UNIVERSITY OF CINCINNATI

Date:______

I, ______, hereby submit this work as part of the requirements for the degree of: in:

It is entitled:

This work and its defense approved by:

Chair: ______

A Study of the Pyrolysis of Tire Derived Fuels and an Analysis of Derived Chars and Oils

A dissertation submitted to the

Division of Research and Advanced Studies of the University of Cincinnati

in partial fulfillment of the requirements for the degree of

DOCTORATE OF PHILOSOPHY (PhD.)

in the Department of Civil and Environmental Engineering of the College of Engineering

2006

by Kessinee Unapumnuk

B.S. Rural Technology Thammasat University, Thailand, 1996

M.S. Environmental Management and Technology Asian Institute of Technology, Thailand, 1999

Committee Chair: Professor Timothy C. Keener, PhD., PE, QEP

ABSTRACT

In this research, the pyrolysis product distribution rates and pyrolysis behavior of tire derived fuels (TDF) were investigated using thermogravimetric (TGA) techniques. A

TGA was designed and built in order to investigate the behavior and products of pyrolysis of typical TDF specimens. Fundamental knowledge of thermogravimetric analysis and principal fuel analysis is applied in this study. Thermogravimetry of the degradation temperature of the TDF confirms the overall decomposition rate of the volatile products during the depolymerization reaction. The principal fuel analysis

(proximate and ultimate analysis) of the pyrolysis char products shows the correlation of volatilization into the gas and liquid phases and the existence of fixed carbon and other compounds that remain as a solid char. The kinetic parameters were calculated using least square with minimizing sum of error square technique. The results show that the average kinetic parameters of TDF are the activation energy, E = 1322 ± 244 kJ/mol, a pre-exponential constant of A= 2.06 ± 3.47 x 1010 min-1, and a reaction order n = 1.62 ±

0.31. The model-predicted rate equations agree with the experimental data. The overall

TDF weight conversion represents the carbon weight conversion in the sample.

Three commercially important pyrolysis by-products from vehicle tires are liquid oil, solid char, and exhaust gas. These products could be contaminated by volatile toxic compounds released from the tire structure before combustion. The release of volatile materials from the tire structure during pyrolysis offers the greatest potential for separation of these compounds from the evolved gases and vapors, as the concentrations of these species are at their greatest in the vapor phase during this period. The influences of heating rate and pyrolysis temperature were investigated. The temperature was studied from 325 to 1000 oC, a range where substantial devolatilization occurs. The results

showed that the heating rate and the pyrolysis temperature were the key factors in

determining the two pyrolysis yields: condensed oil and gas product. However, the overall desulfurization of the pyrolysis reaction was essentially unaffected by the heating rates.

Products from the pyrolysis of TDF were investigated with various analytical

techniques and under various maximum pyrolysis temperatures and heating rates. The

pyrolysis products are classified as char (solid product), pyrolytic oil (liquid) and gas.

Principle functional groups of the TDF and pyrolytic oil were confirmed by Fourier

Transform Infrared Spectrometer, coupled with attenuated total reflectance (FT-IR/ATR).

The components of the pyrolytic oil fraction were individually quantified using gas

chromatography coupled with mass spectrometry (GC-MS). The major products are one-

and two-ring methyl-substituted aromatic isomers. By-product formation mechanisms of

TDF pyrolysis were hypothesized based on the products identified. The mechanisms for

aromatic hydrocarbon formation were found to be associated with polymer degradation,

methyl displacement, and the Diels-Alder reactions. Our study indicated that GC-MS

coupled with FT-IR is sufficient to investigate the semi-volatile and volatile organic species from complex polymeric materials such as tires.

ACKNOWLEDGEMENT

I would like to thank the following people who helped to make this dissertation

possible:

My advisor, Dr. Tim C. Keener, for his valuable guidance, patience, constructive

challenges and encouragement throughout my study. My co-advisor, Dr. Mingming Lu,

for her effort to advice and criticize my work especially on the chromatography

laboratory analysis. My dissertation committee members, Dr. George Sorial and Dr.

Soon-Jai Khang, for their kindly advice on my research. I feel truly appreciated to have

had such distinguished mentors directing this academic process.

My special colleagues during doctoral study, Phirun, Jun, Fuyan, Jacob, Pamela, Sang-

Sup, Ricardo, Sook and Shuang for their encouragement and support energized my hopes

and which whom I shared so many experiences. I would like to thank all Thai friends:

Dr. Sumana, Dr. Supa, Phirun, Atchara, Wassana, Chaichana, and Sarawuth who have

made Cincinnati quite an enjoyable city for me during the past 4 years.

The Royal Thai Government for a scholarship that allowed me to pursue a degree in the

United States.

My loving family for their encouragement, support and numerous sacrifices. Especially,

my mother for her unconditional love, who cry and laugh with me everyday for years.

My two sisters, who have been taking good care of my mother while I concentrated on completing the degree. Lastly, I dedicate my work to my beloved passed away father without his loving spirit I would not have indeed come this far. His loving spirit were a continual source of inspiration to me. TABLE OF CONTENTS

TABLE OF CONTENTS...... I

LIST OF TABLES...... IV

LIST OF FIGURES ...... V

NOMENCLATURE ...... VII

1 Chapter 1 Introduction ...... 1

1.1 Background and Problem Statement...... 1 1.2 Research Objectives...... 4 1.3 Literature Review...... 5 1.3.1 Properties of Tires...... 6 1.3.2 Combustion of Tires ...... 9 1.3.3 Pyrolysis Mechanism of Tires...... 9 1.3.4 Formation of Polycyclic Aromatic Hydrocarbons...... 11 1.4 References...... 14

2 Chapter 2 Research Methodology...... 18

2.1 Properties of Tire Derived Fuel Materials ...... 18 2.2 Experimental Design...... 19 2.2.1 Reactor Design...... 19 2.2.2 Pyrolysis Product Distribution Experiment ...... 23 2.2.3 Proximate and Ultimate analysis ...... 23 2.2.4 Identification of Organic Compounds ...... 23 2.3 References...... 31

3 Chapter 3 Thermogravimetric Study and Pyrolysis Kinetic Mechanisms...... 34

3.1 Abstract...... 34

I 3.2 Introduction...... 35 3.3 Experimental Method...... 35 3.4 Results and Discussion...... 36 3.4.1 Preliminary Experiments to Establish the Reactor Design Criteria...... 36 3.4.2 Investigation of TDF Decomposition ...... 39 3.4.3 Pyrolysis Kinetic Interpretation of Experimental Data...... 43 3.4.4 Decomposition of TDF in Different Atmospheres ...... 50 3.5 Conclusion ...... 52 3.6 References...... 53

4 Chapter 4 The Recovery of Pyrolysis By-products and the Removal of Sulfur...... 55

4.1 Abstract...... 55 4.2 Introduction...... 56 4.3 Experimental Method...... 56 4.3.1 Determination of By-products Recovery...... 56 4.3.2 Determination of Sulfur and Carbon...... 57 4.4 Results and Discussion...... 59 4.4.1 Recovery of Pyrolysis By-Products...... 59 4.4.2 Char...... 61 4.4.3 Oil ...... 62 4.4.4 Gas ...... 65 4.4.5 Distribution of Carbon...... 66 4.4.6 Removal of Sulfur...... 69 4.5 Conclusion ...... 76 4.6 References...... 77

5 Chapter 5 Hydrocarbons Composition from the Pyrolysis Products...... 79

5.1 Abstract...... 79 5.2 Introduction...... 80 5.3 Experimental Method...... 80

II 5.4 Results and Discussion...... 82 5.4.1 Principal Functional Structures of TDF...... 82 5.4.2 Principal Functional Structures of Pyrolytic oil...... 83 5.4.3 Individual Hydrocarbons Composition of Pyrolytic oil...... 86 5.5 Conclusion ...... 95 5.6 References...... 96

6 Chapter 6 Conclusions and Recommendations...... 99

Appendix A: Demonstration of Mathematica Model ...... 102

Appendix B: Supporting Materials for GC-MS Analysis...... 107

Appendix C: Feasibility Study of Burning Tires with Coal ...... 110

1.1 Tire Burning in the United States ...... 110 1.2 Principal Fuel Properties and Environmental Performance...... 111 1.3 Boiler Type and its Operation with TDF ...... 114 1.3.1 Pulverized Coal Boilers...... 114 1.3.2 Cyclone Fired Boilers ...... 114 1.3.3 Fluidized Bed Boilers ...... 115 1.3.4 Stoker Boilers...... 115 1.4 A Case Study: Feasibility study of burning TDF blend with coal at the University of Cincinnati...... 117 1.4.1 Ambient impact of the air emission...... 120 1.4.2 Economics of burning tires with coal ...... 121 1.5 Conclusion ...... 121 1.6 References...... 122

III LIST OF TABLES

Table 1.1 Weight distribution of the various components of a passenger car tire [4]...... 7

Table 1.2 Fuel Properties [6, 7] ...... 7

Table 2.1 Proximate and ultimate analysis of used tires...... 19

Table 3.1 Influence of various heating rate on degradation reaction temperature...... 42

Table 3.2 Kinetic parameters of tire pyrolysis...... 49

Table 4.1 The analysis results of TDF samples subjected to pyrolysis conditions...... 71

Table 5.1 Approximate FTIR frequencies for hydrocarbon-substituted olefins and

heteroaromatic rings...... 85

Table 5.2 The major products in the pyrolytic oil identified by GC-MS ...... 88

Table 5.3 The most abundant isomeric structure identified by GC-MS...... 91

IV LIST OF FIGURES

Figure 1.1 U.S. Scrap Tire Markets 1990-2003 [1]...... 2

Figure 1.2 Cross section of a high performance passenger tire [4] ...... 6

Figure 1.3 Structure formula of rubber compounds [8, 9]...... 8

Figure 1.4 Formation of isoprene from natural rubber compounds [16] ...... 13

Figure 2.1 Schematic diagram of pyrolysis reactor ...... 22

o Figure 3.1 TG-DTG curves of TDF in O2 and N2, at 0.01L/min, heating rate 10 C/min 36

Figure 3.2 TG curve of TDF with the variation of heating rates ...... 37

Figure 3.3 Comparison of the thermogravimetric results ...... 38

Figure 3.4 Weight conversion at different heating rates from TGA...... 40

Figure 3.5 Weight conversion rate at different heating rates from TGA...... 41

Figure 3.6 Plot of TDF pyrolysis versus decomposition time. Symbols are the

experimental data, and lines are the calculated data...... 47

Figure 3.7 Plot of TDF decomposition in inert and non-inert atmosphere versus

decomposition time. Symbols are the inert atmosphere, and lines are the non-inert

atmosphere ...... 51

Figure 4.1 Pyrolysis product distribution with heating rate of 1 oC/min...... 60

Figure 4.2 Pyrolysis product distribution with heating rate of 5 oC/min...... 60

Figure 4.3 Pyrolysis product distribution with heating rate of 10 oC/min...... 61

Figure 4.4 Carbon distribution in pyrolytic by-products as function of heating rates...... 66

Figure 4.5 Carbon distribution in pyrolytic by-products as function of temperature ...... 67

Figure 4.6 Temperature profile and weight conversion of TDF and carbon content in

pyrolytic char ...... 68

V Figure 4.7 Percent sulfur removal at different heating rates...... 70

Figure 4.8 Percent sulfur in derived oil and pyrolyzed char compare to the original parent

sulfur content in TDF versus pyrolysis temperature...... 73

Figure 4.9 Percent sulfur removal versus percent devolatilzation and the rate of sulfur

removal as a function of volatiles ...... 74

Figure 4.10 Percent sulfur removal (%DeS) and percent ash recovery (%Ash) versus

percent devolatilization (%DeV) ...... 75

Figure 5.1 Schematic diagram of the vacuum pyrolysis reactor...... 81

Figure 5.2 IR Spectrum of TDF...... 83

Figure 5.3 IR Spectrum of pyrolytic oil heated at 5 oC/min and, a) maximum pyrolysis

temperature 500 oC, b) maximum pyrolysis temperature 800 oC ...... 84

Figure 5.4 The total ion chromatogram of pyrolytic oils at pyrolysis condition of 5

oC/min and a maximum temperature of 500 oC. The chemical structures of the

numbered group(s) are further identified by GC-MS in Table 5.2...... 87

Figure 5.5 Isomeric mass spectrum of trimethyl benzene and ethylbenzene ...... 90

Figure 5.6 The C4-Cyclohexenes in the pyrolytic oil from selective ion chromatogram:

pyrolysis condition of 5 oC/min and the maximum temperature of 500 oC. 1).

1,5,5,6- tetramethyl, 1, 3-Cyclohexadiene; 2). 1, 5, 5-trimethyl-6-methylene-

Cyclohexene; 3). 1,3,5,5-tetramethyl 1,3-Cyclohexadiene; 4). Terpinolene; 5). 4-

Isopropenyl-1-methyl-1-cyclohexene; 6). Terpinolene; 7). Terpinolene ...... 92

Figure 5.7 Diels-Alder reaction for the formation of naphthalene in scrap tire

pyrolysis[15] ...... 93

VI NOMENCLATURE

TDF Tire derived fuel

TGA Thermogravimetric analyzer

As Area of the characteristic ion for the analyte to be measured

Ca Known concentration of the analyte in the calibration solution (μg/mL)

Ma Mass of each target analytes, including alkylated homologues, μg

Aa Area of the characteristic ion for the analyte measured

RF Response factor for the analyte from the current calibration

C Concentration of each target analyte in a sample, μg/L

DF Dilution factor applied to the pyrolysis oil sample

W Sample volume, L

TG Thermogravimetric graph

DTG Derivative thermal analysis graph n Reaction order

E Activation energy, kJ/mol

A Pre-exponential parameter, min-1

NR Natural rubber

SBR Styrene butadiene rubbers

BR Butyl rubber

ASTM American Society for Testing of Materials

o Ti Initial decomposition temperature, C

o Tf Final decomposition temperature, C

VII o Tm Peak decomposition temperature, C

T Temperature, oC

dW/dT Weight conversation with respect to Temperature

R Universal gas constant, 8.3144 J mol-1K-1

t time, min

X Dimensionless weight conversion

dX/dt Derivative of conversion with respect to time

Wi Initial weight, g

Wf Final weight, g

Wt Weight at any time, g

SSQ Sum of square of the conversion

PAH Polycyclic aromatic hydrocarbon

DeS Desulfurization

DeV Devolatilization

H Heat content, Btu/lb, Btu/kg, kJ/kg

Wchar Weight of the solid char or pyrolyzed char, g;

WTDF Weight of the original TDF, g.

HTDF Heat value of the original TDF, Btu/lb;

Hpyrolyzed Heat values of the pyrolyzed product, Btu/lb.

STDF Total percent by weight of sulfur content in original TDF, wt.%

Spyrolyzed Total percent by weight of sulfur content in pyrolyzed products, wt.%

GC/MS Gas chromatography coupled with mass spectrometry

FT-IR Fourier Transform Infrared Spectrometer

VIII ATR Attenuated total reflectance

KBr Potassium bromide

IR Infrared spectroscopy

PTFE Polytetrafluoroethylene

DCM Dichloromethane

EI Electron ionization

NIST National Institute of Standards and Technology

SIM Selected ion monitoring

TIC Total ion chromatogram

AED Atomic emission detector

S Sulfur

C Carbon

SO2 Sulfur Dioxide

N2 Nitrogen

IX Chapter 1

1 Introduction

1.1 Background and Problem Statement

More than 290 million tires are discarded annually in the United States, or an estimate of one tire per person per year [1]. In the past, environmental concern for tire

disposal has focused on solid and hazardous waste issues. Much information has already

been reported regarding the comparative merits of disposal alternatives – such as

recycling, landfilling and burning for fuel – in minimizing scrap tires and maximizing

recycle markets. One way of removing tires from the waste stream is disposal in valuable

landfill space. Scrap tires present unusual disposal problems, mainly because they are

bulky, non-biodegradable, and vary in durability. Unregulated stockpiles of used tires in

the open create a potential fire hazard as well as contributing to environmental, health,

and safety problems. Dumpsites with warm, stagnant water in scrap tire castings provide

an ideal breeding ground for mosquitoes. Severe illnesses, including encephalitis and

dengue fever, have been contributed to disease-carrying mosquitoes originating from

scrap tire piles [2]. The various characteristics of tires and the increase of unregulated

dumpsites present solid waste management problems that make tire disposal almost

impossible. For these reasons, many locations in the U.S. have prohibited the disposal of

whole tires in landfills and begun to strictly regulate the landfilling of scrap tires.

Landfilling, however, is still the least expensive option and the preferred disposal method

in many areas.

Another solution of removing tires from the waste stream is incineration. A

serious consequence of burning tires is that it may release toxic chemical compounds

1 such as dioxin, furans, and aromatic hydrocarbons into the atmosphere. The final products from the open combustion of waste tires are ash, particulates, tar, and exhaust gases, which are blamed for several physical health problems such as eye irritation and respiratory problems. The high levels of pollution generated from open tire burning are unacceptable to the public.

After years of detailed research, scientists have recognized scrap tires as a potential source of fuel – so-called Tire Derived Fuel (TDF). TDF has been known as a low-cost material with excellent and consistent fuel properties, especially containing significant heating value. Fuel derived from waste tires has an extremely high thermal value: about 12,000-16,000 BTU/lb. From 2001 through the end of 2003, the total number of scrap tires being used in the market slightly increased, as presented in Figure

1.1.

350 Millions of Tires % Scrap Tire Utilization 100

90 300 80

250 70 n s 60 200 50 150 40 Millions of Tire of Millions 30 100 % Scrap Tire Utilizatio 20 50 10

0 0 1990 1992 1994 1996 1998 2001 2003 Year

Figure 1.1 U.S. Scrap Tire Markets 1990-2003 [1]

2 The three major markets for scrap tires in the U.S. – including tire derived fuels, civil engineering, and ground rubber applications – are expected to expand in conjunction with the research on TDF. Legitimate markets for scrap tires are growing, and several industries have expressed significant interest in their fuel capabilities. Some of these industries include utility and process boilers such as electric utilities, pulp and paper mills, cement manufacturers, stoker boilers, and rotary kilns for supplement fuel [3].

Since the 1970s, public concern over the use of scrap tires as a fuel mainly deals with the potential environmental consequences if proper care is not exercised during the combustion process. A number of studies have investigated how to use available technology in order to eliminate the harmful emissions from combustion. In order to gain better knowledge of the potential of using used tire as supplemental fuels, several factors have been taken into consideration.

First, pyrolysis was studied because it is the first step of combustion for any solid fuel. The process can be altered to maximize the yield of char, oil or gas. The higher pyrolysis temperature thermally degrades the higher molecular weight oil species to gas due to its major impact on the nucleation process. However, various reaction steps leading to chemical nucleation of the aromatic compound in the pyrolysis process of TDF are still quite difficult to understand. Rubber compounding essentially depends on the combination of various ingredients to produce a polymer with characteristics sufficient for the end product to perform satisfactorily under its intended conditions. In addition, scrap tires cannot be melted and separated into their basic chemical components. With so many variables, it is almost impossible to know the exact composition of particular used

3 tires, so that a complete knowledge of the formation of many chemical compounds from tire pyrolysis has not yet become available.

Second, several questions remain concerning the use of TDF. The correlation between the TDF decomposition rate and the hydrocarbon production rate in pyrolysis- byproducts has yet to be developed. Also, a significant gap exists in the database of aromatic hydrocarbons that can be formed from the combustion of tires. Questions exist over the fate and level of volatile carbons emission in utility plants that use TDF as a fuel.

The possible production of carcinogens from the use of TDF is still a contentious topic and could possibly limit the future use of TDF as a fuel. Therefore, any understanding that will allow users to minimize the formation of hazardous polycyclic aromatic hydrocarbons and organic volatile carbons during the combustion of shredded tires will certainly increase the potential of using TDF as a fuel for many industries.

1.2 Research Objectives

The fact that tires are thermal-set polymers – meaning that they cannot be melted and separated into their chemical components – presents a challenge in the study of potential hazards emitted from burning used tires. One of the main objectives of this research was to study heat and mass transfer phenomena and kinetics behavior during pyrolysis of

TDF. The experiments were designed under the hypothesis that changes in the pyrolysis temperature and heating rates can alter the pyrolysis kinetics parameter and the mass distribution of the pyrolysis byproducts. A number of experiments were conducted to determine the degradation temperature by varying the pyrolysis conditions – such as carrier gas flow rates, heating rates, and pyrolysis temperatures. A laboratory scale thermogravimetric analyzer has been developed to study pyrolysis mechanisms

4 corresponding to the distribution of the products. The overall kinetic parameters for TDF

pyrolysis mechanisms were generated by dynamic heating method. The experimental

results were applied to validate the mathematical model.

Another objective of this study was to investigate the formation mechanisms of the

individual aromatic hydrocarbon during the pyrolysis of TDF. The particular thermal

decomposition mechanism was determined by using the pyrolysis reactor and the solid

state model. In this way, it was possible to find the relative stability of aromatic

compounds such as benzene and naphthalene. The model of aromatic hydrocarbons, with

the main chain radical propagators was applied. The main purpose of this research lied in

the understanding of the fundamental features of the aromatic compound formation and

pyrolysis mechanisms of TDF. Due to the numerous effects from the pyrolysis

conditions such as the heating rate and the pyrolysis temperature, it was extremely

difficult to simplify the detailed mechanisms of the changes in product yields. Hence, the

pyrolysis of used tire materials was studied in order to acquire the new knowledge about

the interference of TDF pyrolysis conditions with the aromatic hydrocarbons formation.

