thermal methods in petroleum analysis

Contents 1 Introduction 1 Methods and instrumentation 3 Thermal analysis on model substances Experiments using the simultaneous thermal analyzer Differential scanning calorimetry on mode

Thermal Methods in Petroleum Analysis

by Heinz Kopsch CopyrightoVCH Verlagsgesellschaft mbH, 1995

Heinz Kopsch

Thermal Methods

in Petroleum Analysis

Distribution:

VCH, PO Box 10 11 61, D-69451 Weinheim, Federal Republic of Germany

Switzerland: VCH, PO Box, CH-4020 Basel, Switzerland

United Kingdom and Ireland: VCH, 8 Wellington Court, Cambridge CB1 lHZ, United Kingdom

Japan: VCH, E k o w Building, 10-9 Hongo 1-chome, Bunkyo-ku, Tokyo 113, Japan

ISBN 3-527-28740-X

Published jointly by

VCH Verlagsgesellschaft mbH, Weinheim (Federal Republic of Germany)

VCH Publishers, Inc., New York, NY (USA)

Editorial Director: Dr Barbara Bijck

Production Manager: Peter J Biel

The cover photo shows a view of part of the BASF steamcracker in Antwerp

(Courtesy of BASF Aktiengesellschaft Ludwigshafen, Germany)

Library of Congress Card No applied for

British Library Cataloguing-in-Publication Data:

A catalogue record for this book is available from the British Library

Die Deutsche Bibliothek - CIP-Einheitsaufnahme

Thermal methods in petroleum analysis / Heinz Kopsch -

Weinheim ; New York ; Base1 ; Cambridge ; Tokyo : VCH, 1995

ISBN 3-527-28740-X

0 VCH Verlagsgesellschaft mbH, D-69451 Weinheim, Federal Republic of Germany, 1995

Printed on acid-free and low-chlorine paper

All rights reserved (including those of translation into other languages) No part of this book may be reproduced

in any form - by photoprinting, microfilm, or any other means - nor transmitted or translated into a machine language without written permission from the publishers Registered names, trademarks, etc used in this book, even when not specifically marked as such, are not to be considered unprotected by law

Composition: Graph Werkstatten Lehne GmbH, D-41516 Grevenbroich

Printing and Bookbinding: Paderborner Druck Centrum D-33100 Paderborn

Printed in the Federal Republic of Germany

Preface

The monograph Thermal Methods in Petroleum Analysis is based mainly on results of more than twelve years research work on the application of thermoanalytical methods to petroleum and its products during the activities of the author at the German Institute for Petroleum Research It was very interesting to research the application of well defined physical methods, such as thermogravimetry and differential scanning calorimetry, to the multicomponent systems of petroleum and its products, and to understand the limits of those methods on the one hand and the excellent transferability of the results to technical processes on the other The diversity of possible applications of thermoanalytical methods

to various problems in the petroleum laboratory can only be indicated in this mono- graph

Many people supported my work, either by active or by indirect help Thanks are expressed to Mrs Elvira Falkenhagen, who has been a skilful and reliable assistant for many years, as well as to Dr.-Ing Maria Nagel, Dr.-Ing Ulrike Tietz, Mrs Liliane Varoscic, Mrs Regina Bosse, Mrs Gerda Sopalla, and the late Mrs Heidi Gottschalck An

acknowledgement should be made to the directors of the German Institute for Petroleum Research: Professor Dr H H Oelert, Professor Dr H.-J Neumann, and Professor Dr D

Kessel who granted me maximum independent research capacity Some parts of the research work were carried out with financial support from the German Association for Research CD (Deutsche Forschungsgemeinschaft) For several years successful and plea- sant cooperation was established with colleagues of the University of Belgrade, especially with Professor Dr D Skala, Professor Dr M Sokic, and Professor Dr J A Jovanovic Thanks are also expressed to those whose names do not appear in this list All the companies which supplied me with information as well as with illustrations are likewise acknowledged; their names may be found in the appendix

I hope that this monograph will be of some help to colleagues in both academic and industrial research establishments and will encourage them towards further attempts in the application of thermal methods of analysis, even to chemically non-defined multicompo- nent systems The examples presented might represent a stimulation for further experimen- tal work

Heinz Kopsch Oktober 1995

Contents

1 Introduction 1

Methods and instrumentation 3

Thermal analysis on model substances

Experiments using the simultaneous thermal analyzer

Differential scanning calorimetry on model substances

DSC in an inert atmosphere 54

DSC in an oxidizing atmosphere

Reaction kinetics 68

Theoretical basis 68

Method according to ASTM E 698-79 69

Method according to Borchardt and Daniels

Method according to Flynn and Wall 72

Method according to McCarthy and Green 74

Kinetic investigations on model substances 75

DSC experiments according to ASTM E 698-79 heat of vaporization of n-alka- nes 75

