I 2 lfydrocarbon dewpoint calculation Calculation of hydrocarbon dewpoint temperature is complex, as interactions between components must be accounted for in addition to individual comp
Trang 2JOURNAL OF CHROMATOGRAPHY LIBRARY- volume 56
petroleum industry
Trang 3This Page Intentionally Left Blank
Trang 4JOURNAL OF CHROMATOGRAPHY LIBRARY-volume 56
Trang 5ELSEVIER SCIENCE B.V
Sara Burgerhartstraat 25
P.O Box 21 1,1000AE Amsterdam, The Netherlands
Library of Congress Cataloging-in-Publication Data
Chromatography in the petroleum industry / edited b y E.R Adlard
cm (Journal of chromatography library ; v 56)
p
Includes bibliographical references and index
ISBN 0-444-89776-3 (acid-free)
1 Petroleum-Analysis 2 Chromatographic analysis-Industrial
applications I Adlard, E.R 11 Series
TP691.C58 1995
CIP
ISBN 0-444-89776-3
0 1995 Elsevier Science B.V All rights reserved
No part of this publication may be reproduced, stored i n a retrieval system or transmitted i n any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior written permission of the publisher, Elsevier Science B.V., Copyright & Permissions
Department, P.O Box 521,1000 A M Amsterdam, The Netherlands
Special regulations for readers i n the U.S.A - This publication has been registered with the Copyright Clearance Center Inc (CCC), Salem, Massachusetts Information can be obtained from the CCC about conditions under which photocopies of parts of this publication may be made i n the U.S.A All other copyright questions, including photocopying outside of the U.S.A., should be
referred to the publisher
No responsibility is assumed by the publisher for any injury and/or damage t o persons or property as a matter of products liability, negligence or otherwise, or from any use or operation
of any methods, products, instructions or ideas contained in the material herein
This book is printed on acid-free paper
Printed i n The Netherlands
Trang 6V
Contents
Foreword XV List of Contributors XVII
Chapter 1 The analysis of hydrocarbon gases
C.J Cowper 1.1 Introduction
1.2 Natural gas
1.2.1 Analytical requirements
1.2.1.2 Hydrocarbon dewpoint calculation
1.2.2 Analytical procedures
1.2.2.3 C6+ detail
1.2.2.4 Temperature programming
1.2.2.6 Separation in backflush
1.2.3 Quantitative measurement
1.2.1.1 CV measurement
1.2.2.1 Isothermal methods
1.2.2.2 Two detectors
1.2.2.5 Combined systems
1.2.2.7 Rapid analysis
1.2.3.1 Comprehensive analysis
1.2.3.2 Air contamination
1.2.3.3 Resolution
1.2.3.4 Precision
1.2.4.5 Response function
1.3 Refinery gas
1.3.1 Analytical requirements
1.3.2 Analytical procedures
1.3.3 Sample handling
1.4 Conclusions
1.5 Acknowledgements
1.6 References
1 1 2 5 5 8 10 10 16 16 18 19 19 21 25 25 25 26 26 27 28 29 30 37 38 39 40 Chapter 2 Advances in simulated distillation 41
2.1 Introduction 41
2.2 Middle distillates and lube oils 42
2.2.1 Precision 42
D.J Abbott 2.2.2 Capillary columns 44
Trang 7VI Contents
2.2.3 Aromatics and heteroatoms
2.2.4 Multi-element speciation.,
2.3 Gasolines and gasoline fractions
2.4 Heavy lube oils and residues
2.5 Crude oils
2.6 Process control and other applications
2.1 Conclusions
2.8 References
Chapter 3 The chromatographic analysis of refined and synthetic waxes
3.1 3.2 3.3 3.4 3.5 3.6 A Barker Introduction
Gas liquid chromatography
3.2.1 Establishing present technology
3.2.2 The 1980s revolution
3.2.2.1 Sample introduction
3.2.2.2 Detection
3.2.2.3 High temperature GLC columns
Quantitative gas liquid chromatography separation of waxes
3.2.3.1 3.2.3 Carbon number distribution analysis
Supercritical fluid chromatography of waxes
Size exclusion chromatography
3.4 I Early work on the SEC analysis of waxes
3.4.2 Present day technology
Conclusions
References
Chapter 4 4.1 Introduction
4.2 Microcapillary hydrodynamic chromatography
4.2.1 The0 ry
4.2.2 Expressing the size of macromolecules
4.2.3 Instrumentation
4.2.3.1 General aspects
4.2.3.2 Detection
4.2.3.3 The column
4.2.4 Applications
4.3.1 From hydrodynamic chromatography to tubular pinch
4.3.2 Coiling effects
4.3.3 Applications
4.4 4.5 Conclusions
4.6 References
Hydrodynamic chromatography of polymers
J Bos and R Tijssen 4.3 Capillary hydrodynamic chromatography
Hydrodynamic chromatography in packed columns
44
45
46
47
50
51
52
52
55
55
57
57
59
60
64
65
70
75
78
81
82
84
89
90
95
95
96
96
99
100
100
102
103
104
1 1 1
111
117
119
120
124
125
Trang 8Contents VII
Chapter 5 Chromatography in petroleum geochemistry 127
S.J Rowland and A.T Revill 5.1 5.2 5.3 5.4 5.5 Introduction
5.1.1 Recent reviews
Kerogen and other petroleum macromolecules
Geochemistry of petroleum
5.3.1 LC, TLC and TLC-FID
5.3.2 HPLC
5.3.3 GC
5.3.4 GC-MS
5.3.6 LC-MS
5.3.5 GC-isotope ratio-MS
5.3.7 SFC, GPC and SEC
Summary
References
Chapter 6 The O-FID and its applications in petroleum product analysis
Oxygenates as components of motor gasoline
Determination of oxygenates in unleaded fuels
6.3.1 Cracking reactor
6.3.2 Hydrogenation microreactor
6.4 Analytical procedure
6.4.1 Quantitative analysis
6.4.2 Total oxygen determination
Selectivity for oxygenates and sensitivity
6.5 O-FID applications
A Sironi and G.R Verga 6.1 Introduction
6.2 6.3 O-FID analyser
6.3.1.1 Low temperature cracker
6.2.1 6.4.2.1 6.6 Conclusion
6.7 References
Chapter 7 Microwave plasma detectors
Principle of operation of an atomic emission detector (AED)
Historical development of the plasma detector
Description and evaluation of a home-built atomic emission detector
A de Wit and J Beens 7.1 Introduction
7.2 7.3 7.4 7.4.1 Description of the apparatus
7.4.1.1 Microwave cavities
7.4.1.2 Microwave power supply
7.4.1.3 Spectrometer
7.4.1.5 Sample introduction system
7.4.1.4 Optical system
127
127
128
130
131
132
133
135
137
137
138
138
139
143
143
144
146
147
147
150
151
152
152
154
155
156
157
157
159
159
160
162
164
164
164
166
167
168
169
Trang 9VllI Contents
7.4.2 Plasma start-up
7.4.3 Operating limits of the detector
7.4.3.1 Emission line intensity
7.4.3.2 Plasma pressure
7.4.3.3 Microwave power
7.4.3.4 Quartz tube diameter
7.4.3.5 Optical system
7.4.3.6 Slit width
7.4.3.7 Upper limit of detection
7.4.3.8 Type and amount of scavenger gas
7.4.3.9 Linear dynamic range of the detector
7.4.3.10 Linear dynamic range for hydrogen
7.4.4 Selectivity
7.4.4.1 Selectivity to H, 0 and N
7.4.4.2 Selectivity to carbon
7.5 Conclusions
7.6 Description of the Hewlett Packard 592 1 A AED
7.6.1 Gas chromatograph and transfer line
7.6.2 Microwave cavity, discharge tube and gas flow system