The experiments were designed under the hypothesis that changes in the pyrolysis temperature and heating rates can alter the composition of the aromatic compounds and the carbon contents in the pyrolysis oils.

1.3 Literature Review

The description of the properties, combustion, and pyrolysis mechanisms of tires, as well as fundamental knowledge of some aromatic hydrocarbon formation, will be discussed in this chapter.

5 1.3.1 Properties of Tires

A modern passenger car tire is made of more than 100 heterogeneous parts. The main

substrates are vulcanized rubber, rubber filler, rubberized fabric, steel cord, fillers like

carbon black or silica gel, sulfur, zinc oxide, processing oil, fabric belts, steel wire- reinforced rubber beads, and many more additives. The major composition of a typical passenger car tire is presented in Table 1.1. A wide variety of tire parts have been designed to accommodate the different desired material characteristics and the variety of consumer demands. A typical cross-section of a high performance passenger tire is presented in Figure 1.2[4].

Figure 1.2 Cross section of a high performance passenger tire [4]

The most common rubbers used for tires manufacture are natural rubber (NR),

styrene-butadiene rubber (SBR) and butadiene rubber (BR) [5]. Rubber is usually a blend

6 of two or three rubber compounds with additives. For example, the tread components can consist of blends of NR and SBR, compounded with carbon black, oils, and vulcanizing chemicals. The sidewall materials consist of a NR/BR blend. The structural formulas of rubber compounds can be distinguished from one another, but the main structures are similar.

Table 1.1 Weight distribution of the various components of a passenger car tire [4]

Tire components Percentage Natural rubber 15-19 Carbon black 24-28 Synthetic rubber 25-29 Steel cords 9-13 Textiles cords 5-6 Chemical additives 14-15

The comparison of fuel properties derived from various solid fuel sources is summarized in Table 1.2. The TDF shows the highest heating value among coal, residual derived fuel and wood.

Table 1.2 Fuel Properties [6, 7] Residual Fuel Analysis Coal TDF Derived Wood Fuel Moisture, % 12.8 0.8 24.0 39.1 Ash,% 10.4 9.6 12.0 0.4 Volatile% 34.4 66.4 54.0 49.7 Fixed carbon, % 42.4 23.2 10.0 10.8 Heating value(Btu/lb) 11000 15312 5900 5140 C, % 61.1 76.9 H, % 4.1 7.0 S, % 2.8 1.4

7

The principle components of rubbers considered in this study are hydrocarbons consisting solely of carbon and hydrogen. Polymers are high molecular weight compounds made from multiple low molecular weight building units, or monomers.

Polymer structures may consist of 1000-20000 repeating units of single- or double- bonded carbon-hydrogen monomers [5]. For instance, NR is a polymer composed of isoprene, butadiene, styrene and isobutyl units. SBR is a combination of styrene and a number of alkane and alkene units as shown in Figure 1.3.

CH3 H3C H H H H H

HH HH H NR H H NR

a) Natural Rubber (NR)

H HH H BRS H

H H HH SBR

H H

H H

H

b) Styrene-Butadiene Rubber (SBR)

H H BR BR

H H H H

c) Polybutadiene Rubber (BR)

Figure 1.3 Structure formula of rubber compounds [8, 9]

8 1.3.2 Combustion of Tires

The use of scrap tires as a fuel has been recognized as a potential low cost material

with excellent and consistent fuel properties, especially containing a significant heating

value. The current situation of burning waste tires in the U.S. has been reviewed and

reported in a concise, cohesive, and exclusive summary[3, 6]. The extensive review on this

particular subject is based on the application of using used tires as a supplementary fuel

in boiler utility plants. A case study of used tire combustion is presented in Appendix C.

1.3.3 Pyrolysis Mechanism of Tires

Thermogravimetry (TG/DTG) has been used extensively for the study of pyrolysis

behavior and the analysis of elastomer decomposition [10-12]. The reaction order (n), the

activation energy (E), and pre-exponential parameter (A), were determined according to

the heating value and weight change. The kinetic parameters from the other studies were

developed empirically and were valid over the precise selection of the derivative peak and degradation temperature. Pyrolysis has been recognized as a chemical degradation

reaction that is caused by thermal energy alone [13]. The pyrolysis reaction occurs at the

beginning of any combustion process of solid fuels. The reaction is known to be a major

mechanism in the combustion process that generates aromatic hydrocarbons [7, 9, 14-17].

One concern about burning scrap tires centers on the impact of air emissions – in particular, the possibility of producing carcinogenic compounds such as polycyclic aromatic hydrocarbon (PAH) during combustion [18-21]. Lemieux [18] indicated that

volatile organic emissions – such as chloromethane, benzene, and styrene – from TDF

emissions were comparable to those from pulverized coal and oil combustion, although

no significant amounts of semi-volatile organic compounds were reported in his work.

9 Leung [19] reported that the degradation of rubber compounds such as NR, BR and SBR started when the combustion temperature was higher than 350 oC. The same study

concluded that the decomposition rate at temperatures above 350oC was affected by the heating rates rather than the sample sizes. Atal [20] studied the emission of semi-volatile

PAH from the combustion of pulverized bituminous coal and ground automobile tires.

They reported that none of the deuterated PAH, adsorbed on the surface of mixed fuels,

existed after the combustion process. Small amounts of the overall labeled PAH

standards were recovered in other deuterated components under high temperature, ~1150

oC, in the pyrolysis. Atal [20] and Levendis [21] measured PAH emissions from dry-

injected cylindrical streams of coal and waste tire particles in a sealed laminar flow furnace. They claimed that the total amounts of PAH remained unchanged, although they observed a slight change in the distribution between the gas and the condensed PAH phases. Kaminsky [14] reported less than 5% of benzene and toluene by tire weight in the

pyrolysis oil at 750oC via a fluidized bed reactor. Cypress [16] suggested that the

concentrations of naphthalene at 3.8% and phenanthrene at 1.6% by tire weight were

generated via a two-stage pyrolysis processes. Williams [9] analyzed the molecular

weight distribution of pyrolysis oil from a static-batch fixed-bed reactor. They concluded

that over 10% of the total mass of tire-derived oil could be contributed to the

concentration of PAH. Some carcinogenic PAH such as benzo[a]pyrene,

benzo[e]pyrene, chrysene and fluoranthene were present in significant concentrations.

Cunliffe [16] used a nitrogen-purged static-bed batch reactor to pyrolyze three-kilogram

batches of shredded scrap tires at temperatures between 450 and 600 oC. He reported that

the results of the gas analysis support the dehydrogenation of alkanes to alkenes,

10 followed by cyclisation and aromatization. Limonene was found to be a major component in pyrolysis oil. Significant quantities of light aromatics – such as benzene,

toluene, xylene and styrene – were recovered in the same study. Pakdel [17] reported that

the short vapor residence time in a vacuum pyrolysis reactor caused low decomposition

temperature but increased the oil yield.

Tire pyrolysis is technically feasible; however; economics are more favorable

when the pyrolysis by-products can be profitable. The suitability of the usage of oils,

solid char, and gas yielded from the pyrolysis of TDF has been explored for many

applications [7, 14-17, 22]. The derived gases are useful as fuels, and the solid chars – carbon

black and activated carbon – can be used as smokeless fuels [7, 14, 15, 22]. The derived oils

from TDF pyrolysis have a high calorific value similar to that of a medium heating oil, and may be used for substitution of conventional fuels or added to petroleum refinery feedstocks [14-17]. Of highest interest are the derived oils, which have been recognized as

the most valuable products, particularly in the chemical industry. The potential of using

derived oil in some industrial applications has been researched for many years, but

additional research is still necessary. These studies have raised public concerns regarding the potential for derived oil to substitute conventional fuels.

1.3.4 Formation of Polycyclic Aromatic Hydrocarbons

Polycyclic aromatic hydrocarbons are a group of harmful organic substances which have come to be of public interest on account of the existence of hazardous waste deposit sites [23]. The method of measuring on PAH emission and TDF pyrolysis has been the

subject of the numerous studies for several years since the introduction of used tires as a

potential supplemental fuel in a power generation industry [3, 6, 24]. The history of PAH

11 formation from the pyrolysis process started long ago, and TDF has been well known as

one of the very few materials having essentially high heat content in comparison with

coal. Nevertheless, because of the complexity in TDF composition and the factors that

affect PAH formation, some studies on pyrolysis composites have resulted in

contradicting and doubtful observations [16, 20, 25, 26]. Pyrolysis is the process of breaking

up a molecule by heat [18], without air or avoiding oxidation. It is commonly performed

in an inert atmosphere or low pressure. The pyrolysis process of a polymer depends on

the experimental conditions. Violi [25] studied the modeling of particulate formation in combustion and pyrolysis. They applied a detailed chemical kinetic mechanism to explore the effects of C/O ratio, temperature, and pressure on the formation of high

molecular weight aromatic species in premixed flames and shock tubes in a wide range of

operating conditions. As a result, they reported that in very fuel-rich and pure pyrolysis conditions the mechanism involving H-atom abstraction and acetylene addition to aromatics controls the formation of aromatics. They summarized that particulate formation in combustion is based on three steps: (1) production of the first aromatic ring through the oxidation and early pyrolysis of the fuel, (2) planar PAH growth by a sequence of H-atom abstraction and acetylene addition reactions resulting in the formation of primary solid particles, and (3) the formation of solid state clusters by coagulation of surface growth reactions that compete with oxidation reactions. The formation of PAH compounds basically follows the steps of soot formation in the combustion process. The condensed phase material and subsequent absorption of high molecular weight hydrocarbons is formed after the nucleation reaction process. The dehydrogeneration of the primary soot particles and cyclisation and dehydrogenation of

12 the absorbed hydrocarbons took place before agglomeration, conglomeration and thermal

ionization of the small particles [16, 27].

The mechanism of polymer decomposition begins with the first decomposition step or

initiation step, then subsequent steps – including chain scission, side group scission, and

combined reactions – take place. Polymer chain scission is an elimination reaction that

takes place by breaking the bonds that form the polymeric chain. It is also sometimes

called a depolymerization [8, 26, 28], which results in the formation of molecules of lower molecular weight. An example of the formation of isoprene from NR is presented in

Figure 1.4.

CH3 H3C H H H H C H H 3

H HH H H H2C CH2 NR H H NR

Figure 1.4 Formation of isoprene from natural rubber compounds [16]

Sahouli [28] stated that during the pyrolysis of used tires, the long elastomer chains broke down and were adsorbed on the carbon black surface. As a consequence, carbon deposits and small aromatic compounds were formed, depending on the pyrolysis conditions. Streitwieser [8] found that pyrosynthesis is the only mechanism for any

formation of PAH during the combustion of light fuels such as methane, ethylene or

blends of aliphatics such as some grade of gasoline. However, in the case of heavier

fuels, such as diesel and heating oils, coal or coal liquefaction products and tire

13 considered herein some of the PAH emission could be attributed to surviving PAH

constituents of the fuels themselves.

Despite a lot of work that has been done on pyrolysis of TDF, only a few studies

have reported the correlation of carbon decomposition and the morphological changes of

the aromatic hydrocarbons in the pyrolysis-byproducts. It was concluded that the

formation route of PAH might have been initially caused by the depolymerization from

rubber compounding. These studies have raised public concerns regarding the

management and environmental impact from the combustion of TDF. Most

investigations were undertaken to evaluate the use of scrap tires as a supplementary fuel

source in industrial applications.

1.4 References

1. U.S. Scrap tire markets 2003 Edition. 2004, Rubber Manufacturers Association:

Washington, DC.

2. Illegal dumping prevention guidebook. 1998, US.EPA Region 5Waste, Pesticides

Land Toxics Division: Chicago, IL.

3. Clark, C., Meardon, K., and Russell, D., Burning tires for fuel and tire pyrolysis

air implications. 1991, U.S. Environmental Protection Agency Office of Air

Quality Planning and Standards: Research Triangle Park, NC.

4. Mark, J.E., Erman, B., and Eirich, F.R., Science and technology of rubber. 2005,

Elsevier Academic Press: Burlington, MA. p.619-611.

5. Barlow, F.W., Rubber compounding : principles, materials, and techniques.

1993, M. Dekker: New York. p.9-55.

14 6. Dodds, J., et al., Scrap tires a resource and technology evaluation of tire pyrolysis

and other selected alternate technologies. 1983, Dept. of Energy. Idaho

Operations Office.: Idaho Falls, ID.

7. Williams, P.T., Besler, S., and Taylor, D.T., The pyrolysis of scrap automotive

tyres : The influence of temperature and heating rate on product composition.

Fuel, 1990. 69(12): p.1474-1482.

8. Streitwieser, A. and Heathcock, C.H., Introduction to organic chemistry. 3rd ed.

1985, New York: Macmillan.

9. Williams, P.T. and Taylor, D.T., Aromatization of tyre pyrolysis oil to yield

polycyclic aromatic hydrocarbons. Fuel, 1993. 72(11): p.1469-1474.

10. Friedman, H.L., Kinetics of thermal degradation of char-forming plastics from

thermogravimetry. Application to a Phenolic Plastic. J. Polym. Sci., 1964. C6:

p.183-195.

11. Chen, J.H., Chen, K.S., and Tong, L.Y., On the pyrolysis kinetics of scrap

automotive tires. Journal of Hazardous Materials, 2001. 84(1): p43-55.

12. Oh, S.C., Jun, H. C., and Kim, H. T., Thermogravimetric evaluation for pyrolysis

kinetics of styrene-butadiene rubber. Journal of Chemical Engineering of Japan,

2003. 36(8): p.1016-1022.

13. Moldoveanu, S., Analytical pyrolysis of natural organic polymers, in Techniques

and instrumentation in analytical chemistry ; v.20. 1998, Elsevier: Amsterdam ;

New York. p.1-31.

14. Kaminsky, W., Menzel, J., and Sinn, H., Recycling of plastics. Conservation &

Recycling, 1976. 1(1): p.91-110.

15 15. Cypress, R. and Bettens, B., Production of benzoles and actives carbon from

waste rubber and plastics materials by means of pyrolysis with simultaneous post-

cracking, in Pyrolysis and gasification, Ferrero, G.L., Maniatis, K., Buekens, A.

and Bridgwater, A. V., Editor. 1989, Elsevier Applied Science: London ; New

York. p. 209-229.

16. Cunliffe, A.M. and Williams, P.T., Composition of oils derived from the batch

pyrolysis of tyres. Journal of Analytical and Applied Pyrolysis, 1998. 44(2):

p.131-152.

17. Pakdel, H. and Roy, C., Simultaneous gas chromatographic--Fourier transform

infrared spectroscopic--mass spectrometric analysis of synthetic fuel derived from

used tire vacuum pyrolysis oil, naphtha fraction. Journal of Chromatography A,

1994. 683(1): p.203-214.

18. Lemieux, P.M., Pilot-Scale Evaluation of the Potential for Emissions of

Hazardous Air Pollutants from Combustion of Tire-Derived Fuel. 1994, U.S.

Environmental Protection Agency: Washington, DC.

19. Leung, D.Y.C. and Wang, C.L., Kinetic study of scrap tyre pyrolysis and

combustion. Journal of Analytical and Applied Pyrolysis, 1998. 45(2): p. 153-169.

20. Atal, A., et al., On the survivability and pyrosynthesis of PAH during combustion

of pulverized coal and tire crumb. Combustion and Flame, 1997. 110(4): p. 462-

478.

21. Levendis, Y.A., et al., Comparative study on the combustion and emissions of

waste tire crumb and pulverized coal. Environmental Science & Technology,

1996. 30(9): p.2742-2754.

16 22. Senneca, O., Salatino, P., and Chirone, R., A fast heating-rate thermogravimetric

study of the pyrolysis of scrap tyres. Fuel, 1999. 78(13): p. 1575-1581.

23. Hubschmann, H.-J., Handbook of GC/MS : fundamentals and applications. 2001,

Wiley-VCH: Weinheim Chichester. p. 444-458.

24. Unapumnuk, K., Keener, T. C., Khang, S., and Lu, M.,, Pyrolysis Behavior of

Tire Derived Fuels at Different Temperatures and Heating Rates. The Journal of

Air and Waste Management Association, 2006. 56:p. 618-627.

25. Violi, A., Anna, A., and Alessio, A., Modeling of particulate formation in

combustion and pyrolysis. Chemical Engineering Science, 1999. 54(15-16): p.

3433-3442.

26. Cypres, R., Aromatic hydrocarbons formation during coal pyrolysis. Fuel

Processing Technology, 1987. 15: p. 1-15.

27. Yang, J. and Lu, M., Thermal growth and decomposition of methylnaphthalenes.

Environmental Science & Technology, 2005. 39: p. 3077-3082.

28. Sahouli, B., et al., Surface morphology and chemistry of commercial carbon black

and carbon black from vacuum pyrolysis of used tyres. Fuel, 1996. 75(10): p.

1244-1250.

17 Chapter 2

2 Research Methodology

2.1 Properties of Tire Derived Fuel Materials

In general, one passenger car tire contains many heterogeneous parts, such as

vulcanized rubber, rubber filler, rubberized fabric, steel cord, fabric belts, and steel wire-

reinforced rubber beads in order to meet the various demands. A typical passenger car

tire contains 24-28 percent of carbon black, 40-48 percent of natural rubber (NR) and 36-

24 of synthetic rubber including styrene butadiene rubbers (SBR) and butyl rubber (BR).

Various types of rubber are used in the different components of passenger car tires. For

instance, SBR are used in tread, NR and NR/BR are used in belt compounds,

NR/BR/SBR and NR/SBR blend are used in subtread, NR/BR blend are used in sidewall,

and 70-100 % of NR with 30% of SBR or BR are used in bead filler [1]. The chemical

structure of rubber compounding is somewhat complicated. Carbon and hydrogen atoms

are cross-linked and bunched together in a way which will achieve the vulcanization

process. For example, NR itself is a polymer composed of isoprene, butadiene, styrene

and isobutyl units. The chemical composition of each TDF sample set is unique and

needed to be determined prior to pyrolysis.

Used tire samples were obtained at random from shredded car tires. Central Ohio

Contractors, Inc. and Cinergy Inc. supplied the TDF samples. The chemical composition of the TDF samples, as given in the material safety data sheets, showed a composition of

16-36% carbon black, <1% clay, <1.5% titanium dioxide, <2.0% zinc dioxide, <2.0%

sulfur, 5-13% petroleum hydrocarbons, and 40-48% rubber compounds. The specific

gravity of TDF samples was measured and found to be in the range of 1.085-1.331. The

18 shredded tire samples were cut and sieved to isolate samples in the range of 1.00-2.00 mm. A typical analysis of the tires, on a steel- and fabric-free basis, is reported in Table

2.1.

Table 2.1 Proximate and ultimate analysis of used tires

Analysis As received Dry Proximate analysis: Moisture, % 2.58 Ash, % 6.99 7.18 Volatile, % 57.09 58.60 Fixed carbon, % 33.34 34.22 Heat content, Btu/lb 15,314 15,712 Ultimate analysis Carbon, % 82.64 Sulfur, % 1.54 Nitrogen, % 0.31 Hydrogen, % 8.04 H/C 1.17

2.2 Experimental Design

Series of experiments were performed to study the pyrolysis of TDF using a

specific design reactor. The task was divided into two categories: (1) thermogravimetric

experiments including reactor design and pyrolysis product distribution study; (2) laboratory analysis of the pyrolysis products including proximate and ultimate analysis, and identification of organic compounds. Descriptions of the experimental set up and the methods used in this study are described in detail below.

2.2.1 Reactor Design

At the beginning of this study, a series of preliminary experiments was performed with the objective of obtaining results that would allow the optimum design of the pyrolysis kinetic studies. These experiments intended to get an initial idea of the temperature range of decomposition of the TDF, and the behavior of the system, so that

19 the main experiments could be designed to achieve optimum performance. Initial

pyrolysis experiments were performed using a commercially available TGA model 2050 manufactured by TA Instruments, Inc. to obtain initial weight change data with respect to temperature.

A TGA reactor was developed at the University of Cincinnati in order to analyze the heterogeneous TDF samples. A schematic diagram of the experimental unit is shown in Figure 2.1. The reactor is capable of performing pyrolysis on 3-4 grams of TDF sample which was found to be a sufficient amount of the material. During the experiment, the TDF sample was placed in a ceramic crucible capable of holding up to 4 grams of shredded tire. The sample container was hung from the bottom weight balance.