Pyrolysis kinetics according to ASTM E 698-79 82

DSC oxidation kinetics according to ASTM E 698-79

Kinetics according to Borchardt and Daniels

TGA kinetics according to Flynn and Wall 90

TGA kinetics according to McCarty and Green 94

Thermoanalytical investigations on petroleum und petroleumproducts 97 Crude oils (degasified crudes) 99

Refinery residues 11 1

Description and characterization of the samples

Implementation and evaluation of tests

Thermogravimetry in an inert atmosphere

Directly measured index numbers

Derived index numbers 136

Directly measured index numbers 155

Correlations of analytical data with index numbers from thermogravime- try 160

Simulated thermal cracking by TGA

Index numbers from simulated cracking

Correlation of index numbers from simulated cracking with analytical da-

ta 165

Start temperature of the cracking process in an inert atmosphere

Differental scanning calorimetry (DSC) 167

Experiments in argon at atmospheric pressure

Experiments in methane at 10 bar pressure

Reaction enthalpy from tests at 10 bar pressure

Start temperatures of the cracking process at different pressures

Correlation of kinetic parameters with analytical data

Conclusions from experiments on refinery residues

Thermogravirnetry in inert gas 195

Correlation of index number from thermogravimetry with consistency da-

ta 202

Correlation index numbers with analysis data 21 1

Thermogravimetry in air 217

Isothermal aging tests by thermogravimetry

Differential scanning calorimetry (DSC) 233

Test in argon at atmospheric pressure 233

Tests in methane at 10 bar pressure 237

Tempratures of the cracking process 243

Oxidation in air 247

Low temperature behavior of bitumen

Conclusions from experiments on bitumen

Investigations on polymer modified bitumens (PMB)

Description and characterization of the samples 265

Thermogravimetry 269

Dynamic (temperature-programed) thermogravimetry 269

Isothermal gravimetq 275

Reaction kinetics using DSC 283

Low temperature behaviour of PMB using DSC

Aging properties of polymers for the modification of bitumen 287

Investigation on the hydrocracking reaction of heavy residues 296

Investigation on a vacuum residue from Kirkuk 297

Investigation on residues of different origins

Oil shale and shale oil

Investigation using TGA and DSC

Modelling and simulation of oil shale pyrolysis

Fingerprinting of oil shale by oxidation

Lubricants 348

Evaporation behavior of lubrication oils

Oxidation behavior of lubrication oils

Comparison of the oxidation stability of virgin oils, reclaimed oils, and synthe- tic lubrication oils 365

Silicone oils 376

Relation of the kinetics of pyrolysis and oxidation reactions to the system pressure: Investigations on tertiary oil recovery by in situ combustion Pyrolysis tests 405

Oxidation tests 410

Range of low temperature oxidation (LTO)

Range of fuel deposition 415

Range of fuel combustion 421

4.10.1.3 Kinetics according to Flynn and Wall 442

4.10.1.4 Kinetics according to McCarty and Green 453

4.10.2 Oxidation reaction 458

4.10.2.1 Kinetics according to ASTM E 698-79 460

4.10.2.1.1 DSC (DTA) experiments 460

4.10.2.1.2 Experiments using Simultaneous Thermal Analyzer

4.10.2.2 Kinetics according to Borchardt and Daniels 468

ments 439

467

X Contents

4.10.2.3 Kinetics according to Flynn and Wall 469

4.10.2.4 Kinetics according to McCarty and Green 473

4.10.3 Conclusions 477

5 Final consideration 485

5.1 Other applications 485

5.2 Summary on progress of instrumentation (hard and software) and advice 487

6 Appendix: Manufacturers of thermoanalytical instrumention 495

Conradson Carbon Residue (%)

Crackable part of the sample (%)

Dynamic Difference Calorimetry (see DSC)

Deutsches Institut k r Normung e V

(German Institute for Standardization)

Differential Scanning Calorimetry

Differential Thermal Analysis

Differential Thermogravimetry (First differential quotient of weight loss with respect to time) (% min ')

Activation Energy (J Mol-l)

Base of natural logarithm

Exponent with base e

Residual weight at the point of inflexion of the TGA curve (%)

Non-distillable part of the sample (%)

Nuclear Magnetic Resonance Spectroscopy

Reaction order (dimensionless)

XI1 List of Symbols

P Pressure (bar)

PCR Practical thermal crackable part of the sample (%)

Pen Needle penetration at 25 "C (0.1 mmj

PMB Polymer modified Bitumen

Q Quotient of weight loss in air divided by weight loss in inert gas (Isother-

mal Gravimetry)

R Universal Gas Constant (J Mol-' K-')

R600 Residue (%) at 600 "C experimental temperature

R800 Residue (%) at 800 "C experimental temperature

r Coefficient of correlation (dimensionless)

SAR Simulated atmospheric residue (%)