7.6.3 The photodiode array spectrometer
7.6.4 Computerized control and data treatment
7.6.5 Characteristics
7.7 Typical applications
7.7.1 Multi-element SimDist ses
7.7.1.1 Multi-element SimDist s o h a r e
7.7.1.2 Linear dynamic ranges
7.7.1.3 Simdist results
Determination of chlorine-containing compounds in the ethylene oxide process
Determination of oxygenates in gasoline
Determination of metal porphyrins in crude oils
7.7.2 7.7.3 7.7.4 7.7.5 Determination of noble gases in natural gas
7.5 Conclusion
7.9 References
Chapter 8 The sulfur chemiluminescence detector
R.S Hutte 8.1 Introduction
8.2 Sulfur-selective detectors for gas chromatography
8.3 The sulfur chemiluminescence detector
8.4 Performance characteristics of the SCD
8.5 Factors influencing the sensitivity and selectivity of the SCD
8.6 Flameless sulfur chemiluminescence
8.7 Column selection and sampling techniques
8.8 ADdications
169
170
171
172
172
173
173
174
174
176
177
179
180
180
181
184
184
185
187
189
190
190
191
191
192
194
195
197
197
197
197
197
200
201
201
202
203
207
209
211
213
215
Trang 10Contents IX
8.8.1 Refinery gases
8.8.1.2 Gasoline
8.8.1.3 Diesel fuels
8.8.1.4 High temperature gas chromatography
8.9 Conclusions
8.10 Acknowledgments
8.11 References
Chapter 9 Multi-column systems in gas chromatography
H Mahler T Maurer and F Mueller 9.1 Introduction
Limitations of single-column systems
9.1.1 9.1.2 Multi-column systems
9.1.3 Multi-column chromatography
9.1.3.1 Definition
9.1.3.2 Variants of column switching
Selectivity tuning in series-coupled columns
9.2 9.3 Column switching techniques
9.3.1 General
9.3.2 Backflushing
9.3.3 cutting
9.3.4 Distribution cutting
9.3.5 Special switching techniques
Summation of backflushed compounds
9.3.5.2 Stopped flowlstuttering
9.3.5.3 Recycle chromatography
9.3.5.1 9.4 Practical aspects
9.4.1 Valve switching
9.4.2 Valveless flow switching
9.4.2.2 Live-switching
Column-switching as a sampling technique
9.4.3.1 Sampling fiom a capillary pre-column to a main capillary column
9.4.3.2 Coupling of columns of different type with intermediate 9.4.2.1 Deans-switching
9.4.3 trapping
Strategies for the application of multi-column systems
9.5.1 General remarks
9.5.2 Guidelines for the use of single- or multi-column systems
9.6 References
9.5 Chapter 10 Supercritical fluid extraction
T.P Lynch 10.1 Introduction
10.2 Why use supercritical fluid extraction?
215
218
218
218
227
227
227
231
231
231
232
233
233
234
238
241
241
244
245
246
249
249
250
250
250
251
252
253
255
258
260
262
263
263
266
266
269
269
271
Trang 11x
10.3
10.4
10.5
10.6
10.7
Contents
10.2.1 Technical advantages
10.2.1.2 Selectivity
10.2.1.3 Volatility
10.2.1.4 Matrix penetration
Environmental and safety advantages
ms
10.3.2 Sample vessels and temperature control
10.3.4 Extract collection
SFE techniques
10.4.1 Static extraction
10.42 Dynamic extract
10.4.3 Recirculating extraction
10.4.4 Extraction of liquids
10.4.5 Reactive extraction
10.4.6 Off-line extraction
10.4.7 On-line extraction
Petroleum-based applications
10.5.1 Off-line applications
10.5.1.1 Residual oil on drill cuttings
10.5.1.2 Drilling mud characterization
10.5.1.3 Petroleum core and rock extractions
10.5.1.4 Refinery catalysts, deposits and sludges
10.5.1.5 Automotive engine particulates
10.5.1.6 Environmental analysis
10.5.2 On-line applications
10.5.2.1 On-line SFE-GC
10.5.2.2 On-line SFE-HPLC
Conclusion
References
10.2.1 1 Solubility
10.2.2
10.3.3 Depressurization
27 1 271 271 274 274 275 275 276 277 278 278 280 280 280 281 281 283 283 284 284 284 284 286 287 288 288 290 291 291 299 301 301 Chapter 11 Supercritical fluid chromatography
1 Roberts 1 1 1 Introduction
1 1.2 Instrumentation
Mobile phase pumps
1 1.2.1.2 Syringe pumps
I 1.2.2 Ovendtemperature control
11.2.3 Injectors
1 1.2.4 Detectors
11.2.4.1 Universal detectors
1 1.2.4.2 Spectroscopic detectors
1 I 2.1 1 1.2 I 1 Reciprocating pumps
305
305
306
307
308
308
309
309
310
311
314
Trang 12Contents XI
11.2.4.3 Element specific detection
1 1.2.5 Restrictors
11.2.5.1 Fixed restrictors
11.2.5.2 Variable restrictors
1 1.3 Applications
11.3.1 Simulated distillation
1 1.3.2 Hydrocarbon group type analysis
1 1.3.2.1 Gasolines
11.3.2.2 Kerosenes and naphthas
11.3.2.3 Diesel fuels
11.4 References
Chapter 12 HPLC and column liquid chromatography
A.C Neal 12.1 Introduction
12.2 Apparatus
12.2.1 Solvent reservoirs
12.2.2 Pumps
12.2.3 Sample injectors
12.2.4 Columns
12.2.5 Detectors
12.2.6 Selective property detectors
12.2.6.1 UV-visible spectrophotometers
12.2.6.2 Diode array detectors (DAD)
12.2.6.3 Fluorescence detectors
12.2.6.4 Electrochemical detectors
12.2.6.5 Flame ionization detector
12.2.6.6 Mass spectrometers
12.2.6.7 Infrared and NMR
12.2.7 Bulk property detectors
12.2.7.1 Refractive index detector
12.2.7.2 Evaporative light scattering detectors
12.2.7.3 Dielectric constant detector
12.4 Applications
12.4.1.1 Polycyclic aromatic hydrocarbons (PAHs)
12.4.1.2 Other indigenous compounds
12.4.1.3 Additives and contaminants
12.4.1.4 Compound classes
12.5 Preparative HPLC and column liquid chromatography
12.5.1 Standard methods
12.7 Future trends
12.8 References
12.3 Quantitation
12.4.1 Individual compounds
12.6 Individual publications
318
320
320
321
322
322
328
330
334
336
343
347
347
348
348
348
350
351
353
353
353
354
355
355
356
356
357
358
358
359
360
361
362
362
362
365
366
368
370
370
371
372
3 72
Trang 13XI1 Contents
Chapter 13 Modern data handling methods 375
The role ofthe data processor
13.2.1 The modem integrator has four distinct roles 376
13.2.3 Prepare and distribute data, information and reports 378
13.3 Limitations of the detector signal 378
13.3.1 Injector fractionation is uniform 378
13.3.2 All solutes of inte 379
13.3.3 Column resolution is 379
13.3.4 All solutes of interest 379
13.3.5 Signalinoise ratio 379
13.3.6 Detector is linear 380
13.4 Detector signal processing 380
13.4.1 Improving signal to 380
13.4.2 Electronic filtering 381
N Dyson 1 3.1 Introduction 375
13.2 376 13.2.2 Lab economics 378
13.4.3 Software smoothing
13.4.4 Measurement of pe 381
13.4.5 Baseline placement
developments 382
13.4.7 Selective extraction 383
13.4.7.2 Signal subtraction 383
13.4.8 Resolution enhancement 384
13.5.2 Errors crea 385
13.5.4 Errors created by asymmetry 386
13.5.5 Transmission of systematic errors through oups 387
13.5.7 Errors of tangent/perpendicular peak splitting 390
33.5.8 Ease of use 390
13.6 Calibration 390
13.6.1 Use of area or height for quantitation 390