The reactor tube was heated via a Lindberg Ltd., electrical heater. The heating temperature was programmed by a Eurotherm, Ltd. temperature controller. A type-K thermocouple, placed in the middle of the heating tube, monitored the reaction temperature. The sample weight change and temperature change were recorded every 10 seconds via the data acquisition system. The carrier gas was introduced through a top nozzle at a constant flow rate for at least 30 minutes in order to purge any remaining oxygen inside the reaction zone. Several flow rates of carrier gas were tested until an optimum gas flow rate was calculated. A flow rate in which the thermogram from the

TGA reactor showed good agreement with data obtained from the commercially available

TGA was used as an optimum flow rate for the subsequent experiments. In this study, a nitrogen gas flow rate of 0.47 L/min was determined to be optimum from the pretrial experiments. The thermocouple was placed inside the reactor tube, 2-4 cm above the sample basket. The measured temperature value represented the actual temperature of

20 the TDF sample. As suggested by the other investigation, the heat transfer effect can be

reduced by purging with the highest carrier gas velocity possible [2]. In this study, two

methods might be used to reduce any error in temperature measurement due to

temperature gradient: (i) bring the thermocouple in contact with the sample or (ii) reduce

the sample size [3]. As such, the effects of heat transfer and temperature gradient in the

TDF sample were eliminated in the TGA experiments. The heater is programmed to give

the sample heating rates of 1, 5, 10, 20, 30, 50 and 100 °C/min to the final pyrolysis

temperatures of 400, 450, 500, 600, 800 and 1000 oC. The final temperature was held for

at least an hour after the final desired temperature was reached. The change of the

sample mass (TG curve) and the derivatives of the mass change (DTG curve) with

respect to reaction temperature were recorded simultaneously. After the completion of

the experiment, a continuous purge of inert carrier gas was injected in the reactor until the temperature returned to room temperature.

The gaseous products of pyrolysis were passed directly through the oil condensation system. The oil condenser unit was filled with dry ice during the process to maintain a temperature below 0 oC. The oil condenser unit was connected to a glass fiber

filter (Whatman type GF/F filter, diameter 47 mm.). The Millipore filter holder was

controlled at a temperature of 100 oC during the experiment.

21

Figure 2.1 Schematic diagram of pyrolysis reactor

22 2.2.2 Pyrolysis Product Distribution Experiment

At the end of each experiment, the condensed oil, residual ash, and filter paper

were removed. The inside reactor tube was rinsed using dichloromethane to remove any

deposited soot. The filters were dried and desiccated for at least 24 hours. The mass of

the residual ash, soot, and condensed oil were determined by gravitational method. The

mass of the gases was calculated by the difference between residual ash, soot, condensed

oil and the initial weight of the TDF sample.

2.2.3 Proximate and Ultimate analysis

The residual pyrolysis char was removed from the sample basket after each

experiment. The char was dried, desiccated for at least 24 hours, and then sent for

analysis of the principle fuel properties. Proximate analysis was used to determine the

moisture, ash, volatile, and fixed carbon content. Ultimate analysis was used to determine the elemental composition in terms of carbon, hydrogen, sulfur, nitrogen, and

oxygen. The fuel property analysis of the pyrolysis char was performed by OKI analytical laboratory, a certified commercial laboratory in Cincinnati. The analytical methods were performed according to ASTM D3172 Standard Practice for Proximate

Analysis of Coal and Coke, ASTM D5865 Standard Test Method for Gross Calorific

Value of Coal & Coke and ASTM D3176-89 Standard Practice for Ultimate Analysis of

Coal and Coke.

2.2.4 Identification of Organic Compounds

The gaseous product was neglected in this study because only trace amount of gas emission could be recovered from the TGA. The oil and char from pyrolysis can be

23 recovered as valuable byproducts from the TDF pyrolysis. It has been suggested that the

oils derived from the pyrolysis of used tires have a high calorific value similar to medium

heating oil, and can be used directly as fuel or added to petroleum refinery feedstocks[4-7].

Moreover, some studies stated that the recovered oil and char product from the TDF

pyrolysis has made tires into one of the most economical recyclable resources [8]. In this

study, the laboratory analysis was divided into two categories: chemical structural

analysis of pyrolysis oil and proximate and ultimate analysis of pyrolytic char.

2.2.4.1 Gas Chromatography coupled with Mass Spectrometry (GC-MS)

For the pyrolytic oil, a good clean-up and the necessary concentration of the analyte was required for the analysis. A 2-μL portion of the pyrolytic oil was diluted in

10 mL of DCM. The diluted oil was filtered through a polytetrafluoroethylene pre-cut membrane filter. The filtrate was concentrated by gently introducing nitrogen gas to the sample. The final elution volume was 10 mL. The soluble organic fragments in the pyrolytic oil were analyzed by gas chromatography coupled with mass spectrometry (GC-

MS). The GC-MS system consists of a Varian Model CP-3800 GC with a CP-Sil 8 CB

Low Bleed/MS capillary column, a CP-8400 auto sampler and a Varian Saturn 2200 mass

spectrometer. Basically, the specification of the CP-Sil 8 CB capillary column was

comparable to the DB-5 capillary column that was designed for the analysis of the

standard list of 16 PAHs by the EPA. Structural formulas were specified in the GC-MS

by the EPA 525 analytical method. The identification of interesting components present

in the pyrolytic oil sample was first carried out manually by comparison of the spectra

with the National Institute of Standards and Technology (NIST) library. The

24 quantification of interesting components which have been identified by spectra comparison was then carried out with a number of external standard calibration curves.

Prior to calibration, the MS was auto-tuned to acetonitrile using criteria established by the instrument manufacturer. A three-point response factor (RF) calibration curve was established for analytes of interest prior to analysis of samples and quality control samples. An RF was determined for each analyte and calibration level by using the following equation:

As RF = ------(2-1) Ca Where:

As = the area of the characteristic ion for the analyte to be measured

Ca = the known concentration of the analyte in the calibration solution (μg/mL)

A mid-level calibration standard was analyzed at the beginning or the end of each analytical set. The RF for each compound was determined of each time in order to make comparisons to the previous RF values from the initial calibration curve. The percent difference among all analytes must be less than 25%. If the calibration check did not meet this criterion then the initial four-point calibration was repeated.

The initial calibration of the GC-MS must meet the previously described criteria prior to sample analysis. The pyrolytic oil samples were analyzed in analytical sets that consisted of standards, samples and quality control samples. Quality control samples were methods blanks, laboratory duplicates, and standard reference materials. The type and number of quality control samples depended upon the material availability. An autosampler was used to inject 0.5 μL of all samples, standards and QC samples into the

25 capillary column of the GC using the EPA 525 original method under the following

conditions; however, slight modifications were necessary depending upon the analysis.

The injection temperature was 280 oC in splitless mode. The transfer line temperature

was set at 280 oC. The carrier gas was ultra-high purity helium and the flow rate was 1.2

mL/min. The oven temperatures were programmed as follows:

Temp (oC ) Rate (oC/min) Hold (min) Total (min) 40 - 2 2 200 10 0 18 270 5 0 32 300 10 10 45

The effluent from the GC capillary column was routed directly into the ion

sources of the MS. The MS was operated in the Electron Ionization, full scan mode

ranging 100-300 u at a scan rate of 1 second/scan. In order to include the quantification

and confirmation masses for the interesting compounds, the MS was operated in the selected ion monitoring mode by the appropriated selected windows. For all compounds detected at a concentration above the method detection limit, the confirmation ion was checked to confirm the analyte’s presence.

The current profiles of the extracted ions of the primary m/z and the confirmatory ion for each analysis must meet the following criteria [9]:

ƒ The characteristic masses of each analyte of interest must be in the same

scan or within one scan of each other. The retention time must fall within

±5 seconds of the retention time of the authentic compound or alkyl

homologue grouping determined by the analysis of the daily calibration

check or reference external standard, respectively.

26 ƒ The alkylated homologue groupings (e.g. C2-benzene) appear as a group

of isomers. The pattern and the retention time window for each group

were established by the analysis of a reference external standard. Each

group of alkylated homologues was integrated in its entirety and the total

area response was used to determine the concentration of the entire group.

Identification of the isomers was carried out by comparison of retention

times with those of standard substances.

ƒ The relative peak heights of the primary mass ion, compared to the

confirmation or secondary mass ion, must fall within ±30 percent of the

relative intensities of these masses in a reference mass spectrum. The

reference mass spectrum was obtained from the NIST database or the

reference external standard for the parent compounds and alkylated

homologues, respectively. However, in some instances, a compound that

did not meet secondary ion confirmation criteria may still be determined

to be present in a sample after comparison of the data by the Wiley 625

mass spectra library reference book or the eight peak index reference

book. Supportive data included the presence of the confirmation ion, but

a different ratio.

ƒ Data not meeting the criteria established in this section were appropriately

qualified or re-analyzed.

Sample analytes’ concentrations were calculated based on the concentration and response of the external standard compounds. The equations in Section

2.3.1.1 were used to calculate the RF of each analyte in relation to the

27 concentration and area of the external standard in the initial calibration. RFs for

alkylated homologues were assumed equal to the RF of the respective parent

compound.

The mass (Ma) of each target analytes (μg), including alkylated homologues, was

calculated using the following equation:

Aa Ma = ------(2-2) RF Where:

Aa = the area of the characteristic ion for the analyte measured

RF = the response factor for the analyte from the current calibration

The concentration ( C ) of each target analyte in a sample (μg/L) was calculated

using the following equation:

MaDF C = ------(2-3) W Where:

DF = the dilution factor applied to the pyrolysis oil sample

Volume of DCM (μL) DF = Volume of the pyrolytic oil sample used to make dilution (μL)

W = the sample volume (L)

The initial calibration must pass established criteria before sample analysis can begin. All continuing calibration checks must pass established criteria for analysis to continue.

The mass of carbon in each target analyte in an pyrolytic oil sample (μg-C/g-

TDF) was calculated using the following equation:

28 C × DF ×12× C Mass in C-basis = m ------(2-4) Densityofoil × MW Where:

Cm = the number of carbon atom appeared in the target compound

MW = the molecular weight of the target compound

2.2.4.2 Fourier Transform Infrared Spectrometer (FT-IR)

Organic functional groups within the pyrolytic oil sample were confirmed by a

Fourier Transform Infrared Spectrometer, coupled with attenuated total reflectance (FT-

IR/ATR). The FT-IR/ATR (Nicolet Mogna 760) was connected with a ZnSe crystal.

The ATR technique has been known to obtain qualitative spectra of solid materials regardless of thickness. The absorption spectrum was generated when light passed through the sample sheet to a depth of a few micrometers [10]. In this analysis mode, a crystal prism with a high refractive index physically contacts the surface layer of the specimen. Absorption bands were detected from the corresponding vibrational modes within the surface of the materials contacting the prism [11]. Basically, FT-IR measures vibrational excitation of atoms around the bonds that connect them. Each molecule shows a characteristic infrared spectral pattern in the absorption region due to bond strength and bonding motions [11]. Characteristic peaks were absorbed for specific functional groups, a result of stringing bending, other modes of vibration, or combination a combination of both factors [10-13].

The stretching frequency for a bond between two atoms of mass m1 and m2 can be calculated [12]:

29 1 (m + m ) ν ′ = K 1 2 ------(2-5) 2πc m1m2

Where:

ν ′ = absorption frequency (cm-1)

c = speed of light (3 x 1010 cm/sec)

K = bond force constant (dyne/cm or g/sec2)

12 -23 m1, m2 = atomic mass in grams; mass C = 2 x 10 g

In this study, a small droplet of the pyrolytic oil was deposited on a potassium bromide (KBr) pressed disk and the spectrum was acquired via transmission infrared spectroscopy (IR). The microdisk technique permits examination of sample weight as small as 1 μg [10]. A pure spectrum of the pyrolytic oil was obtained since KBr contains no absorption bands in the IR region. The technique for obtaining the infrared spectra of the TDF sheet was more diversified than for the pyrolytic oil. IR vibration frequencies characterized the presence or absence of functional groups for the hydrocarbons.

Principal functional structures of the TDF materials were confirmed.

2.2.4.3 Gas Chromatography equipped with an atomic emission detector (GC-

AED)

The oil produced from pyrolysis was collected in a condenser trap. It had a yellowish color and a distinctive smell. The derived oil was analyzed for total sulfur content using an Agilent (Agilent Technologies, Palo Alto, CA) Model 6890 Series gas chromatograph equipped with an Agilent G2350A atomic emission detector (GC-AED).

The AED has been well described as efficient in the quantification of elements including

30 sulfur and its compounds by selective spectrometric detection technique [14-17]. The sulfur elements were separated on a HP-5MS type capillary column (30 m x 0.25 mm i.d. x 0.25

μm film). Helium was used as a carrier gas at the column flow rate of 1.3 mL/min. The split ratio was 20:1 and an injection volume was 1 μL. The injection temperature and the

AED transfer line were controlled at 280 and 310 oC respectively. The GC oven temperature program was heated from 40 oC to 300 oC at 10 oC/min and held for 10 min.

A combination of 50 psi oxygen and 45 psi hydrogen was used as a reagent gas with a make up flow rate of 66 mL/min. The detector cavity block was kept at 300 oC.

Chromatograms were simultaneously monitored at the emission lines of carbon (179 nm) and sulfur (181 nm).

2.3 References

1. Mark, J.E., Erman, B., and Eirich, F.R., Science and technology of rubber. 2005,

Elsevier Academic Press: Burlington, MA. p. 619-611.

2. Szekely, J., Evans, J.W., and Sohn, H.Y., Gas-solid reactions. 1976, Academic

Press: New York. p. 205-246.

3. Chen, J.H., Chen, K.S., and Tong, L.Y., On the pyrolysis kinetics of scrap

automotive tires. Journal of Hazardous Materials, 2001. 84(1): p. 43-55.

4. Clark, C., Meardon, K., and Russell, D., Burning tires for fuel and tire pyrolysis

air implications. 1991, U.S. Environmental Protection Agency Office of Air

Quality Planning and Standards: Research Triangle Park, NC.

5. Mirmiran, S., Pakdel, H., and Roy, C., Characterization of used tire vacuum

pyrolysis oil: Nitrogenous compounds from the naphtha fraction. Journal of

Analytical and Applied Pyrolysis, 1992. 22(3): p. 205-215.

31 6. Urban, D.L. and Antal, M.J., Study of the kinetics of sewage sludge pyrolysis

using DSC and TGA. Fuel, 1982. 61(9): p. 799-806.

7. Gallagher, P.K. and Brown, M.E., Handbook of thermal analysis and calorimetry.

1998, Amsterdam [Netherlands] ; New York: Elsevier.

8. Kim, S., Park, J.K., and Chun, H.-D., Pyrolysis kinetics of scrap tire rubbers. 1.

using DTG and TGA. Journal of Environmental Engineering-ASCE, 1995. 121(7):

p. 507-514.

9. Hubschmann, H.-J., Handbook of GC/MS : fundamentals and applications. 2001,

Wiley-VCH: Weinheim Chichester. p. 444-458.

10. Silverstein, R.M., Webster, F.X., and Kiemle, D.J., Spectrometric identification of

organic compounds. 2005, J. Wiley & Sons: Hoboken, N.J. p. 3-164.

11. Ege, S.N., Organic chemistry : structure and reactivity. 1994, D.C. Heath:

Lexington, Mass. p. 376-393.

12. Crews, P., Rodr ํguez, J., and Jaspars, M., Organic structure analysis. 1998,

Oxford University Press: New York. p. 317-347.

13. Colthup, N.B., Daly, L.H., and Wiberley, S.E., Introduction to infrared and

Raman spectroscopy. 1990, Academic Press: Boston. p. 247-288.

14. Inoue, S., Takatsuka, T., Wada, Y., Hirohama, S., and Ushida, T., Distribution

function model for deep desulfurization of diesel fuel. Fuel, 2000. 79(7): p. 843-

849.

15. Link, D.D. and Zandhuis, P., The distribution of sulfur compounds in

hydrotreated jet fuels: Implications for obtaining low-sulfur petroleum fractions.

Fuel, 2006. 85(4): p. 451-455.

32 16. Ramil Criado, M., Rodriguez Pereiro, I., and Cela Torrijos, R., Selective

determination of polychlorinated biphenyls in waste oils using liquid-liquid

partition followed by headspace solid-phase microextraction and gas

chromatography with atomic emission detection. Journal of Chromatography A,

2004. 1056(1-2): p. 263-266.

17. van Stee, L.L.P., Brinkman, U.A.T., and Bagheri, H., Gas chromatography with

atomic emission detection: a powerful technique. TrAC Trends in Analytical

Chemistry, 2002. 21(9-10): p. 618-626.

33 Chapter 3

3 Thermogravimetric Study and Pyrolysis Kinetic Mechanisms

3.1 Abstract

Pyrolysis product distribution rates and pyrolysis behavior of tire derived fuels (TDF) were investigated using thermogravimetric (TGA) techniques. A TGA was designed and built in order to investigate the behavior and products of pyrolysis of typical TDF specimens. Fundamental knowledge of thermogravimetric analysis and principal fuel analysis is applied in this study. Thermogravimetry of the degradation temperature of the

TDF confirms the overall decomposition rate of the volatile products during the depolymerization reaction. The principal fuel analysis (proximate and ultimate analysis) of the pyrolysis char products shows the correlation of volatilization into the gas and liquid phases and the existence of fixed carbon and other compounds that remain as a solid char. The kinetic parameters were calculated using least square with minimizing sum of error square technique. The results show that the average kinetic parameters of

TDF are the activation energy, E = 1322 ± 244 kJ/mol, a pre-exponential constant of A=

2.06 ± 3.47 x 1010 min-1, and a reaction order n = 1.62 ± 0.31. The model-predicted rate equations agree with the experimental data. The overall TDF weight conversion represents the carbon weight conversion in the sample.

Keywords: Thermogravimetric analyzer (TGA), Tire Derived Fuels (TDF)

34 3.2 Introduction

The use of TDF as a fuel in utility plants is a contentious topic and the potential risks could possibly limit its use. Therefore, any understanding which will clarify the advantage of TDF and minimize the formation of hazardous organic volatile carbons during the combustion of shredded tires will indeed increase the potential of using TDF as a fuel for many industries. The purposes of this study were to: (1) generate kinetic parameters for TDF pyrolysis mechanisms by dynamic heating methods, and (2) generate a temperature profile of carbon weight conversion in the design of the pyrolytic reactor.

The experiments were designed under the hypothesis that changes in the pyrolysis kinetic parameters due to temperature and heating rate can alter the level and generation rate of the pyrolytic products.

3.3 Experimental Method

Details of experimental system and design of the laboratory scale pyrolysis reactor have been discussed in Chapter 2 and are briefly described here. A TGA reactor has been developed at the University of Cincinnati in order to analyze the heterogeneous

TDF samples. The reactor is capable of performing pyrolysis on 3-4 grams of TDF sample, which was found to be a representative amount of the material. The final temperature was held for at least an hour after the final desired temperature was reached.

The change of the sample mass (TG curve) and the derivatives of the mass change (DTG curve) with respect to reaction temperature were recorded simultaneously. The experimental set up contains two categories of the pyrolysis of the TDF: the preliminary study with a commercially available TGA to design the laboratory scale reactor and thermogravimetric experiments to determine the pyrolysis kinetic parameters.

35 3.4 Results and Discussion

3.4.1 Preliminary Experiments to Establish the Reactor Design Criteria

A thermogravimetric analyzer measures changes in weight of a sample with increasing temperature. A representative sample size for any commercially available

TGA is 50-100 mg [1, 2]. The optimum heating rate for solid devolatilization experiments is 10 oC/min. In most cases, the presented results represent the mean of four pyrolysis experiments. The standard deviation of those experiments fall into the range of ±0.5.

The influence of the carrier gases used in the TGA 2050 was investigated by applying nitrogen and oxygen at a flow rate of 0.01 L/min and at identical heating rates.

Thermogravimetric data from the preliminary experiments is presented in Figure 3.1.

100 1.2 100 0.8 TG TG 90 DTG DTG 90

80 80 0.6 70 0.8 70 60

50 60 0.4 dW/dT dW/dT

40 TDF weight loss, %

TDF weight loss, % TDF loss, weight 50 0.4 30 40 0.2 20

30 10

0 0 20 0 0100 200 300 400 500 600 0100 200 300 400 500 600 Temperature, oC Temperature, oC

a) TG-DTG curve of TDF in O2 b) TG-DTG curve of TDF in N2

o Figure 3.1 TG-DTG curves of TDF in O2 and N2, at 0.01L/min, heating rate 10 C/min

36 The DTG plot of this differential provided information on exothermic and endothermic reactions during the pyrolysis. The DTG curve (Figure 3.1b) showed clearly that there was only one area of weight loss when nitrogen gas was used. Nitrogen carrier gas may reduce the extent of secondary reactions such as thermal cracking, repolymerization and recondensation [3]. Therefore, nitrogen gas was selected as the carrier gas in further experiments with the TGA. The influences of heating rate on the decomposition temperature were also investigated. As a result, after the reacting temperature reached 450-500 oC, the decomposition rate was independent to temperature change. Williams[3] analyzed the thermogravimetric pyrolysis of tires and tire components in a static batch reactor in a nitrogen atmosphere and at heating rates between 5 and 80 oC/min. They showed that the thermal decomposition started at 250 oC. Pyrolysis was essentially complete by 550 oC. Above that temperature there was no further weight loss. The heating rate has a significant effect on decomposition temperature.

