 19
2.1 Biomass: Definition and Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.1.1 Correlation Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.1.2 Triangulation Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.2 Thermogravimetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.3 Pyrolysis Kinetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.4 Optimization Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

3 Reactor Design 31
3.1 Feeding System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.2 Auger System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
3.3 Collection System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.4 Configuration of the Experimental Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

4 Materials and Methods 47


4.1 Biomass Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
4.2 Thermogravimetric Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
4.3 Pyrolysis Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
4.4 Objective Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

5 Results 55
5.1 Biomass Composition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
5.2 Thermogravimetry Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

ix
5.3 Kinetic Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
5.4 Preliminary Testing of Reactor Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
5.4.1 Temperature profile of the reactor . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
5.4.2 Feeding system testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
5.5 Numerical Prediction of Pyrolysis Experiments . . . . . . . . . . . . . . . . . . . . . . . . 64

6 Conclusions 67
6.1 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
6.2 Recommendations for Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70

Bibliography 71

A Volumetric filling level and residence times 77

B Technical drawings of extra components 81

x
List of Figures

1.1 Portuguese final energy consumption balance in 2020 [1]. . . . . . . . . . . . . . . . . . . 4


1.2 Biomass Thermochemical Conversion Technologies [2]. . . . . . . . . . . . . . . . . . . . 5
1.3 Circulating fluidized-bed reactor (CFB) [3]. . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.4 Bubbling fluidized-bed reactor (BFB) [3]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.5 Auger Reactor [3]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.6 Evaluation of fast pyrolysis reactor configurations/technologies for current commercial ap-
plications [4]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.7 Concentration of bio-oil compounds and their relative yield at selected temperature [5]. . . 15
1.8 Most common techniques to identify bio-oil compounds [6]. . . . . . . . . . . . . . . . . . 16

2.1 The pathways of cellulose pyrolysis at different temperature [7]. . . . . . . . . . . . . . . . 22


2.2 The pathways of hemicellulose pyrolysis at different temperature [7]. . . . . . . . . . . . . 22
2.3 The pathways of lignin pyrolysis at different temperature [7]. . . . . . . . . . . . . . . . . . 22
2.4 Example of a TGA device [8]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.5 Example of a thermogravimetric analysis [9]. . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.6 Scheme of the cellulose (CELL) pyrolysis multi-step mechanism [10]. . . . . . . . . . . . . 28
2.7 Scheme of the hemicellulose (HCE) pyrolysis multi-step mechanism [10]. . . . . . . . . . 29
2.8 Scheme of the lignin (LIG) pyrolysis multi-step mechanism [10]. . . . . . . . . . . . . . . . 29
2.9 Scheme of the hydrophilic extractives (TANN) and hydrophobic extractives (TGL) pyrolysis
multi-step mechanisms [10]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

3.1 Scheme of the electric oven including dimensions in mm. . . . . . . . . . . . . . . . . . . 33


3.2 View of the electric oven. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.3 View of the Feeding System. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.4 Biomass feeding system schematic. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.5 Auger system schematic. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
3.6 Technical drawing of the casing of the auger system. . . . . . . . . . . . . . . . . . . . . . 39

xi
3.7 Technical drawing of the first sector of the right-handed auger. . . . . . . . . . . . . . . . 40
3.8 Technical drawing of the assembly of the auger. . . . . . . . . . . . . . . . . . . . . . . . . 40
3.9 Technical drawing of the upstream closing cover. . . . . . . . . . . . . . . . . . . . . . . . 41
3.10 Technical drawing of the downstream closing top part (left) and the downstream closing
bottom part (right). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3.11 Technical drawing of the top part of the bearing housings. . . . . . . . . . . . . . . . . . . 42
3.12 Technical drawing of the bottom part of the bearing housings. . . . . . . . . . . . . . . . . 43
3.13 Exploded view of the Auger System. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.14 Collection system schematic. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
3.15 Reactor configuration schematic. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

5.1 Measured weight and volumes values for the estimation of bulk density (ρ) of the pinewood
sample. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
5.2 Thermogravimetric (TG) and Derivative Thermogravimetric (DTG) mean results from the
pinewood samples. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
5.3 Temperature profile (upper left), mass loss (upper right), mass loss rate (bottom right)
and bio-oil yield (bottom left) of 5 separate numerical pyrolysis experiments of pinewood,
performed with different total reaction times but the same final pyrolysis temperature. . . . 60
5.4 Temperature profile (upper left), mass loss (upper right), mass loss rate (bottom right)
and bio-oil yield (bottom left) of 5 separate numerical pyrolysis experiments of pinewood,
performed with different final pyrolysis temperatures but the same total reaction time. . . . 61
5.5 Temperature profile at a control temperature of 500 ◦ C. . . . . . . . . . . . . . . . . . . . 62

5.6 Temperature profile at a control temperature of 550 C. . . . . . . . . . . . . . . . . . . . 63
5.7 Biomass mass flow vibratory experiment. ∗ The term x is the correlation corresponds to the percent-
age of the maximum voltage of the vibratory system. . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
5.8 Temperature profiles of the particle at 10 rpm. Blue for 500 ◦ C and yellow for 550 ◦ C. . . . 65
◦ ◦
5.9 Temperature profiles of the particle at 20 rpm. Blue for 500 C and yellow for 550 C. . . . 65
5.10 Temperature profiles of the particle at 40 rpm. Blue for 500 ◦ C and yellow for 550 ◦ C. . . . 65
◦ ◦
5.11 Temperature profiles of the particle at 60 rpm. Blue for 500 C and yellow for 550 C. . . . 65
5.12 Summary of the predicted yields at different rotation speeds of operation. . . . . . . . . . 65

B.1 Technical drawing of the reactor opening component. . . . . . . . . . . . . . . . . . . . . . 82


B.2 Technical drawing of the see-through connection component. . . . . . . . . . . . . . . . . 83
B.3 Assembly of the two connection components between the feeding and the auger system. 83
B.4 Technical drawing of the casing support component. . . . . . . . . . . . . . . . . . . . . . 84

xii
List of Tables

1.1 Pyrolysis Methods and Variants [11]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7


1.2 Summary of the reviewed fast pyrolysis studies. . . . . . . . . . . . . . . . . . . . . . . . . 11
1.3 Typical properties of wood pyrolysis bio-oil and of heavy fuel oil [12, 13]. . . . . . . . . . . 16
1.4 Brief description and treatment conditions of the current techniques used for upgrading
bio-oil [13]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.5 Technical feasibility of the current techniques used for upgrading bio-oil [13]. . . . . . . . . 17

2.1 Chemical and structural composition of various biomass groups and sub-groups [14]. . . 24
2.2 Chemical formula and C/H/O mass fraction of the reference components considered in
the pyrolysis CRECK-PM [15]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

3.1 Biomass feeding system description. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35


3.2 Auger system description. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
3.3 Estimated solid residence times for different auger rotation speed. . . . . . . . . . . . . . 38
3.4 Required volumetric gas flow rates at the N2 flow meter to reach target gas residence
times at the control temperature of 500◦ C and 550◦ C, and volumetric biomass filling rate
(f) of 10 % and 36.7 %. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.5 Component description of Figure 3.13 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.6 Collection system description. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.7 Reactor configuration description. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

5.1 Results of USP general test 786 method I. . . . . . . . . . . . . . . . . . . . . . . . . . . . 57


5.2 Comparison of proximate and ultimate analysis of pinewood from different sources. . . . . 57
5.3 Organic composition values of pinewood from different sources. . . . . . . . . . . . . . . . 58
5.4 Full organic composition of the pinewood sample. . . . . . . . . . . . . . . . . . . . . . . . 58
5.5 Biomass mass flow vibratory experiment. ∗ The term x is the correlation corresponds to the percent-
age of the maximum voltage of the vibratory system. . . . . . . . . . . . . . . . . . . . . . . . . . . . 64

xiii
5.6 Predicted yields of the byproducts of pyrolysis at different control temperatures and at
different auger rotation speed. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

A.1 Estimated solid residence times for different auger rotation speed. . . . . . . . . . . . . . 79
A.2 Required volumetric gas flow rates at the N2 flow meter to reach target gas residence
times at the control temperature of 500◦ C and 550◦ C, and volumetric biomass filling rate
(f) of 10 % and 36.7 %. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80

xiv
Acronyms

3PM three parallel reactions model

5PM five parallel reactions model

am as measured

Bi Biot Number

BFB Bubbling fluidized-bed reactor

CELL cellulose

CFB Circulating fluidized-bed reactor

CRECK-PM Chemical Reaction Engineering and Chemical Kinetics – Pyrolysis Mechanism

daf dry ash free basis

db dry basis

DSC Differential Scanning Calorimetry

DTG Derivative Thermogravimetric

FC fixec carbon

fmincon multidimensional constrained nonlinear minimization

FP Fast Pyrolysis

FTIS Fourier Transform Infrared Spectrophotometry

GA genetic algorithm

GC Gas Chromatography

GPC Gel Permeation Chromatography

HAB Herbaceous and Agricultural Biomass

HCE hemicellulose

HAG Herbaceous and Agricultural Grasses

xv
HHV Higher Heating Value

HPLC High-performance Liquid Chromatography

HPLC-MS HPLC-Mass Spectroscopy

HAR Herbaceous and Agricultural Residues

HAS Herbaceous and Agricultural Straws

Hz hertz

LIG lignin

M moisture

P-NMR Phosphorus Nuclear Magnetic Resonance

ppm parts per million

rpm rotations per minute

SFOM single first order reaction model

TANN hydrophilic extractives

TG Thermogravimetric

TGA thermogravimetric analysis

TGL hydrophobic extractives

USP United States Pharmacopeial Convention

VM volatile matter

vpm vibrations per minute

W watt

WWB Wood and Woody Biomass

AAEM alkali and alkaline earth metal

xvi
Introduction
1

Contents
1.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

1.2 Previous Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

1.3 Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

1
2
1.1 Motivation

For the best part of the last century, non-renewable and, in particular, fossil fuels have been the ”go-to”
source of energy for almost all human activities. However, its finite supply and negative impact in our
environment began to outweigh its reliability and the easy solutions fossil fuels provide. This encouraged
governments globally and the scientific community to search for a new cleaner and sustainable energy
source, capable of fulfilling the ever growing energy demand. With both international pressure and
government support to achieve high goals regarding reduction of greenhouse gas emissions, renewable
energy sources became a focal point and, even though most of the demand still depends heavily on non-
renewable sources, lately the increase in use and share of renewable energy has been accentuated.
From 2011 to 2021, the share of low carbon sources in electricity generation went from 31.8% to 39.5%,
which meant in 2019, for the first time, low-carbon sources represented a higher share than coal in
electricity generation [16].
The change to sustainable and efficient energy consumption and production is happening and is
expected to continue with renewable energies as its backbone. By 2030, the European Union targets
to cut greenhouse gas emissions by 40% and for renewable energy’s share in total energy demand to
reach 32% when compared to 1990 levels [17].
From the renewable energy bundle, biomass is seen with great potential as a reliable source to
replace, at least in part, non-renewable energy sources. Being highly available and non-weather de-
pendent, means biomass and bioenergy represent advantages over other renewable energy sources.
Biomass is an abundant resource in developing and developed countries, being a viable choice as any
country’s cornerstone in this new-age energy reform [18, 19]. Biomass and biomass byproducts rep-
resent already a major role meeting the energy requirements of communities worldwide, generating
electricity, providing heating for households, fuelling transportation and producing chemicals for various
industries. Globally, the share of bioenergy in 2018 was 11.3%, representing 55.6% of total renewable
final energy consumption [20].
Figure 1.1 shows that, in Portugal, the share value of biomass alone was 13.1% of final energy
consumption and bioenergy was 14.7%, which represents great potential since final energy consumption
does not account for efficiency. The governmental policies are working and the trend as been set as the
contribution of renewable energies in final energy consumption rouse from 23.6% to 33%, from 2012 to
2018 [1].
Most of the new technologies in the biomass industry are still not economically profitable at an indus-
trial scale and this growth gives space for development along with excellent opportunities for investment.
Also, it allows the industry to tackle problems inherent to biomass such as logistics with harvesting,
transporting, scanning and choosing optimal forms and types of biomass, since it is such a widely di-
verse resource [21]. Biomass comes in all sizes and shapes, vary in moisture content, energy and bulk

3
Figure 1.1: Portuguese final energy consumption balance in 2020 [1].

densities which explain the difficult treatment, usage and transportation. However, through biomass
conversion technologies, techniques and processes, such as pyrolysis, it is possible to transform and
upgrade biomass into more valuable byproducts with numerous advantages. Figure 1.2 illustrates the
market routes of the three main biomass conversion technologies into higher valued products and shows
the potential for pyrolysis to become an important factor in a variety of industries.

Pyrolysis is a thermochemical process which uses heat to decompose plant biomass or organic
materials into gaseous, liquid and solid byproducts in the absence of oxygen. This process, unlike
other thermochemical processes such as combustion or gasification, can be adjusted and conditioned
to maximize the yield of one of the product types - bio-char, bio-oil or gas. The versatility that pyrolysis
provides increases its interest since all of the byproducts have potential for either transportation, heat or
power generation. Fast pyrolysis, in particular, is a type of pyrolysis in which the conditions applied tend
to maximize the liquid product, the bio-oil. The basic conditions of the process of fast pyrolysis are high
transfer and heating rates, short vapour residence time and rapid cooling of the byproducts. High yields
of bio-oil allow for storage and transportation of the liquid products to be easy and economically viable,
increasing the process’s market attractiveness since bio-oil can be used directly in the production of bio-
fuels and chemical feedstock [22,23]. However, both properties and yields of the pyrolysis products vary
with the origin and composition of initial biomass, which is a parameter in a state of constant change.
Further analysis and studies are needed in order to reduce the impact of this variation, consequently

4
Figure 1.2: Biomass Thermochemical Conversion Technologies [2].

standardizing the products and projecting biomass to a reliable source of high valued products.

Nowadays, fluidized bed and rotating cone reactors at an industrial level are commercially available
but the thermal properties and overall characteristics of its products are still too different from traditional,
non-renewable, fossil originated fuels. Also, other types of reactors deserve further investigation since
they display great advantages over the above mentioned reactors such as a higher degree of accept-
ability for feed size, easier scale-up and less construction complexity. Auger reactors are viewed as a
promising type of reactor, managing to achieve a simpler overall operation with less energy consump-
tion and precise residence time, being directly correlated to the screw’s rotating speed [24, 25]. Despite
having a lot of potential, this technology is yet to reach its place in the renewable energies industry.
Portugal is estimated to have from 5630 thousand to 6500 tons of annual dried biomass production,
which provides plenty of prospects and room for development of this up and coming pyrolysis conver-
sion technology, the auger reactor [26, 27]. The marketability of these reactors in Portugal could provide
profitability to many departments and help improve the overall state of the country, not only financially.
The lower maintenance and simpler operational demand of auger reactors suggests it may be used in
remote or less optimal infrastructures and with less technically qualified personnel. For example, the
process of clearing and managing forests and woods to prevent wildfires could became a cost-effective
practice, actively preventing a problem in this country that, with climate change, will only get worse.

5
1.2 Previous Studies

1.2.1 Fast Pyrolysis Fundamentals

In the next sections, focus will be given to pyrolysis and fast pyrolysis in particular since, as mentioned
before, these processes present incredible potential to guide the change to renewable energy sources
and also to become an important factor in other market environments or industries, for example as a
chemical feedstock source.
Pyrolysis is a very complex process, involving hundreds of reactions across the three phases, po-
tentially generating hundreds of volatiles. Generally, pyrolysis is the thermochemical decomposition of
organic materials at temperatures between 200◦ C and 700◦ C, in the total absence of oxygen. In this
process, large hydrocarbon molecules break down into smaller molecules resulting in three types of
products - liquid, gas and solid. During the thermal treatment, the various pyrolysis conversion tech-
niques can be explained sequentially as temperature and decomposition increase. As temperature
rises, around 300◦ C, biomass starts decomposing. Between 300◦ C and 500◦ C, carbonization takes
place. Long residence times are used to achieve high biomass decomposition, favoring bio-char pro-
duction. Further increasing temperature while lowering residence time of the by-products leads to higher
rates of degradation and volatile release, approaching pyrolysis. This process produces higher yields
of liquid and gaseous products when compared to carbonization. For gasification the conditions slightly
differ since the process is not in the total absence of oxygen. An atmosphere with a partial supply of
oxygen is required to prevent combustion and high temperatures are also needed to produce high yields
of syngas during gasification. However, these processes are view as branches or variants of pyrolysis
and their differences can be reduced mostly to the operating conditions of each process, as changing
operation conditions alone may alter the entire process.
At a molecular level, the decomposition process of pyrolysis can also be described sequentially as
temperature increases. The first reactions start from around 200◦ C up to 400◦ C. In this first stage,
hemicellulose, the weaker structure of the major components of biomass, is the first to decompose
between 225◦ C and 350◦ C. As temperature continues to increase, cellulose pyrolysis reactions oc-
cur between 325◦ C and 400◦ C, often overlapping hemicellulose pyrolysis reactions. During this stage,
the highest decomposition rate is reached and most of the overall condensable volatiles (bio-oil) are
formed/released. Afterwards, with increasing temperature, secondary reactions occur. In most of these
reactions, condensable volatiles suffer cracking/fragmentation into smaller molecules, mostly resulting
in non-condensable gases. Depending on the residence time, the volatiles trapped in the solid matrix
may be otherwise re-polymerized into the increasingly porous surface, creating a structure similar to
carbon coke. Thus, residence time and temperature play a crucial role on the properties and yields of
by products [3, 28, 29].