I 3.4.6 Mathematical separation (deconvolution) of peaks and integrator 13.4.7.1 Peak model1 383
13.5 Measurement errors 384
13.5.1 Accuracy 13.5.3 Errors crea 13.5.6 Baseline construction errors 390
13.6.2 13.6.3 Limitations of the standard calculations 391
391
392
13.6.4 Calibration curves 392
Calculations and response factors
1 3.6.3.3 External standard
13.6.5 Empirical correction of analysis errors 393
13.7 Validation and standard chromatograms 393
Trang 14Contents XI11
13.7.1 Meaning of validation
13.7.2 System suitability
13.7.3 Validation and standard chromatograms
Strategies for peak measurement
13.8.1 Noise
13.8.2 Baseline drift
13.8.3 Peak overlap
13.8.4 Asymmetry
13 3.6 Integrators
Checking the analysis results
13.8.7.1 Checking the chromatogram
Checking the analysis rep0 rt
Accepting the results
13.8 13.8.5 Detectors
13.8.7 13 3.7.2 13 X7.3 13.9 References
Chapter 14 Capillary electrophoresis in the petroleum industry
T Jones and G Bondoux 14.1 Introduction
14.2 Separation techniques
14.2.1 Free-zone capillary electrophoresis (FZCE)
14.2.2 Micellar electrokinetic chromatography (MEKC)
14.2.3 Gel filled capillary electrophoresis (GFCE)
14.2.4 Capillary isoelectric focusing (CIEF)
14.2.5 Instrumentation
14.2.6 Capillary
14.2.7 High voltage power supply
14.2.8 Temperature control
14.2.9 Injection
14.2.10 Detection
14.2.10.1 UV detection
14.2.10.2 Fluorescence, indirect fluorescence, laser-induced fluo- rescence
14.2.10.3 Amperometric detection, conductometric detection, MS detection
14.3 14.4 Conclusion
14.5 References
Applications of capillary electrophoresis
393 394 394 395 396 396 396 396 397 397 397 397 398 398 398 401 401 404 404 408 411 412 412 412 413 413 414 415 416 418 419 421 424 425 Subject Index 427
Trang 15This Page Intentionally Left Blank
Trang 16XV
Foreword
Although the spectacular development of gas chromatography in the 1950s and 1960s is now a fading memory of a golden era, there are still advances being made in apparatus, technique and applications The petroleum industry makes use of all the variants of chromatography as should be clear from the present volume, but gas chromatography in particular is the most important analytical technique in petroleum analysis and has been since its first announcement by James and Martin in 1952 Indeed it is no exaggeration to claim that many of the major advances in gas chromatography in that golden era emanated from the laboratories of the petroleum industry
This book is intended primarily for those concerned with the analysis of crude oil and its products but many of the chapters have much broader applications It
is hoped, therefore, that many outside the immediate sphere of petroleum analy- sis will find sufficient of interest to make it a worthwhile purchase
In multi-author books there will be inevitable variations in the style and con- tent of each contribution There is no reason why this should be regarded as a weakness since as William Cowper pointed out “variety’s the very spice of life” Likewise a small amount of overlap between some chapters is not a drawback if
it allows each chapter to be a freestanding account of a particular topic
It was interesting to reread the comments of the editors of the only other book dedicated to the subject of petroleum analysis by chromatography published
15 years ago These editors spent some time describing the reasons for the choice of the title of their book In this context, it is interesting that the original title intended for their book was the one used here
In concluding this foreword, I should like to thank all the contributors and El-
sevier for their efforts to make this both a useful and an interesting volume
E R Adlard
Trang 17This Page Intentionally Left Blank
Trang 18B K), Badhuisweg 3, 1031 CMAmsterdam, The Netherlands
Waters Chromatography Division, Millipore S.A., 6 Rue Jean- Pierre Timbaud, BP 307, 78054 St Quentin-en- yvelines, France
B K), Badhuisweg 3, 1031 CMAmsterdam, The Netherlands
84, West Grove, Walton on Thames, Surrey KT12 5PD, UK
KoninklijkdShell-Laboratorium, Amsterdam (Shell Research
B K), Badhuisweg 3, 1031 CMAmsterdam, The Netherlands
Houghton-le-Spring, Tyne and Wear DH5 Om, UK Sievers Instruments Inc., 1930 Central Avenue, Suite C, Boul- der, CO 80301, USA
Waters Chromatography Division, MiIlipore (CK) Ltd, Winster House, Heronsway, Chester Business Park, Wrexham Road, Chester CH4 SQR, UK
Analytical & Applied Science Division, BP Research & Engi- neering Centre, Sunbury-on-Thames, Middlesex, T W I 6 7LN
UK Siemens AG, Abt AUT 35 CHR, Postfach 21 12 62, 76187 Karlsruhe, Germany
Siemens AG, Abt AUT 35 CHR, Postfach 21 12 62, 76187 Karlsruhe, Germany
Siemens AG, Abt AUT 35 CHR, Postfach 21 12 62, 76187 Karlsruhe, Germany
OX13 6AE UK
Trang 19Analytical & Applied Science Division, BP Research & Engi-
neering Centre, Sunbury-on-Thames, Middlesex, TW16 7LN,
UK CSIRO Division of Oceanography, Castray Esplanade, Hobart, Tasmania, Australia
Petroleum and Environmental Geochemistry Group, Depart- ment of Environmental Sciences, Universily of Plymouth, Drake Circus, Plymouth, PL4 8AA, UK
Fisons Instruments, Strada Rivoltana, 20090 Rodano (Milan), Italy
KoninklijkdShell-Laboratorium, Amsterdam (Shell Research
B V.), Badhuisweg 3, I03 I CM Amsterdam, The Netherlands
Fisons instruments, Strada Rivoltana, 20090 Rodano (Milan), Italy
Trang 20E.R Adlard (Ed.), Chromatography in the Petroleum Industry
Journal of Chromatography Library Series, Vol 56
0 1995 Elsevier Science B.V All rights reserved 1
CHAPTER 1
C.J Cowper”
British Gas pfc, London Research Station, Michael Road, London SW6 ZAD, UK
“That man sat down to write a book, to tell the world what the world had all his life been telling him.”
Boswell’s Life of Johnson
Gas chromatography is the principal analytical method used for hydrocarbon gases Particular components can be measured by spectroscopic or chemical means, but for analysis of the bulk of components, the separating power of chromatography is both essential and well developed
Although gases are often considered to be simple mixtures, their analysis has frequently tested the ability of gas chromatography, either because of the range
of components present (boiling range, or concentration spread, or both), or be- cause of the need to use highly specific stationary phases to separate apparently intractable pairs of components The different separating requirements relating