S.P.R&B Softening Point Ring and Ball ("C)

Temperature of the point of inflexion of the TGA curve ("C)

Temperature ("C) at 1 % weight loss

Temperature ("C) at 5 % weight loss

Fractional conversion (dimensionless)

Heating rate (K min-')

Solubility parameter according to Hildebrandt

Thermal Methods in Petroleum Analysis

by Heinz Kopsch

CopyrightoVCH Verlagsgesellschaft mbH, 1995

1 Introduction

Analytical methods describing the thermal behavior of substances during programmed temperature changes, like thermogravimetry, differential thermoanalysis, or differential scanning calorimetry are old methods, which were applied at first to problems of inorganic chemistry, mainly to minerals The analysis of petroleum and petroleum products has been mentioned relatively late In the literature survey by Weselowski [ 1-11 the first citation dates from 1958 Also, the oldest citation in the research report by Kettrup and Ohrbach [1-21 dates from 1965

Petroleum, especially heavy crudes, is recovered sometimes by the use of thermal processes like steam flooding or by in situ combustion The processing of the recovered crudes in the refineries is usually done by thermal methods at very different temperatures

A review of the temperatures applied in refinery operations is given in Table 1-1 These thermal processes are performed partly by sequential heating until the desired products are obtained The operating parameters for the different processes have been obtained to a large extent by empirical experience or partly by simulation of the processes in laboratory installations or in pilot plants For that reason thermoanalytical methods are considered to

be very useful in obtaining data concerning the thermal behavior i e data describing the

Table 1-1: Temperature Ranges in Petroleum Processing

Process Temperature Range ("C)

High Temperature Pyrolysis

Hydrocracking (Gas Phase)

Hydrocracking (Liquid Phase)

Visbreaking

Reforming (Thermal Treating)

Reforming (Catalytic Treating)

2 1 Introduction

thermal and oxidation stability of petroleum and its products; data predicting the manner and quantity of products gained in the processes; and data concerning reaction kinetics which can be used to optimize the refinery processes

Thermogravimetry (TGA), differential thermoanalysis (DTA), and differential scanning calorimetry (DSC) are the main methods which can be used in the analysis of petroleum and its products DSC is preferred to DTA, because DSC supplies values of energies directly, whereas the DTA supplies only temperature differences

These thermal methods of analysis have been described in several basic books [l-3 to 1-17] The application to polymers is described likewise [l-18, 1-19] So far no compilation on the application to petroleum and its products exists The situation in the field of standards is similar The NormenausschuB Materialpriifung im Deutschen Institut fur Normung (Committee for Testing and Materials of the German Institute for Standardi- zation e V., DIN) has approved only two standards (one of them contains terms of thermal analysis [ 1-20], the other is the standard for thermogravimetry [l-211) Furthermore there are three proposals (principles of differential thermal analysis [ 1-22], determination of melting temperatures of crystalline material by DTA [l-231, and testing of plastics and elastomers by DSC [ 1-24]) The American Society for Testing and Materials (ASTM) has

to date approved forty standards for the application of thermal methods of analysis Among them, seven standards are concerned with the testing of petroleum and its products [l-251

to [l-321, six standards are general methods [l-321 to [l-381, and four standards concerning the testings of polymers are applicable to petroleum and its products too [l-391 to 11-42]

Thermal Methods in Petroleum Analysis

by Heinz Kopsch

CopyrightoVCH Verlagsgesellschaft mbH, 1995

2 Methods and instrumentation

Using thermogravimetry (TGA), the dependence of the change in sample weight (mass)

on the temperature during programmed temperature changes in a chosen gas atmosphere can be measured The first derivative of the weight (mass) signal with respect to time is called derivative thermogravimetry (DTG) and is a criterion for the reaction rate It is usual

to record both the slope of the weight (mass) versus the time or temperature (TGA), and the differentiatoed curve versus the time or temperature (DTG) The heating rate dictates the actual position of the TGA and DTG graphs; it is therefore advisable always to use the same heating rate ( p ) so that different tests may be compared For small sample weights (masses), up to approximately 10 rng, a standard heating rate of 10 K/min is practicable This heating rate is slow enough to avoid any temperature gradient inside the sample while permitting a reasonable utilization of the available workmg time The shift to higher temperatures of the TGA and DTG curves as a consequence of faster heating rates permits calculation of the Arrhenius kinetic parameters and hence investigation of the reaction kinetics (see chapter 3.3) Furthermore, the position of the TGA and DTG curves will be influenced by the shape of the sample pan, especially by the ratio of surface to volume of the sample, and lastly by the quantity of gas flowing through the oven (gas flow rate) Therefore it is important that variations in sample quantity are minimized and that the gas flow rate is maintained as constant as possible However, the gas flow rate must not fall below a certain minimum value in order to avoid condensation of evaporated sample fractions on the hangdown of the sample holder or in the gas outlet tubes The minimum gas flow rate depends on the geometric shape of the oven and the position of the gas inlet and outlet tubes and therefore differs for different instruments If the gas flow rate is sufficient, the evaporated portions of the sample will be discharged immediately and therefore no equilibrium between liquid and vapor will be attained As a consequence the boiling (evaporation) temperature of the sample will decrease adequately That can be used

to perform a simulated distillation (see chapter 3.1.2) However, the application of ther- moanalytic methods is limited to substances having a start temperature of evaporation at

atmospheric pressure not far below 200 "C Otherwise there is the risk that evaporation in

the gas flow will begin at room temperature and thus the correct start temperature of evaporation (zero point of the TGA curve) cannot be ascertained