6
Specifically, fast pyrolysis seeks to increase the yield of bio-oil. Fast pyrolysis’s intent is to prevent
over-cracking of vapors into non-condensable gas molecules and prevent recombination of the vapors
with larger molecules, such as char. Since secondary pyrolysis reactions occur when contact between
by products is verified, hence, residence times must be reduced. This way, secondary reactions are kept
to a minimum, maintaining maximum condensable volatiles.

Table 1.1: Pyrolysis Methods and Variants [11].

Types of pyrolysis Pyrolysis conditions Technology (commonly used reactors) Main products
Volatiles residence time Solid residence time Heat transfer Temperature (◦ C)
BFB, CFB,
Fast Up to 2 seconds Short High 400-600 spouted bed, entrained flow Liquid
and auger
Between 10 and From seconds Auger, rotary kiln Liquid, solid
Intermediate Medium 400-600
30 seconds to minutes and fixed bed and gas
From minutes From minutes Auger, rotary kiln
Slow-Torrefaction Low 250-350 Solid
to hours to hours and fixed bed
Auger, rotary kiln
Slow-Carbonization Long From hours to days Very low 300-500 Solid
and fixed bed

As Table 1.1 indicates, fast pyrolysis occurs at higher heating rates, where biomass decomposes in
a few seconds, resulting in mostly condensable vapors but also some char and non-condensable gases.
After rapid extraction and cooling of the products, the condensed volatiles compose a black tarry liquid,
which is bio-oil. For fast pyrolysis, typical values for bio-oil yield from woody feedstock, in laboratory
scale reactors, is around 60-70% on dry feed basis [22]. This process usually means that the small
percentage of char and gas produced is enough to reach an energetic input threshold, elevating it to a
energy self-sufficient process and extending further pyrolysis’ potential going forward [2, 22, 30].
As mentioned earlier, there are three types of products after performing fast pyrolysis - solids, liquids
and gaseous. Solid content of the products of fast pyrolysis is small in yield and is mostly char and car-
bon. Biochar, as it is commonly called, besides being a noticeable upgrade to biomass regarding energy
density, it also shows valuable soil-amendment qualities, representing a renewable and sustainable path
for water, nutrients and contaminant retention in soil [31]. The high HHV also makes biochar a compet-
itive choice over coal in fuel applications. The gaseous content released during pyrolysis is composed
by carbon dioxide (CO2 ), carbon monoxide (CO), hydrogen (H2 ), low carbon number hydrocarbon such
as methane (CH4 ), ethane (C2 H6 ), ethylene (C2 H4 ), and may contain small amounts of gases such as,
propane (C3 H8 ), ammonia (N H3 ), nitrogen oxides (N Ox ), sulphuroxides (SOx ) and alcohols of low car-
bon numbers [32]. However, both solid and gas products represent transportation issues, inconvenient
problems related to supply chain management and underachieve to reach the usefulness of the liquid
products.
Liquid products contains mostly tars, heavy hydrocarbons and water. The bio-oil is the primary
product of interest and its yield is maximized when the conditions of fast pyrolysis are applied. Chemi-
cally, bio-oil may contain ”water and hundreds of organic compounds, such as acids, alcohols, ketones,
aldehydes, phenols, ethers, esters, sugars, furans, alkenes, nitrogen compounds and miscellaneous

7
oxygenates” [32].
In the remainder of this section, a number relevant of studies will be reviewed. In the available liter-
ature, some studies spotlight auger or screw reactors, however, pyrolysis as a thermochemical process,
is product of its operating conditions, i.e. it is only the result of the set of conditions that are applied to
the biomass. There are many meaningful conditions to be concerned about regarding pyrolysis:

• Temperature of Pyrolysis

• Heating Rate

• Residence Time

• Particle Size

• Biomass Type

Some conditions and parameters applied for pyrolysis may be reproduced using different reactor
configurations, maintaining studies regarding reactors, other than auger, relevant. In this review focus
was given to studies that analyse variable pyrolysis parameters and their effects of the overall process
and byproducts. Special attention was given to research on maximization of bio-oil production and
quality. Below, Table1.2 summarizes briefly the reviewed studies, where the type of biomass chosen,
heating rate applied, reactor or mechanism used, effects evaluated and main conclusions are specified.
Furthermore, the effects and importance of the main controlling parameters in the pyrolysis behavior are
discussed.

Temperature

Chen et al. [33] concluded through experiments that temperature had a more significant impact
on bio-oil yield and quality when compared to heating rate. Yield of bio-oil was increased 3.91wt.%,
1.34wt.% and 1.71wt.% using heating rates of 10◦ C/min, 30◦ C/min and 50◦ C/min, respectively, when
the temperature was increased for 400◦ C to 500◦ C. However, when further increasing the tempera-
ture to 600◦ C the yields decreased, which may be caused by secondary cracking reactions of the
volatiles, which places optimal operating temperature for bio-oil production at around 500◦ C. Demir-
bas [34], Akhtar et al. [35], Kan et al. [32], Anca-Couce [11] and Perkins et al. [24] also concluded that
the temperature range from 400◦ C to 550◦ C is the best for bio-oil mass yield while lower temperatures
favor char production and higher temperatures increase gas content. Additionally, Demirbas [34] ob-
served that the best quality bio-oil, with highest concentration of valued components, was formed at
452◦ C.

8
Particle Size

Shen et al. [36] studied effects of particle size on bio-oil yield in woody biomass and concluded
that bio-oil yields decreased as average particle size increased from 0.3 mm to 1.5 mm. However, the
variation in yield was marginal and no change was verified while further increasing the average particle
size. Significant differences were found in the composition of bio-oil with different particle sized feeding
biomass, with heavy components being lowered in relative quantity to lighter components as average
particle size increases. Moreover, Vamvuka et al. [37] observed that particle size, between 250 µm and
1000 µm, had no significant influence on pyrolysis or its products. On the other hand, Anca-Couce [11]
and Akhtar et al. [35] denoted the importance for smaller particle sizes, below 1 mm, not only to reach
high biomass conversion rate but also to minimize Biot Number (Bi), therefore minimizing intra-particle
temperature gradient and displaying a uniform heat distribution.

Heating Rates

Alongside pyrolysis temperature, heating rate is the most important parameter in the process of
biomass pyrolysis. Various studies conclude that medium to high heating rates improve bio-oil produc-
tion and, therefore, are applied for fast pyrolysis [11, 33, 38]. Nanda et al. [38] concluded high heating
rates lead to higher yields and higher HHV bio-oils, indicating superior fuel properties. Despite also
observing rise in HHV with the increase in heating rate, Chen et al. [33] found it not to be a significant
change when heating rate is increased from 10◦ C/min to 50◦ C/min. Analysing DTG profiles, several
authors observed that biomass components (hemicellulose, cellulose, lignin and extractives) had dif-
ferent ranges of decomposition temperatures and that increasing the heating rate would delay overall
thermal decomposition to higher temperatures. As the transition temperatures increase and DTG pro-
files shift, the individual decomposition peaks of the components overlap and the temperature range
of each component become indistinguishable [39, 40]. Furthermore, lower heating rates may increase
higher volatile matter production since it takes more time to reach a certain temperature, allowing for
other thermochemical reactions, such as dehydration, to occur. This may indicate that, even though
high heating rates are optimal for bio-oil production, there is a limit, past which increasing heating rate is
counterproductive in the production of bio-oil [38]. Also, heat and mass transfer limits may be reached
as heating rate increases, possibly lowering biomass conversion rate and percentage [35].

Residence Times

Akhtar et al. [35], Kan et al. [32], Perkins et al. [24] and Anca-Couce [11] concluded that higher
residence times allow more time for secondary reactions and cracking to take place, providing higher

9
yields of char and reducing volatile matter, consequently lowering bio-oil production. Residence vapour
time is, therefore, directly related to sweeping gas flow. Once again, in order to reduce residence time,
and more importantly vapour residence time, higher sweeping gas flows are recommended to separate
volatile matter and reach high yields of bio-oil during fast pyrolysis. However, Akhtar et al. [35] concluded
that changing the flow of the sweeping gas demonstrate negligible to low effects on bio-oil yield and
production.

Biomass Type

Biomass feedstock has very broad sizes, shapes and characteristics. Effects of different biomass
compositions in the process of pyrolysis were studied to properly model the procedure, preview the
results and maximize the outcome. Through thermogravimetric analysis, Grønli et al. [39] concluded
that softwood and hardwood decompose at slightly different temperature ranges, leading to more of an
overlap between cellulose and hemicellulose in softwoods. Additionally, valid and consistent predic-
tions of DTG profiles were reached fixating the activation energy of each of the macro-components of
the biomass [37, 39, 40]. Furthermore, cellulose and hemicellulose decompose in a smaller range of
temperatures, impact positively the production of bio-oil and are usually consumed totally while lignin
decomposes in a wide temperature range and typically into char [11]. Akhtar et al. [35] and Kan et
al. [32] also observed that the yield of char increases as biomass composition exhibits more lignin.

10
Table 1.2: Summary of the reviewed fast pyrolysis studies.

Reference Sample β(◦ C/min) Mechanism Studied Effects Conclusions


• Increase in heating rate represents an increase in mass yield of bio-oil
• Optimal HHV, water content and mass yield reached around 500-550°C
• Heating Rate
Chen et al. [33] Poplar Wood Sawdust 10-50 Lab-scale fixed-bed reactor • Increase in heating rate decreases water content and increases HHV,
• Pyrolysis Temperature
however not significantly
• Temperature influences products production and quality more than heating rate
• Peak of different components overlap in DTG profiles at high heating rates due
Maiti et al. [41] Mustard Stalk 5-50 Arrhenius Plot Method • Heating Rate to resistances to mass and heat transfer inside the particles
• Higher heating rates represent a noticeable increase followed by a decrease in activation energy
Wheat Straw • With high heating rate pryolysis, bio-oil achieves higher HHV and yield percentage values
• Heating Rate
Nanda et al. [38] Timothy Grass 2-450 Lab-scale fixed-bed reactor • High carbon and hydrogen content bio-oil when compared
• By-product Quality
Pinewood to slow heating rate bio-oils suggests superior fuel properties
Alder
• Softwood starts decomposition at lower temperatures
Beech
• Five parallel model is required for species with high extractive content at lower temperatures
Birch
• Fixing the activation energy for cellulose, hemicellulose and lignin proves
Oak TGA
to accurately describe high temperature degradation behavior for all species
Grønli et al. [39] Douglas fir 5 Least Squares • Biomass Type
• Overlapping of cellulose, hemicellulose and extractives
Redwood 5PM
• Lignin decomposes in a large range of temperatures while cellulose and
Spruce
hemicellulose react at a lower and shorter range of temperatures
Pine A
•Cellulose and hemicellulose zone overlap more in softwoods
Pine B
Olive Husk
• 790K is the optimal pyrolysis temperature for liquid yield
Hazelnut Shell • Pyrolysis Temperature
Demirbas [34] 600 Pyrolysis batch reactor • Liquid yield increases with increase in lignin content at the cost of biomass conversion rate
Spruce Wood • By-products Quality
• Bio-oil formed at 725K had high overall concentration of useful compounds
Beech Wood
• Temperature is the most influencing parameter
• Both biomass type and pyrolysis temperature highly influence pyrolysis
• Pyrolysis Temperature
• A combination of moderate temperature (500-550ºC),
• Vapor Residence Time
high heating rates and short residence times increase liquid oil yield
• Particle Size
• Initial moisture content and heating rate increase liquid yield but decrease oil quality
Akhtar et al. [35] • Heating Rate
• Small particles are preferable for uniform heat distribution and higher biomass conversion rate
• Sweeping Gas
• Sweeping gas helps reduce volatile residence time and ultimately reduce secondary reactions
• Biomass Type
but demonstrate negligible to low effects on liquid yield
• Initial Moisture Content
• Generally cellulose and hemicellulose tend to produce more volatiles
• Moderate to low sweeping gas rates are helpful to obtain high liquid yields
• Optimal pyrolysis temperature is between 400-550ºC for liquid yields
• Pyrolysis Temperature
• Higher heating rates increase liquid yields until mass and heat transfer limits are reached
• Heating Rate
Kan et al. [32] • Shorter vapor residence times favor oil yields because it reduces time for secondary reactions to occur,
• Vapor Residence Time
which increase char and gas yields at the expense of oil yield
• Biomass Type
• Cellulose and hemicellulose increase bio-oil production yield but lignin may decrease water concentration in the bio-oil
• Bio-oil yields decrease as the average particle size increases from 0,3mm to 1,5mm but
• Particle Size
Shen et al. [36] Oil Mallee 10 Fluidised-bed reactor no change is bio-oil yield was verified while further increasing particle size
• By-products Quality
• Small-particle pyrolysis could give lower yields of lighter bio-oil components and higher yields of heavier components
• Pyrolysis Temperature • Pyrolysis temperature around 500ºC maximazes liquid yield
• Heating Rate • Particle size smaller than 1mm is recommended to lower Bi number and,
Anca-Couce [11] • Vapor Residence Time therefore, minimize intra-particle temperature gradient
• Particle Size • It is also necessary but not sufficient condition to achieve pure kinetic regime,
• Biomass Type which is challenging to do at high heating rates
• Pyrolysis Temperature • Bio-oil yield from fast pyrolysis is maximized at 450-550ºC
Perkins et al. [24] • Pressure • Higher temperatures lead to more gas content, while lower temperatures increase char production
• Vapor Residence Time • Lower pressures and low vapor residence times are also important parameters to assure high liquid yields
11
1.2.2 Fast Pyrolysis Technologies

In the last few years, the process of fast pyrolysis has been intensely studied and better understood.
Technologies, techniques and configurations associated to FP where not an exception to this evolution.
There are plenty of different configurations in which pyrolysis can occur [11, 24]:

• Bubbling fluidized-bed reactor (BFB)

• Circulating fluidized-bed reactor (CFB)

• Fixed bed reactor

• Auger/Screw reactor

• Vacuum reactor

• Ablative reactor

• Entrained flow reactor

• Rotating cone reactor

• Cyclonic reactor

Although all of the above-mentioned reactors or configurations have technical and economical up-sides
and down-sides, given the high heat transfer the process of FP demands, only a few are seen as com-
petitive for this purpose. Alongside Bridgwater [2], Raza et al. [25], Perkins et al. [24] and Sharifzadeh
et al. [6] made extensive reviews on the state-of-the-art of reactors applied for fast pyrolysis, including
major advantages and disadvantages of each of the most popular configurations. Specific applications
of these technologies were also briefly described. The BFB, CFB and auger reactors are acknowledged
by the scientific community to be the better configurations to produce bio-oil and make it at reliable rate
while keeping all the economic advantages [24, 42], as it is displayed in Figure 1.6.
Fluidization is a phenomenon in which solid particles are in such conditions that they represent fluid
behavior. At this point, a turbulent flow is created with the velocity of the passing fluid and high chemical
reaction and heating rates are achieved. In both CFB reactors and BFB reactors, this environment is
attained.
A typical CFB reactor consists in a reactor bed where the biomass is fluidized with an inert gas to
avoid combustion. The reactor bed is usually linked to multiple cyclones, one of which is connected in
series. This configuration creates, with the fluidization and movement of the circulating gas, a circulation
motion of the biomass between the reactor bed and the first cyclone during pyrolysis, as Figure 1.3
illustrates. After the cyclone cycles, bio-oil and gaseous products are condensed and collected.

12
Figure 1.3: Circulating fluidized- Figure 1.4: Bubbling fluidized-bed
bed reactor (CFB) [3]. reactor (BFB) [3]. Figure 1.5: Auger Reactor [3].

Similarly to the CFB reactor, the BFB also includes a fluidized bed. However, in this type of reactor,
the biomass is not transported cyclically to a cyclone. Instead, a bubbling effect is created as the biomass
is fluidized and pyrolysed with the passing inert gas. Afterwards, alike CFB reactors, the products are
directed to a cyclone to separate solid products and volatiles, followed by the condensation and collection
of the latter. A typical BFB is illustrated in Figure 1.4.

CFB reactors and BFB reactors, Figure 1.3 and Figure 1.4 respectively, are the most commonly
used types of reactors. In order to achieve proper function and optimal performance in CFB and BFB
reactors, biomass feedstock needs pre-treatment that involves drying and milling. After pre-treatment,
biomass is lead into the reactor with the aid of a feeding system, which usually comprises a screw. Once
biomass leaves the feeding system, it enters a fluidized environment where the passing fluid heats up
the particles causing decomposition of the biomass. When it comes to auger reactors biomass fed into
a tubular decomposition chamber, which is usually heated via electric resistances and, in some cases,
aided with the combustion of some of the pyrolysis products, where the mechanical movement of the
screw or screws transport and mix the biomass, increasing heat transfer coefficient, biomass conversion
rate, providing a continuous process and controlling solid and gas flows. A simple example of an auger
reactor configuration is displayed in Figure 1.5. Afterwards, in both fluidized and auger reactors, the
charcoal is collected and the volatiles exit the reactor bed/decomposition chamber, passing through
various cycles of condensation to further separate the products. Even though CFB and BFB reactors
are viewed as robust and fill a significant part of the market space, these configurations belong along
the most complex to operate, require large quantities of inert gas to create the fluidizing media and
becomes hard to achieve enthalpy threshold for pyrolysis, while increasing scale size of these reactors.
Furthermore, the high velocities of the media, that are used to reach very high heat transfer coefficients,
generate solid to solid collisions which create fine particles and hinder the separation process of the
by-products. Since particulate-free bio-oil is desirable for every possible application, this phenomena
cripples the process by leading to a less optimal final product [42].