120

o 110 5 C/min 10 oC/min o 100 20 C/min 25 oC/min 90

80

70

60 TDF weight loss, % loss, weight TDF

50

40

30

20 0100 200 300 400 500 600 700 800 Temperature, oC

Figure 3.2 TG curve of TDF with the variation of heating rates

37 Shown in Figure 3.2, the TG curves shifted to higher temperatures as the heating rate increased. The influence of the nitrogen flow rate to the phase transition of TDF sample obtained from the TGA (N2 0.47 L/min) and the TGA 2050 (N2 0.01 L/min) is shown in Figure 3.3.

1.2 120 0.01 L/min 110 0.01 L/min 0.47 L/min 0.47 L/min 1

100

90 0.8

80 0.6

70 dW/dT

60

TDF weight loss, % 0.4

50

0.2 40

30 0 0 200 400 600 800 1000 20 Temperature, oC 0200 400 600 800 1000 1200 Temperature, oC

a) TG curve b) DTG curve

Figure 3.3 Comparison of the thermogravimetric results

In the commercial TGA, the pyrolysis experiment was performed using nitrogen at a flow rate of 0.01 L/min and at a heating rate of 10 oC/min. A comparison thermogram was prepared from the TGA at the same heating rate of 10 oC/min but at a nitrogen flow rate of 0.47 L/min. As the gas flow rate was increased, the thermal degradation temperature shifted to higher temperature. Qualitative pyrolysis behavior and temperature ranges for TDF degradation in the TGA were in agreement with data obtained from the commercial TGA. The impact of the carrier gas flow rates was similar

38 from both the TGA 2050 and the TGA. Therefore, the use of thermogravimetric data in the TGA was a valid method for the pyrosynthesis study of TDF sample.

3.4.2 Investigation of TDF Decomposition

Thermogravimetry measures the mass loss during the decomposition of the TDF sample as it is heated. A sample pan is usually specifically designed for a small mass of solid samples for thermal decomposition studies. The size of the sample pan on the commercially available TGA limited its application for determining the distribution of pyrolytic products. Pre-trial experiments were operated using the TGA from TA

Instruments, Inc. This proved unsatisfactory since TDF is a heterogeneous material and larger sample sizes were required for analyses. A nitrogen gas flow rate of 0.47 L/min. was determined to be optimum from the pretrial experiments. The thermogravimetric experiments were performed using the TGA on TDF sample sizes of 3-4 g. The gas residence time is the time interval during which a representative element of material actually remains in the reactor [4], mainly determined by the gas velocity and length of the reactor tube [5]. For this section of the experimental work, the gas residence time was determined by the gas velocity (0.24 m/min) and the length of the reactor (0.78 m). As a result, the inert carrier gas flow rate of 0.47 L/min gave the maximum gas residence time of approximately 3.25 min within the reactor.

39

Figure 3.4 Weight conversion at different heating rates from TGA

Figure 3.4 shows the thermo depolymerization curves of the TDF pyrolysis for a sample size of 1-2 mm. The repeatability and precision of the entire analysis process was demonstrated by conducting at least three sets of the experiment for each individual conditions. The TG curves show continuous loss in weight as the temperature increases and become level at temperatures greater than 600 oC. The TGA experiment indicates completion of the pyrolysis process after 600 oC. From the thermogravimetric data, at the temperatures below 300 oC, the temperature was not high enough for the reaction to become significant and thus no weight loss was observed. As the temperature increased from 300 to 600 oC, the evaporation and devolatilization processes were essentially completed. No additional weight loss as a function of temperature was found after 600 oC. The parallel curves indicate that the reaction mechanism was not changed.

40 Williams[6] reported that it resulted from the heat transfer limitation of the sample.

Typically, the elementary consecutive or parallel reactions from the pyrolysis of TDF takes place in the thermal cracking of alkenes[3].

Figure 3.5 shows the derivative of the TG curve which helps to pinpoint the maximum weight loss. An identical peak was evident within the range of 300-500 oC on the derivative curves. Corresponding results were also reported by William and Besler

[3], Chen[7] and Senneca[8]. The data in Figure 3.5 show that low heating rates cause a slower decomposition and the reaction takes place in a more restricted temperature range than at higher heating rates. The consequence is that heat transfer limitations at low heating rates are much lower compared to higher heating rates. The result of this is that kinetic data can be approximated using the lower heating rate data [9].

Figure 3.5 Weight conversion rate at different heating rates from TGA

41

The influence of the heating rate on the degradation temperature was observed at the heating rates: 1, 5, 10, 20, 30, 50 and 100 oC. These degradation temperature results are presented in Table 3.1. The decomposition reaction temperature shifts forward when the heating rate is increased. Increasing the heating rates reduces the overall weight loss, as shown in Table 3.1. The change of thermal degradation has been attributed to the combined effects of heat transfer at different heating rates and the kinetics of decomposition, which results in delayed decomposition [5, 8].

Table 3.1 Influence of various heating rate on degradation reaction temperature

o o o o Heating rate, C/min Ti, C Tm, C Tf, C Wf/Wi, %

1 291.7 351.9 446.4 46.0 5 312.1 369.5 482.8 44.5 10 327.9 378.4 497.3 41.5 20 355.5 418.2 529.1 41.0 30 388.4 459.0 571.3 41.4 50 409.2 526.6 597.2 41.4 100 477.4 555.1 675.7 40.2

From Table 3.1, the initial decomposition temperature (Ti) of the TDF sample is

o o about 290-477 C and the final reaction temperature (Tf) is in the range of 446-676 C for the TDF sample. The peak temperature (Tm) at which the conversion rate is maximum or dW/dT =0 is about 352-555 oC. The results in this work are consistent with the decomposition temperature of processing oil and other rubber compounds which have been reported elsewhere [7, 10, 11].

42 Once the weight loss of the sample was determined, the single rubber decomposition fraction and the kinetic parameters were easily calculated by selecting the

Ti, Tm, Tf from the DTG experiments. However, the position of Tm shows significant overlap between the combined components. It would be difficult to identify Ti, Tm and Tf for individual rubber compounds. However, the overall kinetic parameters, which is found to be more appropriated[11] can be determined from the TG and DTG shown in

Figures 3.4 and 3.5 respectively.

3.4.3 Pyrolysis Kinetic Interpretation of Experimental Data

The mechanisms of pyrolysis include a wide range of different reactions. Not every reaction causes the release of a volatile molecule. Some only cause a change in mechanical properties. During thermogravimetric analysis, only the weight loss is measured during heating, meaning that only those reactions causing the weight loss are considered. Furthermore, because of the system complexity, all reactions are lumped together into an overall reaction. The primary rate expression used to describe depolymerization reaction is the Arrhenius expression using the following equation:

E dX − = Ae RT ()1− X n ------(3-1) dt

Where:

A = The Arrhenius pre-exponential factor

E = The activation energy of the reaction

n = The overall rate constant

R = The universal gas constant; 8.3144 J mol-1 K-1

T = The temperature, K

43 t = Time, min

dX = The derivative of conversion with respect to time; dt

Where:

X = dimensionless weight conversion;

()W −W X = i t ------(3-2) Wt −W f

Where:

Wi = Initial weight

Wt= Weight at any time t

Wf = Final Weight

This rate expression has previously been used successfully to account for the pyrolysis of similar compounds, such as commodity plastics, using data from dynamic weight loss experiments using a thermogravimetric analyzer [9]. Under the dynamic method, the sample is heated at a constant rate. Thereby, the weight loss and the decomposition rate are obtained as a function of temperature and conversion. The advantage of this method is that the kinetic parameters can be obtained from just one experiment; however, these parameters are only able to describe the overall decomposition reaction. It can not be used to study the reaction in detail, like with isothermal methods, because the parameters are determined for the whole temperature range rather than at a specific temperature. Therefore, the parameters have a temperature-average character. This method is more realistic when the experimental reaction conditions are similar to those of a practical application (e.g., a pyrolysis unit).

44 Thus, it is more reasonable to predict the degradation behavior of TDF in a pyrolysis unit by performing experiments with dynamic measurements. Furthermore, this method is also very suitable for a fast overall characterization of decomposition behavior.

By taking natural log on both sides of the equation 3-1, a linear least square regression was used to find the initial kinetic parameters A , E and n:

dX E 1 ln= ln A − + nln()1− X ------(3-3) dt R T E From equation 3-3, the three unknowns – lnA, − , and n – from the R experimental results can be obtained in the linear form:

Y = a + bX1 +cX2 ------(3-4) dX 1 Where, the individual heating rate with ln , and ln(1− X ) were used as the dt T variable parameters. Calculation procedures were performed on Mathematica, a mathematical computing package. The Mathematica program performs a definite and efficient procedure to find which linear or nonlinear combination of a list of functions gives the best least-squares fit to the experimental data[12]. The result was the best linear combination of the functions 1, X1 and X2 . As such, the three variables – lnA, -E/R, and n – were obtained directly from the experimental data. A demonstration of the mathematical program is shown in Appendix A.

The least-square regression technique was applied to determine the initial E, A, and n values. The least-square method was used afterward – minimizing the sum of square of the conversion (SSQ) techniques based upon the initial results from the linear regression. This technique was used to minimize the quantity of the difference between each original data point and the model fitted value. The activation energies of

45 degradation of the TDF samples were determined by adjusting the parameters E, n, and lnA with the best-fit techniques[9]. First, the activation energy was fixed, then the best fit values of n and lnA were estimated by correlation of the least-square method and SSQ technique.

1 N SSQ = ∑()Experimentalconversion − Calculatedconversion 2 ------(3-5) N i=1 Where N was the number of data points used. The one simple Arrhenius expression and the constant heating rate experiments for TDF pyrolysis in the TGA resulted in good agreement with each other. The analysis results for different heating rates obtained from the model and the experimental data are shown in Figure 3.6.

Analysis of Figure 3.6 shows that the initial decomposition time shifts toward the right as the heating rate increases. In other words, increasing the heating rate delays the decomposition step of the TDF due to heat transfer effects.

46 Model 1 oC/min o 0.8 5 C/min 10 oC/min 20 oC/min 30 oC/min o

sion 50 C/min r 100 oC/min

Experiment 1 oC/min 5 oC/min Weight Conve 0.4 10 oC/min 20 oC/min 30 oC/min 50 oC/min 100 oC/min

0

0.1 1 10 100 1000 Decomposition Time, min

Figure 3.6 Plot of TDF pyrolysis versus decomposition time. Symbols are the experimental data, and lines are the calculated data.

Table 3.2 shows the results obtained from the best fit set of parameters that can be found from the minimum SSQ between the calculated conversions and the respective measured conversions. A comparison of pyrolysis kinetics parameters with the various analytical methods is given in Table 3.2. The reaction order and pre-exponential data agree with published data [7, 10, 11, 13]. The difference in activation energies may be due to the fact that those calculated in our tests are based on manufacturer combinations of SBR,

NR, BR and other unknown binding and inhibitor agents, as opposed to pure compound activation energies as shown in the Table 3.2.

47 Notice that Kim [10] and Yang [11] initially assumed the reaction order n = 1 for BR and SBR and n= 2 for NR composition. The activation energy and pre-exponential value were averaged from the various heating rate and the assumed reaction order. Therefore, the activation energy and the pre-exponential data were based on the assumption of reaction order. Chen[7] applied the Arrhenius relation and used the least square method to develop overall kinetic parameters of scrap automotive tire. In his study, the activation energy was averaged from the data collected at various heating rates. Meanwhile, the reaction order and pre-exponential value resulted from the calculated activation energy.

The kinetic parameters that resulted from his studies excluded the effect of different heating rates and cannot give the overall reaction order of thermal degradation at any time. Oh[13] used the multiple heating rate experiment to determine the pre-exponential factor. Although the kinetic parameters for each heating rate were obtained, the results were only slightly affected by the heating rates.

In this study, the weight loss and the degradation rate were obtained as a function of pyrolytic temperature at the individual heating rate. Additionally, the kinetic parameters can be obtained directly from the TGA experimental data.

48 Table 3.2 Kinetic parameters of tire pyrolysis Reaction Pre-exponential factor Activation energy, E Tire sample order, n (A), min-1 (kJ/mol) TDF(least-square &minimizing error square) Heating rate 1 oC/min 1.07 1.32 x 103 751 Heating rate 5 oC/min 1.52 1.29 x 106 1273 Heating rate 10 oC/min 1.46 1.37 x 107 1532 Heating rate 20 oC/min 2.36 1.26 x 1011 1754 Heating rate 30 oC/min 1.94 1.87 x 1010 1523 Heating rate 50 oC/min 1.37 5.73 x 107 1330 Heating rate 100 oC/min 1.60 1.17 x 105 1095 Mean 1.62 2.06 x 1010 1322 Standard deviation 0.42 4.68 x 1010 329 95%, Confidence Interval 0.31 3.47 x 1010 244

Car[7] 1.98 7.57 x 1010 148 Truck[7] 1.63 5.02 x 1010 148

SBR[13] 0.9-1.2 4.3 x 1013 - 4.7 x 1010 238-241 NR[10] 1 3.89 x 1016 207 BR[10] 1 6.32 x 1014 215

Sidewall rubber[10] 1 2.04 x 1014 204 1 2.08 x 1015 195 1 1.44 x 103 42 Tread rubber[10] 1 8.75 x 108 127 1 3.78 x 1016 209 1 9.34 x 102 39

49 The activation energy was calculated throughout the entire degradation reaction.

Therefore, this method is capable of providing the changes in the pyrolysis mechanisms due to a constant heating rate. Nevertheless this method assumes that the degradation occurs via a single nth order reaction mechanism over the entire degradation of a wide range of compounds. The method in this work may be used to predict the degradation behavior of a carbon-hydrogen aggregate in a pyrolysis unit by performing experiments with dynamic measurements. Although the depolymerization throughout this wide temperature range is dependent upon the decomposition of carbon and hydrogen compounds, the ingredients of tires are complicated and contain a number of additives and cross-linkage agents. The kinetics of tire decomposition are initially rate-controlled as surface reactions predominate. However, in the final stages it becomes more diffusion-controlled due to the morphological changes of hydrocarbons [9]. The pyrolysis of numerous long-chain species of rubber compounds, such as polyisoprene in NR and

SBR, may consist of multiple pyrolytic reactions occurring simultaneously and sequentially.

3.4.4 Decomposition of TDF in Different Atmospheres

Average values of kinetic parameters, as in Table 3.2, indicated that these numbers could be used for the design of inert and non-inert atmospheres. However, industrial units may have significantly higher heating rates than those tested here, and one could expect to see differences in kinetic parameters in those units due to heat transfer effects.

The pyrolysis reactor was modified to support combustion experiments. The nitrogen gas was removed. The TDF sample was put inside the heated reactor tube when the temperature reached 600, 800 and 1000 oC. Figure 3.7 compares the results obtained in

50 the different atmospheres at the different heating rates. As shown, the absence of nitrogen accelerates the decomposition taking place at lower temperatures. As presented in Figure 3.7, the decomposition shifts forward to the right as the combustion temperature decrease. In other words, the low temperature delays the degradation process. There is not a great difference between the runs performed under three different temperatures (600,

800 and 1000 oC). From these observations, the decomposition of TDF in non-inert atmosphere appears in two stages. The first-stage decomposition rates in inert and non – inert atmosphere are comparable.

0.8

Pyrolysis 1 oC/min 5 oC/min 10 oC/min 20 oC/min 30 oC/min 50 oC/min Weight Conversion 0.4 100 oC/min Non-Inert 600 oC Non-Inert 800 oC Non-Inert 1000oC

0

0.1 1 10 100 1000 Decomposition Time, min Figure 3.7 Plot of TDF decomposition in inert and non-inert atmosphere versus decomposition time. Symbols are the inert atmosphere, and lines are the non-inert atmosphere

51 However, the second decomposition process in the absence of an inert atmosphere starts at 70% weight conversion and shows a lower decomposition rate than that of the inert atmosphere. With respect to the two fractions in non-inert atmosphere, a high value of decomposition rate at the beginning was obtained as a consequence of the degradation of the rubber and some additive materials in TDF. The slow decomposition rate in the second stage of decomposition corresponds to the carbon black degradation.

3.5 Conclusion

Pyrolysis kinetics of TDF in a TGA have been investigated thermogravimetrically for the constant heating rates of 1, 5, 10, 20, 30 50 and 100 oC/min with nitrogen as the carrier gas. The results show that the degradation temperature range of the TDF is from

352 to 555 oC. The initial reaction temperature increases when the heating rate is increased. Three kinetic parameters – activation energy, pre-exponential factor and reaction order for the TDF – can be determined directly from the Arrhenius equation and the TGA experiment. Given that the thermograph for each particular heating rate of these experiments is unique, the model for the kinetic parameters can be solved by the least- square technique in conjunction with the minimized sum of error square method. Given that the chemical composition of TDF varies, conclusions drawn here may be limited to the specific TDF used in the experiments. The average kinetics parameters of the TDF are: E = 1322 ± 244 kJ/mol, A= 2.06 ± 3.47 x 1010 min-1, and n = 1.62 ± 0.31. These results can be used for the design and operation of a pyrolysis or combustion system that uses shredded tires. However, industrial pyrolysis units may have significantly higher heating rates than those tested here, and differences in kinetic parameters could be expected to be seen in those units due to heat transfer effects.

52 3.6 References

1. Haines, P.J., Principles of thermal analysis and calorimetry, in RSC paperbacks.

2002, Royal Society of Chemistry: Cambridge. p. 10-54.

2. Hatakeyama, T. and Liu, Z., Handbook of thermal analysis. 1998, Wiley:

Chichester ; New York. p. 42-46.

3. Williams, P.T., Besler, S., and Taylor, D.T., The pyrolysis of scrap automotive

tyres : The influence of temperature and heating rate on product composition.

Fuel, 1990. 69(12): p. 1474-1482.

4. Dodds, J., Domenico, W.F., Evans, D.R., Fish, L.W., Lassahn, P.L., and Toth,

W.J., Scrap tires: a resource and technology evaluation of tire pyrolysis and

other selected alternate technologies. 1983, US. Department of Energy: Idaho

Falls, Idaho.

5. Dai, X., Yin, X., Wu, C., Zhang, W., and Chen, Y., Pyrolysis of waste tires in a

circulating fluidized-bed reactor. Energy, 2001. 26(4): p. 385-399.

6. Williams, P.T. and Brindle, A.J., Temperature selective condensation of tyre

pyrolysis oils to maximise the recovery of single ring aromatic compounds. Fuel,

2003. 82(9): p. 1023-1031.

7. Chen, J.H., Chen, K.S., and Tong, L.Y., On the pyrolysis kinetics of scrap

automotive tires. Journal of Hazardous Materials, 2001. 84(1): p. 43-55.

8. Senneca, O., Salatino, P., and Chirone, R., A fast heating-rate thermogravimetric

study of the pyrolysis of scrap tyres. Fuel, 1999. 78(13): p. 1575-1581.

53 9. Heinzel, A., Keener, T. and Khang, S., The pyrolysis behavior of mixtures of

commodity plastics with polyvinyl chloride in a thermogravimetric analyzer.

Archives of Environmental Protection, 2001. 27(3): p. 11-33.

10. Kim, S., Park, J.K., and Chun, H.-D., Pyrolysis kinetics of scrap tire rubbers. 1.

using DTG and TGA. Journal of Environmental Engineering-ASCE, 1995. 121(7):

p. 507-514.

11. Yang, J., Kaliaguine, S., and Roy, C., Improved quantitative-determination of

elastomers in tire rubber by kinetic simulation of DTG curves. Rubber Chemistry

and Technology, 1993. 66(2): p. 213-229.

12. Wolfram, S., The mathematica book. 2003, Wolfram Media: Champaign, IL. p.

722-982.

13. Oh, S.C., Jun, H.C., and Kim, H.T., Thermogravimetric evaluation for pyrolysis

kinetics of styrene-butadiene rubber. Journal of Chemical Engineering of Japan,

2003. 36(8): p. 1016-1022.

54 Chapter 4

4 The Recovery of Pyrolysis By-products and the Removal of Sulfur

4.1 Abstract

Three commercially important pyrolysis by-products from vehicle tires are liquid oil, solid char, and exhaust gas. These products could be contaminated by volatile toxic compounds released from the tire structure before combustion. The release of volatile materials from the tire structure during pyrolysis offers the greatest potential for separation of these compounds from the evolved gases and vapors, as the concentrations of these species are at their greatest in the vapor phase during this period. The influences of heating rate and pyrolysis temperature were investigated. The temperature was studied from 325 to 1000 oC, a range where substantial devolatilization occurs. The results showed that the heating rate and the pyrolysis temperature were the key factors in determining the two pyrolysis yields: condensed oil and gas product. However, the overall desulfurization of the pyrolysis reaction was essentially unaffected by the heating rates.

Keywords: pyrolysis, Tire Derived Fuels (TDF), pyrolysis products, char, oil, gas, desulfurization, devolatilization, sulfur, carbon

55 4.2 Introduction

Since the 1970s, a number of studies have investigated available technology to combust used tires in many industrial scenarios, especially those applications that relate to the use of tire as supplemental fuels. The growing interest in pyrolysis mechanisms of tires has been promoted along with the study of combustion. Pyrolysis has generally been known as the first step of combustion for any solid fuel. Often times, the yields of the three pyrolysis by-products – char, oil and gas – are associated with the pyrolysis conditions. Consequently, market demand of char and oil has increased in recent years.