In principle all except very corrosive gases can be passed through a thermobalance; in practice the inert gases nitrogen, helium, and argon and the reactive gases air, oxygen, and hydrogen will be used

The weight calibration of thermobalances is done using standard weights The tempera- ture calibration is more difficult The method using the Curie point temperature, as

4 2 Methods and Instrumentation

described in ASTM E 914-83, does not work if a magnetic field from outside the oven is prevented from reciprocal action with the standard inside the oven, by the construction or the material of the oven Calibration using calcium oxalate monohydrate for standard is very common, since it has exhibited three clearly-defined steps of weight loss during heating (Fig 2-1 to 2-3).:

The thermogravimetric experiments are run using open platinum sample pans Pans made from aluminium, platinum, quartz, glass, stainless steel etc were also available The

Fig 2-1: Thermogravimetry of CaC,O, H,O

Plot of STA 780: TGA and DTA

Atmosphere: Argon 30 + 20 cm3/min

Heating Rate p: 10 K/min

Fig 2-2: Thermogravimetry of CaC,O, H,O

Plot of STA 780: TGA and DTG

6 2 Methods and Instrumentation

catalytic effect of the pan material on the pyrolysis reaction could not be ascertained when comparing the reaction in platinum and quartz pans, however, it could not be completely excluded All thermogravimetric experiments carried out by the author were run in plati- num pans Argon was used as the inert atmosphere Oxidation experiments were run in air because the reactions are too fast in oxygen

The first stage of experiments was carried out using a Stanton-Redcroft TG 750 thermo- balance connected to a three-pen recorder, recording weight (mass) loss (TGA), derivative thermogravimetry (DTG), and temperature (q For documentation the graphs of weight (mass) versus temperature were drawn manually Later on, the experiments were perfor- med using a simultaneous thermal analyzer Stanton-Redcroft STA 780 (STA 1 000), which

is equipped with a personal computer for control, data sampling, and data evaluation (Table

2-1) Using this device the curves of TGA, DTG, and DTA (differential thermal analysis) versus temperature can be plotted Furthermore, the PC is equipped with extensive softwa-

re to evaluate the results under varying conditions

0.5 100 K/min

TGA empirical index numbers evaporation

pyrolysis oxidation simulated distillation kinetics according to ASTM E 698-79 DTG empirical index numbers

STA 780 Stanton-Redcroft (STA 1000) TGA + DTG + DTA simultaneous up to 1000°C

normal pressure and vacuum

0.5

PC

PC TGA

DTG DTA

50 K/min

empirical index numbers evaporation

pyrolysis oxidation simulated distillation kinetics according to Flynn & Wall kinetics according to McCarty & Green empirical index numbers

kinetics according to ASTM E 698-79 specific heat

conversion temperatures kinetics according to Borchardt & Daniels kinetics according to ASTM E 698-79

FURNACE LIFTING SYSTEM’ 11

Fig 2-4: Diagram of the Thermobalance Stanton-Redcroft TG 750

1 Balance glass housing 8 Cooling water flow meter

2 Glass protection tube 9 Furnace

2a Brackets 10 Furnace lifting system

3 Glass protection tube 11 Spirit level

4 Counter weight glass housing 12 Support for glass protection tube

5 Gas inlet 13 Lower cover

6 Protection lid (Figure by Stanton-Redcroft Ltd.)

7 Gas flow meter

8 2 Methods and Instrumentation

A schematic digaram of the TG 750 is shown in Fig 2-4, of the STA 780 in Fig 2-5

The recorder script of an experiment with a hydrocarbon using the TG 750 is depicted schematically in Fig 2-6 Curve I represents the weight (mass) signal (TGA), curve I1 that

of the first derivative (DTG), and curve I11 the temperature (T> of the thermocouple directly below the sample pan Point A marks the start of the weight (mass) loss 1 % and the corresponding temperature T1 %; point B is the weight (mass) loss 5 % and the correspon- ding temperature T5 % Point C corresponds to the weight (mass) loss at 400°C (AG400) This is the temperature limit of the thermal stability of most non-aromatic hydrocarbons and of the heterocompounds Point D marks the weight (mass) of the coked residue at