Butler et al [4] made a comprehensive and thorough review of the development and commercial appli-
cations of the most important fast pyrolysis technologies. As it is illustrated in Figure 1.6, auger reactors

13
are, alongside CFB and BFB reactors, a contender for the strongest and most attractive technology
going forward. The lack of high velocity components and relatively simple design of auger reactors,
represent a clear advantage in longevity and build robustness over other configurations, such as the
ablative reactor and the rotating cone, specially when scaling up. Moreover, the absence of catalysts in
the process also represent significant leverage towards fluidized bed reactors. Furthermore, the flexible
nature of the auger configuration makes the modifications to adapt to diverse feedstock easy. This way,
the same auger reactor system may achieve high levels of performance for different feedstocks that
require slightly distinct conditions and even perform other variations of pyrolysis (slow and intermediate
pyrolysis), if the targeted final product shifts.

Figure 1.6: Evaluation of fast pyrolysis reactor configurations/technologies for current commercial applications [4].

1.2.3 Bio-Oil Properties And Applications

Overall, the pyrolysis process entails the separation and fragmentation of the carbon to carbon linkages
present in lignocellulosic biomass, followed by the formation of carbon to oxygen bonds. In simple terms,
chemically, pyrolysis consists in a large number of oxidation-reduction reactions. During the degradation
of the biomass particles, some are reduced to carbon while some are oxidized and hydrolysed. There-
fore, the aftermath of fast pyrolysis, bio-oil, is a combination of highly oxygenated compounds, carboxylic
acids and trace water [23, 43].
Although water is the most abundant compound in bio-oil, other major compound groups are found.
Thangalazhy-Gopakumar et al. [5] analysed the physiochemical properties of bio-oil produced at dif-
ferent temperatures from pine wood. Upon chemical analysis, the compounds found were grouped
as carbohydrates, aromatics, furans, phenols and cresols, and guaiacols. Further specifications and
concentrations of each compound at various temperatures are presented in Figure 1.7.

14
Figure 1.7: Concentration of bio-oil compounds and their relative yield at selected temperature [5].

Biomass fast pyrolysis liquid products can be either an homogeneous single phase oil with large
quantities of water or, more commonly, a heterogeneous separated fluid. Bio-oil is a micro-emulsion, in
which the continuous phase is the aqueous phase, constituted by holocellulose (cellulose and hemicel-
lulose) deterioration products and some small molecules from lignin decomposition. The discontinuous
part consists mostly in heavy pyrolytic lignin-derived macro molecules and presents a very oily feel and
behavior [23]. The aqueous part stabilizes the discontinuous part through mechanisms such as hydro-
gen bonding. The instability, typically detected in bio-oils, is believed to be related to the breakdown of
the emulsion and of these mechanisms (among other factors like extractive content quantity), the overall
biomass composition and origin, and the unwanted continuation of secondary reactions [2, 43].
In order to properly access the potential of bio-oil(e.g. as a source for chemical feedstock), a rig-
orous characterization must be executed. However, a broad chemical characterization is difficult. Bio-
oils have a vast range of molecular sizes that require distinct methods to analyse. The most common
analytic tools are: High-performance Liquid Chromatography (HPLC) and Gas Chromatography (GC)
for the identification of the compounds, Gel Permeation Chromatography (GPC) and HPLC-Mass Spec-
troscopy (HPLC-MS) for larger compounds, Phosphorus Nuclear Magnetic Resonance (P-NMR), Fourier
Transform Infrared Spectrophotometry (FTIS), Karl Fischer titration for water content determination and

15
Thermogravimetric (TG) Analysis. Figure 1.8 illustrates the necessity of having multiple analytic tools to
accurately evaluate the chemical properties of bio-oil [6].

Figure 1.8: Most common techniques to identify bio-oil compounds [6].

From a physical standpoint, there are significant differences between standard fossil fuels and bio-oil.
Some of the undesirable properties of bio-oil are directly related to some of the constituent compounds.
The highly oxygenated nature of most of the compounds present in bio-oil justifies the contrasting oxygen
content of the fuels. In addition, it grants bio-oil its high polarity and hydrophilic properties, which disallow
its mixing with hydrocarbon liquids [13]. Table 1.3 compares conventional petroleum fuel properties to
typical pyrolysis wood derived bio-oil.

Table 1.3: Typical properties of wood pyrolysis bio-oil and of heavy fuel oil [12, 13].

Properties Pyrolysis bio-oil from wood Heavy petroleum fuel oil


Moisture content (wt.%) 15-30 0.1
pH 2.5 -
Specific gravity 1.2 0.94
Elemental composition (wt.%) C 54-58 85
H 5.5-7.0 11
O 35-40 1.0
N 0-0.2 0.3
Ash 0-0.2 0.1
HHV (MJ/kg) 16-19 40
Viscosity (at 50◦ C)(cP) 40-100 180
Solids (wt.%) 0.2-1 1
Distillation residue (wt.%) Up to 50 1

Water content in bio-oil is typically very high when compared to other fuels, and even though it has a
stabilization effect at low content and lowers viscosity of the oil, at high content may lead to phase sepa-
ration and heavily reduces Higher Heating Value (HHV). Water originates from initial biomass moisture
and dehydration reactions during pyrolysis. Removing or reducing moisture weight percentage after
the production of bio-oil is difficult since heating it may lead to the continuation of secondary reactions,
also known as ”aging”, and the oil tends to polymerize. Therefore to reduce moisture content, biomass
should be dried to an acceptable degree before being pyrolyzed [2, 23]. Beyond water and oxygen con-
tent, bio-oil differs from conventional heavy petroleum oil in density, pH, viscosity and HHV substantially,
as Table 1.3 shows. All of these factors dismiss the possibility to simply replace fossil fuels. In order for
bio-oils to be used directly, special attention must be payed to the design of the equipment (e.g. boilers

16
and engines) [2, 44].
Despite some undesirable physical and chemical characteristics, bio-oil maintains high potential
since fast pyrolysis is considered a technology capable of producing bio-oil from biomass waste [44].
Therefore, a sizable effort was made to develop upgrading and post-processing methods for bio-oil.
Physical and chemical techniques were developed to convert bio-oil into feedstock ready for innumerous
applications [2]. Xiu et al. [13] made a review of these procedures, which are summarized in Tables 1.4
and 1.5.

Table 1.4: Brief description and treatment conditions of the current techniques used for upgrading bio-oil [13].

Upgrading methods Treatment conditions/ Requirements Reaction mechanism/ Process description


Mild conditions (±500ºC/low pressure),
Hydrogenation without simultaneous cracking
Hydrotreating/Hydrofining chemicals needed: H2 /CO,
(eliminating N, O and S as N H3 , H2 O and H2 S)
catalyst(e.g. CoMo, HDS, NiMo, HZSM-5)
Severe conditions (>350ºC,100–2000 Psi),
Hydrogenation with simultaneous cracking
Hydro-cracking/Hydrogenolysis/Catalytic cracking chemicals needed:H2 /CO or H2 donor solvents,
Destructive (resulting in low molecular product)
catalyst (e.g. Ni/Al2 O3 − T iO2 )
Promotes the reaction by its unique transport properties:
Mild conditions, organic solvents needed such as alcohol, gas-like diffusivity and liquid-like density,
Sub/super-critical fluid
acetone, ethyl acetate, glycerol thus dissolved materials not soluble in either liquid
or gaseous phase of solvent
Reduces oil viscosity by three mechanisms:
Solvent addition(direct add solvent or esterification Mild conditions, polar solvents needed such as water, physical dilution
of the oil with alcohol and acid catalysts) methanol, ethanol, and furfural molecular dilution or by changing the oil microstructure
chemical reactions like esterification and acetalization
Combines with diesel directly.
Emulsification/Emulsions Mild conditions, need surfactant (e.g., CANMET)
Bio-oil is miscible with diesel fuels with the aid of surfactants
Chemical extracted from the bio-oils Mild conditions Solvent extraction, distillation, or chemical modification
Steam reforming High temperature (800–900ºC), need catalyst (e.g. Ni) Catalytic steam reforming and water–gas shift

Table 1.5: Technical feasibility of the current techniques used for upgrading bio-oil [13].

Upgrading methods Technique feasibility


Pros. Cons.
Hydrotreating/Hydrofining Cheaper route, commercialized already High coking (8–25%) and poor quality of fuels obtained
Needs complicated equipment, excess cost,
Hydro-cracking/Hydrogenolysis/Catalytic cracking Makes large quantities of light products
catalyst deactivation, reactor clogging
Higher oil yield, better fuel quality
Sub/super-critical fluid Solvent is expensive
(lower oxygen content, lower viscosity)
The most practical approach (simplicity,
Solvent addition(direct add solvent or esterification
the low cost of some solvents and their Mechanisms involved in adding solvent are not quite understood yet
of the oil with alcohol and acid catalysts)
beneficial effects on the oil properties)
Emulsification/Emulsions Simple, less corrosive Requires high energy for production
Chemical extracted from the bio-oils Extract valuable chemicals Low cost separation and refining techniques still needed
Steam reforming Produces H2 as a clean energy resource Complicated, requires steady, dependable, fully developed reactors

Hu et al. [44] made an extensive review on the state-of-the-art of bio-oil applications and studied the
techno-economic situation and prospects of each of the conversion routes. Several applications and
end-products were taken into account:

• Use of bio-oil as boiler fuels and in heavy-duty engine:

– Direct burning

– Co-firing

– Burner technology for bio-oil

• Bio-oil to biofuel by hydrodeoxygenation (HDO)

• Bio-oil to hydrogen

• Chemicals from bio-oil:

17
– Distillation

– Extraction by solvent

• Carbonaceous materials from bio-oil

• Bio-oil to binder

• Bio-oil to binder for asphalt

• Wood vinegar as a pesticide and fertilizer from bio-oil

• Bio-oil to polyurethane foams

• Bio-oil to plastics

Thorough analysis of each case were made, even though, for some of the routes such as the produc-
tion of polyurethane foams and plastics, only a theoretical discussion was made since the methods are
still being developed or represent a very small scale of the production. No direct conclusion was drawn
for which is the single best path to take since most revealed advantages and disadvantages, either eco-
nomically or technically. Further fundamental studies, developments and optimizations in the process of
each application are still needed to raise pyrolysis technology to a real commercial contender [44].

1.3 Objectives

Against the given background, the main objective of this work is to develop and design a rotating screw
system in a laboratory scale reactor, able to have controllable operating conditions(e.g. temperature,
residence time), to perform the fast pyrolysis of pinewood. To this end, preliminary experimental and
kinetic studies were performed to predict the optimal operating conditions that provide the highest yield
of bio-oil. The specific objectives are as follows:

• Perform thermogravimetric experiments to estimate the organic composition of pinewood, to be


used as input in the kinetic tool;

• Investigate the impact of the operating conditions in the fast pyrolysis behavior and the production
of bio-oil;

• Design and develop the main reactor systems (feeding, auger, products collection) and perform
preliminary experiments;

• Use the kinetic tool to predict the pyrolysis outcome, given the preliminary experimental operating
conditions;

This study ultimately provides an experimental tool and a methodology for future research in biomass
conversion technologies, specifically auger reactors, on how the organic biomass composition can affect
fast pyrolysis byproducts and their biochemical analysis.

18
Theoretical Foundations
2
Contents
2.1 Biomass: Definition and Characterization . . . . . . . . . . . . . . . . . . . . . . . . . 21

2.2 Thermogravimetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

2.3 Pyrolysis Kinetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

2.4 Optimization Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

19
20
2.1 Biomass: Definition and Characterization

Biomass is a renewable organic material, a biological material derived from plants and animals. It origi-
nates from many sources such as wood and wood residues, agricultural crops and waste materials, farm
residues, biogenic materials in municipal solid waste, animal manure and human sewage. Biomass is
extremely diverse, the different conditions of its growth i.e., weather conditions, livestock feed, chemi-
cals and fertilizers used, geographic location and soil properties, all play a part in its formation. Thus,
biomass properties differ from biomass to biomass, especially in its moisture, structural components
and inorganic constituents [45]. The major constituents of biomass are Carbon (C), Oxygen (O) and
Hydrogen (H). Nitrogen (N ), Sulphur (S), Calcium (Ca), Potassium (K), Silicon (Si), Magnesium (M g),
Aluminum (Al), Iron (F e), Phosphorus (P ), Chlorine (Cl), Sodium (N a), Manganese (M n), Titanium
(T i) also generally appear in the constitution of biomass, though usually present in oxide form and take
a minor role, since the content of each of these elements is small [14]. Lignocellulosic biomass, i.e.,
originated from agricultural waste and forest debris, is composed by three macro structural components:
cellulose, hemicellulose and lignin [7]. In addition, hydrophobic and hydrophilic extractives also take part
representing non-structural organic and inorganic matter in the composition of biomass [46]. Cellulose
is a linear macromolecular polysaccharide which basically comprises of long glucose polymeric chains
bounded to each other by hydrogen bounds, contributing to its strong crystalline structure and providing
chemical and thermal stability. Hemicellulose is described as a short heterogeneous polysaccharide
chain structure, usually containing xylose, mannose, glucose, galactose, some acids and acetyl prod-
ucts to name a few constituents. Hemicellulose presents an irregular and shapeless structure, which
conceding low thermal stability to the macro component. Thus, hemicellulose is the component to begin
biomass thermal degradation at the lowest temperature (200◦ C). Lignin is a non-sugar, irregular struc-
tured polymer mostly composed by aromatic rings, alcohols and some acids. Its high chemical activity
turns its decomposition process very difficult, and therefore takes a large temperature range to fully de-
grade (100◦ C - 900◦ C). As mentioned above, extractives correspond to non-structured constituents that
do not make an integral part of the biomass structure, are usually composed by carbohydrates, proteins,
hydrocarbons, oils, aromatics, lipids, fats, starches, phenols, waxes and inorganic materials, and are ex-
tractable or removable with solvents (e.g. water, benzene, ethanol, hexane, etc.) [7,10,14,29]. Reaction
paths, at increasing temperature, of cellulose, hemicellulose and lignin are illustrated in Figure 2.1 , 2.2
and 2.3 , respectively.

21
Figure 2.1: The pathways of cellulose pyrolysis at different temperature [7].

Figure 2.2: The pathways of hemicellulose pyrolysis at different temperature [7].

Figure 2.3: The pathways of lignin pyrolysis at different temperature [7].

22
Vassilev et al. [14] made an extensive compilation of chemical characteristics and structural compo-
nents composition. A total of eighty-six species of biomass were examined. The different species of
biomass were grouped and categorized as such:

1. Wood and Woody Biomass (WWB)

2. Herbaceous and Agricultural Biomass (HAB)

(a) Herbaceous and Agricultural Grasses (HAG)

(b) Herbaceous and Agricultural Straws (HAS)

(c) Herbaceous and Agricultural Residues (HAR)

3. Aquatic Biomass

4. Animal and Human Biomass Wastes

5. Contaminated Biomass and Industrial Biomass Wastes (Semi-biomass)

6. Biomass Mixtures

Table 2.1 shows the chemical and structural composition of 47 species of distinct groups. Impor-
tant conclusions can and must be withdrawn from such data. Relative weight of major components or
constituents are compared and presented in a decreasing order as follows:

• Moisture - WWB > HAB > HAS

• Volatile matter - WWB > HAB > HAS

• Ash - HAS > HAB/WWB

• Carbon - WWB > HAB > HAS

• Oxygen - HAS > HAB > WWB

• Hydrogen - WWB/HAB > HAS

• Cellulose - HAB > HAS > WWB

• Hemicellulose -WWB > HAS > HAB

• Lignin - WWB > HAB > HAS

• Extractives - HAS > HAB > WWB

The ash content is a very influential parameter in the pyrolysis process. As shown above, HAS
has more ash content than both HAB and WWB. High ash contents are undesirable for the catalytic
reactions it instigates, leading to numerous negative factors such as reduction of oil yield, increase
in particle emissions and even possible release of dangerous and highly corrosive compounds [47].
Giudicianni et al. [47] made and extensive review on the effects of inherent metal elements in biomass
pyrolysis, especially the effects of alkali and alkaline earth metals (AAEMs) during the process and the
predicted effects of its presence in pyrolysis products.
Once full biochemical analysis is performed, it is possible to directly derive from the data the biomass
composition in terms of its cellulose, hemicellulose, lignin, moisture, and ash content. However, when

23
Table 2.1: Chemical and structural composition of various biomass groups and sub-groups [14].

Biomass WWB mean HAB mean HAS mean


Proximate analysis (wt.%, am)
Moisture 4.7 – 62.9 19.3 4.7 – 62.9 19.3 7.4 – 16.8 10.2
Volatile Matter 30.4 – 79.7 62.9 41.5 – 76.6 66.0 58.0 – 73.9 66.7
Fixed Carbon 6.5 – 24.1 15.1 9.1 – 35.3 16.9 12.5 – 17.8 15.3
Ash 0.1 – 8.4 2.7 0.1 – 8.4 2.7 4.3 – 18.6 7.8
Ultimate analysis (wt.%, daf)
C 48.7 – 57.0 52.1 42.2 – 58.4 49.9 48.5 – 50.6 49.4
O 32.0 – 45.3 41.2 34.2 – 49.0 42.6 40.1 – 44.6 43.2
H 5.4 – 10.2 6.2 3.2 – 9.2 6.2 5.6 – 6.4 6.1
N 0.1 – 0.7 0.4 0.1 – 3.4 1.2 0.5 – 2.8 1.2
S 0.01 – 0.42 0.08 0.01 – 0.60 0.15 0.08 – 0.28 0.15
Structural components (wt.%, daf)
Cellulose 12.4 – 65.5 39.5 23.7 – 87.5 46.1 18 – 54.8 45.4
Hemicellulose 6.7 – 65.6 34.5 12.3 – 54.5 30.2 18 – 39 31.5
Lignin 10.2 – 44.5 26.0 0.0 – 54.3 23.7 14.9 – 35.3 23.1
Extractives (wt.%, daf) 1.0 – 9.9 3.1 1.2 – 86.8 13.7 3.8 – 21.7 13.6
Ash analysis (wt.%, db)
Cl 0.01 – 0.05 0.02 0.04 – 0.83 0.21 0.03 – 0.64 0.41
SiO2 1.86 – 68.18 22.22 8.73 – 84.92 46.18 7.87 – 77.2 43.94
CaO 5.79 – 83.46 43.03 2.98 – 44.32 11.23 2.46 – 30.68 14.13
K2 O 2.19 – 31.99 10.75 2.93 – 53.38 24.59 12.59 – 38.14 24.49
P2 O5 0.66 – 13.01 3.48 3.14 – 20.33 6.62 0.98 – 10.38 4.13
Al2 O3 0.12 – 15.12 5.09 0.67 – 2.59 1.39 0.1 – 5.57 2.71
M gO 1.1 – 14.57 6.07 1.42 – 8.64 4.02 1.67 – 14.1 4.66
F e2 O3 0.37 – 9.54 3.44 0.58 – 1.73 0.98 0.41 – 2.82 1.42
SO3 0.36 – 11.66 2.78 0.83 – 9.89 3.66 1.18 – 4.93 3.01
N a2 O 0.22 – 29.82 2.85 0.09 – 6.2 1.25 0.16 – 3.52 1.35
T iO2 0.06 – 1.2 0.29 0.01 – 0.28 0.08 0.02 – 0.33 0.16
Mn (ppm) 775 – 35740 13160 – 3100 155 – 2790 865

only elemental composition, being carbon (C), hydrogen (H) and oxygen (O) content, is available, the
composition in terms of cellulose, hemicellulose and lignin can be derived from the atomic balances
of carbon, hydrogen and oxygen. Two methods can be used to estimate the organic composition of
biomass: correlation method and triangulation method [48, 49]. The correlation method considers only
cellulose, hemicellulose and lignin, including the extractives content in the latter. In contrast to the
first method, the triangulation considers cellulose, hemicellulose, lignin and hydrophilic and hydrophobic
extractives.