to groups of components within the same gas mixture has led to the use of multi- column systems, with columns being isolated or reversed, or their order changed
* Current address: 84, West Grove, Walton on Thames, Surrey KT12 SPD, UK
Trang 212 Chapter I
by means of valves This complexity, which is more daunting in prospect than in use, has led to a number of ready-configured chromatographic systems for many
of the application areas
The thermal conductivity detector (TCD) and flame ionization detector (FID) are the two most commonly used for hydrocarbon gases in the petroleum indus-
try Because many of the gases contain non-hydrocarbon components, the TCD,
as a universal detector, is essential Its dynamic range allows it to be used also for all the major and many of the minor components of most mixtures The FID, while the most commonly used detector in gas chromatography generally, can be regarded as a specialist and specific detector in gas analysis Process chroma- tographs frequently use the TCD alone, to reduce the need for the extra facilities needed for the use of the FID
1.2 NATURAL GAS
Hydrocarbon gases arise naturally from a variety of sources Bacterial fermen- tation under anaerobic conditions produces methane or marsh gas in great pro- fusion, about 109 tonnes per year worldwide Small accumulations of this type of gas can be found during tunnelling or other operations, and the same mecha- nisms produce landfill gas from waste Mine drainage gas is a methane-rich mixture found where coal measures have been worked However, the term natu- ral gas is normally taken to refer to the fossil-based gaseous equivalent to oil and coal, abstracted from ancient, large, deeply buried accumulations This is the sense in which the term is used in this chapter
Natural gases can vary considerably in composition, from nearly pure nitrogen
to nearly pure carbon dioxide to nearly pure methane Fortunately for the indus- try and the consumer, most natural gases consist mainly of methane, with small amounts of inert gases (helium, nitrogen and carbon dioxide) and ethane and higher alkanes in concentrations which diminish as their carbon number in- creases
By far the largest use of natural gas is as a fuel, where its accessibility via wide-ranging distribution systems and its cleanness in terms of handling and combustion products make it a popular choice for both domestic and industrial/ commercial markets Other uses are as a chemical feedstock, as a source of pure single hydrocarbon gases or (if present in sufficient quantities) of helium, and as
a moderator in nuclear reactors
Current estimates indicate that the world has more reserves of natural gas than
of oil at the present rate of consumption Recent measures of worldwide produc- tion give a figure of around lo9 tonnes per year, which is comparable to the bac- terial production referred to earlier
Trang 22The analysis of hydrocarbon gases 3
Natural gas is part of a continuum of hydrocarbons, ranging from methane to the heaviest ends of oil, which are found in geological accumulations Pressure and temperature conditions in the reservoir are such that there is no distinction between what we regard as gases and liquids; this only occurs when the fluid has been extracted and is subject to conditions at which this discrimination is possi- ble Whether an accumulation is regarded as a gas or oil field is only a matter of the relative proportions of the hydrocarbons Natural gas fields always contain liquids, usually in the form of a lightish condensate, and oil fields always contain associated gases
Gas separated from a natural gas field will burn in that form, but is usually treated to remove or to control the levels of particular components, for opera- tional, or contractual, or legislative reasons Hydrogen sulphide, being toxic and corrosive, is invariably subject to very low (parts per million) specification lim-
its, and is typically removed in an amine plant Carbon dioxide is less acidic, but
still potentially corrosive at the pressures used for gas transmission, and its con- centration is also controlled, usually to low percentage levels It can be removed
by an alkali scrubbing process Water is removed by glycol scrubbing, since the presence of liquid water increases the corrosive effect of acid gases, and because
it can form solid methane hydrate, a clathrate compound, under certain pressure and temperature conditions Potential hydrocarbon liquids are also removed, usually by chilling, sometimes by adsorption This is to prevent their condensa- tion downstream of the processing plant
The fact that natural gas, once processed at the wellhead or reception termi- nal, is in the form which virtually every consumer can accept without modifica- tion has given rise to very complex and detailed pipeline systems, which cross international boundaries and finally enter the consumer’s premises In Western Europe, most countries have access to pipeline supplies from Holland, the North Sea, Siberia and Algeria in addition to their own indigenous sources In the United States, which is the home of long-distance natural gas transmission, pipeline systems include Canada and Mexico as well as extensive offshore net- works
Properties and behaviour of natural gas have been reviewed by Melvin [ 13 A large number of papers on quality specifications, physical properties, sampling, odorization and analysis of natural gas, and on calibration gases and standardi-
zation are collected in the proceedings of the 1986 Gas Quality Congress [2]
Analysis of natural gas is carried out for a range of purposes, and the choice
of analytical method is often dictated by the reason for the analysis being re- quired There are three basic purposes for analysis:
- identification of source,
- calculation of physical properties, and
- measurement of specific minor components because of their particular characteristics
Trang 234 Chapter I
For identification of source, the concentrations of the inert components and the ratios of a small number of hydrocarbons are good indicators; the analysis need not be detailed An example of specific minor component analysis is the measurement of odorants; the analysis is clearly targeted upon a few compo- nents, probably using a selective detector, and the composition of the main com- ponents is without interest, except insofar as they may interfere with the meas-
urement Calculation of properties is the most common need for analysis, with
calorific value the most usual target