600 "C (R600) or at 800 "C (R800) Point E represents the maximum of the DTG curve

Fig 2-5: Cross-Section of Water-cooled Furnance for STA 1 000 (STA 780)

A Water cooled cold finger

G STA hangdown assembly

(Figure by Rheometnc Scientific, Polymer Laboratories GmbH)

Start of weight loss (T1 %)

Start of weight loss (T5 %)

Weight loss up to 400 "C (AG400)

Residue at 600 "C (R600) or at 800 "C (R800)

Maximum of DTG curve (T-)

with the corresponding temperature T- The amplitude of the DTG curve corresponds to

the reaction rate The temperature of the DTG maximum shows whether the reaction remains in the evaporation (distillation) range (Tmx < 400 "C) or if a pyrolysis (cracking)

reaction has occurred (T- > 400 "C)

An example of rescaling the plot of weight versus time to weight versus temperature is shown in Fig 2-7 Here, the point of intersection of the tangents (offset point) represents the weight (mass) Gw of generated coke at the temperature Tw at the point of inflexion of the TGA curve This happens only during experiments in inert gas Using ash-free substan- ces in experiments in air, a TGA curve passing through zero weight is obtained, while ash-containing substances give a constant residual weight

The DTG graph of the experiment in air always shows more than one maximum, the first

of which can represent vaporization as well as oxidation In this case the TGA graph in protecting gas must be consulted for comparison

Figs 2-1 to 2-3 demonstrate possible evaluations using the STA 780 in an experiment with calcium oxalate monohydrate In Fig 2-1 the TGA curve is evaluated with respect to the weight (mass) losses of 1 %, 5 %, 10 %, and further in 10 % steps, whereas in the DTA curve the peak maximum temperature and the corresponding residual weight (mass) are plotted Fig 2-2 again shows the TGA and DTG curves with peak maximum temperatures and corresponding residual weights Fig 2-3 demonstrates the onset and offset temperatu-

10 2 Methods and Instrumentation

Using DSC, the position of the energy flows versus temperature curve as well as the rate

of an event were influeced by the heating rate, too Therefore the DSC tests were run likewise, using a standard heating rate p= 10 K/min with the exception of the investiga-

tion of reaction kinetics There the shift of the maxima of the energy flow curve to higher temperatures as a result of increasing heating rates permits the calculation of the Arrhenius lanetic parameters (see Chapter 3.3) All the influences, such as oven geometry, shape of the sample pan, position of gas inlet and gas outlet, on the results of DSC are the same as in TGA The gas purge with a minimum flow rate is also necessary in DSC to avoid condensations, when petroleum and its products or generally volatile substances are tested

As a consequence, the boiling (evaporation) temperature will decrease in a similar way to that in TGA The gases used in TGA can be used also in DSC Some additional experi- ments have been carried out in methane to study the influence of a hydrocarbon atmos- phere

For calibration, the melting point of indium were measured, which has a temperature of fusion (MP) = 156.4"C and a heat of fusion Hf= 28.46 J/g (Fig 2-8) Because reactions at

higher temperatures occur in experiments with petroleum refinery residues, additional calibration runs were performed using pewter (MP = 231.84"C, H,= 59.61 J/g) and lead

(MP = 327.40°C, Hf = 26.47 J/g) If a calibration at higher temperatures is necessary, potassium perrhenate KReO, (MP = 550°C, H,= 294.8 J/g) can be used

All DSC tests carried out by the author were run using open aluminium pans For reference an empty pan were used Comparative tests with platinum pans gave no indica-

12 2 Methods and Instrumentation

tion that the pan material had any influence, neither for pyrolysis nor for oxidation reactions The first experiments were carried out with the help of a DuPont 990 Thermo- analysis System connected to a 910 DSC This system used a two pen x-y recorder; the resultant graphs were evaluated manually Later, a DuPont 9900 Thermoanalysis System was used, which is equipped with a PC for control, data sampling, and data evaluation (Table 2-2) A cross-section of the DSC cell is shown in Fig 2-9

Table 2-2: Differential Scanning Calorimetry

DSC RT +650 "C vacuum till bar pressure up to 70 bar

0.5 50K/min

0.5 5K/min

PC

PC specific heat conversion temperatures reaction enthalpy heat of conversions kinetics according to ASTM E 698-79

Petroleum and its products are multicomponent systems of varying chemical composi- tion They are predominantly a mixture of hydrocarbons, usually accompanied by a small quantity of heterocompounds which contain in addition to carbon (C) and hydrogen (H) other atoms such as sulfur (S), nitrogen (N), and/or oxygen (0) Metals are present in very small concentrations, such as vanadium and nickel in organically bound forms The average elementary composition of petroleum in weight-% lies between the following limits [2-11:

- alkanes (unbranched n- and branched i-alkanes)