2.1.1 Correlation Method

The correlation method can be used to estimate cellulose (CELL), hemicellulose (HCE) and lignin (LIG).
LIG is calculated by difference, while CELL and HCE are estimated with 90 % and 81 % precision,
respectively. For this the following equations are used where O/C and H/C are the molar fractions and

24
VM is the volatile matter in wt.%, dry ash free basis (daf) [49]:

CELL = −1019.07 + 293.810(O/C) − 187.639(O/C)2 + 65.1426(H/C)

− 19.3025(H/C)2 + 21.7448(VM) − 0.13212(VM)2 (2.1)

HCE = 612.099 + 195.366(O/C) − 156.535(O/C)2 + 511.357(H/C)

− 177.025(H/C)2 − 24.3224(VM) + 0.1453063(VM)2 (2.2)

LIG = 1 − CELL − HCE (2.3)

This method was developed for a large range of biomass samples with molar ratios O/C from 0.56 to
0.83, H/C from 1.26 to 1.69, and volatile matter from 73 % to 86 % daf [49].

2.1.2 Triangulation Method

The triangulation method can be used to estimate the contents of the major components of biomass
based on its elemental composition (C/H/O) in wt.%. This method considers seven different reference
components: hemicellulose (HCE), cellulose (CELL), three structural lignins (LIG-C, LIG-H and LIG-O),
hydrophilic extractives (TANN) and hydrophobic extractives (TGL). The organic composition is calcu-
lated through correlations derived using a non-linear regression. Correlations for grass and woody
lignocellulosic biomass were developed to predict values for the reference components using the ele-
mental composition, specifically from the mass balances of C, H and O [48]. Correlations for woody
lignocellulosic biomass are as shown below:

CELL = 11.814 − 46.121C + 47.910C 2 + 33.934H + 22.352H 2 − 79.165C · H (2.4)

HCE = 5.124 − 22.635C + 23.302C 2 + 25.813H + 19.902H 2 − 34.972C · H (2.5)

LIGH = −14.550 + 61.143C − 70.524C 2 − 75.695H − 49.969H 2 + 212.321C · H (2.6)

LIGO = −16.513 + 72.616C − 75.716C 2 − 16.957H − 70.433H 2 + 19.080C · H (2.7)

LIGC = 18.745 − 84.434C + 96.927C 2 + 68.397H + 35.378H 2 − 165.043C · H (2.8)

T GL = −1.574 + 4.203C − 3.779C 2 + 10.470H + 10.271H 2 − 7.296C · H (2.9)

T AN N = −2.045 + 15.229C − 18.120C 2 − 45.961H + 32.499H 2 + 55.075C · H (2.10)

Correlations for grassy lignocellulosic biomass are as shown below:

CELL = 6.153 − 21.875C + 20.483C 2 + 6.803H + 3.925H 2 − 11.744C · H (2.11)

25
HCE = .178 − 18.098C + 18.187C 2 + 30.983H + 7.512H 2 − 42.778C · H (2.12)

LIGH = −1.411 + 4.574C − 3.935C 2 + 5.047H + 10.413H 2 − 7.619C · H (2.13)

LIGO = −10.653 + 53.328C − 62.901C 2 − 85.335H − 54.471H 2 + 173.332C · H (2.14)

LIGC = 12.623 − 52.992C + 56.568C 2 + 6.285H − 2.564H 2 − 20.184C · H (2.15)

T GL = −3.898 + 11.971C − 10.847C 2 + 14.575H + 3.230H 2 − 7.411C · H (2.16)

T AN N = −5.993 + 23.091C − 17.554C 2 + 21.642H + 31.954H 2 − 83.595C · H (2.17)

Additionally, since the use of the specific correlations is limited to a small portion of the existing biomass
compositions, three extra reference species are defined as linear combinations of the reference compo-
nents CELL, HCE, LIG-C, LIG-H and LIG-O, therefore, broadening the range of biomass types in which
the method is useful and can be applied [10].

2.2 Thermogravimetry

Thermogravimetry is a thermal analysis method, in which typically, weight changes and heat rates of a
sample can be measured while time and temperature are controlled, in an atmosphere of nitrogen (N ),
oxygen (O), argon (Ar), helium (He), air, other gases, or vacuum. Thermogravimetric (TG) analysis, or
Differential Scanning Calorimetry (DSC), is extremely useful to analyse any exothermic or endothermic
reaction that can be associated to weight loss, such as thermal degradation and decomposition. To
study the process of pyrolysis, TGA is helpful since it also provides information on the composition of
the sample and on the effects of atmosphere [50]. The procedure starts with small sample being placed
in a crucible which is then placed on a high precision balance inside the furnace. The temperature is
measured by a thermocouple placed near the crucible and, for the purpose of pyrolysis, an inert gas,
usually nitrogen (N ) or argon (Ar), is injected in the furnace as temperature changes to create an inert
atmosphere. Temperature, time and mass variations are stored and processed into Thermogravimetric
(TG) and Derivative Thermogravimetric (DTG) curves. TG and DTG profiles are mass loss and mass
loss rate curves, respectively, and are then studied where reactions of the main components can be
identified. Olatunji et al. [50] made a review of the process which include further information on the
usefulness and potential of the TGA method. An example of a TGA device is illustrated in Figure 2.4
and, in Figure 2.5 a typical TGA is exemplified. In the latter, the degradation of the different organic
components of biomass is illustrated.

26
Figure 2.4: Example of a TGA device [8]. Figure 2.5: Example of a thermogravimetric analysis [9].

2.3 Pyrolysis Kinetics

In order to improve efficiency and to maximize the potential of pyrolysis and its products, knowledge
of pyrolysis kinetics is essential. In simple terms, pyrolysis kinetics stands for the study of reaction
rates, order of reaction and other influential parameters to the thermal degradation evolution. Given the
large conditions in which pyrolysis is performed and diversity of biomass, most data on kinetic rates is
either unreliable or even inadequate, when parameters differ too much. Therefore, both empirical and
predictive pyrolysis models were developed in order to better predict these parameters and end prod-
ucts. Empirical models are based in apparent kinetics, using estimation methods to find global kinetic
parameters. Predictive models, on the other hand, attempt to describe the decomposition process in
detail without any fitting procedure. In the literature, different empirical models, with different degrees of
complexity, can be found. The simplest is the single first order reaction model (SFOM), which considers
a single decomposition stage and a single reaction to model pyrolysis. Further on the spectrum of com-
plexity, there are the three parallel reactions model (3PM) and five parallel reactions model (5PM), which
use three and five single order parallel reactions to describe the decomposition of cellulose, hemicellu-
lose and lignin, and cellulose, hemicellulose, lignin, hydrophilic extractives and hydrophobic extractives,
respectively [9, 23, 39, 46]. The reaction rate constant of each component/decomposition stage, is de-
scribed by the Arrhenius law:
Ei
ki (Tp ) = Ai TpΥ exp(− ) (2.18)
RTp

where ki (Tp ) is the reaction rate constant of the ith component, Ai the pre-exponential factor (s−1 ), Υ
is the temperature coefficient of the Arrhenius law, Ei the activation energy (J.mol−1 ) and R is the ideal
gas constant (J.K −1 .mol−1 ) [51].
On the other end of the spectrum, the most complex, updated and accurate model is the Chemical

27
Reaction Engineering and Chemical Kinetics – Pyrolysis Mechanism (CRECK-PM). This mechanism,
as a consequence of its modular nature, is able to store highly detailed models of important functional
groups, creating kinetic mechanisms that reach hundreds of species and thousands of reactions. In
sum, similar species with similar behavior are lumped into pseudo components, reducing the number of
reaction pathways while retaining a detailed, descriptive and accurate scheme [52]. CRECK-PM con-
siders biomass to be a combination of cellulose (CELL), hemicellulose (HCE), lignin (LIG), hydrophilic
extractives (TANN) and hydrophobic extractives (TGL). Referenced components are considered to stan-
dardize and normalize the otherwise complex structures of organic components. Particularly, due to its
complex and irregular structure, LIG is described, depending on its elemental composition (C,H and
O), as three functional groups: lignin-C, lignin-H, and lignin-O, which are rich in carbon, hydrogen and
oxygen respectively. Table 2.2 shows the chemical formula and the elemental mass fraction in terms of
C, H and O [15].

Table 2.2: Chemical formula and C/H/O mass fraction of the reference components considered in the pyrolysis
CRECK-PM [15].

Component Hemicellulose Cellulose Lignin-C Lignin-H Lignin-O Triglycerides Tannins


Formula C5 H8 O4 C6 H10 O5 C15 H14 O4 C22 H28 O9 C20 H22 O10 C57 H100 O7 C15 H12 O7
C (wt.%) 45.5 44.4 69.8 60.6 56.9 76.3 59.3
H (wt.%) 6.1 6.2 5.4 6.4 5.2 11.2 3.9
O (wt.%) 48.4 49.4 24.8 33 37.9 12.5 36.8

Overall, the CRECK-PM mechanism returns information on the decomposition of biomass into solid
residue, condensable volatiles and permanent gases from the provided biomass organic composition,
using 50 considered species and 29 reactions to describe the pyrolysis processes of the reference com-
ponents. Also, it is considered that catalytic reactions and interactions between reference components
do not happen. Nonetheless, the mechanism present further advantages such as allowing accurate
predictions for diverse conditions since competing reactions are considered. Figures 2.6 through 2.9
show the multi-step mechanism of the pyrolysis of each reference component. The G{...} species
represented in each scheme, apart from the TGL’s scheme, correspond to lumped pseudo species,
previously condensed or captured in the solid matrix, that are released as volatiles, during secondary
pyrolysis reactions, at high temperatures [15]. Additionally, the full list of considered species, reactions,
thermodynamic and transport properties, and kinetic parameters can be found in [53].

Figure 2.6: Scheme of the cellulose (CELL) pyrolysis multi-step mechanism [10].

28
Figure 2.7: Scheme of the hemicellulose (HCE) pyrolysis multi-step mechanism [10].

Figure 2.8: Scheme of the lignin (LIG) pyrolysis multi-step mechanism [10].

Figure 2.9: Scheme of the hydrophilic extractives (TANN) and hydrophobic extractives (TGL) pyrolysis multi-step
mechanisms [10].

2.4 Optimization Methods

The use of optimization methods is important to find and estimate organic composition values. In order
to accurately and quickly validate the experimental values for the organic composition of biomass, a 2-
step fitting procedure is applied in which two MATLAB solvers operate sequentially: the genetic algorithm
(GA) and the multidimensional constrained nonlinear minimization (fmincon).
The genetic algorithm (GA) is a method that mimics the natural selection process and to function
properly, it usually needs a genetic description of the solution domain and a fitness function, used to
evaluate the fitness level of the solution domain. The GA is a iterative process, and can be set to
find global minimums for nonlinear problems. Within each iteration, fitness levels of the population are
analysed and some are selected to be used as parents to create the next generation. Typically, the
selection of the parents is made through a filter, where the individuals with better fitness levels are

29
more likely to be selected. After the selection, the children of the next generation are generated. Elite
individuals, with the best fitness values, automatically are generated as children in the next generation,
while the rest of the parents population is mixed and matched, by means of a process called crossover,
to create children with mixed characteristics. Children from the new generation may also present random
characteristics called mutations and the generation is made until an adequate population size is reached.
Overall, this process ensures that the mean fitness level continues to increase with each iteration while
maintaining genetic diversity in the gene pool. The algorithm ends when it reaches a pre-defined criteria
(e.g. maximum number of generations) or the conversion of the objective function to a predetermined
minimum value is reached.
The fmincon is an optimization method that solves constrained nonlinear multivariable problems by
finding the minimum for an objective function through a methodical directional analysis.
This method is based on an initial guess, therefore, instead of a random initial guess is made, the GA
output is used to provide a competent start to the process. The use of a two-step optimization method
was applied in previous studies to reach accurate and reliable results [54, 55].

30
Reactor Design
3
Contents
3.1 Feeding System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

3.2 Auger System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

3.3 Collection System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

3.4 Configuration of the Experimental Setup . . . . . . . . . . . . . . . . . . . . . . . . . 45

31
32
In this chapter it will be presented the project design of a double auger reactor. Its overall configura-
tion represents a typical double auger reactor design and it is based on previous studies [56–59]. The
starting point of the reactor design was the heat source. The heat source of the system is a TERMOLAB
electric oven. It represents the foundation of the project since it was a requirement that the components
of the reactor were to be designed around it. The oven is capable of reaching a maximum of 1200 ◦ C
and 6 kW of heating power. Figure 3.1 illustrates the shape and dimensions of the oven and Figure 3.2
is a the real oven. The heat supplier is always a critical component in pyrolysis reactors, however in
this case it is the backbone of the project, thus, its dimensions are one of the most important physical
constraints for the design of the Auger System.

Figure 3.1: Scheme of the electric oven including dimensions in mm.

Figure 3.2: View of the electric oven.

33
For the purpose of simplicity and organization, this chapter will be split in three sections: the feeding
system, the auger system and the collection system. Each branch of this project will be explained and
described in the remainder of this chapter.

3.1 Feeding System

The project of the double auger fast pyrolysis reactor begins with the feeding system. This system is
responsible for accurately and continuously feeding biomass to the auger reactor system. The feeding
of biomass must be reliable, measurable, stable and controllable. In addiction, it also must be airtight
to unsure a inert atmosphere inside the reactor. In this system, nitrogen (N2 ) is used to create the inert
atmosphere for pyrolysis to be performed and, in addiction, nitrogen (N2 ) also acts as a carrier gas to
transport the released volatiles out of the reactor. Figure 3.3 shows a real image of the feeding system.

Figure 3.3: View of the Feeding System.

The system begins with a sealed hopper of 1.5 dm3 capacity, where pre-treated biomass is deposited.
The hopper is connected to an electrical vibrating system which drops the biomass through a tube and
into the auger system. The vibrating system consists of a inclined vibrating plate that provides a biomass
feed rate depending on the vibrating power selected. The system reaches a maximum of 80 W of power
and 3000 vpm (50 Hz). If necessary, the plate may be replaced by a larger plate and assembled to
the system to provide a higher mass feeding rate. Furthermore, the whole system is placed on top of
a scale, which allows real-time mass feeding rate measurements. This is specially important, not only
to enable a correct analysis of the pyrolysis product’s distribution after the process, but also to allow the

34
coordination between of the biomass feeding rate to the auger rotational speed, preventing malfunction
and poor performance. Additionally, both the lid of the hopper and the tube that is connected to the
auger system are made from clear plastic polymer material and are equipped with gas inlets, where
nitrogen can be injected. This is to allow visualization and supervision over the process and to provide
a slight positive relative pressure, discouraging backflow at the tube and assuring smooth operational
performance. Below, the feeding system is illustrated by Figure 3.4 and Table 3.1 describes the important
components present in the system schematic.

Figure 3.4: Biomass feeding system schematic.

Table 3.1: Biomass feeding system description.

Component Description
1 N2 container
2 Vibration system controller
3 Biomass hopper
4 Vibration system support
5 Vibration plate
6 Clear polymer tube
7 Airtight container
8 Connection with the Auger System
9 Inert gas valve
10 Scales
11 N2 gas flow meter

35
3.2 Auger System

The double Auger System includes a variety of components. Figure 3.5 illustrates broadly the system
and Table 3.2 shortly describes each of the components.

Figure 3.5: Auger system schematic.

Table 3.2: Auger system description.