However, little attention has been given to the influence of heating rate and pyrolysis temperature to the release of others pyrolysis by-products such as sulfur dioxide and volatile hydrocarbons. This study focuses on an understanding of the fundamental features and influences of heating rate and pyrolysis temperature in relation to the recovery of pyrolysis products, the overall carbon distribution, and the desulfurization mechanism.

4.3 Experimental Method

4.3.1 Determination of By-products Recovery

Details of the experimental system and the design of the laboratory scale pyrolysis reactor have been discussed in Chapter 2. A series of experiments was performed for the heating rates of 1, 5 and 10 oC/min. At the end of each pyrolysis experiment, the condensed oil, residual ash, and filter paper were removed. The mass of the residual ash, soot, and condensed oil were determined gravitationally. The mass of the gases was calculated by the difference between residual ash, soot, condensed oil, and the initial

56 weight of the TDF sample. The percent devolatilization for each TDF sample was calculated using the thermal gravimetric data[1, 2] as follows:

W −W %devolatilization = Char TDF ×100 ------(4-1) WTDF Where:

Wchar = the weight of the solid char or pyrolyzed char, g;

WTDF = the weight of the original TDF, g.

Desirable heat recovery in the pyrolyzed char is expressed as following[3]:

%Yield × H %Heat Recovery = pyrolyzed ×100 ------( 4-2) 100× HTDF Where:

Yield = the proportion of the total weight of pyrolyzed product to the total weight of the original TDF;

HTDF = the heat value of the original TDF, Btu/lb;

Hpyrolyzed = the heat values of the pyrolyzed product, Btu/lb.

Similar to the heat recovery, the ash recovery was calculated using the same equation, but the total weight of residual ash and the percent ash of the original TDF were used instead of the heat value.

4.3.2 Determination of Sulfur and Carbon

The pyrolyzed solid char was analyzed for carbon and sulfur and compared to the original parent carbon and sulfur content in the TDF using the ASTM D3172, Standard

Practice for Proximate Analysis of Coal and Coke. The ASTM method was written specifically for the determination of carbon and sulfur in coal and coke, but should be

57 applied to the same analysis of TDF and pyrolyzed TDF as well. The proximate analysis were performed by OKI analytical laboratory; a certified commercial laboratory in

Cincinnati. The derived oil was analyzed for total sulfur content using an Agilent

(Agilent Technologies, Palo Alto, CA) Model 6890 Series gas chromatography equipped with an Agilent G2350A atomic emission detector (GC-AED).

The percentage of removed sulfur was calculated for each sample by comparing the final total sulfur content of each pyrolyzed sample with an initial sulfur value of the original TDF. TDF does not have a fixed composition and its properties vary, therefore, the sulfur content from one TDF sample may differ from one another. The original parent sulfur content in TDF was 2.5% which was found to be the highest value employed in this study. The desulfurization is the percentage change in the amount of sulfur in the TDF when the TDF is pyrolyzed. The percent desulfurization for each pyrolyzed char and derived oil can be calculated using the following equation[1, 3]:

(S − S ) %desulfurization = TDF pyrolzed ×100 ------(4-3) STDF Where:

STDF = the total percent by weight of sulfur content in original TDF;

Spyrolyzed = the total percent by weight of sulfur content in pyrolyzed products.

The initial sulfur value for each sample was calculated by multiplying the weight of the sample prior to pyrolysis by the 2.5% sulfur concentration. In order to determine the relationship between desulfurization and devolatilization, the thermogravimetric data from our previous publication was used. The repeatability and precision of the entire analysis process was demonstrated by conducting at least three sets of the experiment for each individual conditions. The standard derivations for each data set are used to indicate

58 error for that set. Other than that, the presented results represent the mean value of that data set and the 5% standard error.

4.4 Results and Discussion

4.4.1 Recovery of Pyrolysis By-Products

The decomposition of TDF occurred at the degradation temperature range of 350-

500 oC. In addition, all organic compounds except the carbon black were completely released after heating the polymeric materials continuously in a nitrogen atmosphere to about 600 oC [4]. Therefore, the final pyrolysis temperature at 400 -1000 oC was selected for the study of heat transfer effect. The material balances on pyrolysis products in TDF have been shown to vary with respect to the temperature and the heating rate. These correspond to the greater effect of the volatilization process at higher temperatures.

Figure 4.1, 4.2 and 4.3 illustrate the effect of pyrolysis temperature on the mass balance of four pyrolysis products – oil, char, soot, and gas – at a heating rate of 1, 5, and 10 oC/min, respectively. Each bar represents an average of the experimental results from at least 3 replicated experiments based on error bar plots from the standard error of the triplicate data set. From our experiments, the major decomposition of TDF contributed to solid, gas, and oil products, respectively.

59 Oil Char 0.90 Soot Gas 0.80

0.70

0.60

0.50

0.40

0.30

Product Yield, g/g-TDF Yield, Product 0.20

0.10

0.00 400 450 500 550 600 800 1000 Temperature, oC

Figure 4.1 Pyrolysis product distribution with heating rate of 1 oC/min

0.90 Oil Char 0.80 Soot Gas 0.70

0.60

0.50

0.40

0.30

Product Yield, g/g-TDF Yield, Product 0.20

0.10

0.00 400 450 500 600 800 900 1000 Temperature, oC

Figure 4.2 Pyrolysis product distribution with heating rate of 5 oC/min

60

Oil Char 0.90 Soot Gas 0.80

0.70 0.60 0.50 0.40 0.30

Product Yield, g/g-TDF Yield, Product 0.20

0.10 0.00 400 450 500 550 600 800 1000 Temperature, oC

Figure 4.3 Pyrolysis product distribution with heating rate of 10 oC/min

4.4.2 Char

The percent yield of solid char decreases in a linear fashion as the temperature increases. The maximum percentage of char yield was found (~62% wt.) at 400 oC of 10 oC/min, as the maximum weight loss of TDF reported herewith. The average percentage of char yield was approximately 40 wt%, which may refer to 36-45 % wt. of carbon black used in tire manufacturing [5]. The pyrolytic char yield was found to remain fairly constant with a mean of 37-40 % wt. In comparison of Figure 4.1, 4.2 and 4.3, char yield increased as the pyrolysis heating rate decreased. The decrease of pyrolytic char was attributed to the increase in either devolatilization of solid hydrocarbons in the char or the partial gasification of the char. The formation of char from tire pyrolysis is clearly linked to a number of factors, including: temperature, heating rate, and system specific

61 parameters such reactor size, the efficiency of heat transfer from the hot reactor surfaces to and within the tire’s mass, and gas residence time in the hot zone[6]. More char was produced at a lower heating rate because of the non-volatilisation of fixed carbon[12].

4.4.3 Oil

The oil derived from pyrolysis was collected in a condenser trap. It had a yellowish color and a distinctive smell. The mass of oil was found to vary with respect to the temperature. The contribution may correspond to the greater effect of the volatilization process at higher temperatures. As shown in Figure 4.1, the mass of oil drastically dropped almost 50 wt.% when the temperature increased from 400 to 450 oC.

This figure represents the maximum decomposition rate of TDF to pyrolytic oil through the temperature range of 500 to 600 oC. At a temperature range of 400 to 550 oC, the decomposition of waste tire sample was suspected to be the result of the degradation of polymeric materials. At temperatures beyond 600 oC, the production of pyrolytic oil shows almost no difference in mass. This finding can be explained by the thermo- depolymerization aspect of the parent materials of tires. Considering that the degradation temperature of different rubber compounds varies between 300 – 550 oC, the thermo- depolymerization should be completed after 600 oC.

From the experiments, the pyrolytic oil yielded the maximum weight at approximately 550 – 600 oC. The analysis repeatability shows standard deviations in the range of 0.15-0.4. Increasing the temperature from 350 to 600 oC increased the oil yield from 0.15 to 0.4 g/g of initial TDF weight as rubber depolymerization began. From the pyrolysis experiments, more than 50 wt.% of the initial TDF weight was converted to the depolymerization products, which contributed to 30-40 wt.% in oil and 15-25 wt.% in

62 gas. The oil fraction is hypothesized to be splitting into lower molecular weight gas products since the oil yield slightly decreased over the temperature range of 600 oC to

1000 oC. These results agree with another observation that the yield in light depolymerization products increases with respect to heavier compounds as the pyrolysis temperature and heating rate increases. Nonetheless, the oil yield has been reported to be considerably increased under vacuum conditions when compared to atmospheric pressure conditions [7]. The vacuum pyrolysis reactor will result in a short gas and vapor residence time in the reactor.

Analysis of Figure 4.1-4.3 indicated an independence of the depolymerization reactions to the heating rate after the completion of the decomposition at 500 oC. The quantity of pyrolytic oil was much dependent on the heating rate or heat transfer effect at the temperature range of 500 to 800 oC. These findings imply that the depolymerization reactions for a specific temperature range can be one of the major reactions that break down the polymer chains and/or cross-linked molecules of rubber compounds like natural rubber and synthetic rubber.

The effect of heating rates is not clear for the distribution of the other three pyrolysis products: char, gas, and soot. The average fraction of pyrolytic oil increases significantly with an increasing of heating rate. The yield of oil from pyrolysis reached the maximum at 600 oC when heated at 5 oC/min with a high yield of more than 20 wt.%.

Then, the maximum yield of oil was found at 1000 oC when heated at 10 oC/min. Among the three different heating rates, almost twice as much derived oil was obtained at 1 oC/min versus 5 and 10 oC/min. Although the heat transfer effect can not be seen clearly in the mass distribution behavior, the analysis in molecular structure levels of

63 hydrocarbons may become useful in the explanation of this mechanism. It has been reported that a major component of car tires is hydrocarbons, but with significant differences in their chemical molecular structures and mechanical and thermodynamic properties[4, 8]. Heat and energy changes partially depend on the molecular structural change and the intermolecular reactions in the deformation of the original compounds[8].

For example, the internal energy increases during deformations of NR, but varies during that of filled styrene-butadiene rubber vulcanizates[9]. Therefore, the morphology changes of the molecular structure in pyrolysis products should be studied in the future to further understand how to optimize the production of oils.

Similar high yields of oil obtained at the lower temperature have been reported by other investigators. Dai [6], Dodds[8] and Chang[12] reported that pyrolytic oil yields generally decrease with increasing temperature. Dai[6] showed the maximum yield of

50wt.%. to oil to occur at 450 oC in a circulating fluidized-bed reactor. Since there was no obvious mechanism for carbon loss with increasing temperature, it was suggested that a higher pyrolysis temperature volatilizes some of the solid hydrocarbon content of the char[8]. Sharma[13] observed that the conversion of tires was about 69 wt.% in a batch bomb tube heated at 400 oC under 1000 psi hydrogen atmosphere. This corresponds to almost the entire organic portion of the tires. William[4] produced a maximum of 58.8

o wt.% pyrolytic oil at 600 C at the nitrogen flow rate of 0.11 L/min and a heating rate of

80 oC/min. However, this was for a static fixed-bed with very high heating rate compared to the present work.

It was suggested that high heating rates with short hot-zone residence times and rapid quenching of the products favored the formation of liquid products, since the

64 pyrolytic gases and vapors were condensed before further reactions break down the higher molecular weight species into gaseous products[14]. It was also concluded that, to maximize the liquid yield, the primary products should be removed from the hot zone to prevent secondary reactions from occurring. Cunliffe and Williams[15] concluded that the characteristics of primary vapors produced in the pyrolysis process were the most influenced by heating rate. The higher temperature volatizes some of the solid hydrocarbon of the char and decomposed some oil vapors to gases.

4.4.4 Gas

As presented in Figure 4.1-4.3, there was a corresponding increase in the yield of gas from 10 wt.% at 400 oC to 22 wt.% at 600 oC with the heating rate of 10 oC/min.

After the formation of liquid in the pyrolysis processes, the primary vapors degraded to secondary tars and gases. Therefore, at the higher temperature the pyrolytic gas was increased. The decrease in oil yield – which corresponded with increasing temperature and an increase in pyrolytic gas – was previously documented[12]. Leung[16] reported that the increasing yield of gas with higher temperature can be attributed to the thermal cracking of more oil or char as the pyrolysis temperature increased.

The experiment began with samples at room temperature. Some gases may be produced when the temperature of the sample itself is raised. Due to the high heating rate and short residence time, most of the gases are produced at the maximum temperature[17].

The removal of pyrolytic gases from the hot zone reduces the extent of secondary reactions, which are known to increase the yield of char at the expense of oil formation.

65 4.4.5 Distribution of Carbon

Carbon is one of the most important components of manufactured tires. It is also one of the most valuable components of pyrolysis products, especially solid char products. Carbon distribution in the pyrolysis products, with respect to the pyrolysis final temperature and heating rates, was determined in this work. All product quantities are reported herein as the mass of the TDF input on a carbon basis for consistency and ease of comparison. Carbon distribution in TDF show that only approximately 20% of the total carbon from the TDF is found in the derived oil, and the remainder exits the reactor in solid and gas form. A summary of the carbon balances is shown in Figures 4.4 and

4.5.

0.60 Oil Char 0.50 Gas

0.40

0.30

0.20 g-Carbon/g-TDF 0.10

0.00 01234567891011

Heating rate, oC/min

Figure 4.4 Carbon distribution in pyrolytic by-products as function of heating rates

66 0.60 Oil 0.50 Char Gas 0.40

0.30

0.20 g-Carbon/g-TDF 0.10

0.00 400 500 600 700 800 900 1000

Pyrolysis Temperature, oC

Figure 4.5 Carbon distribution in pyrolytic by-products as function of temperature

The deference of total carbon content in the pyrolytic char and in the pyrolytic oil to the total carbon in the TDF sample was considered as the total carbon content in the pyrolytic gas. Analysis of Figure 4.4 shows that the carbon content in the pyrolytic char dominated at all heating rates. The majority of carbons in the TDF were contributed to the solid, liquid oil and gas phase, respectively. The maximum carbon content was derived from the pyrolytic oil at 5 oC/min. Regardless of the heating rate difference or the heat transfer effect, Figure 4.5 shows that the carbon content of the oil tends to decrease as the temperature increases from 600 to 1000 oC. On the other hand, the carbon content of the gas phase tends to increase over the same temperature range.

Similar to Figure 4.1-4.3, the carbon content of the solid char is the highest among the three pyrolysis products, yet shows almost no change with the temperature increase. The carbon decomposition at higher temperatures may result from thermal depolymerization.

The change in carbon content at different heating rates could be the consequence of the

67 heat transfer effect. The carbon distribution shows no significant change with respect to heat transfer effect and thermal depolymerization reaction for the conditions reported here.

Figure 4.6 Temperature profile and weight conversion of TDF and carbon content in pyrolytic char

Figure 4.6 shows the pyrolysis temperature profile of the carbon content in pyrolytic char and the TDF weight conversion at different heating rates. From this limited data set, it is seen that the mass of carbon content in pyrolytic char is not dependent upon heating rates. Therefore, the production of pyrolytic char and carbon is independent of heat transfer for the conditions reported here. However, the overall mass of the products decreases with increasing temperature. In other words, the mass of

68 carbon in the pyrolytic char is determined by thermal depolymerization rather than by heat transfer. The carbon content of pyrolytic char is closely related to the overall depolymerization reaction. From the above results, the rate of carbon conversion on a weight basis can possibly be expressed by the Arrhenius expression.

4.4.6 Removal of Sulfur

TDF samples were subjected to pyrolysis conditions and the percentage removals of sulfur were determined. The analysis results are listed in Table 4.1. Data in Table 4.1 indicated that the percent of sulfur removal drops with the increase in temperature from

350 oC to 400 oC and then increases with the increase in temperature. Up to 65wt.% of the total sulfur was removed at a temperature of 350 oC and a heating rate of 10 oC/min.

The minimum percent of sulfur removal (36-46 wt.%) was found at 400 oC. The total sulfur removal for 1 oC/min was maintained at 45-50 wt.% from 350 to 1000 oC. Lin[2] reported that when pyrolysis temperatures were high, the pore structure of coal collapses, changing the solid fuel matrix structure which inhibits further release of sulfur. The same hypothesis may apply to the release of sulfur from TDF when pyrolysis temperatures were higher than 447, 483 and 497 oC , corresponding to the final devolatilization temperature reported for 1, 5 and 10 oC/min, respectively[10]. It was observed that the volatilization yield increases as the pyrolysis temperature increases, but the heat recovery and ash recovery from the analysis of pyrolyzed char decreases as the pyrolysis temperature increases. The experimental results indicate that there was almost no change in the overall heating value at the different heating rates. However, there was a slight decrease in the overall heating value when pyrolysis temperatures were grater than 400 oC. Figure 4.7 reveals additional experiments for different heating rates. These results

69 show that percent removal of sulfur in the TDF is initially affected by the pyrolysis temperature. However, the results indicate that the overall desulfurization, as well as the rate of desulfurization of the pyrolysis reaction, shows only minor influences of the heating rate. There was a small influence of heating rate, but a significant change in the percent removal of sulfur was only observed when the temperature was higher than 350 oC.

80

70

l 60

50

40

30 1 deg C/min 5 deg C/min Percent sulfurremova 20 10 deg C/min 10

0 325 350 400 500 600 800 1000

Temperature, oC

Figure 4.7 Percent sulfur removal at different heating rates

70 Table 4.1 The analysis results of TDF samples subjected to pyrolysis conditions

Heating rate, Temperature, oC/min oC Desulfurization, % Devolatilization, % Heat Recovery, % Ash Recovery, % 350 49.4 40.1 54.4 52.9 400 46.1 51.0 38.0 42.0 500 49.3 55.0 29.7 34.2 1 600 53.0 60.1 28.2 33.7 800 50.1 61.8 27.7 31.4 1000 52.5 61.2 28.5 32.0 350 55.5 41.4 45.5 57.7 400 38.9 37.9 50.4 59.0 500 45.3 58.2 30.7 34.4 5 600 34.5 51.5 34.1 41.7 800 48.0 60.6 28.1 32.4 1000 47.7 60.5 29.6 31.9 350 64.3 40.8 45.6 57.3 400 36.0 33.6 58.7 59.6 500 44.8 56.3 31.4 37.0 10 600 41.0 53.6 31.3 39.2 800 48.6 60.1 29.0 32.4 1000 46.2 60.4 28.5 33.7

71 The graphical representations of the sulfur contents in the solid char, oil, and TDF are shown in Figure 4.8, and those of the sulfur removals as a function of devolatilization are shown in Figure 4.9. Our investigation indicated that the overall derived-oil yield was essentially unaffected by the heating rate of the pyrolysis condition[11]. However, the maximum oil yields were found at the heating rate of 5 oC/min, and so the pyrolysis experiments for the derived oil were performed at that specific heating rate. In the GC-

AED experiment, the sulfur contents of derived oil with regardless of speciation were determined. Analysis of Figure 4.8 shows that the initial percent of sulfur was determined to be less than 0.2 wt.% at 400 oC and increased rapidly to about 2 wt.% with the increase in temperature. Figure 4.8 reveals that within 350-850 oC, also called the devolatilization range of the TDF, the majority of sulfur remains in the pyrolyzed char rather than being released into the condensed vapor phase. The percent of sulfur in the derived oil and pyrolyzed char increases with the increase in temperature, maximizes at

450-500 oC, and then seems to remain constant. When the reaction temperature increases, the amount of decomposable organic compound, including sulfur and gaseous products, increases. As such, when the reaction temperature was higher than 500 oC, the decomposition of fixed carbon increased slightly, and the derived-oil production was increased. Sulfur in the TDF derived oils has also been reported by other researchers.

For example, Cunliffe and Williams[4] reported concentrations of 1.3-1.4 % by weight of

TDF for sulfur in the oil derived from the pyrolysis of TDF at 20 oC/min in the range

450-600 oC. Murena [12] reported only 1 wt.% of sulfur present in the liquid product derived from the pyrolysis of TDF in an autoclave at 380 and 400 oC. The same study also showed that as the pyrolysis temperature increased, there was an increase in the

72 concentration of sulfur in the pyrolyzed char. At two stages, thermal cracking of TDF as a function of temperature may be used to explain this phenomenon[13]. First, the TDF was depolymerized at a temperature of less than 500 oC. In this phase, the rapid dissolution of the TDF occurred and the weakest bonds were broken. Second, the rubber was devolatilized at a temperature of more than 600 oC. At this stage, the release of volatile compounds, including sulfur compounds, took place in the derived oil. Mainly sulfur retained in pyrolyzed char but maximized in the derived oil at lower temperature, whereas the post-cracking reaction explained the decrease of sulfur compounds in the derived oil at higher temperatures. The consumption of sulfur compounds at higher

[12] temperatures has also been shown to produce H2S during coal liquefaction .