- cycloalkanes (naphthenes, unsubstituted and substituted)

- aromatics (unsubstituted and substituted)

- complex hydrocarbons (naphthenoaromatics)

Alkenes (olefins) and alkynes (acetylenes) are not found in petroleum (crude oils) Howe- ver, they were formed during the processing of petroleum at high temperatures

With regard to the boiling behavior, the full range of substances occur, from those which evaporate early during the recovery, as a result of pressure decrease, through to substances which cannot evaporate without decomposition It is possible to separate individual chern- cally-defined substances from the low boiling fractions From medium and high boiling fractions and from the non-distillable residues only multicomponent systems can be obtained, which can be separated into groups characterized by a similar chemical and physical behavior Separation into individual compounds is almost impossible

Under these circumstances, it seems reasonable to study the thermal reactions such as boiling, pyrolysis, and the oxidation behavior of defined model substances first, in order to understand the behavior of petroleum and its main products and to draw some analogous conclusions

Thermal Methods in Petroleum Analysis

by Heinz Kopsch

CopyrightoVCH Verlagsgesellschaft mbH, 1995

3 Thermal analysis on model substances

3.1 Thermogravimetry (TGA)

3.1.1 Thermogravimetry in an inert atmosphere

The series of n-alkanes from n-decane up to n-hexacontane (Table 3-1) and some chemicals mentioned in ASTM D 2887-84 (Table 3-2) were used for modelling purposes

Table 3-1: Boiling Points of n-Alkanes

Table 3-2: Substaiices according to ASTM D 2887-84

168 - 170

172 - 173 183.1

193 213.4 218.05

187 - 188

225 - 221

223 - 226

241 - 242 244.6 277.9 314.8

322

342

360 394.8

be 25 cm3/min Using the STA 780, a gas purge of 30 cm3/min through the oven, and an additional gas flow of 20cm3/min through the side tube of the hangdown protecting cylinder were needed In the STA 780, the gas flow causes complete vaporization of all the n-alkanes The STA plot of n-hexacontane is shown in Fig 3-1 as an example The chemicals mentioned in Table 3-2 were also completely vaporized, with the exception of

1, 3,5-triphenylbenzene, di-p-tolylsulfone, and palmitic acid methyl ester The latter start

to decompose when about 70-80 % of the sample has been evaporated The smaller gas flow rate in the TG 750 (25 cm3/min) causes a reduction in the mass loss rate In conse- quence, a pyrolysis reaction occurs, sooner or later depending on the purity of the indivi- dual substance Whereas pyrolysis of n-hexacontane starts above 400 "C as expected (Fig 3-2), the start of pyrolysis of n-hexatriacontane is as low as 330 'C, an unusually low temperature for n-alkanes (Fig 3-3) The substances which pyrolyze at relatively low temperatures in the STA 780, do the same in the TG 750: i e 1, 3, 5-triphenylbenzene

Fig 3-1: Thermogravimetry of n-Hexacontane

Plot of STA 780: TGA and DTG

Atmosphere: Argon, 30 + 20 cm3/min

Heating Rate f i : 10 K/min

At the beginning there is an induction period, the length of which is determined by the evaporation start temperature of the sample The induction period is followed by the evaporation The temperature T I % of a weight (mass) loss of 1 % or of 5 % (T5 %)

characterizes the initial boiling temperature As expected, the coefficient of variation k V

of the mean X is smaller for T5 % than for T1 %, if a statistical evaluation of at least ten experiments of the same sample had been carried out The coefficients of variation T1 %

and T5 % will also decrease with increasing boiling points (BP) of the samples The ratio T5 %/T1 % also decreases with increasing boiling point of the sample and attains a constant final value of T5 %/T1 % = 1.143 for boling points above 500 "C (Fig 3-4) However, a boiling temperature above 500 "C is an extrapolated value because no organic substance can be distilled at such temperatures without undergoing decomposition For example BP = 500 "C corresponds to the boiling point of n-hexatriacontane at atmospheric pressure With increasing temperature the slope of the TGA curve becomes steeper and passes through the final weight value zero, unless another chemical reaction occurs in the sample sooner Samples which disproportionate forming different volatile fragments (py- rolysis), exhibit a point of inflexion in the curve of weight versus temperature From this

Residue (%)

100

Fig 3-2: Thermogravimetry of n-Hexacontane

TG 750

Atmosphere: Argon, 25 cm3/min

Heating Rate p: 10 K/min

1

Atmosphere: Argon, 25 cm3/min

Heating Rate p : 10 K/min

I

Fig 3-4: Quotient of the Temperatures of 5 % Weight Loss

Divided by 1 % Weight Loss versus Boiling Point

Temperatures (BP) of n-Alkanes

point the slope of the graph becomes less steep The temperature at the point of inflexion