The following is a list of some of the properties of natural gas which are cal- culable from analysis It is not comprehensive, but describes those most fre- quently used Most properties can be measured directly, but independently of each other; a properly configured analytical method allows calculation of all
1 Culorijic value (CV): Natural gas is bought and sold in units of volume, as
a source of energy, hence the importance of CV as energy per unit of volume
2 Relative density (RD): This is the density of a gas relative to dry air (= 1.000) It is used in metering calculations and for the Wobbe index (see be- low)
3 Wobbe index (WI): Gases from different sources must be assessed for their inter-changeability, which represents the effectiveness with which a gas of com- position B will burn on an appliance designed for a gas of composition A WI is
an empirical measure of the ability to supply heat to a burner, and is the most important characteristic in determining interchangeability It is calculated by di- viding the CV by the square root of the RD
4 Compression factor (Z): Compression factor appears in the modified ideal gas equation P V = nZRT, and arises from gas phase interactions For hydro-
carbon gases and their mixtures over normal temperature and pressure ranges, Z
is always less than 1 , which means that a defined volume of gas at a defined pressure will contain more moles than predicted from ideal behaviour by a factor
of 1/Z At ambient conditions, Z for most natural gases is around 0.997, but the correction is much more significant at higher pressures At 70 bar, typical of transmission pressures, Z is usually less than 0.9 Metering at high pressure is
therefore very dependent upon accurate measurement or calculation of Z
5 Hydrocarbon dewpoint: Retrograde condensation is the phenomenon
whereby a liquid phase can separate from a hydrocarbon gas mixture as it is de- pressurized at a constant temperature It is another feature of gas phase interac- tions, and may be regarded as a form of “gas phase solubility”, with components coming out of solution as the pressure binding the molecules together is re- leased
6 .Joule-Thomson coefficient: This property influences the extent of cooling
as a gas is expanded As the pressure of natural gas is reduced, the amount of
pre-heating necessary to avoid hydrocarbon condensation can be calculated
Trang 24The analysis of hydrocarbon gases 5
1.2.1.1 CVmeasurement
The CV of a gas mixture is an additive property, with inert gases contributing zero, and flammable gases contributing in proportion to their concentration and individual CV A small correction is necessary for compression factor (2) at ambient conditions
Figure 1.2 shows, for a typical North Sea gas, the component contributions in terms of molar %, and of CV and RD as percentages of the total Nitrogen, pres-
ent at 2.5%, contributes nothing to the CV, but 4% to the RD The Y-axis of the
figure is limited to 6% so that component contributions can be clearly seen Methane, of course, contributes far more than the figure indicates
Trang 25I M o i a r % B C a l Value =Re1 Density
Fig 1.2 Component contributions
Figure 1.3 expands the information for the higher hydrocarbons (C, and above) It is clear that the relative contribution to CV and RD is greater with in- creasing carbon number, but the diminishing concentrations means that the ac-
mMolar % B C a l Value B R e l Denslty
Fig 1.3 Component contributions
Trang 26The analysis ofhydrocarbon gases 7
Fig 1.4 Calorific value errors
tual contribution is small Figure 1.4 shows the error involved in MJ/m3 if a component or group of components is missed (The total CV would be around
38 MJ/m3) This is shown for components or groups, and also cumulatively, from
a particular carbon number upwards Thus, if the Clo hydrocarbons are not measured, but their molar contribution is assumed to be included with that of methane, the CV will be underestimated by only 0.003 MJ/m3 C, hydrocarbons,
if not measured, would cause an underestimate of 0.009 and C, hydrocarbons of
0.015 MJ/m3 It is, of course, much more likely that if the C, hydrocarbons have been missed, so also would the C, and C,, hydrocarbons, giving a cumulative error In this case, missing C, and above would give an underestimate of
0.027 MJ/m3, and if C, and all higher hydrocarbons are not measured, the error will be about 0.075 MJ/m3
A calculation uncertainty of 0.1 MJ/m3 is a realistic target for a properly con-
figured and accurately calibrated analyser, and so to minimize the bias error arising from undetermined components, the analysis should include C, and pref- erably C, hydrocarbons One of the common methods of analysis backflushes all
C, and higher hydrocarbons to the detector, where they are measured as a com- posite C,+ peak This includes all higher hydrocarbons in the measurement, but raises two further uncertainties: unless there is independent information about the detailed composition of this peak, its response factor must be relatively ill-
Trang 27x Chapter 1
defined, and so must its contribution to CV or other properties In fact, the CV of the C,+ fraction of many gases can be approximated by that of n-hexane without significant error Components such as benzene and toluene, and to a lesser extent the cyclo-alkanes have lower CVs than alkanes of equivalent carbon number, and if present in reasonable proportion can counteract the higher CV contribu- tions of C, and higher alkanes
1.2 I 2 lfydrocarbon dewpoint calculation
Calculation of hydrocarbon dewpoint temperature is complex, as interactions between components must be accounted for in addition to individual component properties, Higher hydrocarbons make a considerable contribution, because of their relatively low vapour pressures For CV calculation, it is normal to con- sider all alkanes of a particular carbon number as a group Since the CVs of al- kane isomers are very similar, this is realistic and involves virtually no loss of accuracy The same approach is incorrect for hydrocarbon dewpoint calculation,
as the contributions of isomers differ This creates two problems: computer packages for these calculations cannot handle as many components as a detailed analysis can measure, and even the most detailed analysis cannot definitely identify all the peaks which it separates, nor find the appropriate properties for those components through a database
A typical computer program will handle 30 components, and one approach has been to group alkane isomers as if their sum was represented by the n-alkane