Component Description
1 Electric motor
2 Coupling system
3 Clear polymer Feeding System connector
4 Electric heating resistances
5 Double counter rotating screws
6 System support
7 Volatiles exit to Collection System
8 Char exit to Collection System
9 Casing support

As previously mentioned, to successfully perform fast pyrolysis and obtain the desired products,
certain conditions must be satisfied. In section 1.2, a brief review of the optimal parameters for fast
pyrolysis was performed. Thus, for the construction and assembly of most of the components in the
Auger System several parameters must be considered, while achieving these operational conditions.
Firstly, similarly to any other auger reactor, given its mechanical nature, the auger or augers are the
focal point of the system. The auger’s properties and design directly influence residence times, and heat
and mass transfer coefficients, which are essential to reach high bio-oil yields as well as reassuring the
products’ quality. Without proper design the reactor may display unwanted effects such as unsteady
flow rate, inaccurate metering and dosing, heterogeneity of each byproduct, product degradation (due
to secondary reactions), high start-up torque, high equipment wear, variable residence time and high
energy consumption [60]. The typical main auger physical characteristics are the outside and in-side
(shaft) diameters, the pitch (distance between adjacent flights), the clearance and the length. To extract
the potential of the seemingly simple auger configuration for fast pyrolysis applications, a number of
parameters must be considered [11]:
• Pitch-to-diameter ratio

36
• Number of augers (single or double)

• Flight geometry

• Mixing properties

• Auger rotation orientation

• Operating conditions (residence times, rotation speed, volumetric fill level)

Several of these factors are not completely independent from each other, which means, between
some, a compromise must be made to improve the reactor performance. This emphasizes the impor-
tance of case-by-case approach and that auger design is not a simple or one-size-fits-all process. As
previously stated, heat transfer is the bottleneck of conventional pyrolysis. Also, the ability for fluidized
bed reactors to achieve high heating rates is one of the only advantages over auger reactors. Therefore,
of the above mentioned parameters, those which tend to increase and facilitate higher heating rates
(e.g. mixing properties and number of augers) were regarded as most important. In the next sector of
this section, the importance of the parameters to consider in the design will be briefly explained.
First and foremost, double auger reactors represent real and considerable advantages over single
auger reactors, i.e. the rotation of the intermeshing screws prevents adhesion and clogging of feedstock,
higher percentages of biomass conversion are reached since higher levels of mixing are obtained and,
therefore, shorter residence times may be established for the same degree of biomass conversion.
Hence, secondary reactions are minimized and bio-oil yields are maximized, which closes further the
bridge between auger and fluidized bed reactors for fast pyrolysis aplications. Furthermore, twin-auger
configurations also represent lower power requirements and overall higher devolatilization levels. In the
available literature, the optimal pitch-to-diameter ratio is reviewed and analysed but a precise consensus
is not reached. However, most authors report on a margin for the dimensionless pitch that typically is
around 1 [11]. In addition, Camp et al. [61] noticed that radial mixing can be further improved with a
pitch-to-diameter ratio between 0.25 and 0.5. Kingston et al. [62] applied a four-way ANOVA procedure
to study the importance and statistical influence of the screw rotation speed, pitch-to-diameter ratio,
screw rotation orientation and material injection configuration on granular flow mixing in a double screw
mixer. Both screw rotation orientation and pitch-to-diameter ratio were found to be the parameters
which provide the best degree of mixing, while screw rotation speed had minimal effects on the level of
mixing. The overall best configuration was a counter-rotating down-pumping screw rotation orientation
with a dimensionless pitch (pitch-to-diameter ratio p/D) of 1.75 at a rotation speed (ω) of 60 rpm. Qi et
al. [63] also observed that, in his study, the particle distributions did not differ with different screw rotation
speeds when the volumetric fill level remained constant. Load factor, or volumetric fill level, is the ratio
between the volume occupied by the biomass and the volume of the screw pitch [64]. Auger reactors and
conveyors typically operate with a load factor between 0.15 and 0.45, in search for a balance to efficiently
transport biomass, since at small load factors the biomass flow rate is too limited and at high load factors

37
the power requirements peak strongly [11]. Qi et al. [63] modeled the heat transfer in granular flow in
a double screw reactor. A DEM particle-scale heat transfer model was developed which considered
conductive and radiative heat transfer between particles. Additionally, another model was also proposed
and validated to resolve the particle-wall conductive and radiative heat transfer. It was verified that
the average temperature of the biomass increases as the volumetric fill level decreases. The most
common geometry of an auger conveyor constitutes a shaft with a continuous helix, with an arbitrary
pitch, around it. However, many different designs exist for the auger flight geometry. Different process
requirements and final utilization may lead the geometry to change accordingly. In auger pyrolyzers,
few to no studies or information have been completed and gathered on the effects of different flight
geometries on pyrolysis, therefore, standard flight remains the most common choice. The simplicity,
trustworthiness and robustness of the design also denote and justify the popularity of the standard flight
for auger reactors [11]. In the section 1.2 residence times were also mentioned. However, in that section,
focus was directed towards gas residence times. While gas residence time is mostly a product of the
sweeping gas flow rate, the solid residence time, in auger pyrolyzers, is essentially a relation between
the biomass feeding flow rate plus the screw rotation speed, and the length of the heated zone. In fast
pyrolysis, solid residence times must be enough to ensure total biomass conversion as long as the inert
sweeping gas flow rate is calculated and used to control gas residence time and minimize secondary
pyrolysis reactions [2, 43, 65, 66]. In appendix A, both residence times are calculated taking into account
the geometry properties of the system and are displayed in Tables A.1 and A.2.

Table 3.3: Estimated solid residence times for different auger rotation speed.

Augers rotation speed (rpm) 10 20 40 60


Solid residence time (s) 129 64 32 21

At the control temperature of 500 ◦ C At the control temperature of 550 ◦ C


Target gas residence time (s) Volumetric gas flow rate (L/min) Volumetric gas flow rate (L/min)
at f=10 % at f=36.7 % at f=10 % at f=36.7 %
1 0.87 0.62 0.87 0.62
5 0.17 0.12 0.17 0.12
10 0.09 0.06 0.09 0.06
30 0.03 0.02 0.03 0.02

Table 3.4: Required volumetric gas flow rates at the N2 flow meter to reach target gas residence times at the control
temperature of 500◦ C and 550◦ C, and volumetric biomass filling rate (f) of 10 % and 36.7 %.

Ultimately, mixing properties and other geometric design parameters, and operational conditions
must be considered simultaneously to ensure the global fast pyrolysis parameters. At this point, the con-
struction and feasibility constraints must also be contemplated for the whole process to be successful.
Thus, pitch-to-diameter ratio, screw rotation speed, solid residence time, volumetric fill level are parame-
ters that must be modified and systematically readjusted to provide the best reactor performance within
the 50 cm of maximum predicted length of the heating zone, the feeding system range of biomass flow

38
rate, the electric motor rotation speed and power, and the practicality of construction of the necessary
components. In the remainder of this section, the components which were designed and built for the
auger system are briefly described.
The reactor was designed to minimize the amount of biomass that can be trapped and accumulated
inside the casing while in operation, consequently increasing efficiency and decreasing maintenance
requirements. Thus, the casing, inside of which the augers rotate, was constructed from two tubes.
The tubes were cut and welded to each other so that a gap of 5 mm would distance the edge of the
augers and the inside wall of the casing. This gap helps prevent the augers scraping the walls of the
casing, which would decrease the longevity of both parts, but also to prevent the screws from jamming or
clogging if biomass of a bigger size were to be used in the reactor [11]. Figure 3.6 shows the dimensions
of the casing.

Figure 3.6: Technical drawing of the casing of the auger system.

The augers must be coupled with the motor at the upstream end and be supported by the bearings
at both ends. Therefore, the total length of the shaft of each auger is 86 cm, of which 67.5 cm of them
are threaded. Given the length of the augers, each one of them had to be split into three and machined
one at a time before being assembled with dowel pins. Each section of each auger was milled out of a
cylinder of AISI 316 stainless steel in a DMG CTX 310 ecoline machine. This material was chosen given

39
the design conditions it will be subjected to while in operation, i.e. high temperatures, heavy contact with
water and highly corrosive substances as it was observed in previous studies [11, 38]. The pitch of the
thread is constant and equal to 2.8 cm and the geometry of the thread is described in Figure 3.7 along
with the main dimensions of the screw. The only difference between each screw is the orientation of the
thread along the screw, since it is required that one screw presents a right-handed thread and another
presents a left-handed screw to create a counter-rotating down-pumping movement. Figure 3.8 details
the assembly of each auger.

Figure 3.7: Technical drawing of the first sector of the right-handed auger.

Figure 3.8: Technical drawing of the assembly of the auger.

For the designed screw rotation speed, i.e. from 0 to 60 rpm, fast pyrolysis residence times can
be reached with this auger geometry. In addition, at these rotation speeds, this auger geometry grants
the possibility of operating at various volumetric filling rates, throughout the biomass mass feeding rate
range provided by the feeding system, which may contribute for better overall results.
To complete the setup of the Auger System, other accessory components were designed, such as:

• Closing covers, for upstream and downstream

40
• Bearing holders, to support and fix the bearings of the augers

• Connection between the Auger and Feeding Systems

• Supporting element for the casing

The closing covers were designed to allow the connection of the auger’s shaft with the bearing and
with the motor upstream. Downstream, the closing covers, were to concede the shaft to reach the
bearings, but also to allow the biomass byproducts to exit the reactor. Therefore, the downstream cover
has an opening for the solid products to exit into the Collection System and two holes of 1cm of diameter
for the volatiles to exit into the Collection System. The upstream cover consists of two identical parts
that are then screwed to the casing. The downstream cover consists of a top part and a bottom part.
Figures 3.9 and 3.10 show, respectively, the dimensions of the closing covers of the system.

Figure 3.9: Technical drawing of the upstream closing cover.

41
Figure 3.10: Technical drawing of the downstream closing top part (left) and the downstream closing bottom part
(right).

The rotation of the augers is supported by one K6809-ZZ Koyo bearing in each end of each auger
shaft. Bearing housings were design to support and protect the bearings. Each housing consists of two
parts, a bottom one and a top one, which are screwed together. Figures 3.11 and 3.12 illustrate the
dimensions and geometry of the top and bottom parts of each of the two bearing housings, respectively.
The technical drawings of the connection between the Auger and Feeding Systems component and the
supporting element for the casing can be consulted in appendix B.

Figure 3.11: Technical drawing of the top part of the bearing housings.

42
Figure 3.12: Technical drawing of the bottom part of the bearing housings.

For better upstanding of the assembly of the Auger System, an exploded view is illustrated in Figure
3.13 followed by the list of components in Table 3.5:

Figure 3.13: Exploded view of the Auger System.

43
Table 3.5: Component description of Figure 3.13

Component Description
nº1 Augers
nº2 Casing
nº3 Connection casing-Feeding System
nº4 Upstream cover (top half)
nº5 Upstream cover (bottom half)
nº6 Spur gears
nº7 Upstream bearing housing (top part)
nº8 Upstream bearing housing (bottom part)
nº9 Bearings
nº10 Downstream cover (top part)
nº11 Downstream cover (bottom part)
nº12 Downstream bearing housing (top part)
nº13 Downstream bearing housing (bottom part)
nº14 Casing supports (stabilizers)

3.3 Collection System

The collection system is connected at the exit of the auger and begins with the insertion of a cyclone
to ensure separate paths for the solid products and the volatiles. The bio-char is collected directly
after it reaches the end of the auger into a char container, whereas volatiles are lead through a triple
condensation setup to collect the bio-oil. The refrigerant fluid used in the condensers is a mixture of
ethylene glycol with water and it is kept at 5 ◦ C. The refrigerated circulator is a model Haake C10-K15. It
has a maximum flow rate capacity of 12.5 L/min, a cooling capacity of 200 W at 0 ◦ C and a temperature
accuracy of 0.04 ◦ C. The remaining gas flow goes through a cotton filter and a silica gel trap to ensure
that the remaining condensables are collected and to dry as much as possible the outflow gas. Finally,
the gas products can be measured using a paramagnetic pressure analyzer for O2 , a non-dispersive
infrared analyzer for CO2 and CO, while the remaining gas products can be collected into sampling
bags to be posterior analysed in a Claurus 500 gas chromatograph. Excess gas products are lead
into the exhaust system. The collection system is illustrated in Figure 3.14 and Table 3.6 labels its
components.

Table 3.6: Collection system description.

Component Description
1 Cyclone
2 Volatiles collection path
3 Char collection container
4 Condenser
5 Volumetric glass flask
6 Sampling gas bag
7 Exhaust

44
Figure 3.14: Collection system schematic.

3.4 Configuration of the Experimental Setup

The overall reactor configuration is an assembly of all three systems mentioned in sections 3.1, 3.2 and
3.3. Figure 3.15 illustrates a schematic of the double auger setup used for the experiments. Apart from
the elements already mentioned, Table 3.7 labels other important elements in the configuration of the
reactor.

Figure 3.15: Reactor configuration schematic.

With this configuration, previously mentioned important pyrolisys parameters, i.e. temperature, solid
and volatiles residence times, can be controlled. The temperature is constantly monitored through two
type-K thermocouples (Cr+, Al−) from Omega distributed along the reactor wall: thermocouple 1 and
thermocouple 2. Their signal is processed by a NI-9211 Data Logger into temperature (◦ C). In order

45
Table 3.7: Reactor configuration description.

Component Description
1 Weight measurement data acquisition and motor control
2 Thermocouple 1
3 Thermocouple 2
4 Furnace temperature controller
5 Heavily insulated area

to estimate the error, the acquired temperature values were compared with a reference temperature
acquired by a high precision auxiliary thermocouple (Omega KMQSS-IM025U-300). The relative error
was found to be ±3 ◦ C. Thermocouple 1 is connected to the controller of the furnace which can be set to
temperatures up to 900◦ C. Thermocouple 2 is not fixed so that other temperatures at a slightly upstream
or downstream position can be measured. The solid residence time is a function of the auger rotation
speed and, therefore, can be altered through the motor controller. Additionally, the combination of the
biomass feeding rate of the vibratory system and the auger rotation speed allow experiment tests with
the same volumetric fill level but different biomass feeding rate and solid residence time. The volatiles
residence time is a function of the inert gas residence time, hence, a function of the inert gas mass flow
rate and can, therefore, be controlled with the nitrogen flow meter. This configuration design represents
overall advantages such as:

• Ability to reach high temperatures and high heating rates

• Parameter flexibility, which provide the option to tune the configuration for various types of pyrolysis

• Easy monitoring of the parameters

46
Materials and Methods
4
Contents
4.1 Biomass Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

4.2 Thermogravimetric Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

4.3 Pyrolysis Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

4.4 Objective Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

47
48
4.1 Biomass Characterization

The biomass chosen for the purpose of this work was pine wood. The biomass was presented in pellets
and, before every experiment or analysis procedure, it was milled with a Retsch Cutting Mill SM100. To
reduce the size of the pellets, two approaches were experimented: 1) a batch of biomass was milled
three times, twice with a sieve of 1 cm sided square holes and once with a sieve of 250 µm trapezoid
holes, and 2) a batch of biomass was milled four times, twice with the sieve of 1cm sided square holes
and twice with a sieve of 250 µm trapezoid holes. Afterwards, each batch was sieved separately using
a USP general test 786, method I, to establish a particle size distribution. Different mesh sieves are
weighted and stacked on top of each other, beginning with a container with the smallest mesh, followed
by increasingly coarser sieves. After the biomass is deposited on top, the sieves are then agitated for five
minute periods until the weight change of any of the sieves does not exceed a 5 wt.% change between
measurements.

The proximate analysis and organic composition of pinewood was performed and compared with
literature and with analytical results performed in Instituto Superior de Agronomia, in Lisbon. The bulk
density was also determined. A brief explanation of the standards and procedures used for the three
procedures is presented below.

Proximate analysis

Moisture content was determined according to standard EN 14774-3 for solid biofuels. The biomass
sample was dried in an oven at 105 ◦ C until mass stabilized. Stabilization is achieved once the weight
change does not exceed 1 mg between heating periods of 60 minutes at 105 ±2 ◦ C [67]. The following
equation is used to determine the moisture (M) content:

m2 − m3
M= · 100 (4.1)
m2 − m1

where m1 is the mass (g) of the empty crucible and lid, m2 is the mass (g) of the crucible and lid plus
test sample before drying and m3 is the mass (g) of the crucible and lid plus test sample after drying.

Volatile matter was determined according to standard EN 15148:2009 for solid biofuels. A test was
performed for this purpose were a biomass sample was heated for seven minutes at 900±10 ◦ C without
ambient air contact. To achieve these conditions, the air intake of the muffle furnace was closed and the
crucible is which the sample was heated was closed with a lid. The percentage of volatile matter (VM) is
calculated from the loss in mass of the test portion after deducting the loss in mass due to moisture [68],
as the following equation demonstrates:

m4 − m5
VM = · 100 − M (4.2)
m4 − m1

49
where m1 is the mass (g) of the empty crucible and lid, m4 is the mass (g) of the crucible and lid plus
test sample before heating, m5 is the mass (g) of the crucible and lid plus test sample after heating and
M corresponds to the pre-calculated mass moisture percentage.

Ash content was determined according to standard EN 18122 for solid biofuels. Ash mass percent-
age was calculated from the loss of mass after a particular heating procedure to which the sample was
submitted. The procedure included two heating zones, two temperature plateaus, the air intake was kept
open during the whole duration and the lid of the crucible was also taken out. Firstly, the sample was
heated to 250 ◦ C over a 30 minute period and kept at 250 ◦ C for 60 minutes to allow the volatiles to
be released before ignition. Next, the sample is heated to 550 ◦ C over a period of 30 minutes and this
final temperature is maintained for 120 minutes. The sample is then weighted and reheated to 550 ◦ C
for an additional 30 minute period. This last part of the procedure is repeated until the weight change
between measurements is under 0.5 mg [69]. The ash (ASH) mass percentage is finally calculated with
the following equation:
m8 − m6
ASH = · 100 − M (4.3)
m7 − m6

where m6 is the mass (g) of the empty crucible, m7 is the mass (g) of the crucible plus test sample
before heating, m8 is the mass (g) of the crucible plus test sample after heating and M corresponds to
the pre-calculated moisture mass percentage.