T , characterizes the pyrolysis behavior of the sample The slope of the TGA curve from this point of inflexion becomes less step and meets the zeroweight value at relatively high temperatures only if the pyrolysis results in the formation of products of different boiling points If the pyrolysis produces a coke residue as well as low boiling fragments, then the part of the TGA curve beyond the point of inflexion runs straight, nearly parallel to the temperature axis, demonstrating a constant residual weight (mass)

The model substances tested mostly exhibit only one maximum in the curve of weight (mass) loss rate versus temperature (DTG) The maximum temperature will be determined

by the boiling behavior of the sample The same is valid for the DTA curves measured by the Simultaneous Thermal Analyzer (STA 780) Each of the temperatures, which describe the thermal behavior in an inert atmosphere, can be repeated within relatively narrow limits

as shown in Table 3-3 by the example of four a-alkanes, and in Table 3-4 for five aroma-

large surface compared to the volume, so that the experimental conditions resemble those

of a thin-layer vacuum distillation

Nevertheless there are differences in the temperatures of equal evaporated portions, and

in the DTG maximum temperatures between the two thermobalances TG 750 and STA 780

as a consequence of the different shape of the sample pans:

TG 750 Pan STA 780 Pan Diameter (mm)

Wall Height (mm)

Ratio Surface: Volume

5.0 2.0 0.5

5.3 3.9 0.26

Table 3-4: Temperature Statistics ("C) for Aromatics

Table 3-5 Boiling Point ("C) BP

Temperature of I % Weight LOSS T1 % for TG 750 and ST* 780 (Q

Temperature of 5 % Weight Loss T5 - % I

TG 750 STA 780 Substance BP TI % T5 % T1% T5 %

n-Undecane 195.9 26.9 42.1 47.9 77.0 n-Eicosane 343.6 118.4 151.3 174.9 200.2 n-Triacontane 449.9 189.4 216.9 237.5 256.1 n-Tetracontane 526.3 230.6 263.5 278.6 309.8 n-Pentacontane 585.7 261.5 299.0 327.5 346.9 n-Hexacontane 634.2 286.3 327.8 340.9 358.4 n-Butylbenzene 183.1 22.8 38.9

Naphthalene 218.1 43.2 60.8

Anthracene 342.0 115.7 138.4

Indole 253.5 64.0 83.0

1,1,2,2-Tetraphenylethylene 417.5 158.4 184.2

3.1 Thermogravimetry (TGA) 23 Even if the ratio surface: volume seems to be nearly equal for equal sample sizes, the higher walls of the STA pan cause a difference in the removal of the evaporated portions Nor did the use of TG 750 pans in the STA 780 yield identical results, because the geometric conditions of the two ovens are different

Mixtures of model substances give TGA graphs with steps or some levelling, which prevent the exact identification of the components of the mixture, even when their boiling points were considerably separated from each other For instance Fig 3-5 demonstrates this on a mixture of equal parts of n-undecane (BP = 196 "C), n-pentacosane

(BP = 402 "C), and n-hexatriacontane (BP = 499 "C) The TGA curve of TG 750 exhibits

an evaporation start temperature T1 % = 46.8 "C The first step is terminated at an offset

of n-Undecane, n-Pentacosane, and n-Hexatricontane

TG 750

Atmosphere: Argon, 25 cm3/min

Heating Rate p : 10 K/min

Fig 3-5: Thermogravimetry of Test Mixture of Equal Parts

temperatue of 115 "C The second step has an onset temperature of 225 "C and an offset temperature of 305 "C The same experiment carried out using STA 780 gives the follow- ing temperatures: T1 % = 56.5 "C, offset1 = 157 "C, onset2 = 290 'C, and off- set2 = 390 "C The DTG curve from TG 750 shows two distinct peaks, at 99 "C and at

256 OC, and an additional shoulder at 327 "C From STA 780, a DTG curve exhibits three distinct peaks, at 129 "C, 296 OC, and 375 "C In the DTA curve two melting peaks were found, between 36 and 38 "C and between 69 and 71 "C The first one is almost 10 degreees below the fusion temperature of n-pentacosane, but the second one corresponds nearly to the fusion temperature of n-hexatriacontane (MP = 71 "C) Moreover three distinct maxi-

ma are present in the DTA curve, at 129 O C , 296 OC, and at 376 "C The statistical evalua- tion of five tests demonstrates that the data described previously can be repeated with small deviations as proved by the coefficients of variation (Table 3-6)

Linear polyethylenes containing minimal chain branching (ratio of CH, groups to 1 000 chain C atoms < 1) can be regarded as very long chain n-alkanes Thermogravimetry proves that they contain minimal portions of vaporizable substances (depending on the mean molecular weight Mw) and depolymerize nearly quantitatively at temperatures above

400 "C A small coked residue amounting to only 3-5 % was found in the TGA curve after

Table 3-6: Temperature Statistics ("C)