of that carbon number Since the n-alkane has the highest boiling point of the series, this approach will over-estimate the contributions to dewpoint tempera- ture, and so has the advantage of a built-in safety margin A more accurate ap- proach is to input data for groups of components as that of fractions rather than components, assuming that the program allows components and fractions to be mixed
Detailed separation of higher hydrocarbons is most likely to be on the basis of boiling point, as in simulated distillation Each peak in the chromatogram, with- out being identified, can have a boiling point allocated to it based upon its reten- tion time relative to bracketing n-alkanes, a carbon number based upon its posi- tion in the chromatogram, and hence an FID response factor and molar percent- age It is therefore practicable to consider a group of hydrocarbons, such as the
C, alkanes, not as n-C, but as the C, fraction This fraction has a defined molecu- lar weight and density, a molar concentration and a calculated average boiling point This is sufficient information to be able to input the C, data as a fraction with properties which more realistically represent its contribution The same ap- proach can be used for C,, C,, C,, and any higher hydrocarbon groups which may be measured
Figure 1.5 shows the errors involved in dewpoint temperature calculation if components or groups of components are not measured, and their molar contri-
Trang 28The analysis of hydrocarbon gases 9
Fig 1.5 Dewpoint temperature errors
bution taken up by that of methane The failure to measure individual compo- nents or groups produces errors measured on the left hand scale The significant role of the higher hydrocarbons is clear, with a C,, error of -3.7"C and a C, error
of -5°C By contrast, ignoring propane will only give a -0.4"C error, and ignor- ing ethane or the inert components hardly any
As with the CV errors above, it is likely that if C, is missed from the analysis,
so also will Clo This cumulative error, where failure to measure hydrocarbons
of any carbon number also includes those of higher carbon number, is shown with reference to the right hand scale, where the range is 10 times greater It is evident that the cumulative errors are larger than would be assumed by addition
of the individual contributions Failure to measure C, and higher components will cause an underestimate in calculated dewpoint temperature of more than
50°C
Analysis for dewpoint calculation must, therefore give details of higher hy- drocarbons The example is for a gas where C,, and higher components are not detectably present, which is normal for gases treated to a transmission specifica- tion If they were detectably present, they must be measured and included in the calculation While it appears that the analysis need not measure lighter compo- nents particularly well, if at all, it would be bizarre and unusual to configure an analytical scheme to be poor for the easy to separate light components and good for the more difficult traces of heavy ones Also, a particular analytical result can
Trang 291.2.2 I Isothermal methods
IP 337 [3] recommended the use of three different separations, a molecular sieve column with argon carrier gas for He, 0, and N,, a porous polymer column with helium carrier operated at 50°C for CO, and C,, and a porous polymer col- umn with helium carrier operated at 140°C for C,, C, and C, hydrocarbons The analysis went no further than C,, and methane was measured by difference
ASTM D 1945 [4] sought to achieve measurement of more components in a single separation, which included a 10-m column with a high loading of silicone oil on Chromosorb P This separated N,, C,, CO, and C, to C, hydrocarbons in- dividually A molecular sieve column was still necessary for measuring air com- ponents
Figure 1.6 shows a separation on a boiling-point column (6 m of 28% silicone oil DC 200/500 on Chromosorb P at 100°C) C,, C, and C, hydrocarbons are well separated, but the light components, N,, C,, CO, and C,, while distinct, are not sufficiently well separated for best quantitative measurement In particular,
CO, at low concentration can be difficult to discern between C, and C, Hydro- carbons above C, are slow to elute, and the combination of their decreasing con- centrations and increasing peak widths makes their measurement more and more difficult (In this and subsequent figures, only the n-alkanes are identified, to avoid clutter)
Backflushing to recombine all hydrocarbons above C, and pass the combined peak (C,+) to the detector has two advantages: the recombined peak will be larger than the individual ones, and the analysis time will be reduced Against this, we cannot make exact allowance for the contribution of all the individual components, but must make some assumptions about the bulk properties Figure 1.7 shows the valve system which allows rapid backflushing and measurement of C,i The original boiling point column is separated into a short (0.75 m) and a long (5.25 m) section The valve both alters the sequence of these sections, and
Trang 30The analysis of hydrocarbon gases 11
fAlten
Fig 1.6 Boiling point separation Column: 6.0 m X 2 mm id., 28% DC 200/500 on 45/60 mesh
Chromosorb P-AW Temperature: 100°C Carrier gas: helium at 28 ml/min
the direction of carrier gas flow through the short section The column lengths are chosen so that after injection (Fig 1.7a), n-C, will have passed into the longer section before the lightest component has emerged from it Reversing the valve after n-C, has passed this point (Fig 1.7b), a time which is found by trial and error, causes all higher hydrocarbons to recombine and emerge from the short column as a sharp C,+ peak The normal, forward eluted components then follow, as shown in Fig 1.8 After measurement of n-C,, the system is immedi-
ately ready for another analysis Figure 1.7 shows the use of a single 10-port valve for both sample injection and backflushing If preferred, two six-port valves, one for injection, one for backflush, may be used instead
Figure 1.6 illustrates the problem with the wide boiling range of the mixture