Fixed carbon FC percentage was determined by difference, as the following equation shows:

FC = 100 − M − VM − ASH (4.4)

where M, VM and ASH correspond respectively to the moisture, volatile and ash mass percentage.

Organic composition

For the organic composition, a series of sequential procedures were executed to determine hy-
drophilic extractives (TANN), hydrophobic extractives (TGL), lignin (LIG), hemicellulose (HCE) and cel-
lulose (CELL).

The extractives were determined using a soxhlet extractor in which the biomass was cyclically
washed with a solvent. A dried and weighted biomass sample (minimum of 3 g) was placed inside
an extraction thimble, which was then put inside an extraction chamber. A boiling flask was connected
to the bottom of the extraction chamber, which was then connected to a condenser. After the heater
reached the boiling point of the solvent, the vapours would exit the boiling flask and would condense into
the extraction chamber. Once the chamber was full, the solvent would flow back into the boiling flask
and the process would restart. The process ran cyclically and continuously for a determined amount of
time depending of the solvents. The setup was kept working for 11 hours with demineralized water and
6 hours for ethanol and hexane. After each of the three procedures, the sample was dried and weighted.

50
The mass loss measured after the demineralized water and ethanol procedures was attributed to TANN
and the mass loss observed after the hexane procedure was attributed to TGL [70].
To determine the lignin content, a weighted sample of extractive-free biomass (2.5 g) was used. Dur-
ing the procedure, the temperature was kept at 70◦ C. Firstly, the biomass was mixed with demineralized
water (80 ml), acetic acid (0.5 ml) and sodium chlorite (1 g). For the next six hours, sodium chlorite
(1 g) and acetic acid (0.5 ml) were added hourly. After the first 6 hours, the solution was kept at the
same temperature for 24 hours. After the procedure, the solution was washed with demineralized water,
filtered, oven dried and weighted. The lignin content was determined with the mass loss measured after
the procedure [71].
After the lignin and extractive-free biomass was dried, a weighted sample (1 g) was mixed with a
sodium hydroxide solution (0.5 mol/L) and heated to 80 ◦ C for 3.5 hours. Afterwards, the solution was
washed with demineralized water, filtered, oven dried and weighted. The mass lost due to the procedure
was deemed hemicellulose and cellulose was determined by difference [72].
Bulk density
For bulk density determination, a container with marked volumes was weighed and subsequently
tared. Biomass samples were then poured into the container from a constant height and levelled off.
The mass of the biomass on was recorded on a scale and this process was performed for four marked
volumes. After, a linear regression was used with the recorded values of weights and volumes to esti-
mate a bulk density value, ρ.

4.2 Thermogravimetric Study

The thermogravimetric tests of pine wood were performed in a Hitachi model STA 7200. To ensure
reliability and reproducibility of the results, two tests were performed. For each test, a biomass sample
was placed in a small platinum measuring crucible and was then heated in an inert atmosphere. The
heating profile was a ramp from room temperature (30 ◦ C) to 900 ◦ C at a constant rate of 5 ◦ C/min.
The inert gas used was nitrogen at a constant flow rate of 200 mL/min. For each second of the test, the
weight and temperature of the sample were measured, recorded and processed into TG and DTG form.
The sensitivity of the scales is 0.2 µg and the temperature precision is ±2 ◦ C.

4.3 Pyrolysis Modeling

For the pyrolysis modelling a 0-D model developed in a previous work was used. This model was
implemented in python using the Cantera reaction kinetics library and considers the CRECK-PM to
predict biomass pyrolysis behavior [10]. Its simpler design allows it to describe the conversion of biomass

51
in the absence of mass or heat transfer limitations, thus, it is assumed that an equilibrium is verified, in
every instance, in regards to operational conditions and products of pyrolysis. For the model described
in this section, the required inputs are the desired/experimental temperature profiles and the biomass
composition in macro-reference components (CELL, HCE, LIG-C, LIG-H, LIG-O, TANN and TGL). Also,
as mentioned in section 2.3, catalytic reactions and interactions between reference components are not
considered, and neither are spatial temperature gradients. In the CRECK-PM model, the considered
chemical reactions are first order reactions, thus, the rate of the reactions is directly proportional to the
concentration of the reacting species. Therefore, firstly, reaction rate constant (k) must be calculated in
order to then determine the rate of each reaction, and ultimately, the yields of the species. The reaction
rate constant (k) is given by the Arrhenius law, similarly to what was previously shown in equation 2.18,
for each reaction:
 
Υ Ea(rx)
k(rx, T ) = A(rx)T exp − , rx = 1, ..., Nrx (4.5)
RT
where A(rx) is the pre-exponential factor (s−1 ) which represents the number of particle collisions per
second in reaction number rx, Ea(rx) the activation energy (J.mol−1 ), Υ is the temperature coefficient
of the Arrhenius law and R is the ideal gas constant (J.K −1 .mol−1 ). The rx variable symbolizes the
number of one particular reaction and Nrx corresponds to the total number of reactions [10].

At the first time step, the model uses the inputted organic composition to set the mass fractions of
the species of the first iteration (composition of the species at t = 0). Afterwards, for each time step,
the total sum of yields is verified to be 100% (equation 4.6), the production of each species is calculated
(equation 4.7) and then integrated in time to determine its mass fraction (equations 4.8 and 4.9), as
demonstrated below:
Nsp
X
Y (sp, t) = 1 (4.6)
sp=1

where Nsp is the total number of species, sp is a variable that symbolizes the number of one particular
species, t is the time and Y (sp, t) is the concentration of the species sp and the time t.
  
Mw (sp)
ω̇(sp, t) = [[matrixrx,sp ] × [k(rx, t)]] × · Y (spsolid,react , t) (4.7)
ρg (t)

where ω̇(sp, t) is the production rate of species sp at time t, matrixrx,sp is a matrix comprised of stoi-
chiometric coefficients where the lines correspond to species and the columns correspond to reactions,
spsolid,react represents the solid reactant species, k(rx, t) is the reaction rate constant determined with
equation 4.5, Mw (sp) is the molar weight of the species and ρg (t) is the gas density.

Z tf
Y (sp, t) = ω̇(sp, t) · Y (spsolid,react , t) dt (4.8)
t

Y (sp, tj ) = Y (spsolid,react , tj−1 ) + ω̇(sp, t) · ∆t (4.9)

52
From these variables, thermogravimetric profiles, such as mass loss and mass loss rate can be
plotted in time after being determined through the following equations:

Nsp,solid
X
massloss = Y (spsolid , t) (4.10)
spsolid =1
Nsp,solid
X
masslossrate = ω̇(spsolid , t) (4.11)
spsolid =1

Ultimately, the yields of the pyrolysis products (wt.%) are determined as follows:

Nsp,solid
X
yieldchar = 100 · Y (spsolid , tend ) (4.12)
spsolid =1
Nsp,liquid
X
yieldliquid = 100 · Y (spliquid , tend ) (4.13)
spliquid =1
Nsp,gas
X
yieldgas = 100 · Y (spgas , tend ) (4.14)
spgas =1

On a final note, it is worth mentioning that equation 4.9 is valid only when the time interval ∆t =
tf − t is small enough to ensure the production of the species (ω̇) is constant. Due to the number of
reactions and species, a stiff ordinary differential equation solver is used (CVODE). The solver applied,
uses an implicit Euler method based on the finite difference approximating derivatives. Specifically, the
backward difference formula is used to estimate the slope of the function at tj using the line that connects
(tj−1 ,f (tj−1 )) and (tj ,f (tj )):
f (tj ) − f (tj−1 )
f ′ (tj ) = (4.15)
tj − tj−1

4.4 Objective Function

The main goal of the optimization procedure is to find the biomass organic composition that minimize
the deviation between experimental and predicted pyrolysis yields, while maintaining the characteristics
of the thermochemical conversion processes. To evaluate this goal, experimental and predicted values
of the initial temperature of the mass loss profile (Ti ), 100 points (equidistant in temperatures) from the
mass loss profile and the final yield of residual char (Y (spsolid , tend )) were used to elaborate an objective
function. Ti , which is the beginning of the thermogravimetric profile, was determined by searching the
temperature value where the first local maximum gradient of the mass loss rate profile occurs. The

53
objective function for each parameter (δi ) is defined as:

Pi,pred − Pi,exp
δi = (4.16)
Pi,exp

where Pi,exp is the experimental value of parameter i and Pi,pred is the predicted value of parameter
i. To calculate the global objective function, a weighted sum of the relative error of each parameter is
performed as shown:
X
δ= ln (1 + δi ) · wi (4.17)
i

where δi the relative error of each parameter and wi is the weight of the given parameter. The logarithm
was used to improve and enhance the convergence of the method [10].

54
5 Results

Contents
5.1 Biomass Composition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

5.2 Thermogravimetry Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

5.3 Kinetic Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

5.4 Preliminary Testing of Reactor Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62

5.5 Numerical Prediction of Pyrolysis Experiments . . . . . . . . . . . . . . . . . . . . . . 64

55
56
5.1 Biomass Composition

Initially, the biomass was in pellet form. Since average particle size is advised to be, at least, smaller
than 1 mm [11,35] (see section 1.2), a milling pre-treatment was performed. Table 5.1 exhibits the results
of the milling tests. Approach I shows good results, enough to undermine the usefulness of approach II
since the extra work shows no significant improvement and would decrease the efficiency of the process.
Hence, for the biomass used as feedstock in the reactor operation, approach I will be the chosen as the
milling pre-treatment.

Table 5.1: Results of USP general test 786 method I.

Weight content of each size interval(g)


Geometry of the opennings Sieve size opening
Approach I Approach II
Square holes s=1 mm 0.19 0.18
Trapezoid holes w=0.4 mm,d=0.25 mm 46.51 32.39
Trapezoid holes w=0.25 mm,d=0.16 mm 18.83 19.63
Final container (content smaller than final sieve) 34.42 47.81
Total weight before sieving 100.04 101.05
Total weight after sieving 99.95 100.01

Afterwards, proximate analysis and organic composition of the biomass was determined. These
values were later used as inputs on the kinetic model. In Table 5.2 the proximate and ultimate analy-
sis results obtained are compared to the results from an independent laboratory, which performed an
analysis of the pinewood sample and to previous bibliographic pinewood results from the literature. The
ultimate analysis values were obtained from the weight content distribution of the experimental organic
composition values. In the CRECK-PM mechanism, the molecular composition of each component is
described in number of carbon, hydrogen and oxygen atoms, thus, once the organic composition is de-
termined, the ultimate analysis can be estimated. In addition, the corresponding experimental standards
used for the analysis are also mentioned.

Table 5.2: Comparison of proximate and ultimate analysis of pinewood from different sources.

Parameter Experimental values Calculation methods Laboratory values Vassilev et al. [14]
Moisture (wt.%) 6.5 EN 14774-3 6.0 4.7<x<7.6
Proximate analysis (wt.%, dry basis)
Volatile matter 86.0 EN 15148:2009 84.0 72.4<x<73.7
Fixed carbon 13.6 Determined by difference 15.7 21.6<x<24.4
Ash 0.4 EN 18122 0.3 1.9<x<6.0
Ultimate analysis (wt.%, dry ash free)
Calculated from experimental
Carbon 54.68 54.2 51.0<x<53.8
organic composition values
Calculated from experimental
Hydrogen 5.85 7.3 5.9<x<6.3
organic composition values
Calculated from experimental
Oxygen 39.47 38.5 39.9<x<41.3
organic composition values

From Table 5.2, a cross examination between the values is possible. Significant differences are
verified between in-house experimental proximate analysis. The volatile matter is the parameter that

57
diverge the most, about 13.6 %, from the experimental value of 86.0 wt.% to from the bibliographic
value of 72.4 wt.%. Even when comparing two pinewood samples, the origin and shape of the sample
has great influence on these values, which justify the disparity acknowledged between the laboratory/in-
house experimental values and the bibliographic values. Nonetheless, proximate and ultimate values
match, at least, within approximately 2 % with the values of the same pine sample from the independent
laboratory. For the most part, the experimental values are in agreement with the rest of the values.

Table 5.3: Organic composition values of pinewood from different sources.

Organic components cellulose (CELL) hemicellulose (HCE) lignin (LIG) hydrophobic extractives (TGL) hydrophilic extractives (TANN)
Experimental values 35.77wt.% 9.32wt.% 44.33wt.% 0.86wt.% 9.72wt.%
Laboratory values 37.72wt.% 25.91wt.% 26.73wt.% 3.49wt.% 6.15wt.%

Table 5.3 compares the experimental in-house organic composition distribution with the values from
the independent laboratory. The mismatch between the LIG and HCE distribution values is due to an
error in the characterization method for the determination of LIG, in which after the fourth addition of
the solution of sodium chlorite and acetic acid, the LIG of the sample is already fully consumed and the
HCE begins to be disintegrated during the procedure, increasing the LIG calculated value at the cost of
the HCE content [73]. In addition, the values of LIG plus HCE in both results match within 1.01 % which
further confirms this effect. In the determination of the extractives, a mismatch between values is also
found. This discrepancy derives from the use of the different solvents in comparison to the method used
at the laboratory, which resulted in an inflated value for TANN and a deflated value for TGL. However,
the sum of extractives in the distinct procedures matched within 0.94 wt.%.

Once the distribution of the major organic components was established, the weighted percentage of
reference components of LIG, i.e. LIG-C, LIG-H and LIG-O, must be discovered in order to correctly
implement the kinetic model. The fitting procedure described in section 2.4 was used to minimize the
objective function described in section 4.4 to estimate the referred values. These results are shown in
Table 5.4.

Table 5.4: Full organic composition of the pinewood sample.

Component CELL HCE LIG-C LIG-H LIG-O TGL TANN


Distribution (wt.%) 37.72 25.91 0.22 19.01 7.5 3.49 6.15

The bulk density was calculated through the process described in section 4.1. The measured weights
and volumes are illustrated in Figure 5.1. After a linear regression was applied to the values, an estimate
of 0.4377 kg/dm3 was determined for the bulk density value (ρ) of the pinewood sample used in this
project.

58
Figure 5.1: Measured weight and volumes values for the estimation of bulk density (ρ) of the pinewood sample.

5.2 Thermogravimetry Experiments

Figure 5.2: Thermogravimetric (TG) and Derivative Thermogravimetric (DTG) mean results from the pinewood
samples.

Figure 5.2 shows the mean values of the two thermogravimetric analysis experiments described in
section 4.2. The mean sample weight was 5.99 mg of pinewood biomass. The profile shows a sharp
mass loss between 250 ◦ C and 450 ◦ C. Afterwards, the TG profile smooths out until the final tempera-
ture. At the final temperature of 900 ◦ C, the final residual weight recorded is 14.37 wt.%, which should
correspond to the fixed carbon content since it is not supposed to be consumed during a pyrolysis ex-

59
periment. This value matches the previously determined value from the experimental proximate analysis
procedure of 13.55 wt.%. In the DTG curve, at approximately 300 ◦ C and 375 ◦ C, the ”shoulder” and the
peak occur, which represents the degradation peaks of HCE and CELL, respectively. This is consistent
with previous studies [39, 41] since HCE is structurally the weakest major component.

5.3 Kinetic Analysis

In this thesis, an extensive study of the optimal fast pyrolysis conditions and parameters was performed.
Temperature of the pyrolysis was established to be the most important condition, where 550 ◦ C was
concluded to be optimal [11, 24, 34]. Therefore, using the kinetic model described in section 2.3, a
parametric test was performed where the reaction time was the variable at a constant final pyrolysis
temperature. The final pyrolysis temperature was kept at 550 ◦ C for the 5 tests and the reaction time
was increased 20 s between tests. The fastest reaction time was 5 s and the longest 85 s. The range of
these pyrolysis reaction times would allow to analyse the behavior of the sample in typical fast pyrolysis
conditions as well as intermediate pyrolysis (see Table 1.1). All the composition values determined and
calculated in section 5.1 were used as model inputs. Figure 5.3 shows the temperature profile, mass
loss, mass loss rate and bio-oil yield of each test.

Figure 5.3: Temperature profile (upper left), mass loss (upper right), mass loss rate (bottom right) and bio-oil yield
(bottom left) of 5 separate numerical pyrolysis experiments of pinewood, performed with different total
reaction times but the same final pyrolysis temperature.

60
The mass loss profiles in Figure 5.3 show that the final wt.% of mass decreases with reaction time.
Although higher heating rates should provide high bio-oi yields, the biomass decomposition percentage
suggests that reaction times smaller than 65 s may not display high enough yields of liquid byproducts
for the process to be efficient. In order to verify the optimal temperature, a similar parametric study was
performed, where final pyrolysis temperature was increased between runs at a constant 65 s of total
reaction time. The temperature profiles, mass loss, mass loss rate and liquid yield values are shown in
Figure 5.4. The tests reach 300 ◦ C, 400 ◦ C, 500 ◦ C, 600 ◦ C and 700 ◦ C of final temperature at 65 s,
respectively.

Figure 5.4: Temperature profile (upper left), mass loss (upper right), mass loss rate (bottom right) and bio-oil yield
(bottom left) of 5 separate numerical pyrolysis experiments of pinewood, performed with different final
pyrolysis temperatures but the same total reaction time.

In Figure 5.4, only the runs performed with a final temperature equal or above 500 ◦ C represent high
yields of bio-oil, confirming the bibliographic analysis. Additionally, the increase in bio-oil yield after the
final temperature of 550 ◦ C is negligible, which suggests performing pyrolysis at a higher temperature
would not be profitable nor energetically efficient.