5 Derived Oil Char TDF 4

3

Sulfur, % Sulfur, 2

1

0 250 350 450 550 650 750 850

Pyrolysis Temperature, oC

Figure 4.8 Percent sulfur in derived oil and pyrolyzed char compare to the original parent sulfur content in TDF versus pyrolysis temperature

73

Figure 4.9 Percent sulfur removal versus percent devolatilzation and the rate of sulfur removal as a function of volatiles

Figure 4.9 shows that, generally, the percent of sulfur removal tends to increase with the increase in percent of devolatilization. By assuming the first order reaction of sulfur decomposition rate similar to the percent devolatilization[1, 2, 12], the rate of sulfur removal as a function of volatiles was developed. Fitting an exponential equation to the plot of experimental data [1] resulted in the rate of removal of sulfur as a function of volatiles as presented in Figure 4.9. By maximizing the regression coefficient of a specific set of conditions, the slope of the linearization provides an approximate value of

74 1. Upper limit value of desulfurization was obtained assuming that all sulfur present in the TDF released into pyrolyzed char. During the 325-400 oC, dissolution of TDF in the liquid phase took place by rupturing of the weakest bonds[12]. This agrees with the results

[12] in Figure 4.7 that many S-bonds were broken with the release of H2S in the gas phase , so that the sulfur content was found more in the pyrolyzed char than that in the derived oil.

Figure 4.10 Percent sulfur removal (%DeS) and percent ash recovery (%Ash) versus percent devolatilization (%DeV)

75 The data in Figure 4.9 indicate that the rate of sulfur evolution as a function of volatiles decreases with as increase of pyrolysis temperature. The maximum rate occurred as a level of devolatilization less than 10%. Figure 4.10 reveals that the percent ash recovery decreases exponentially with the increase in the percent devolatilization. The percent removal of sulfur was proportional to the percent of devolatilization. This may be due to the fact that desulfurization products were entrained less at lower temperature, whereas at higher temperatures devolatilization increased desulfurization.

4.5 Conclusion

Overall recovery rates for solid char, gas, and derived oil at different temperatures are 40% , 22% and 15% , respectively. The percentage of the char and gas products increases as the pyrolysis temperature increases. However, after the reaction temperature approaches 450 to 500 oC, the degradation rate of TDF sample in the thermogravimetric experiments was independent of the temperature. The distribution and percent removal of sulfur in char and derived oil obtained from vacuum pyrolysis of used tires were investigated. The interesting factors in this study were heating rate and temperature. The majority of sulfur in TDF remained as solid char rather than being released as condensed vapor within the range of 350 to 850 oC. The overall desulfurization of the pyrolysis reaction was essentially unaffected by the heating rate. The release of sulfur from the tire structure during pyrolysis offers the greatest potential for the separation of sulfur dioxide from the evolved gases and vapors.

76 4.6 References

1. Merdes, A.C., Keener, T.C., Khang, S.-J., and Jenkins, R.G., Investigation into

the fate of mercury in bituminous coal during mild pyrolysis. Fuel, 1998. 77(15):

p. 1783-1792.

2. Lin, L., Khang, S.J., and Keener, T.C., Coal desulfurization by mild pyrolysis in a

dual-auger coal feeder. Fuel Processing Technology, 1997. 53(1-2): p. 15-29.

3. Kawatra, S.K. and Eisele, T.C., Coal desulfurization : high-efficiency preparation

methods. 2001, Taylor & Francis: New York. p. 9-43.

4. Cunliffe, A.M. and Williams, P.T., Composition of oils derived from the batch

pyrolysis of tyres. Journal of Analytical and Applied Pyrolysis, 1998. 44(2): p.

131-152.

5. Waddell, W.H., Bhakuni, R.S., Barbin, W.W., and Sandstrom, P.H., Pneumatic

Tire Compounding, in The Vanderbilt rubber handbook, Ohm, R.F., Editor. 1990,

Published by R.T. Vanderbilt Company Inc.: Norwalk, CT. p. 596-611.

6. Senneca, O., Salatino, P., and Chirone, R., A fast heating-rate thermogravimetric

study of the pyrolysis of scrap tyres. Fuel, 1999. 78(13): p. 1575-1581.

7. Pakdel, H. and Roy, C., Simultaneous gas chromatographic--Fourier transform

infrared spectroscopic--mass spectrometric analysis of synthetic fuel derived from

used tire vacuum pyrolysis oil, naphtha fraction. Journal of Chromatography A,

1994. 683(1): p. 203-214.

8. Godovskii, Y.K. and Bessonova, N.P., The entropies and energies of deformation

of elastomers based on natural rubber blends with ethylene-propylene

copolymers. Polymer Science U.S.S.R., 1977. 19(12): p. 3155-3163.

77 9. Suzuki, N., Ito, M., and Yatsuyanagi, F., Effects of rubber/filler interactions on

deformation behavior of silica filled SBR systems. Polymer, 2005. 46(1): p. 193-

201.

10. Unapumnuk, K., Keener, T. C., Khang, S., and Lu, M., Pyrolysis Behavior of Tire

Derived Fuels at Different Temperatures and Heating Rates. The Journal of Air

and Waste Management Association, 2006.56:p. 618-627..

11. Unapumnuk, K., Lu, M. and Keener, T. C., Carbon Distribution from Pyrolysis of

Tire Derived Fuels. Industrial & Engineering Chemistry Research, 2006. (In

reviews).

12. Murena, F., Kinetics of sulphur compounds in waste tyres pyrolysis. Journal of

Analytical and Applied Pyrolysis, 2000. 56(2): p. 195-205.

13. Cypress, R. and Bettens, B., Production of benzoles and active carbon from waste

rubber and plastic materials by means of pyrolysis with stimultaneous port-

cracking, in Pyrolysis and gasification, Ferrero, G.L., Editor. 1989, Elsevier

Applied Science: London ; New York. p. 209-229.

78 Chapter 5

5 Hydrocarbons Composition from the Pyrolysis Products

5.1 Abstract

Products from the pyrolysis of Tire Derived Fuels (TDF) were investigated with various analytical techniques and under various maximum pyrolysis temperatures and heating rates. The pyrolysis products are classified as char (solid product), pyrolytic oil (liquid) and gas. Principal functional groups of the TDF and pyrolytic oil were confirmed by

Fourier Transform Infrared Spectrometer, coupled with attenuated total reflectance (FT-

IR/ATR). The components of the pyrolytic oil fraction were individually quantified using gas chromatography coupled with mass spectrometry (GC-MS). The major products are one- and two-ring methyl-substituted aromatic isomers. By-product formation mechanisms of TDF pyrolysis were hypothesized based on the products identified. The mechanisms for aromatic hydrocarbon formation were found to be associated with polymer degradation, methyl displacement, and the Diels-Alder reactions.

Our study indicated that GC-MS coupled with FTIR is sufficient to investigate the semi- volatile and volatile organic species from complex polymeric materials such as tires.

Keywords: Tire Derived Fuels (TDF), Polycyclic Aromatic Hydrocarbons (PAHs),

Hydrocarbons

79 5.2 Introduction

The purposes of this study were: (1) to generate a temperature profile of aromatic hydrocarbons and isomers using a laboratory scale pyrolysis unit; (2) where possible, to render fundamental knowledge related to aromatic ring formation from the controlled pyrolysis of shredded tires; and (3) in order to utilize the pyrolytic oils as refining feedstock, to obtain comprehensive data on the chemical properties of these oils. Despite numerous studies on the pyrolysis of TDF, very few have reported the temperature series of the product decomposition, mass balance, and speciation changes of the aromatic hydrocarbons in the pyrolytic oils. Therefore, our experiments were designed to study changes in the pyrolysis temperature, heating rates, the resultant changes in the composition of the aromatic compounds in pyrolytic oils, and the overall mass balance on carbon basis.

5.3 Experimental Method

Details of the experimental system and the design of the laboratory-scale pyrolysis reactor have been reported in Chapter 2 and are briefly described here. A laboratory- scale vacuum pyrolysis reactor was developed at the University of Cincinnati for this experiment. The schematic diagram of the pyrolysis reactor is shown in Figure 5.1. The reactor was designed to perform the pyrolysis of the heterogeneous TDF samples. A representative amount of material for this study was found to be 3-4 grams of TDF on a steel wire-free basis[1].

80

Figure 5.1 Schematic diagram of the vacuum pyrolysis reactor

The carbon content of the pyrolysis products was measured. The carbon content of the TDF materials and the pyrolytic char was determined using ultimate and proximate analysis. The pyrolytic oil and TDF samples were instrumentally examined for organic hydrocarbons and other compounds. Organic functional groups in the TDF and pyrolysis samples were obtained by a Fourier Transform Infrared Spectrometer (FT-IR, Nicolet

Mogna 760) coupled with attenuated total reflectance (ATR). The soluble organic fragments in the pyrolytic oil were analyzed via gas chromatography coupled with mass spectrometry (GC-MS). The system consisted of a GC (Varian Model CP-3800) with a capillary column (CP-Sil 8 CB Low Bleed/MS), an auto sampler (CP-8400) and a mass

81 spectrometer (Saturn 2200). The EPA standard analytical method 525 was modified slightly for our analysis.

5.4 Results and Discussion

5.4.1 Principal Functional Structures of TDF

Today, tires are made from a mixture of natural rubber (NR), which is polyisoprene[2], and synthetic rubber, which is mostly styrene-butadiene rubber (SBR).

Butyl rubber is generally used in liners due to its impermeability to air[3]. Tires are insoluble in organic solvents and, therefore, the individual components can not be identified by our experimental setup. Only the principal functional groups can be obtained from FTIR. Figure 5.2 shows an IR spectra of the TDF materials using FT-

IR/ATR. Analysis of Figure 5.2 indicates that the main component of TDF is a combination of hydrocarbons at the fingerprint region from 400 – 1600 cm-1. A medium absorption peak in the region 675 to 730 revealed a –CH=CH-(cis) structure, and a medium absorption peak in the region 960 to 970 indicated a –CH=CH-(trans) in a ring structure[4]. A variable absorption peak in the region 1000-675 confirmed a C-H bond type for alkenes[5]. The FTIR results are consistent with the raw material composition.

82

Figure 5.2 IR Spectrum of TDF

5.4.2 Principal Functional Structures of Pyrolytic oil

Figure 5.3 shows an example of IR spectra of the pyrolytic oil obtained at 500 and 800 oC at a heating rate of 5 oC/min. Figures 5.3a and 5.3b indicate that the main component of pyrolytic oil is a combination of olefinic groups and aromatic rings, which are represented by the C=C bond, C-H stretching vibration, and C-H bending vibration. The approximate frequencies for hydrocarbon-substitute olefins and hydrocarbon-substitute heteroaromatic rings are given in Table 5.1.

The observation of weak absorption peaks at 965 cm-1 revealed a strong –

CH=CH-(trans) in conjugation to polyenes[6]. A strong absorption peak in the region of

2960-2850 cm-1 and an absorption peak in the region of 1470-1350 cm-1 revealed the C-H stretching vibration and the C-H bending vibration respectively[7]. A medium absorption peak in the region of 1600-1450 cm-1 indicated a C=C in aromatic ring[4]. A weak or absent absorption peak in the region of 1678-1668 cm-1 indicated a C=C stretch in IR[7].

83

a) Pyrolysis temperature at 500 oC

b) Pyrolysis temperature at 800 oC

Figure 5.3 IR Spectrum of pyrolytic oil heated at 5 oC/min and, a) maximum pyrolysis temperature 500 oC, b) maximum pyrolysis temperature 800 oC

84 A variable absorption peak in the region 1000-675 cm-1 confirmed a C-H alkene bond type[5]. However, the presence of a weak absorption peak in the region of 1600 to1700 cm-1 and 3737 cm-1 may have resulted from either the -OH impurity in the sample or the

KBr[8].

Table 5.1 Approximate FTIR frequencies for hydrocarbon-substituted olefins and heteroaromatic rings

Functional group Bond Type Wave range, cm-1

CH-CH C-H stretch alkanes 2960-2850 (s)[5] C=C C=C stretch alkenes 1680-1620 (v) [5] C=C 1600-1450(v) Aromatic C-H 1470-1350 (s)[5]

[7] C-H CH3 or CH2 bend alkanes 1380-1370 (v)

C H CH stretch 3050-3000 (m) C=C stretch 1678-1668 (w) H C tran CH 980-965(s)[8] CH stretch 3050-3000 (m) C C C=C stretch 1662-1631 (m) CH bend 1429-1397(m) H H cis CH 650-730(m,s)[8] C=C-H out of plane bend alkenes 790-840 (s) [9]

R 1604±3

R R mono, ortho, and meta 1510-1470 benzene 770-730

[8] 710-690 s = strong absorption w = weak absorption m = medium absorption v = variable absorption

85 IR analysis has identified many more functional groups in the pyrolytic oil than in the

TDF, which clearly suggests the existence of the aromatic structures. The individual compounds in the pyrolytic oil were identified by GC-MS.

5.4.3 Individual Hydrocarbons Composition of Pyrolytic oil

The total ion chromatogram (TIC) of pyrolytic oil is shown in Figure 5.4. The identified individual compounds and isomers are numbered, and the chemical structures are listed in Table 5.2. The identified individual compounds and isomers were quantified based on the two most abundant ions, known as selective ion count mode (SIC).

However, the quantification of individual compounds based on the total ion count mode

(TIC) was presented in Appendix B. The compounds listed were classified based on molecular weight and functional groups. For example, C2-benzene referred to benzene with two carbon substitution (molecular weight: 106), which included three possible isomers: o-xylene, p-xylene and ethybenzene. Pure chemical compounds were acquired to obtain calibration curves. In cases where a pure standard was not available, the response factor of the isomer, or a compound with a similar structure, was used. The relative mass of the individual compounds (on a carbon basis) was calculated and divided by the mass of TDF input to become a relative quantity that is comparable to others throughout the temperature range.

86

Figure 5.4 The total ion chromatogram of pyrolytic oils at pyrolysis condition of 5 oC/min and a maximum temperature of 500 oC. The chemical structures of the numbered group(s) are further identified by GC-MS in Table 5.2.

87 Table 5.2 The major products in the pyrolytic oil identified by GC-MS

Concentration of Carbon, μg-C/g-TDF Peak No.* Functional Group Compound Name Molecular Weight Molecular Formula 400 oC 500 oC 600 oC 800 oC 1000 oC

Benzene, 1,2 dimethyl 106 C8H10 39 NA 1354 NA NA

1 C2-Benzene (3) EthylBenzene 106 C8H10 NA NA 594 527 141

Benzene, 1,4 dimethyl 106 C8H10 NA NA 4714 4263 1008

2 C0-Styrene (1) Styrene 104 C8H8 NA 3473 841 896 326

Benzene, (1-methylethyl) 120 C9H12 NA 2427 332 310 144

trimetylBenzene (isocumene) 120 C9H12 NA 389 114 114 65

trimethylBenzene (cumene, Isopropylbenzene) 120 C9H12 NA NA NA 662 274

trimethylBenzene (Benzene, 1-ethyl-2-methyl) 120 C9H12 NA 5444 1508 1971 752 3 C3-Benzene (8) trimethylBenzene (Benzene, 1 methylethyl) 120 C9H12 NA NA 312 338 156

trimethylBenzene (Bennzene, 1-ethyl-4-methyl) 120 C9H12 NA 775 224 226 125

trimethylBenzene (Bennzene, 1,2,3-trimethyl) 120 C9H12 22 3536 1154 1194 649

trimethylBenzene (Bennzene, 1,2,4-trimethyl) 120 C9H12 5 2669 904 838 462

1-Isopropyl-2-methylbenzene 134 C10H14 447 44052 17521 18456 10038

methylpropyl-Benzene 134 C10H14 NA 1133 305 384 237 4 C4-Benzene (4) methylpropyl-Benzene 134 C10H14 27 6391 2320 2460 1402

methylpropyl-Benzene 134 C10H14 NA 2469 807 937 578

1,3-Cyclohexadiene, 1,5,5,6 tetramethyl 136 C10H16 NA 6741 1643 1683 545

1,5, 5-Trimethyl-6-Methylene-Cyclohexene 136 C10H16 52 11158 3386 3821 1696

1,3-Cyclohexadiene, 1,3,5,5 tetramethyl 136 C10H16 83 14842 4429 4807 2057

5 C4-Cyclohexene (7) Terpinolene 136 C10H16 71 14583 4483 5233 2839

4-Isopropenyl-1-methyl-1-cyclohexene 136 C10H16 1009 148413 46189 50884 29404

Terpinolene 136 C10H16 360 34668 13706 12946 9430

Terpinolene 136 C10H16 NA NA 2655 2232 1901 *The peak numbers denote compound(s) marked in Figure 5.4, and the numbers in parentheses refer to the number of isomers identified.

88 Table 5.2 The major products in the pyrolytic oil identified by GC-MS (Con’t)

Concentration of Carbon, μg-C/g-TDF Peak No.* Functional Group Compound name Molecular Weight Molecular Formula 400 oC 500 oC 600 oC 800 oC 1000 oC

Styrene, 3,4-dimethyl 132 C10H12 50 NA NA NA NA

Styrene, α 2 dimethyl 132 C10H12 NA 15125 5710 5668 3503

6 C2-Styrene (5) Styrene,dimethyl (Benzene, 4-ethenyl-1,2-dimethyll) 132 C10H12 NA NA 1286 1210 846

Styrene,dimethyl (Benzene, 1 -ethenyl-4-ethyl) 132 C10H12 17 NA NA NA NA

Styrene,dimethyl (Benzene, 2-methyl-1-propenyl) 132 C10H12 NA 4790 2082 1790 1222

Benzene,1-methyl-4-(1-methylpropyl) 148 C11H16 NA 1347 486 389 230

Benzene, 1-ethyl-4-isopropyl 148 C11H16 NA 680 282 NA NA

1,4-Dimethyl-2-isopropylbenzene 148 C11H16 NA NA 190 NA NA 7 C5-Benzene (6) Benzene,1-ethyl-4-(1-methylethyl) 148 C11H16 NA NA NA 212 NA

1,3-Dimethyl-4-isopropylbenzene 148 C11H16 NA NA NA NA 137

C5-Benzene 148 C11H16 NA NA NA 372 307

Indan, 2,2-dimethyl 146 C11H14 NA 2778 1280 1215 787 Indan, dimethyl 146 C11H14 38 5652 2679 2399 1636 8 C2-Indan (5) Indan, 4,7-dimethyl 146 C11H14 66 5787 2684 2565 1739 Indan, 4,7-dimethyl 146 C11H14 NA 2612 1222 1047 NA C2-Indan 146 C11H14 NA NA NA NA 756

C3-Indene 158 C12H14 NA NA NA 912 789

9 C3-Indene (3) trimetylindene 158 C12H14 NA NA 6850 5134 3697

1,1,3-Trimethylindene 158 C12H14 NA 4410 2500 1701 1199

Naphthalene,1,4-dimethyl 156 C12H12 NA NA NA NA NA

10 C2-Naphthalene (3) Naphthalene,1,3-dimethyl 156 C12H12 NA 1166 757 NA NA

Naphthalene,1,2-dimethyl 156 C12H12 NA NA NA 558 407

11 C3-Naphthalene (1) Naphthalene,2,3,5-trimethyl 170 C13H14 52 7965 7341 5340 3709

12 C4-Naphthalene (1) Naphthalene, tetra-methyl 184 C14H16 13 1707 2416 1779 1617

Total 2,359 357,183 147,260 147,473 86,809 *The peak numbers denote compound(s) marked in Figure 5.4, and the numbers in parentheses refer to the number of isomers identified. 89 Styrene and benzene isomers mainly eluted during the first 10 minutes of the chromatogram (m/z 78 is the characteristic fragment ion peak), followed by C2-, C3-, and

C4-, substituted benzenes (m/z 91, 105, and 119, respectively), cyclohexene isomers (m/z

121, 136), indan (m/z 131), and finally, naphthalene derivatives (m/z 156, 170, 189).

Isomers of substituted aromatic hydrocarbons usually had similar mass spectra and were easily grouped by the selective ion search method. The mass spectra of trimethyl- benzene and ethylmethyl-benzene are shown in Figure 5.5 as an example .

Figure 5.5 Isomeric mass spectrum of trimethyl benzene and ethylbenzene

90

The complexity of the pyrolytic oil is indicated by the many isomers identified.

Oftentimes, the pattern shows one prominent isomer peak followed by several smaller peaks. Table 5.3 lists the most abundant isomers for all the temperatures and their relative fractions. The most abundant isomers contributed, on average, 64% of all of the functional groups. Exceptionally, 1-ethyl-2-methylbenzene, contributed only 34% of all of the functional groups. The C4-cyclohexene isomers are shown in Figure 5.6 as an example of this pattern. More than 80% of the C4-cyclohexene isomers are identified as

4-Isopropenyl-1-methylcyclohexene (also known as limonene), which was the most abundant of all. One of the isomers was identified as trimethyl-methylene-cyclohexene, which may be formed by the addition reaction of alkenes during pyrolysis[10].

Table 5.3 The most abundant isomeric structure identified by GC-MS

Contribution percentage of all the Peak No.* Most Abundant isomer possible isomers

1 Benzene, 1,4 dimethyl 79 2 Styrene NA 3 Benzene, 1-ethyl-2-methyl 34 4 Benzene, 1-Isopropyl-2-methyl 82 5 Cyclohexene-4-Isopropenyl-1-methyl 87

6 Styrene, α 2 dimethyl 69 7 Benzene,1-methyl-4-(1-methylpropyl) 53 8 Indan, 4,7-dimethyl 48 9 Indene, trimethyl 58 10 Naphthalene,1,3-dimethyl 67 11 Naphthalene,2,3,5-trimethyl NA 12 Naphthalene, tetra-methyl NA

NA: Not applicable, one isomer is identified. * The peak numbers denote compound(s) marked in Figure 5.4.