Mixture of Equal Parts of n-Undecane, npentacosane, n-Hexatriacontane

128.0 296.0 375.4

6.68 2.26 1.38 2.11 1.03 0.69 2.50 0.79 0.84

Table 3-7: Thermogravimetry of Polyethylene TG 750

Atmosphere: Argon 25 cm3/min

Heating rate p = 10 K/min

Mw Average of TI % T5 % T50 % TDTG

Chain Lenght ("C) ("C) ("C) ("C) C-Atoms

3.1 Therrnogravirnetvy (TGA) 25

a point of inflexion in the vicinity of 500 "C (Fig 3-6) The DTG curve demonstrates only one maximum around 500 "C A dependence of the evaporation start temperature on the mean molecular weight M , can only be recognized on polymers having small molecular weights The temperature of 50 % weight (mass) loss (T50 %) and the temperature of the DTG maximum do not exhibit any dependence on the molecular weight (Table 3-7) This indicates that depolymerization always follows the same reaction mechanism regardless of

the C chain length (molecular weight)

The aromatics tested (Tables 3-4 and 3-5) normally exhibit a levelling of the TGA curve

at weight (mass) losses more than 85-95 % There is doubt whether this is a real pyrolysis because the point of inflexion T, is in the range from 230 "C up to 300 "C Nevertheless the length of an aliphatic side chain influences the thermal behavior of substituted pyrenes

Atmosphere: Argon, 25 cm3/min

Heating Rate p : 10 K/min

Table 3-8: Thermogravimetry of Pyrenes and Alkylpyrenes (TG 750)

Atmosphere: Argon 25 cm3/min

Heating Rate p= 10 K/min

Fig 3-7: Thermogravimetry of Pyrene and Alkylpyrenes

TG 750

Temperature of 1 % Weight Loss (T1 %)

Temperature of 5 % Weight Loss (T5 %)

Temperature of 50 % Weight Loss (T50 %)

Temperature of the DTG Maximum (TDTG)

3.1 Thevmogravirnetry (TGA) 27

1 0 0 150 200 250 3 00

Fig 3-8: Boiling Point Temperatures (BP) of Aromatics versus Molecular Weights (M)

and trisubstituted alkylbenzenes, where the boiling point versus the chain length (C num- ber) of the alkyl substituent also gives a steady function Even the boiling points of the series of condensed aromatics versus the molecular weight results in a steady curve (Fig 3-

8)

The behavior of the substances specified in ASTM D 2887-84, Table X 1-1, (this

Chapter, Table 3-2) is different because these substances are not members of an homolo- gous series but represent olefins, cycloalkanes, aromatics, and heterocompounds The condensed aromatics and the polar compounds containing heteroatoms such as oxygen, sulfur, or nitrogen diverge distinctly from the boiling behavior of the non-aromatics of equal molecular weight (Fig 3-9) Therefore it is understood that the evaporation behavior

in the thermobalance ist also different

Fig 3-9: Boiling Point Temperatures (BP) of Substances according to ASTM 2887-84, Table X1.l

(Table 3-2) versus Molecular Weights ( M )

The fact that volatile substances will vaporize at temperatures far below their atmosphe- ric boiling point, due to the gas flow through the oven of the thermobalance, can be used to perform a simulated distillation [3-1 to 3-41 In contrast to the simulated distillation by gas chromatography GC [3-5, 3-61 no partition or rectification effect can be assumed during evaporation in a thermobalance Therefore calibration can only take place using individual substances of known boiling points It seems reasonable to use the substances which are named for calibration in the standards of simulated distillation by GC [3-5, 3-61

3.1 Thermogravimetry (TGA) 29

For calibration, the temperatures of the weight losses of 1 %, 5 %, 10 % and up to 100 %

or up to evident cracking symptoms were assembled from the TGA graphs of individual substances, and drawn versus the evaporated portion, similar to the Engler curves in

DIN 51 572 [3-71 Fig 3-10 demonstrates this in the diagram for n-hexatriacontane It can

be seen that cracking of the sample starts at a temperature above 320 "C when 70 % has already evaporated In the TG 750 the lower n-alkanes, up to a carbon number of approxi- mately 40 to 44, do not crack at all The higher ones start cracking when about 70 or 80 %

of the sample has evaporated In the STA 780 every n-alkane can be vaporized without decomposition except n-pentacontane and n-hexacontane Fig 3-1 1 shows the Engler curves of the n-alkanes measured using TG 750 (extrapolated to 100 % weight loss) Fig 3-12 shows the Engler curves of the substances presented in Table 3-2, also acquired

by TG 750 The repeatability is worse at small weight losses than at higher ones However, the repeatability is much better for the homologous series of n-alkanes than for the group of chemicals from ASTM D 2887-84 (Fig 3-13)

Fig 3-10: Thermogravimetry of n-Hexatriacontane

Diagram analogous to DIN 5 1 75 1

TG 750