61
5.4 Preliminary Testing of Reactor Setup

5.4.1 Temperature profile of the reactor

Since the temperature profile during pyrolysis is one of the input for the 0-D kinetic model, the correct
temperature profile is fundamental for an accurate prediction of liquid yields and efficiency analysis of the
system. Therefore, the real temperature profile was measured. After insulating the casing of the augers,
two type-K thermocouples (Cr+, Al−) from Omega were placed in the reactor:1) the first thermocouple
was placed on top of casing to record the temperature between the outer casing wall and the inner oven
wall; 2) the second thermocouple was placed in different axial coordinates, inside the casing. For the
measured profile to be accurate, at the beginning and ending of the casing, the temperature was mea-
sured in each centimeter, while at the to the center of the reactor, the temperature was measured every
two centimeters. This way, more measurements are performed at the points in which the temperature
raises quickly along the length of the casing and less measurements are performed at the center where
the temperature is closer to constant. In each singular coordinate, the temperature was measured for
30s and the mean value was recorded. Additionally, every measured coordinate was performed twice to
reassure the reliance of the values. The temperature profiles inside the casing were measured using two
different oven control temperatures, 500 ◦ C and 550 ◦ C. These temperatures were chosen, since both
the kinetic discussion made in the previous section and the bibliographic analysis concluded that be-
low these temperatures low bio-oil yields would be achieved and that increasing the temperature above
these values would become energetically inefficient. The mean values for the temperature profile at 500

C and 550 ◦ C are illustrated in Figures 5.5 and 5.6, respectively.

Figure 5.5: Temperature profile at a control temperature of 500 ◦ C.

62
Figure 5.6: Temperature profile at a control temperature of 550 ◦ C.

5.4.2 Feeding system testing

In order to analyse the byproducts of pyrolysis, the biomass feeding rate must be determined correctly.
In this system, there are two ways of assuring that the mass feeding rate is at the desired point. The
first method is through an experimental correlation between the vibratory system’s voltage input and the
mass flow the system provides. To establish this correlation, a fixed mass was deposited in the hopper
and once steady flow of biomass was verified, the time value for the flow of the entire fixed amount of
biomass was measured. Once this experiment was completed for the full spectrum of possible inputs of
the vibratory system, a linear regression was applied and a correlation was established. Table 5.5 and
Figure 5.7 show the results of the experiment, as well as the determined experimental correlation. For
the second method, the scales are tared after filling the hopper, the systems’ weight is measured peri-
odically and automatically as the reactor operates and the mass flow is calculated through the following
equations:

dm mt − mt+∆t
= ṁ (5.1) ṁ = (5.2)
dt ∆t
where ∆t is the periodic interval of each weight measurement, ṁ is the biomass mass flow rate, mt
and mt+∆t correspond to the weight before and after the time interval.

63
Table 5.5: Biomass mass flow vibratory experiment.
∗ The term x is the correlation corresponds to the percentage of the maximum voltage of the vibratory system.

Percentage of the maximum voltage (%) Time (s) Biomass weight (g) Mass flow (g/s) Mass flow (kg/h)
20 10000 107,4 0,01 0,04
30 3759,7 107,4 0,03 0,10
40 757,5 107,4 0,14 0,51
50 243,9 107,4 0,44 1,59
60 64,9 107,4 1,66 5,96
70 35,5 107,4 3,03 10,90
80 24,81 107,4 4,33 15,58
90 21,5 107,4 4,50 17,98
100 19,7 107,4 5,44 19,59
Experimental correlation∗ ṁ = 0.0006 · x2 + 0.0014 · x − 0.6235

Figure 5.7: Biomass mass flow vibratory experiment.


∗ The term x is the correlation corresponds to the percentage of the maximum voltage of the vibratory system.

5.5 Numerical Prediction of Pyrolysis Experiments

The reactor design can be used to determine the solid residence times given an arbitrary rotation speed
to the motor. Therefore, once the temperature profile of the reactor is established, the precise tempera-
ture and location of a particle is only a function of the operating rotation speed of the motor. Hence, the
temperature profile measured in subsection 5.4.1 was implemented at different motor rotation speeds in
the 0-D kinetic model. Figures 5.8, 5.9, 5.10 and 5.11 display the temperature profiles, at 500 ◦ C and
550 ◦ C control oven temperature, at 10 rpm, 20 rpm, 40 rpm and 60 rpm, respectively.

64
Figure 5.8: Temperature profiles of the particle at Figure 5.9: Temperature profiles of the particle at
10 rpm. Blue for 500 ◦ C and yellow for 20 rpm. Blue for 500 ◦ C and yellow for
550 ◦ C. 550 ◦ C.

Figure 5.10: Temperature profiles of the particle at Figure 5.11: Temperature profiles of the particle at
40 rpm. Blue for 500 ◦ C and yellow for 60 rpm. Blue for 500 ◦ C and yellow for
550 ◦ C. 550 ◦ C.

Figure 5.12: Summary of the predicted yields at different rotation speeds of operation.

65
Table 5.6: Predicted yields of the byproducts of pyrolysis at different control temperatures and at different auger
rotation speed.

Auger rotation speed (rpm) 10 20 40 60


Byproduct type Char Bio-oil Gas Char Bio-oil Gas Char Bio-oil Gas Char Bio-oil Gas
Yield at oven temperature control of 500 ◦ (wt.%) 24.5 55.6 19.9 25.8 54.5 19.7 27.7 53.1 19.1 29.7 51.8 18.5
Yield at oven temperature control of 550 ◦ (wt.%) 21.5 56.9 21.7 22.5 55.9 21.7 24.0 54.5 21.5 25.0 53.7 21.3

Figure 5.12 illustrates the predicted yields of char, liquids and gases of the eight tests described
above. Table 5.6 summarizes the results obtained.
As rotation speed increases, the yields of liquids and gases decrease slightly and the char yield
increases significantly. Although the highest liquid yield is reached at 550 ◦ C of control temperature
and at 10 rpm, the reactor should perform at a higher speed since the predicted decrease in bio-oil
content is 3.8 wt.% with a 600 % rotation speed increase. Also, it is verified no significant increase
in heating rate as the rotation speed increases and therefore, the yield is not supposed to increase
substantially. It is, however, confirmed a small increase in yield while increasing the control temperature
since, in that case, a higher heating rate is achieved. This effect derives from the fact that, assuming a
negligible temperature gradient inside the particle, the highest heating rate that the particle is submitted
to occurs between its exit of the feeding system and its entry to the auger system. Thus, and according
to the temperature profile recorded, the temperature, once inside the oven volume, does not show large
variation. Overall, the predicted bio-oil yields at these rotation speeds and conditions are in line with
what is previewed in the literature.

66
Conclusions
6

Contents
6.1 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

6.2 Recommendations for Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70

67
68
6.1 Conclusion

It is well-known that global geopolitical issues as well as heath and safety guidelines implemented in the
last few years affected multiple industry sectors, postponing deadlines and delaying material deliveries
worldwide. The material stock (e.g. steel, wood) and manufacturing previewed deadlines in Portugal
were not an exception. Due to these conditions, the experimental reactor testing was not possible within
the thesis work duration. However, the rest of the proposed objectives were successfully achieved.
Firstly, pinewood was chosen as feedstock and it was analysed in terms of its organic composition. Since
the experimental organic composition of biomass is difficult to obtain, a thermogravimetric analysis was
performed. Through these analysis, the complete organic composition was optimized using a fitting tool
to be used as input in the kinetic tool. The kinetic tool, with the proper inputs, was used to conduct a
parametric study to assess the optimal conditions that could favor the bio oil production (e.g. residence
time and temperature of pyrolysis). Afterwards, the key guidelines for the reactor design were described
and used for the final reactor setup build. Finally, with the dimensions from the reactor components and
the final organic composition, the kinetic tool was used to predict the yields of the reactor in its range of
operating conditions. The main conclusions of this work are as follows:

• The experimental organic composition of the pinewood sample was found to be 37.72 wt.% of
CELL, 25.91 wt.% of HCE, 26.73 wt.% of LIG and 9.64 wt.% of total extractive content in dry
basis. The fitting tool with the two-step optimization method was used to compare the experimental
thermogravimetric analysis and estimate the weighted distribution of the different LIG reference
components: 0.22 wt.%, 19.01 wt.% and 7.5 wt.% for LIG-C, LIG-H and LIG-O, respectively;

• The parametric studies showed that residence times smaller than 65 s did not provide sufficient
liquid yields (<50wt.%) and that the liquid percentage of byproducts stabilized after 65 s of res-
idence time. Additionally, optimal pyrolysis temperature was confirmed to be at least 500 ◦ C, at
which point the bio-oil yield remains constant with further temperature increase;

• Although the reactor was not possible to be completely assembled, preliminary results were pos-
sible, by performing temperature profiles. Since 500 ◦ C was found to be the minimum temperature
for bio-oil yield production, the temperature profiles of the reactor were recorded at 500 ◦ C and
550 ◦ C of control temperature. This led to the conclusion that the best temperature is 550 ◦ C since
it represented a small increase in performance;

• The kinetic tool was used to predict the yields of the pyrolysis products at 10 rpm, 20 rpm, 40 rpm
and 60 rpm. The predicted liquid yields decrease slightly with rotation speed and increase slightly
with control temperature. The maximum predicted bio-oil yield value was 56.85 wt.%, at 10 rpm
and 550 ◦ C of control temperature. All the yields predicted were comprised between 51.8 wt.%

69
and 56.85 wt.%. These values fall within the expected range of yields of bio-oil for an auger reactor
of these proportions.

6.2 Recommendations for Future Work

Primarily, the configuration of the reactor should undergo tests to assure the setup is airtight in order
to reduce oxygen inside the reactor and avoid combustion of the biomass. Afterwards, the validation
of the numerical results could be executed. Hence, the first step could be to utilize the configuration
setup at the operation conditions in which the numerical experiments were performed (e.g. control tem-
perature of 500 ◦ C at multiple auger rotation speed values). Subsequently, the bio-oil and the rest of
the byproducts could be analysed extensively to establish the potential of the process. For example, a
chemical characterization of the bio-oils through a chromatography analysis could identify their major
chemical compounds and access the possibility of the process being used for chemical feedstock pro-
duction. Furthermore, a porosimetry analysis of the chars could also assess their potential as industrial
adsorbents.
Other experiments, with different conditions and biomass feedstock (after its characterization) could
also be interesting, to test their influence on the yields and the quality of the byproducts.
In regards to the configuration, a few optimizing changes could be interesting. One example would
be to preheat the inert gas before it is injected into the system. This is mentioned in the literature as
a possible way to increase bio-oil yield since some volatiles could condense inside the reactor when in
contact with the unheated inert gas, possibly at the beginning of the reactor. Another more ambitious
optimization of the configuration could be to develop a heat carrier system, where heated sand or metal
carriers would be mixed with the biomass inside the reactor, improving the homogeneity of temperature
between the biomass particle, reassuring a better thermochemical conversion.

70
Bibliography

[1] “Estatı́sticas rápidas das renováveis,” 2021. [Online]. Available: www.dgeg.gov.pt

[2] A. V. Bridgwater, “Review of fast pyrolysis of biomass and product upgrading,” Biomass and Bioen-
ergy, vol. 38, pp. 68–94, 3 2012.

[3] V. Dhyani and T. Bhaskar, “A comprehensive review on the pyrolysis of lignocellulosic biomass,”
Renewable Energy, vol. 129, pp. 695–716, 12 2018.

[4] E. Butler, G. Devlin, D. Meier, and K. McDonnell, “A review of recent laboratory research and
commercial developments in fast pyrolysis and upgrading,” Renewable and Sustainable Energy
Reviews, vol. 15, pp. 4171–4186, 10 2011.

[5] S. Thangalazhy-Gopakumar, S. Adhikari, H. Ravindran, R. B. Gupta, O. Fasina, M. Tu, and S. D.


Fernando, “Physiochemical properties of bio-oil produced at various temperatures from pine wood
using an auger reactor,” Bioresource Technology, vol. 101, pp. 8389–8395, 11 2010.

[6] M. Sharifzadeh, M. Sadeqzadeh, M. Guo, T. N. Borhani, N. V. M. Konda, M. C. Garcia, L. Wang,


J. Hallett, and N. Shah, “The multi-scale challenges of biomass fast pyrolysis and bio-oil upgrading:
Review of the state of art and future research directions,” Progress in Energy and Combustion
Science, vol. 71, pp. 1–80, 3 2019.

[7] Y. K. N, P. D. T, S. P, K. S, Y. K. R, S. Varjani, S. AdishKumar, G. Kumar, and R. B. J, “Lignocellulosic


biomass-based pyrolysis: A comprehensive review,” Chemosphere, vol. 286, 1 2022.

[8] D. Granados, H. Velasquez, and F. Chejne Janna, “Energetic and exergetic evaluation of residual
biomass in a torrefaction process,” Energy, vol. 74, 09 2014.

[9] C. D. Blasi, “Modeling chemical and physical processes of wood and biomass pyrolysis,” Progress
in Energy and Combustion Science, vol. 34, pp. 47–90, 2 2008.

[10] A. I. Ferreiro, “Impact of inorganic elements on the pyrolysis and gasification of various biomass
fuels: experiments and modeling,” 2022.

71
[11] A. Anca-Couce, “Reaction mechanisms and multi-scale modelling of lignocellulosic biomass pyrol-
ysis,” Progress in Energy and Combustion Science, vol. 53, pp. 41–79, 3 2016.

[12] Q. Zhang, J. Chang, T. Wang, and Y. Xu, “Review of biomass pyrolysis oil properties and upgrading
research,” Energy Conversion and Management, vol. 48, pp. 87–92, 1 2007.

[13] S. Xiu and A. Shahbazi, “Bio-oil production and upgrading research: A review,” Renewable and
Sustainable Energy Reviews, vol. 16, pp. 4406–4414, 9 2012.

[14] S. V. Vassilev, D. Baxter, L. K. Andersen, and C. G. Vassileva, “An overview of the chemical com-
position of biomass,” Fuel, vol. 89, pp. 913–933, 5 2010.

[15] E. Ranzi, P. E. A. Debiagi, and A. Frassoldati, “Mathematical modeling of fast biomass pyrolysis
and bio-oil formation. note i: Kinetic mechanism of biomass pyrolysis,” ACS Sustainable Chemistry
and Engineering, vol. 5, pp. 2867–2881, 4 2017.

[16] I. E. Agency, “Review 2021 assessing the effects of economic recoveries on global energy demand
and co 2 emissions in 2021 global energy,” 2021. [Online]. Available: www.iea.org/t&c/

[17] E. Comission, “A 2030 framework for climate and energy policies,” 2013. [Online]. Available: https:
//ec.europa.eu/clima/eu-action/climate-strategies-targets/2030-climate-energy-framework en

[18] A. H. Demirbas and I. Demirbas, “Importance of rural bioenergy for developing countries,” Energy
Conversion and Management, vol. 48, pp. 2386–2398, 8 2007.

[19] B. Ozcan and I. Ozturk, “Renewable energy consumption-economic growth nexus in emerging
countries: A bootstrap panel causality test,” Renewable and Sustainable Energy Reviews, vol. 104,
pp. 30–37, 4 2019.

[20] W. B. Association, “Global bioenergy statistics 2020.”

[21] M. I. Jahirul, M. G. Rasul, A. A. Chowdhury, and N. Ashwath, “Biofuels production through biomass
pyrolysis- a technological review,” Energies, vol. 5, pp. 4952–5001, 2012.

[22] C. V. Stevens, J. Dewulf, H. Van, L. Biofuels, W. Soetaert, E. Vandamme, T. Bechtold, R. Mussak,


M. Kjellin, I. Johansson, C. Bergeron, D. J. Carrier, S. Ramaswamy, C. E. Wyman, S. Kabasci, D. D.
Stokke, Q. Wu, G. Han, D. L. Karlen, J. H. Clark, and F. Deswarte, Thermochemical Processing of
Biomass Wiley Series in Renewable Resources Handbook of Natural Colorants Surfactants from
Renewable Resources Aqueous Pretreatment of Plant Biomass for Biological and Chemical Con-
version to Fuels and Chemicals Introduction to Wood and Natural Fiber Composites Introduction to
Chemicals from Biomass, 2nd Edition.

[23] P. Basu, Biomass Gasification, Pyrolysis, and Torrefaction, 2nd Edition, 2013.

72
[24] G. Perkins, T. Bhaskar, and M. Konarova, “Process development status of fast pyrolysis technolo-
gies for the manufacture of renewable transport fuels from biomass,” Renewable and Sustainable
Energy Reviews, vol. 90, pp. 292–315, 7 2018.

[25] M. Raza, A. Inayat, A. Ahmed, F. Jamil, C. Ghenai, S. R. Naqvi, A. Shanableh, M. Ayoub, A. Waris,
and Y. K. Park, “Progress of the pyrolyzer reactors and advanced technologies for biomass pyrolysis
processing,” Sustainability (Switzerland), vol. 13, 10 2021.

[26] V. Silva, E. Monteiro, and A. Rouboa, “An analysis on the opportunities, technology
and potential of biomass residues for energy production in portugal.” [Online]. Available:
https://www.researchgate.net/publication/257925675

[27] J. Dias, “Utilização da biomassa: Avaliação dos resı́duos e utilização de pellets en caldeiras
domésticas,” 2002.

[28] M. V. de Velden, J. Baeyens, A. Brems, B. Janssens, and R. Dewil, “Fundamentals, kinetics and
endothermicity of the biomass pyrolysis reaction,” Renewable Energy, vol. 35, pp. 232–242, 1 2010.