91

Quantitatively, the total identified products in the derived oils (as shown in Table 5.2) increased from 0.24 wt. % at 400 oC to 35.7 wt. % at 500 oC, then decreased drastically to about 15 wt.% between 600 and 800 oC, then further reduced to less than 10 wt.% at 1000 oC. The GC-MS results have indicated that the majority of compounds were methyl- substituted aromatic hydrocarbons with multiple isomers. All of the compounds reached their maximum quantity at 500 oC and tend to decrease as temperature increases.

Cyclohexene (~58%) and benzene (~26%) were the most abundant functional groups in the pyrolytic oil.

Figure 5.6 The C4-Cyclohexenes in the pyrolytic oil from selective ion chromatogram: pyrolysis condition of 5 oC/min and the maximum temperature of 500 oC. 1). 1,5,5,6- tetramethyl, 1, 3-Cyclohexadiene; 2). 1, 5, 5-trimethyl-6- methylene-Cyclohexene; 3). 1,3,5,5-tetramethyl 1,3-Cyclohexadiene; 4). Terpinolene; 5). 4-Isopropenyl-1-methyl-1-cyclohexene; 6). Terpinolene; 7). Terpinolene

92

The most abundant group of isomers contains the seven C4-cyclohexene isomers, as quantified in Table 5.3 and Figure 5.6, which contribute to more than 50% of the total identified products. Limonene is the most abundant of all. Limonene and other cyclohexene isomers suggests the depolymerization of the natural rubber, which is mainly polyisoprene, followed by the Diels-Alder reactions to form the ring structure from aliphatic components[11-13].

The aromatization caused by a Diels-Alder-type reaction – involving pyrolysis of alkanes to alkenes and subsequent cyclisation to aromatic compounds – has been studied as a route to PAH formation in the post-cracking reaction of the rubber materials [10, 14, 15].

A decrease in mono-ring aromatic compounds between 600 oC and 900 oC was reported as a result from the post-pyrolysis cracking of olefins by the Diels Alder reaction [16, 17].

Cunliffe [14] supported the Diels-Alder mechanism of alkane dehydrogenation to alkene, followed by cyclisation and aromatization by the results of pyrolytic gas analysis. The pyrolysis model for aliphatic compounds was proposed as a route to naphthalene formation by the Diels-Alders reaction in post-pyrolysis cracking reactions, as illustrated in Figure 5.7 [15].

H2C H2C CH2 + + 2 H2

H2C +

CH2

CH 2

Figure 5.7 Diels-Alder reaction for the formation of naphthalene in scrap tire pyrolysis[15]

93

The additional reactions took place at a carbon-carbon multiple bond, or carbon- hetero atom. Multiple bonds, in conjunction with some rearrangement, cause the formation of unsaturated cyclical compounds with six carbons, such as cyclohexene.

Further dehydrogenation at higher temperatures results in a more stable compound, such as saturated aromatic compounds like benzene. In these reactions, the unsaturated ring may further react with the carbon-carbon multiple compounds to produce aromatic derivatives corresponding to polycyclic hydrocarbons. As a result of further aromatization reactions, this may lead to the formation of naphthalene. Several studies suggested that the continuous formation of naphthalene, followed by condensation reactions, may continue to produce higher amounts of PAH [14, 15, 18].

The decomposition of natural rubber also provides a pool of methyl radicals that can add to other aromatic byproducts to result in a series of methylated isomers. The methyl radicals may be initially formed through bond scission, followed by hydrogen displacement of the methyl group from the base compound as the pool of hydrogen builds up from the pyrolytic process. The second most abundant product at 500 oC was the C4- benzene isomers. C4-benzene isomers were formed in high concentrations at 500 oC and decreased as temperature increased. Similar trends have been observed for C2-benzenes and C3-benzenes, which may be the result of the thermal degradation of limonene in oils[14], or the loss of methyl groups from the C4-benzenes. Styrene has been identified in minor quantities. The formation of styrene can be resultant from the depolymerization of the synthetic rubber fraction, which is mainly the styrene butadiene polymer.

Naphthalene, indan, and indene could be produced from cyclisation reactions of alkyl benzenes[19]. These compounds, together with the alkylbenzenes, were the principal tire-

94 pyrolysis oil products derived from the degradation of the styrene butadiene rubber.

Styrene, naphthalene, and indan contributed to approximately 16% of the total functional groups in the pyrolytic oils.

Different operating conditions, such as temperature, pressure, heating rate, and residence time can result in vastly different byproduct formation. In this study, we were not able to identify PAH compounds with more than three aromatic rings. This may be due to several factors. As indicated by other similar studies, the vacuum pyrolysis process results in less residence time, as well as less pyrolysis by-products [20, 21]. The reduction of these monoaromatic products tends to result in lower quantities of the heavier PAH compounds. Although TDF pyrolysis has been performed at higher temperatures, such as 800 and 1000 °C, the pyrolytic oil collected has been much lower in quantity. It is possible that the depolymerization process at higher temperatures results in increasingly higher soot formation (which is measured as char). As with our studies of the methylnaphthalenes, the starting material with alkylated components (such as TDF) tended to result in less heavy PAH, partly due to the reactivity of the aliphatic components (such as the methyl and ethyl groups)[18].

5.5 Conclusion

The pyrolysis of TDF yielded a complex mixture of hydrocarbons in the pyrolytic oil, along with gaseous and char products. The principal functional groups of the oils appeared to be the olefinic groups and aromatic hydrocarbon rings and were fairly independent of the heating rates. However, the pyrolytic oil composition depended quantitatively on the final pyrolysis temperature. The quantity of the byproducts reached the maximum concentration at 500 oC and then started to decrease with temperature. The

95 major products include alkylated isomers of cyclohexene, benzene, and naphthalene, as well as indan and styrene. This study has shown that a pyrolysis reactor can be designed and operated to produce specific products from used tires under certain operating conditions, such as temperature, pressure, and residence time, etc. Pyrolysis conditions can be altered to maximize the quantity of the pyrolytic oil but minimize the quantity of the aromatic ring hydrocarbons composition.

5.6 References

1. Unapumnuk, K. and Keener, T. Kinetic Study of the Degradation Temperature vs.

Weight Loss in the Pyrolysis of Tire Derived Fuels. in A&WMA's 97th Annual

Conference &Exhibition. 2004. Indianapolis, IN.

2. http://www.pslc.ws/mactest/isoprene.htm, accessed Dec. 2005.

3. http://www.dnr.state.oh.us/recycling/awareness/facts/tires/thruwwii.htm, accessed

Dec. 2005.

4. Crews, P., Rodr ํguez, J., and Jaspars, M., Organic structure analysis. 1998,

Oxford University Press: New York. p. 317-347.

5. Ege, S.N., Organic chemistry : structure and reactivity. 1994, D.C. Heath:

Lexington, Mass. p. 376-393.

6. Colthup, N.B. and Orloff, M.K., Calculation of infrared group frequencies from

molecular orbital theory--I : Out-of-plane CH2 and CH bending vibrations in

substituted ethylenes. Spectrochimica Acta Part A: Molecular Spectroscopy,

1971. 27(8): p. 1299-1314.

7. Silverstein, R.M., Webster, F.X., and Kiemle, D.J., Spectrometric identification of

organic compounds. 2005, J. Wiley & Sons: Hoboken, N.J. p. 3-164.

96

8. Colthup, N.B., Daly, L.H., and Wiberley, S.E., Introduction to infrared and

Raman spectroscopy. 1990, Academic Press: Boston. p. 247-288.

9. Arjunan, V., Subramanian, S., and Mohan, S., Fourier transform infrared and

Raman spectral analysis of trans-1,4-polyisoprene. Spectrochimica Acta Part A:

Molecular and Biomolecular Spectroscopy, 2001. 57(13): p. 2547-2554.

10. Moldoveanu, S., Analytical pyrolysis of natural organic polymers, in Techniques

and instrumentation in analytical chemistry ; v. 20. 1998, Elsevier: Amsterdam ;

New York. p. 1-31.

11. Groves, S. and Lehrle, R., Pyrolysis mechanisms of natural rubber deduced from

the dependence of product yields on sample size. European Polymer Journal,

1992. 28(4): p. 373-378.

12. Groves, S.A., Lehrle, R.S., Blazso, M., and Szekely, T., Natural rubber pyrolysis:

Study of temperature-and thickness-dependence indicates dimer formation

mechanism. Journal of Analytical and Applied Pyrolysis, 1991. 19: p. 301-309.

13. Pakdel, H., Pantea, D.M., and Roy, C., Production of dl-limonene by vacuum

pyrolysis of used tires. Journal of Analytical and Applied Pyrolysis, 2001. 57(1):

p. 91-107.

14. Cunliffe, A.M. and Williams, P.T., Composition of oils derived from the batch

pyrolysis of tyres. Journal of Analytical and Applied Pyrolysis, 1998. 44(2): p.

131-152.

15. Cypress, R. and Bettens, B., Production of benzoles and activated carbon from

waste rubber and plastics materials by means of pyrolysis with simultaneous post-

cracking, in Pyrolysis and gasification, Ferrero, G.L., Maniatis, K., Buekens, A.

97

and Bridgwater, A. V., Editor. 1989, Elsevier Applied Science: London ; New

York. p. 209-229.

16. Williams, P.T. and Taylor, D.T., Aromatization of tyre pyrolysis oil to yield

polycyclic aromatic hydrocarbons. Fuel, 1993. 72(11): p. 1469-1474.

17. Levendis, Y.A., Atal, A., Carlson, J., Dunayevskiy, Y., and Vouros, P.,

Comparative study on the combustion and emissions of waste tire crumb and

pulverized coal. Environmental Science & Technology, 1996. 30(9): p. 2742-

2754.

18. Yang, J. and Lu, M., Thermal growth and decomposition of methylnaphthalenes.

Environmental Science & Technology, 2005. 39: p. 3077-3082.

19. Giray, E.S. and Sonmez, O., Supercritical extraction of scrap tire with different

solvents and the effect of tire oil on the supercritical extraction of coal. Fuel

Processing Technology, 2004. 85(4): p. 251-265.

20. Mirmiran, S., Pakdel, H., and Roy, C., Characterization of used tire vacuum

pyrolysis oil: Nitrogenous compounds from the naphtha fraction. Journal of

Analytical and Applied Pyrolysis, 1992. 22(3): p. 205-215.

21. Pakdel, H. and Roy, C., Simultaneous gas chromatographic--Fourier transform

infrared spectroscopic--mass spectrometric analysis of synthetic fuel derived from

used tire vacuum pyrolysis oil, naphtha fraction. Journal of Chromatography A,

1994. 683(1): p. 203-214.

98

Chapter 6

6 Conclusions and Recommendations

In this research, the pyrolysis mechanisms, the recovery and the removal of pyrolysis by-products of the TDF were investigated. For all the scenarios investigated, experimental results revealed significant changes of the pyrolysis kinetic parameters and the generation rate of specific pyrolysis-by-products from TDF to the designed and operated pyrolysis conditions. Important observations included: First the kinetic parameters of the TDF pyrolysis mechanisms were determined thermogravimetrically.

The decomposition model or the thermogravimetrical model was run to simulate various scenarios of the influences of heating rates and temperature based upon the Arrhenius equation. The influences of heating rate and pyrolysis temperature were investigated from 350 – 1000 oC, a range where substantial devolatilization occurs. The results of this study demonstrate that pyrolysis mechanism of TDF could indeed be adequately expressed by the Arrhenius relation.

Second, the mass distribution of pyrolysis by-products along with the removal of sulfur were determined gravimetrically. The experiments were designed under the hypothesis that changes in the pyrolysis parameters due to temperature and heating rate can alter the level and generation rate of the pyrolytic products. Throughout the results obtained in the investigation into the removal of sulfur, it appeared that all the findings agree well with the following:

• The majority of sulfur in TDF remain in the pyrolysed char rather than release to

the condensed vapor phase within 350 -850 oC;

99

• The overall desulfurization of the pyrolysis reaction is essentially unaffected by

the heating rates;

• The release of sulfur from the tire structure during pyrolysis offers the greatest

potential for the separation of sulfur dioxide from the evolved gases and vapors.

Finally, the hydrocarbons composition of derived oil were identified in conjunction with the study of PAHs formation. The experiments were designed to study changes in the pyrolysis temperature and heating rates and the resultant changes in the composition of the aromatic compounds in pyrolysis oils and the overall mass balance on carbon basis. As a result, the major composition of hydrocarbons from the derived oil could be sufficiently identified by GC-MS and FT-IR. The principal functional groups of the oils appeared to be the olefinic groups and aromatic hydrocarbon rings and were fairly independent the heating rates. However, the formation of major products in the derived oil such as alkylated isomers of cyclohexene, benzene and naphthalene as well as indan and styrene depended quantitatively on the final pyrolysis temperature.

This study has shown that a pyrolysis reactor can be designed and operated to produce specific products from used tires under certain operating conditions. It is recommended that the pyrolysis of used tires be considered in decision making regarding the management and the utilization of used tires which might involve activities such as incineration, power generation and the use of scrap tires as a fuel.

The following topics are recommended for potential future works:

100

• Model system utilizing parent materials of rubber such as styrene, polystyrene

might be recommendable to simulate the formation of heavy molecular weight

polycyclic aromatic hydrocarbon from the pyrolysis reaction of TDF.

• Combustion might be proper characterization tool in determining the

composition of the exhaust gas formed during the devolatilization process and

the depolymerization at high heating rate and temperature.

• More structural investigation of the elemental metal composition such as Zn,

S, Fe, etc., under various pyrolysis conditions might be needed to obtain the

basic knowledge on the mechanism of the desulfurization and the release of

metal process in the tire structure.

• Similar to the batch pyrolysis reactor, continuous feeding of TDF blending

with coal might be possible with the rotating pyrolysis-combustion reactor

with the various percentages mixture, which probably represent the real

situation of traveling gate fuel feeder system in the power plant. The design

might be modify for the gas sampling system as well.

101

Appendix A: Demonstration of Mathematica Model

1. Download the build-in function for nonlinear curve fitting technique:

<< Statistics`NonlinearFit`

2. Import data from text file:

data0 = Import "C:\\d55.txt", "Table" ; Length data0 790 @ D d55 = Take data0, 1, 790, 30 ; @ D data = d55 All, Range 1, 2 ; maxData = Length data @ 8

0.8 @ 8 @ D @ D

0.4

0.2

50 100 150 Graphics data2 = d55 All, Range 5, 7 ;

3. Determine the equation and the statistic data that best represent the input data by executed the nonlinear@@ curve fitting@ command:DDD

Nonlinear5 = NonlinearFit data2, a + b ∗ x1 + c ∗ x2, x1, x2 , a, b, c −3.00367 − 2335.87 x1 + 0.762353 x2 regress5 = NonlinearRegress data2, a + b ∗ x1 + c ∗ x2, x1, x2 , a, b, c @ 8 < 8

@ 8 < 8

102

BestFitParameters → a→−3.00367, b →−2335.87, c →0.762353 , Estimate Asymptotic SE CI a −3.00367 1.0544 −5.17984, −0.827496 ParameterCITable → , 9 b8 −2335.87 722.875 <−3827.81, −843.928 c 0.762353 0.179319 0.392256, 1.13245 8 < 8 DF SumOfSq< MeanSq Model8 3 1321.01< 440.337 EstimatedVariance →1.82059, ANOVATable → Error 24 43.6942 1.82059, Uncorrected Total 27 1364.71 Corrected Total 26 76.6023 1. −0.938401 0.867035 AsymptoticCorrelationMatrix → −0.938401 1. −0.754396 , 0.867035 −0.754396 1. i Curvature y Max Intrinsic 0 FitCurvatureTable → k { Max Parameter−Effects 0 95. % Confidence Region 0.576507 lnko =−3.003674; e = 2335.87 ∗ 8.3144;= n = 0.76; maxTime = data maxData, 1 ∗ 1.05 204.75 R = 8.3144; To = 480; a = 5; @@ DD

4. Interpolate the obtained parameters and the experimental data by check the residual error technique:

lnko− e sol =NDSolve x' t E R To+a∗t UnitStep 1−xt 1−UnitStep xt xt ^n, x 0 0,x, t, 0, maxTime x → InterpolatingFunction 0., 204.75 , <> classicFit5 = Plot x Ht L. sol, t, 0.1, maxTime , PlotRangeA9 →@D 0, maxTime , 0,@ 1.1@DD,H PlotStyle@@→DDDashing@DL @D0.05,= 0.058

0.4

0.2

50 100 150 200 Graphics

103

Show Graphics experiment5 , classicFit5, PlotRange → 0, maxTime , 0, 1.1 1 @ @ D 0.8 88 < 8 <

0.4

0.2

50 100 150 200 Graphics x data 5, 1 . sol 1 0.0225473 data 5, 2 @ @@ DDD ê @@ DD 0.0183383 ko1 =−3.003674; e1 = 2335.87@@ DD∗ R; n1 = 0.76; e =.; n =.;

5. Determine the best fit parameters by least square method with minimizing the sum of square of the conversion (SSQ) techniques: residual e_ := FindMinimum ko− e sol =NDSolve@ D x' t E R To+at UnitStep 1−x t 1−UnitStep x t x t ^n, x 0 0 ,x, t, 0, maxTime ; maxData A data i, 2 −x data H i, 1L . sol 1 ^2, ko, ko1, 0.95∗ko1 , n, n1, 0.95∗n1 i=1 A9 @ D @ @ DD H @ @ DD @ DL @ D = 8

104

50000, 0.0318456 , 55000, 0.0268902 , 60000, 0.0228919 , 65000, 0.0196801 , 70000, 0.0171084 , 75000, 0.0150696 , 80000, 0.0130656 , 85000, 0.0114898 , 90000, 0.0102566 , 95000, 0.00927228 , 100000, 0.00850135 , 105000, 0.00791024 , 110000, 0.00747421 , 115000, 0.00717212 , 88 < 8 < 8 < 8 < 8 < 120000, 0.00698588 , 125000, 0.00690073 , 130000, 0.00690413 , 135000, 0.00698379 , 8 < 8 < 8 < 8 < 8 < 140000, 0.00712964 , 145000, 0.0073329 , 150000, 0.00758485 , 155000, 0.007879 , 8 < 8 < 8 < 8 < 160000, 0.00820841 , 165000, 0.00856742 , 170000, 0.00895112 , 175000, 0.00935498 , 8 < 8 < 8 < 8 < 180000, 0.00977531 , 185000, 0.01021 , 190000, 0.0106543 , 195000, 0.0111057 , 200000, 0.0115635 8 < 8 < 8 < 8 < ListPlot resTable 8 < 8 < 8 < 8 < 0.038 < 8 < 8 < 8 < 8 <<

0.025 @ D

0.02

0.015

80000 100000 120000 140000 160000 180000 200000

Graphics resFunction = Interpolation resTable InterpolatingFunction 50000., 200000. , <> Plot resFunction e , e, 100000, 140000 @ D 0.0076 @88 << D 0.0075 0.0074 @ @ D 8

110000 120000 130000 140000 0.0069

Graphics FindMinimum resFunction e , e, 120000, 140000 0.00689204, e → 127266. e = % 2 1 2 @ @ D 8

0.4

0.2

50 100 150 200 Graphics Show Graphics experiment5 , simulation5 1

0.8 @ @ D D 0.6

0.4

0.2

50 100 150 200 Graphics

106

Appendix B: Supporting Materials for GC-MS Analysis Table B-1 Standard Calibration Curves

Total Ion Chromatograms (TICs) Selected Ion Chromatograms (SICs) Standard Solution Molecular Weight Respond Factor R2 Most abundant Ions Respond Factor R2 Styrene 104.15 200409.36 1.00 104,103140462.73 1.00 χ-Xylene 106.17 137598.87 1.00 106, 9179210.39 1.00 Indene 116.16 169332.75 1.00 115, 116134142.27 1.00 1,3,5 TrimethylBenzene 120.2 273125.15 0.94 105, 120 180605.16 0.94 1H-Indene-2-Methyl 130 346004.47 1.00 NA NA NA α, 2 DimethylStylene 132.2 86008.00 1.00 132, 11741591.43 1.00 ο-Cymene 134.22 98268.01 0.98 119, 13458419.64 0.98 ρ-Cymene 134.22 289223.45 0.93 NA NA NA 1,2,4,5 TetramethylBenzene 134.22 236475.70 1.00 NA NA NA Terpinolene 136.24 49039.78 0.99 121, 13620140.51 0.99 1, MethylNaphthalene 142.2 197730.97 0.96 NA NA NA 2,4,6 Trimethylstylene 146.24 97706.83 1.00 146, 131 54680.84 1.00 2,2 DimethylproplyBenzene 148.24 144797.36 0.97 91, 92, 57 63866.38 0.98 1,8 DimethylNapthalene 156.23 180772.94 1.00 NA NA NA 1,4 DimethylNaphthalene 156.23 426853.95 0.96 156, 141 262344.44 0.96 2,6 DimethylNaphthalene 156.23 213909.48 0.99 NA NA NA 1,3 DimethylNaphthalene 156.23 391647.36 0.98 NA NA NA 1-Phenyl-1-cyclohexene 158.24 115652.92 1.00 158, 143 40524.94 1.00 2,3,5 TrimethylNaphthalene 170.25 135059.86 1.00 170, 155 53400.29 1.00 2,6 DiethylNaphthalene 184.28 124258.00 0.96 186, 169 71100.16 0.96