C, to n-C, (boiling range 78°C) are well separated in a reasonable time, while the lighter gases are somewhat overlapped and the C, and heavier components are slow to elute and difficult to detect Backflushing of C,+ solves that part of the problem, although introducing uncertainty about composition; the lighter gases need different conditions for good separation Since they emerge rapidly and in a group, it is possible to divert them to a separate column, more suitable for their separation, and then to allow the C, to C, components to emerge and be detected as before A porous polymer bead column will give good separation of
Trang 31Fig 1.7 Accelerated backflush: (a) valve 1 position 1, inject and forward flow; (b) valve 1 posi-
tion I sample load and backflush
these light gases at the same temperature as the boiling point column uses for the C3 to C, separation
Figure 1.9 shows the configuration which achieves this, with valve 2 serving
to include or isolate column 3 , the porous polymer column Restrictor A is ad- justed to give the same pneumatic resistance as column 3 , so that the carrier gas
Trang 32The analysis of hydrocarbon gases I3
I
20 min
Fig 1.8 Boiling point separation with backflush Column 1: 0.75 m X 2 mm i.d Column 2:
5.25 m X 2 mm i.d Both containing 28% DC 200/500 on 45/60 mesh Chromosorb P-AW Tem-
perature: 100°C Carrier gas: helium at 28 ml/min
flow remains constant With column 3 in series, the sample is injected via valve
1 As before, C,+ is backflushed to the detector by returning valve 1 to the load
position As soon as all the C, has passed into column 3 (found by trial and er-
ror), valve 2 is switched to isolate the light gases, N,, CO,, C, and C, in that col-
Trang 3314 Chapter I
Fig 1.10 Boiling point and polymer bead column Column 1: 0.75 m X 2 mm i.d Column 2:
5.25 m X 2 mm i.d Both containing 28% DC 200/500 on 45/60 mesh Chromosorb P-AW Col-
umn 3: 2.4 m x 2 mm i.d 15% DC 200/500 on S O B 0 mesh Hayesep N Temperature: 100°C Car- rier gas: helium at 28 ml/min
umn C,, C, and C, hydrocarbons emerge from columns 2 and 1 to the detector After n-C, has eluted, column 3 is returned on-line by switching valve 2, and the light gases elute and are measured Figure 1.10 shows a typical chromatogram The above configuration, with possible minor variations, is widely used for on-line natural gas analysers, where the sample stream is connected in such a way that the possibility of contamination of the sample by air is minimal How- ever, any sample returned for analysis to a laboratory is prone to air contamina- tion, which means that 0, and N, must be separated if the presence of air is to be recognized, and accurately measured if the air-free composition is to be recalcu- lated 0, may also be present in a transmitted natural gas if air or N, ballasting is used as a means of controlling CV or WI
Porous polymer bead columns will not separate air components at the tem- peratures used for normal analysis; molecular sieves are the only materials able
to do this, and they in turn retain CO, for so long as to make it unmeasurable The solution then is to cut the N, and CH,, with any 0, which may be present, onto a molecular sieve column, leaving the CO, and C, on the porous polymer
Figure 1.1 1 shows the configuration, where the additional valve 3 includes or isolates the molecular sieve
The procedure is similar to the previous one, starting with all columns in se- ries After backflush of C,+, the light gases pass into the polymer bead column,
Trang 34The analysis of hydrocarbon gases 15
Fig 1.1 1 Four column analyser Valve 1, position 2; valve 2, position 1; valve 3, position 1
but now the N, and CH, (and 0, if present) are allowed to go forward to the molecular sieve column Both columns are now isolated, with CO, and C, in column 3, the porous polymer, and 0,, N, and CH, in column 4, the molecular sieve After elution of C,, C, and C, hydrocarbons, column 3 is reconnected for elution and measurement of CO, and C,, and finally column 4 is connected for measurement of 0,, N, and CH, Figure 1.12 is a typical chromatogram
Fig 1.12 Boiling point, polymer bead and molecular sieve Column 1: 0.75 m X 2 mm i.d Col- umn 2: 5.25 m x 2 mm i.d Both containing 28% DC 200/500 on 45/60 mesh Chromosorb P-AW
Column 3: 2.4 m x 2 m m i.d 15% DC 200/500 on 50/80 mesh Hayesep N Column 4:
2.4 m x 2 mm i.d 45/60 mesh Molecular Sieve 5A Temperature: 100°C Carrier gas: helium at
28 ml/min
Trang 3516 Chapter 1
The above chromatograms were generated, as is evident from the conditions,
on a single chromatograph, configured as in Fig 1.1 I The configuration of Fig 1.9 (chromatogram in Fig 1.10) was achieved by isolating column 4 throughout, and that of Fig 1.7 (chromatogram in Fig 1.8) by isolating columns 3 and 4
throughout It would be possible to optimize columns for these configurations which gave somewhat better N,/CH, separation, but it is clear that the baseline separations of all light components seen in Fig 1.12 could not be matched The setting-up procedures described above sound more complicated than they
i n fact are and most instrument suppliers provide ready configured systems Once set up, their performance is usually extremely stable, since none of the coiumns are used anywhere near their temperature limit The molecular sieve column will lose separation gradually due to slow adsorption of moisture, but its performance does not influence other timings or separations within the sys-
tem
1.2.2.2 Two detectors
The chromatogram in Fig 1.12 illustrates the use of two detectors, TCD and FID in series The presence of N, and CO, makes the use of the TCD necessary, and it is sufficiently sensitive for measurement of the lighter hydrocarbons To include pentanes, however, means choosing a larger sample size than would be desirable for linear measurement of the major components Using an FID in series with the TCD avoids this problem, as the FID has much higher sensitiv- ity for hydrocarbons A smaller sample size can be used, satisfying linear detec-
tion requirements for major components and ample sensitivity for minor compo-
n en ts
1.2.2.3 C 6 ~ detail
Separation and measurement of the individual higher hydrocarbons, repre- sented by the C,+ backflushed peak, is needed both to define the composition and hence the properties of this group, and also to provide detail for other calcu- lations such as hydrocarbon dewpoint temperature The complexity of the minor alkane isomer components increases dramatically with carbon number, and this, with the presence of cyclo-alkanes and aromatics, means that high resolution chromatography is required