[29] F. X. Collard and J. Blin, “A review on pyrolysis of biomass constituents: Mechanisms and compo-
sition of the products obtained from the conversion of cellulose, hemicelluloses and lignin,” Renew-
able and Sustainable Energy Reviews, vol. 38, pp. 594–608, 2014.

[30] D. Mohan, C. U. Pittman, and P. H. Steele, “Pyrolysis of wood/biomass for bio-oil: A critical review,”
Energy and Fuels, vol. 20, pp. 848–889, 5 2006.

[31] P. R. Yaashikaa, P. S. Kumar, S. J. Varjani, and A. Saravanan, “Advances in production and appli-
cation of biochar from lignocellulosic feedstocks for remediation of environmental pollutants,” Biore-
source Technology, vol. 292, 11 2019.

[32] T. Kan, V. Strezov, and T. J. Evans, “Lignocellulosic biomass pyrolysis: A review of product proper-
ties and effects of pyrolysis parameters,” Renewable and Sustainable Energy Reviews, vol. 57, pp.
1126–1140, 5 2016.

[33] D. Chen, Y. Li, K. Cen, M. Luo, H. Li, and B. Lu, “Pyrolysis polygeneration of poplar wood: Effect of
heating rate and pyrolysis temperature,” Bioresource Technology, vol. 218, pp. 780–788, 10 2016.

[34] A. Demirbas, “The influence of temperature on the yields of compounds existing in bio-oils obtained
from biomass samples via pyrolysis,” Fuel Processing Technology, vol. 88, pp. 591–597, 6 2007.

[35] J. Akhtar and N. S. Amin, “A review on operating parameters for optimum liquid oil yield in biomass
pyrolysis,” Renewable and Sustainable Energy Reviews, vol. 16, pp. 5101–5109, 9 2012.

73
[36] J. Shen, X. S. Wang, M. Garcia-Perez, D. Mourant, M. J. Rhodes, and C. Z. Li, “Effects of particle
size on the fast pyrolysis of oil mallee woody biomass,” Fuel, vol. 88, pp. 1810–1817, 10 2009.

[37] D. Vamvuka, N. Pasadakis, E. Kastanaki, P. Grammelis, and E. Kakaras, “Kinetic modeling of


coal/agricultural by-product blends,” Energy and Fuels, vol. 17, pp. 549–558, 5 2003.

[38] S. Nanda, P. Mohanty, J. A. Kozinski, and A. K. Dalai, “Physico-chemical properties of bio-oils


from pyrolysis of lignocellulosic biomass with high and slow heating rate,” Energy and Environment
Research, vol. 4, 7 2014.

[39] M. G. Grønli, G. Várhegyi, and C. D. Blasi, “Thermogravimetric analysis and devolatilization kinetics
of wood,” Industrial and Engineering Chemistry Research, vol. 41, pp. 4201–4208, 8 2002.

[40] M. R. Guerrero, M. M. D. S. Paula, M. M. Zaragoza, J. S. Gutiérrez, V. G. Velderrain, A. L. Ortiz, and


V. Collins-Martı́nez, “Thermogravimetric study on the pyrolysis kinetics of apple pomace as waste
biomass,” vol. 39. Elsevier Ltd, 10 2014, pp. 16 619–16 627.

[41] S. Maiti, S. Purakayastha, and B. Ghosh, “Thermal characterization of mustard straw and stalk in
nitrogen at different heating rates,” Fuel, vol. 86, pp. 1513–1518, 7 2007.

[42] F. Campuzano, R. C. Brown, and J. D. Martı́nez, “Auger reactors for pyrolysis of biomass and
wastes,” Renewable and Sustainable Energy Reviews, vol. 102, pp. 372–409, 3 2019.

[43] A. V. Bridgwater, “The production of biofuels and renewable chemicals by fast pyrolysis of biomass,”
International Journal of Global Energy Issues, vol. 27, pp. 160–203, 2007.

[44] X. Hu and M. Gholizadeh, “Progress of the applications of bio-oil,” Renewable and Sustainable
Energy Reviews, vol. 134, 12 2020.

[45] U. Fernandes and M. Costa, “Potential of biomass residues for energy production and utilization in
a region of portugal,” Biomass and Bioenergy, vol. 34, pp. 661–666, 5 2010.

[46] A. I. M. Ferreiro, M. M. G. da Costa, M. E. R. F. Rabaçal, and V. S. de Almeida Semião, “Pyrolysis


of pine bark, wheat straw and rice husk: Thermogravimetric analysis and kinetic study mechanical
engineering examination committee,” 2015.

[47] P. Giudicianni, V. Gargiulo, C. M. Grottola, M. Alfè, A. I. Ferreiro, M. A. A. Mendes, M. Fagnano,


and R. Ragucci, “Inherent metal elements in biomass pyrolysis: A review,” Energy & Fuels, vol. 35,
no. 7, pp. 5407–5478, 2021. [Online]. Available: https://doi.org/10.1021/acs.energyfuels.0c04046

[48] P. E. A. Debiagi, C. Pecchi, G. Gentile, A. Frassoldati, A. Cuoci, T. Faravelli, and E. Ranzi, “Extrac-
tives extend the applicability of multistep kinetic scheme of biomass pyrolysis,” Energy and Fuels,
vol. 29, pp. 6544–6555, 10 2015.

74
[49] C. Sheng and J. L. T. Azevedo, “Modeling biomass devolatilization using the chemical percolation
devolatilization model for the main components,” Proceedings of the Combustion Institute, vol. 29,
pp. 407–414, 2002.

[50] O. O. Olatunji, S. A. Akinlabi, M. P. Mashinini, S. O. Fatoba, and O. O. Ajayi, “Thermo-gravimetric


characterization of biomass properties: A review,” vol. 423. Institute of Physics Publishing, 11
2018.

[51] S. Badzioch and G. W. Hawksley, “Kinetics of thermal decomposition of pulverized coal particles,”
Industrial & Engineering Chemistry Process Design and Development, vol. 9, no. 4, p. 521–530,
1970.

[52] M. Pelucchi, C. Cavallotti, A. Cuoci, T. Faravelli, A. Frassoldati, and E. Ranzi, “Detailed kinetics of
substituted phenolic species in pyrolysis bio-oils,” Reaction Chemistry and Engineering, vol. 4, pp.
490–506, 3 2019.

[53] The creck modeling group. [Online]. Available: http://creckmodeling.chem.polimi.it/menu-kinetics/


menu-kinetics-detailed-mechanisms

[54] O. Authier, E. Thunin, P. Plion, and L. Porcheron, “Global kinetic modeling of coal devolatilization in
a thermogravimetric balance and drop-tube furnace,” Energy and Fuels, vol. 29, pp. 1461–1468, 3
2015.

[55] A. I. Ferreiro, M. Rabaçal, and M. Costa, “A combined genetic algorithm and least squares fitting
procedure for the estimation of the kinetic parameters of the pyrolysis of agricultural residues,”
Energy Conversion and Management, vol. 125, pp. 290–300, 2016, sustainable development of
energy, water and environment systems for future energy technologies and concepts. [Online].
Available: https://www.sciencedirect.com/science/article/pii/S0196890416303648

[56] L. Kapoor, A. Mekala, and D. Bose, “Auger reactor for biomass fast pyrolysis: Design and operation.”
Institute of Electrical and Electronics Engineers Inc., 9 2017.

[57] E. Henrich, N. Dahmen, F. Weirich, R. Reimert, and C. Kornmayer, “Fast pyrolysis of lignocellulosics
in a twin screw mixer reactor,” Fuel Processing Technology, vol. 143, pp. 151–161, 3 2016.

[58] S. Sirijanusorn, K. Sriprateep, and A. Pattiya, “Pyrolysis of cassava rhizome in a counter-rotating


twin screw reactor unit,” Bioresource Technology, vol. 139, pp. 343–348, 2013.

[59] J. N. Brown, “Development of a lab-scale auger reactor for biomass fast pyrolysis
and process optimization using response surface methodology,” 2009. [Online]. Available:
https://lib.dr.iastate.edu/etd

75
[60] M. Bortolamasi and J. Fottner, “Design and sizing of screw feeders.”

[61] D. W. Camp, T. T. Coburn, and P. H. Wallman, “Coal gasification-product separation pilot-unit sup-
port, twin screw heat transfer, and h2s evolution.”

[62] T. A. Kingston and T. J. Heindel, “Granular mixing optimization and the influence of operating con-
ditions in a double screw mixer,” Powder Technology, vol. 266, pp. 144–155, 2014.

[63] F. Qi and M. M. Wright, “Particle scale modeling of heat transfer in granular flows in a double screw
reactor,” Powder Technology, vol. 335, pp. 18–34, 7 2018.

[64] M. T. Morgano, H. Leibold, F. Richter, and H. Seifert, “Screw pyrolysis with integrated sequential
hot gas filtration,” Journal of Analytical and Applied Pyrolysis, vol. 113, pp. 216–224, 5 2015.

[65] J. Solar, I. de Marco, B. M. Caballero, A. Lopez-Urionabarrenechea, N. Rodriguez, I. Agirre, and


A. Adrados, “Influence of temperature and residence time in the pyrolysis of woody biomass waste
in a continuous screw reactor,” Biomass and Bioenergy, vol. 95, pp. 416–423, 12 2016.

[66] W. T. Tsai, M. K. Lee, and Y. M. Chang, “Fast pyrolysis of rice husk: Product yields and composi-
tions,” Bioresource Technology, vol. 98, pp. 22–28, 1 2007.

[67] BS EN ISO 18134-3 Solid biofuels - Determination of moisture content - Oven dry method Part 3:
Moisture in general analysis simple, vol. 44, 2014.

[68] BS EN 15148:2009 Solid biofuels - Method for the determination of the content of volatile matter,
vol. 2009, 2009.

[69] BS EN ISO 18122 Solid biofuels - Determination of ash content, vol. 44, 2014.

[70] A. Sluiter, R. Ruiz, C. Scarlata, J. Sluiter, and D. Templeton, “Determination of extractives in


biomass: Laboratory analytical procedure (lap); issue date 7/17/2005,” 2008. [Online]. Available:
http://www.nrel.gov/biomass/analytical procedures.html

[71] I. Andersone, B. Andersons, G. Avramidis, M. Barbu, L. Bergland, C. Clemons, M. Dietenberger,


P. Evans, C. Frihart, R. Ibach, M. Irle, J. Jakes, H. Militz, J. Peltonen, R. Pettersen, D. Packett,
R. Réh, C. Risbrudt, R. Rowell, and J. Winandy, Handbook of wood chemistry and wood compos-
ites, second edition, 09 2012.

[72] H. Yang, R. Yan, H. Chen, D. H. Lee, and C. Zheng, “Characteristics of hemicellulose, cellulose and
lignin pyrolysis,” Fuel, vol. 86, pp. 1781–1788, 8 2007.

[73] R. M. Rowell, Handbook of Wood Chemistry and Wood Composites, 1st ed. CRC Press, 2005.

76
A
Volumetric filling level and residence
times

77
Vbiomassperpitch
f (%) = ∗ 100 (A.1)
Vf reeperpitch
f (%) ∗ Vf reeperpitch
vbiomassperpitch = (A.2)
100
Vf reeperpitch = ((Acrosssecreactor − Ainnershaf ts ) ∗ p) − Vf light (A.3)

Acrosssecreactor = 2 ∗ (Acircle − Asegmentof circle ) (A.4)

Vf light = lengthpitch ∗ (Ageometryf light ) (A.5)

Ainnershaf ts = 2 ∗ (π23.12 ) = 3352.71mm2 (A.6)


2
42.45
Acrosssecreactor = 2 ∗ ((π42.452 ) − ( (2 ∗ 0.682 − (sin 2 ∗ 0.682)))) = 10615.69mm2 (A.7)
2
142
Ageometryf light = ∗ 2 = 196mm2 (A.8)
2
p
lengthpitch = p2 + (23.1 + (14/3))2 = 39.43mm (A.9)

Vf reeperpitch = ((10615.69 − 3352.71) ∗ 28) − (196 ∗ 39.43) = 195635.16mm3 = 0.1956dm3 (A.10)

where Vbiomassperpitch is the amount of biomass transported per pitch, Vf reeperpitch is the volume occu-
pied by gas per pitch, f is the volumetric filling level, Acrosssecreactor is the cross section area of the inner
casing walls, Ainnershaf ts is the sum of the cross section areas of the shafts of the augers, Acircle is the
cross section area of the inner wall of each tube of the casing if it was a circle, Asegmentof circle is the
area missing for cross section of inner wall of each tube to be a circle, Vf light is the volume occupied
by the flights of the augers in a 2.8cm section, Ageometryf light is the geometry of the flight of the augers
and lengthpitch is the length of the flight for a 2.8cm section of the augers.

Therefore, the necessary biomass feeding rate for each volumetric filling level is as follows:

ρ = 0.4377Kg/dm3 (A.11)
rpmreactor/motor
ṁ = ρ ∗ vbiomassperpitch ∗ = (A.12)
60
f (%) ∗ Vf reeperpitch rpmreactor/motor
ṁ = ρ ∗ ∗ , inKg/s (A.13)
100 60

where ρ is the bulk density, rpmreactor/motor is the rotation speed of the augers in rpm and ṁ is the
biomass feeding rate.

Assuming that, given the low rotation speed, the biomass transported in each pitch is not lost as it is
axially transported, then the solid residence time of the biomass is a function of the auger rotation speed
as follows:

rlength ∗ 60
∆t = (A.14)
p ∗ rpmreactor/motor

78
where p is the pitch of the auger flight in cm, rlength is the heated reaction zone length in cm and
rpmreactor/motor is the rotation speed of the augers in rpm.
Therefore, since the distance between the entry of the biomass and the exit of the casing is 60cm,
for different auger rotation speed, the solid residence time will be as shown in Table A.1

Table A.1: Estimated solid residence times for different auger rotation speed.

Augers rotation speed (rpm) 10 20 40 60


Solid residence time (s) 128.57 64.29 32.14 21.43

Assuming the pressure inside the reactor is approximately the atmospheric pressure, the 50 cm of
the reactor at highest temperature as the reaction zone and the room temperature is 23 ◦ C, then the gas
residence times can be calculated as a function of the geometry of the reactor and the inert gas (N2 )
volumetric flow rate as follows:

rzone
tgasres = (A.15)
ugas
mgf˙ low
ugas = (A.16)
ρN 2reactor ∗ Af ree
Af ree = (Vf reeperpitch /p) ∗ (1 − f ) (A.17)
Vgf˙low ∗ ρN 2room
mgf˙ low = (A.18)
1000 ∗ 60
patm ∗ MN 2
ρN 2reactor = (A.19)
R ∗ Treactor
patm ∗ MN 2
ρN 2room = (A.20)
R ∗ Troom

where tgasres is the gas residence time, rzone is the reaction zone length, ugas is the average axial
gas velocity, mgf˙ low is the average mass gas flow rate inside the reactor, ρN 2reactor is the estimated
density of the N2 inside the reactor, Af ree is the cross section area occupied by gas inside the reactor,
Vgf˙low is volumetric flow rate of the N2 at the flow meter, ρN 2room is the density of the N2 at the room
temperature, Treactor is the average temperature inside the reactor, Troom is the room temperature, R is
the ideal gas constant and patm is the atmospheric pressure.
To estimate the volumetric gas flow rate necessary for an arbitrary gas residence time, the following
equation can be used:

6000 ∗ Af ree ∗ Troom ∗ rzone


Vgf˙low = , inL/min (A.21)
tgasres ∗ Treactor

where Af ree is in m2 , tgasres is in s and rzone is in m.


Table A.2 summarizes the volumetric gas flow rates at different filling rates in order to reach certain
thresholds in terms of gas residence times at the control temperature of 500 ◦ C and 550 ◦ C. The val-

79
ues for the average temperature in the 50 cm of reaction zone are 450 ◦ C and 496 ◦ C for the control
temperature of 500 ◦ C and 550 ◦ C, respectively.

At the control temperature of 500 ◦ C At the control temperature of 550 ◦ C


Target gas residence time (s) Volumetric gas flow rate (L/min) Volumetric gas flow rate (L/min)
at f=10 % at f=36.7 % at f=10 % at f=36.7 %
1 0.87 0.62 0.87 0.62
5 0.17 0.123 0.17 0.123
10 0.087 0.062 0.087 0.062
30 0.029 0.02 0.029 0.02

Table A.2: Required volumetric gas flow rates at the N2 flow meter to reach target gas residence times at the control
temperature of 500◦ C and 550◦ C, and volumetric biomass filling rate (f) of 10 % and 36.7 %.

80
B
Technical drawings of extra
components

81
To assemble the complete setup of the reactor, extra supporting components were needed. Two
components were created to properly connect the feeding and the auger systems and a third component
was constructed to support the weight of the casing while the reactor is in operation.
The components that carry out the job of connecting the feeding and auger systems are illustrated
in Figures B.1 and B.2. The first component act as the opening of the reactor. Its shape matches the
shape of the casing so that it can be welded and secured. The second component is made from a clear
see-through polymer so that it is possible to supervise the entry of the biomass into the reactor. In the
setup, these two components are connected to each other as Figure B.3 shows.

Figure B.1: Technical drawing of the reactor opening component.

The third component supports the weight of the casing, as well as the weight of the connection
components between the auger system and the other systems. This component was design with a
threaded shaft so that it can be attached to the metal structure in which the oven is fixed. Additionally,
these supports grant the user of the reactor the option of inclining the casing and, therefore, the reactor,
if the experiment would so require. The dimensions of the casing supports are illustrated in Figure B.4

82
Figure B.2: Technical drawing of the see-through connection component.

Figure B.3: Assembly of the two connection components between the feeding and the auger system.

83
Figure B.4: Technical drawing of the casing support component.

84
85

Você também pode gostar