107

Table B-2 The major products in the pyrolysis oil identified by GC-MS using total ion chromatograms technique

Concentration of Carbon, μg-C/g-TDF Compound Name Molecular Weight 400 oC 500 oC 600 oC 800 oC 1000 oC Styrene 104.15 NA 5818 1545 1297 559 Benzene 1,2 dimethyl 106.17 37 NA NA NA NA Benzene, ethyl 106.17 NA NA 1844 1714 428 χ-xylene 106.17 NA 21532 5779 4723 1258 Benzene, (1-methylethyl) 120 NA 1940 577 213 248 trimetylBenzene (isocumene) 120 NA 2442 524 213 300 trimethylBenzene (cumene, Isopropylbenzene) 120 NA 4134 1205 1210 598 trimethylBenzene (Benzene, 1-ethyl-2-methyl) 120 24 8860 2457 2485 1250 trimethylBenzene (Benzene, 1 methylethyl) 120 10 3792 1156 758 602 trimethylBenzene (Bennzene, 1-ethyl-4-methyl) 120 NA 2034 539 374 339 trimethylBenzene (Bennzene, 1,2,4-trimethyl) 120 NA 6116 1960 1666 1110 Styrene, 3,4-dimethyl 132 NA 24459 9184 NA NA Styrene, 3,5-dimethyl 132 124 NA NA 5953 5548 Styrene,dimethyl 132 NA 11200 4740 2544 2950 Styrene,dimethyl 132 44 16786 7155 3822 4504 1-Isopropyl-2-methylbenzene 134.22 606 54918 21983 NA 14647 1-Isopropyl-4-methylbenzene 134 NA NA NA 21976 NA methylpropyl-Benzene 134 18 5152 1765 2433 1106 methylpropyl-Benzene 134.22 26 5724 1912 2064 1149 methylpropyl-Benzene 134 NA 7502 1623 2695 1776 1-Isopropyl-4-methyl-1,4-cyclohexadiene 136 208 31384 12237 9333 4280 1,3-Cyclohexadiene, 1,3,5,5 tetramethyl 136 167 35064 10332 8905 5766 Terpinolene 136 59 15896 4931 5422 2876 4-Isopropenyl-1-methyl-1-cyclohexene 136 2739 355285 112572 87106 71583 Terpinolene 136 471 47884 18413 13873 12113 Indan, 1,6-dimethyl 146 2 2239 1513 346 891 Indan, dimethyl 146 15 5826 2087 1222 1876 Indan, 4,7-dimethyl 146 28 4551 2107 3024 1395 Indan, 4,7-dimethyl 146 1662 NA 2241 861 NA

108

Table B-2 The major products in the pyrolysis oil identified by GC-MS using total ion count technique (Con.’t)

Concentration of Carbon, μg-C/g-TDF Compound Name Molecular Weight 400 oC 500 oC 600 oC 800 oC 1000 oC Naphthalene,1,2,3,4-tetrahydro-1-methyl 146.24 NA NA NA NA 4803 2-Phenyl-3-methylbutane 148 56 7801 3053 1830 1982 Benzene,1-methyl-4-(1-methylpropyl) 148.24 49 8227 3208 1058 2089 Benzene, 1-ethyl-4-isopropyl 148.25 41 7076 NA NA NA 1,4-Dimethyl-2-isopropylbenzene 148 NA NA 3024 NA NA Benzene,1-ethyl-4-(1-methylethyl) 148 NA NA NA 2537 NA 1,3-Dimethyl-4-isopropylbenzene 148 NA NA NA NA 1870 Naphthalene,1,4-dimethyl 156 52 NA NA NA 2187 Naphthalene,1.3-dimethyl 156 NA 6388 4103 NA NA Naphthalene,1.2-dimethyl 156 NA NA NA 3306 NA 1,1,3-Trimethylindene 158.24 133 20392 11217 5434 6099 Naphthalene,2,3,5-trimethyl 170 NA 18232 15950 6813 8595 Naphthalene,1,4,6-trimethyl 170 173 NA NA NA NA Azulene, 7-ethyl-1,4-dimethyl 184.28 NA NA 11912 4526 6899 Naphthalene, 1-tert-butyl 184 29 7392 NA NA NA Total 6772 756045 284845 211735 173672

*NA: Not Available

109

Appendix C

1 Feasibility Study of Burning Tires with Coal

1.1 Tire Burning in the United States

In 2003, approximately 290 million tires were generated in the U.S [1]. The major scrap-tire market was tire-derived fuel (TDF). Scrap tires have been used as a cleaner and a more economical alternative to coal in cement kilns, pulp and paper mills, and industrial and utility boilers. In these applications, almost 130 million scrap tires were consumed in the U.S. in 2003, or nearly 45% of the total scrap tires generated [1]. Due to elevating fuel prices and improvements in the quality and reliable delivery of scrap tires, the use of TDF may also replace a significant percentage of higher-value fuels such as natural gas, oil, or coal.

Export, 3.1%

Electric Arc Furnances, 0.2% Unknown, 10.3% Landfill, 9.3%

Tire Derived Fuel, 44.5%

Civil Engineering, 19.3%

Micellaneous Agriculture, 1.7%

Ground Rubber, 9.7%

Cut/Punched/Stamped, 2.0%

Figure C 1 U.S. tires disposition 2003[1] Scrap tires have been recognized as a potential source of fuel due to its low cost and excellent and consistent fuel properties, especially containing significant heating value.

110 Legitimate markets for scrap tires are growing and several industries have expressed significant interest in the use of TDF for fuel. Some positive factors promoting the application of TDF include increased importance of NOx reduction and perhaps other air emissions advantages, cost-competitiveness with natural fuels, increasing energy prices, and increased acceptance as a standard alternative fuel.

However, one of the major challenges facing the use of TDF is the potential for air emission issues. Public concern over the use of scrap tires as a fuel mainly deals with the potential environmental consequences if proper care is not exercised during the combustion process, which could possibly produce carcinogenic compounds, such as polycyclic aromatic hydrocarbons. Other concerns, such as competition from natural gas and the increasing opposition by competitors to TDF[2], still need to be explored.

The purpose of this chapter is to summarize data on the potential of burning tires or

TDF as supplemental fuel from utility and process boilers. The primary area of interest in this investigation is the effect on emissions from the process, including its environmental impact, combustion efficiencies, and economics. In addition, this section discusses the feasibility of burning used tires with coal in an existing stoker boiler in the

University of Cincinnati. The performance of combusting used tires blended with coal has been studied for many years but still offers potential for additional research and development for this particular application.

1.2 Principle Fuel Properties and Environmental Performance

With the proper operational system, blending TDF with coal may lead to improved steam capacity and boiler performance. The principle fuel properties of coal and TDF are compared in Table C 1. The uniform and proper blending of TDF with coal can improve

111 the combustion efficiency of the boiler because of the high heating value of TDF ~37,500

kJ/kg (~16,000 Btu/lb).

Table C 1 Fuel Analysis of TDF and Coal

Coal, wt. % TDF, wt. %

As received Dry As received Dry

Proximate Analysis Moisture 2.21 0 0.6-3 0 Ash 0.56 10.4 4-6 6.16 Volatile Matter 36.96-37.62 37.8-38.39 64-67 65 Fixed Carbon 52.27 53.45-51.27 26-29 27 Ultimate Analysis Moisture 0 0 0.6-3 0 Ash 8.75 10.4 4-6 6.1 Carbon 53.45 61.1 82-84 78.74 Hydrogen 4.1 6.75-8.25 6.82 Nitrogen 1.26-2.61 1.26-2.62 0.24-0.5 0.45 Sulfur 0.68-0.70 0.73-2 1.6-2.1 2.44 Oxygen (by difference) - - 0.25-1 5.3

Heat value, Btu/lb 13,018-13,457 13,761-13,485 15,000-15,500 16,136 FC/VM 1.41-1.34 1.27-1.10 0.41-0.43 0.42 *FF 664-505 586-377 201-192 187

** ER, lb SO2/MMBtu 1-3 1-3 2-3 3

*From Equation (a)

**From Equation (b)

Jung et al., [3] developed a relationship based on the coal’s composition factors including

the coal nitrogen, volatile matter, and the fixed carbon as following:

112 ⎛ FC ⎞ FF = 550⎜ ⎟ − 92()No ------(d) ⎝VM ⎠ Where: FF is the coal’s fuel factor; FC is the coal’s fixed carbon; VM is the volatile matter;

No is the nitrogen all in percentage on a dry, ash-free basis. The relationship in equation (a) was obtained by linear regression associated with the best-fit curve techniques. The ratio of the fixed carbon to the volatile matter is an indication of the coal’s reactivity and is related to its combustion efficiency. The relationship is applied to evaluate the TDF’s fuel factor in this study.

Table C 2 Comparison of stack emissions between the baseline concentrations of coal only and when burning a 20% blend of TDF[4]

Stack Emission from Stoker Boiler Coal & TDF CO No change

SO2 Increase HF Decrease HCl Decrease NOx Decrease VOCs Decrease PM Decrease Cr Decrease Hg Decrease Zinc Decrease Cd Decrease Pb Decrease The lower value of the fixed carbon to volatile matter indicated that TDF have more reactivity than coal. As the consequence of low nitrogen content and low reactivity in TDF, the combustion of TDF may inhibit NOx formation in the combustion system.

113 These results agree with the other study presented in Table C 2 [4].

1.3 Boiler Type and its Operation with TDF

A number of different types of boilers and fuel-feed equipment have been used to generate steam, and some are more amenable than others for the use of TDF as a fuel.

The following section describes several boiler types and summarizes their limitations with TDF.

1.3.1 Pulverized Coal Boilers

For pulverized-coal boiler applications, the fuel may be co-pulverized and injected into the boiler with the coal, separately pulverized and fed with the coal, or injected separately without further size reduction and burned partially in suspension and partially on a dump grate, then added to the bottom of the furnace[4] . The cost of pulverizing tires into a very fine powder could inhibit their use as a supplemental fuel.

1.3.2 Cyclone Fired Boilers

Another type of boiler is the cyclone-fired boiler. A cyclone boiler burns coal that is crushed to a size of less than a quarter of an inch. Smaller particles of TDF, or of an equal size to the coal particles are required in this system, which can make combusting the TDF uneconomical. Too large size TDF may cause incomplete combustion and block the boiler or dust collection system.

However, in July 2002 Asbury Power Plant in Missouri decided to add TDF to the

212 MW-cyclone boiler’s fuel mix[5]. The only piece of equipment that had to be modified for the application was the coal conveyer and the installation of a new hopper

114 downstream of the crushed coal. The plant used bituminous coal crushed to a sized of less than 1/8 inch mixed with TDF prior to entering the cyclone. They reported that air enters perpendicular to the cyclones, creating cyclonic action and increasing retention time sufficiently to consume the well-mixed fuel inside of the burner. In addition, the small surface area of the fuel chips showed an improvement in combustion efficiency of the cyclone boiler.

1.3.3 Fluidized Bed Boilers

A fluidized-bed combustion system accommodates a high temperature inert material – such as sand, ash, or limestone – as primary bed material at the bottom of the chamber. Wire removal from a circulating fluidized bed in the system with TDF applications has been a design challenge. If the wire was not successfully removed from the TDF, it resulted in poor air/fuel distribution and eventually caused the system to shut down. Moreover, problems reported by some pilot scale tests involved wire clogging the boiler-grate opening and ash drawdown, and overloading of the particulate control device[2].

1.3.4 Stoker Boilers

Smaller industrial spreader-stoker coal-fired boilers seem particularly suited for using TDF because of the size of the coal which these systems require. The spreader- stoker units offer the advantage that there are so many of them (estimates are ~10,000 in the U.S.), and in most cases, larger-sized coal and/or other solid fuels such TDF may be burned. In these units, fuel is thrown from fuel-receiving hoppers to the back of the furnace onto a traveling grate. During combustion, the fuel travels to the front where the

115 ash is automatically discharged to the ash storage hopper and is periodically transferred to the ash removal system. As the fuel moves forward the amount, of air required for combustion may be varied by the adjustment of the underfire-overfire air. The traveling grate design allows the individual grate bars to open at the lower portion of the catenary, facilitating air admission to the fuel bed and discharging any siftings that may have accumulated on the lower stand[6]. A typical overfire air spreader-stoker is shown in

Figure C 2.The fuel feed rates are important and should be adjustable to meet the existing air emission permit ratio.

Figure C 2 Traveling Grate Stoker Boiler [6]

However, to burn TDF successfully in a spreader-stoker furnace, the particle size of the TDF should be slightly smaller than the largest coal permitted so that the TDF falls on top of a layer of primary fuel. This is to prevent the TDF from directly contacting the

116 grate since oils from the pyrolysis process may flow into the grate opening, carbonize, and plug the grate. The recommended size of TDF for this application is 2 to 4 inches in diameter. Air quality changes from the use of TDF blending with coal in the utility boilers, has been seriously concerned in the governmental restrictions. For this reason, the use of TDF as supplement fuel is still a debatable topic and is limited in many utility plants and process boilers. This controversial issue could be easily overcome by the proper design and operation procedure of the feeding grate speed. The fuel feeders and fuel-receiving hoppers should be specially designed to prevent the build-up of fine particles of fuel which would interfere with proper fuel distribution [6]. The designed grate speed and air-to-fuel ratio should allow enough time for rubber compounds and carbon to decompose such that the subsequent oxidation step produces a minimum amount of undesirable air pollutants. A suitable designed grate speed should minimize the release of aromatic hydrocarbons from the combustion process. Pyrolysis is the first step in the combustion of any solid fuel and is well-known as the releasing stage for potential hazardous compounds. It is important to characterize and understand how the pyrolysis process can be optimized for minimizing air pollution in the combustion of

TDF.

1.4 A Case Study: Feasibility study of burning TDF blend with coal at the University of Cincinnati

In 2003, 25,000,000 scrap tires were reportedly collected in stockpiles in Ohio[1].

The coal-fired boiler was used in the utility plant at the University of Cincinnati, equipped with a continuous ash discharge traveling grate spreader-stoker and a Keeler

Dorr Oliver, MKB boiler. The plant has the capability of producing 100,000 pounds of

117 steam per hour, or 135.6 MMBtu/hr with fabric filter. According to the specific operation for the existing coal-fired spreader-stoker boiler with fabric filter, the plant has applicable emissions limitations of 1.45 lb SO2/MMBtu, 0.05 lb PM/MMBtu and 0.70 lb

NOx/MMBtu of actual heat input. In addition, the visible emissions should not exceed

20% opacity, as a six-minute average. The representative sulfur dioxide emission rate for each analyzed sample of coal, can be calculated using the following formula from OAC rule 3745-18-37[7]:

(1×106 )× S ×1.95 ER = ------(e) H Where:

ER = the emission rate in pounds of sulfur dioxide per MMBtu;

H = the heat content of the solid fuel in Btu per pound; and

S = the decimal fraction of sulfur in the solid fuel. The local air agency allows the university to assume that only 95% of the sulfur in the coal goes to sulfur dioxide.

Tire burning feasibility calculation for the existing system based on 2.5% of sulfur at 16,000 Btu/lb of TDF and 0.73% of sulfur at 13,400 Btu/lb of coal are the following:

135.6 x 106 = 16,000X + 13,400Y ------(f)

Where X = lb/hr of TDF and Y = lb/hr of coal

From equation (b), the following relationship based on the SO2 limitation can be developed. ⎛ ⎛ 2.5X ⎞ ⎛ 0.73Y ⎞ ⎞ ⎜ ⎜ ⎟ + ⎜ ⎟ ⎟ 6 ⎝ 100 ⎠ ⎝ 100 ⎠ ⎛ ()()16,000X + 13,400Y ⎞ 1.45 = ()1.95 ()1×10 ⎜ ⎟/⎜ ⎟ ------(g) ⎜ X + Y ⎟ ⎝ X + Y ⎠ ⎜ ⎟ ⎝ ⎠

118 Figure C 3 shows a calculation of the SO2 emission rate for combustion of TDF with a 2.5 % sulfur level (the highest reported) blended with coal having a 0.73 % sulfur content, which is similar to that currently being burned at the University of Cincinnati facility. The upper solid line represents the permitted SO2 emission rate limit for the facility. The capacity of the plant would remain at 135.6 MMBtu/hr with a fabric filter air pollution control device. The boiler has a maximum capacity to produce 100,000 pounds of steam per hour. As a result, the existing boiler can burn up to 1,655 lb/hr of

TDF mixed with 8,140 lb/hr of coal. In other words, the existing system is capable of burning a 20% blend of TDF. By properly blending TDF with coal, the air emissions and ash disposal characteristics can be accurately assessed to meet the governmental restrictions. One air emission change from burning a combination of TDF will potentially be an increase in SO2 emissions if the sulfur content of the TDF is greater than that of the coal being burned.

119 1.6

SO2 limitation

1.4 /MMBtu 2 1.2

1 Stack emission,lb SO 2 SO

0.8

0.6

0 102030 TDF blending, %

Figure C 3 SO2 emission from the burning of TDF blend

1.4.1 Ambient impact of the air emission

Our review of the literature indicates that the combustion of tires in a properly designed and operated boiler system such as at the University of Cincinnati has the potential to reduce stack emissions, thereby improving ambient air quality. Most air emission limits – including particulate matter, nitrogen oxide, carbon monoxide, and organic compounds – are reduced when TDF is used. In addition, the modification will not result in an increase of SO2 above the current permitted level. This modification will

120 not interfere in the attainment of any National Ambient Air Quality Standard (NAAQS) or result in the release of additional hazardous air pollutants (HAPs). In summary, the ambient impact of the emissions from this source – burning a combination of coal and

TDF – will result in decreased emissions of most criteria pollutants and HAPs. The percent blend of the TDF will be maintained (and determined from the testing period) at levels such that the permitted level of SO2 emissions will not be violated.

1.4.2 Economics of burning tires with coal

At the present time, the 2-inch size TDF costs approximately $1.77-$1.80 per

MMBtu or ~$50 per ton, while the Kentucky coal costs $407 per MMBtu, or $106 per ton. The plant can save about $471,934 per year, or 11% of fuel cost, with the 20% blend of TDF. In 1997, Ferro[8] studied the economics of tire remanufacturing. The study focused on the use of tires in electric plants and in cement kilns. In the results, each tire produced 20 kilowatt hours (kWh) of electric energy. The generated energy was sold to the local utility company for rates ranging from 6 to 8.3 cents/kWh. Therefore, it implies a contribution of $0.80 -$1.10 per passenger car tire. The same study reported that one ton of scrap tire fed into a thermostatic plant could generate $76 worth of electricity, considering a conservative market price of $50 per MWh. Despite this, heat generation appeared to be the only economically viable alternative process for tire remanufacturing.

1.5 Conclusion

From the above analysis, the total saving fuel costs and the availability of existing system to utilize TDF are significant and should make the application achievable. A major challenge to the use of TDF is its potential to increase sulfur dioxide emission.

121 One of our suggestions to limit the sulfur dioxide emission is to limit the mixing percentage of TDF to coal by evaluating the amount of sulfur content in the parent materials. Another suggestion is to separate the release of sulfur from the solid and gas phases by optimizing the pyrolysis conditions. Our findings indicated that the majority of sulfur in TDF remains in the solid char, rather than being released into the condensed vapor phase when pyrolysis occurs in the range of 350-850 oC, where extensive devolatilization occurs.

1.6 References

1. U.S. Scrap tire markets 2003 Edition. 2004, Rubber Manufacturers Association:

Washington, DC.

2. Clark, C., Meardon, K., and Russell, D., Burning tires for fuel and tire pyrolysis

air implications. 1991, U.S. Environmental Protection Agency Office of Air

Quality Planning and Standards: Research Triangle Park, NC.

3. Jung, K.S., Keener, T.C., and Khang, S.-J., Compositional factors affecting NOx

emissions from Ohio coals. Fuel Processing Technology, 2001. 74(1): p. 49-61.

4. Stopek, D.J. and Justice, A.L., Cofiring Tire-Derived Fuel and Coal for Energy

Recovery, in Clean energy from waste and coal, ACS symposium series, 515.,

Khan, M.R., Editor. 1992, American Chemical Society: Washington, DC. p. 104-

116.

5. Wicker, K., Top Plants 2005: Asburry Power Plant. Platts: Power Business and

Technology for the Global Generation Industry, 2005. 149(6): p. 34-37.

6. DetroitStokerCompany, Detroit Rotograte Stoker Bulletin no.85. 1985.

122 7. OhioEPA, Air Pollution Regulations. 2005, Division of Air Pollution Control,

http://www.epa.state.oh.us.

8. Ferrer, G., The economics of tire remanufacturing. Resources, Conservation and

Recycling, 1997. 19(4): p. 221-255.

123