Capillary columns are widely used for liquid hydrocarbon samples, and are equally adaptable to gas analysis The sample injected from a conventional gas sampling valve can be split without fear of sample discrimination, as there is no phase change on injection Alternatively, the capillary column can be connected directly into a micro-volume gas sampling valve, fitted with a sample loop of some tens of microlitres Chromatograms of natural gases with very different isomer distributions are shown in Figs 1.13 and 1.14 The limit of detection for individual components in this instance is around 5 parts per million molar
Trang 36The analysis of hydrocarbon gases 17
Fig 1.14 Capillary separation Column: 50 m X 0.2 mm i.d capillary coated with OV-101 Tem- perature: 35°C for 5 min, then 6"C/min to 220°C Carrier gas: helium at 1.3 bar
Trang 3718 Chapter I
I 2.2.4 Temperacure programming
Figure 1.6 shows a chromatogram with good resolution and peak shapes for
C, to C, hydrocarbons, insufficient resolution for lighter components and long analysis time and low, broad peaks for C, and heavier components The solution consisted of diverting groups of components onto different columns, chosen to optimize their separation at the selected operating temperature Another ap- proach is to extend the range by programming the temperature of a single col- umn
The boiling point column used for Fig 1.6 would only give good separations
of the light gases by using a low sub-ambient column temperature, and so Stufkens and Bogaard [ 5 ] proposed the use of a porous polymer bead column, which has inherently longer retentions They chose Porapak R, a material with intermediate polarity which optimizes the separation of C,, CO, and C, Using a
-.IO"C to 230°C programme, N,, CO, and C, to C, hydrocarbons are well sepa- rated for individual measurement, and C,, C, and C, hydrocarbons are measured
as groups n-C, elutes at the upper temperature limit for the column, and any heavier components cannot feasibly be measured A TCD and an FID were used
in series, with ethane acting as a bridge component to link the detector re- sponses Methane was measured by difference, although there is no reason why it could not be measured directly The method is intended for CV determi- nation
I
24 min
Fig 1.15 Temperature programmed separation Column: 3 m X 2 mm i.d 50180 mesh Porapak R
Temperature: -50°C for 2 min, then I S W m i n to 240°C Carrier gas: helium at 30 ml/min
Trang 38The analysis of hydrocarbon gases 19
Using this method, but with a sub-ambient (-50°C) start to the temperature programme, O,/N, separation is achieved within the same analysis Figure 1.15
is a typical chromatogram I S 0 6974 [6] is based on this separation, with a starting temperature of 35"C, which eliminates the need for sub-ambient equip- ment, but an additional separation on a molecular sieve column is required for helium, oxygen and nitrogen
1.2.2.5 Combined systems
The capillary chromatogram in Fig 1.13 separates the majority of components
in a natural gas, but not those few lighter components which have the largest concentrations Figure 1.13 has been optimized for good detection limits over a wide range, giving quantitative measurement from C, to C,, and beyond Used in conjunction with the column switching system illustrated in Fig 1.1 1, either C,
If the capillary conditions are optimized in order to include quantitative meas- urement of C,, then the packed column TCD system for light gases can be sim- plified to a porous polymer/molecular sieve combination Both this and the capil- lary/FID system can be fitted into the same chromatograph [7] The sample is
injected first into the packed column system for isothermal separation of 0,, N,,
C,, CO, and C,, and then into the capillary for temperature programmed separa- tion of the higher hydrocarbons There is no backflush provision to remove the heavier hydrocarbons from the packed columns, as they are forward eluted (but not measured) during the temperature programmed part of the cycle Figure 1.16 illustrates the column arrangement and Fig 1.17 a typical chromatogram
The analysis is comprehensive, but there is no provision for a bridge compo- nent C, and C, can be distinguished on the capillary column, but it is doubtful whether C, could be measured with sufficient accuracy Although the analysis takes 30 min, the cycle time, allowing for cool down and re-stabilization, is longer In a similar application, three sample aliquots are injected, one onto mo- lecular sieve for O,/N, separation, and from which C, and above are backflushed
to vent, one onto porous polymer for (0, + N,), C,, CO, and C,, and the third onto a capillary for temperature programmed separation [ 81
In a novel application intended for process use, the natural gas is injected onto
a single, long packed column containing porous polymer and operating with a large pressure drop N,, C, and CO, are measured normally, then the entire col- umn is backflushed to the detector for measurement of the other components To understand how this works requires careful consideration of the mechanism of backflushing in gas chromatography
Backflushing is assumed to recombine separated or partially separated com- ponents because they have to travel equally far in the reverse direction from that
Trang 40The analysis of hydrocarbon gases 21
in which they have been undergoing separation This is an over-simplification In gas chromatography, the mobile phase is compressible, and is driven through the column under the influence of a pressure gradient Since each section of the col- umn has the same mass flow of mobile phase passing through it at any time, the actual volumetric flow at each point must vary according to the pressure at that point As the pressure decays from the beginning to the end of the column, so the
volumetric flow rate increases, and so does the linear velocity
The mobile phase linear velocity controls the rate of travel of components, and so if this increases along the length of the column, components, even in iso- thermal analysis, will accelerate between injection and detection Looked at an- other way, a component with a retention time of x will not have reached the mid point of the column at an elapsed time of x/2 The larger the pressure gradient, the greater the acceleration
When the whole column is backflushed, the pressure gradient which exists in forward flow is reversed, and so those parts of the column where the components travelled most slowly become the high velocity areas, and vice versa A compo- nent which has nearly reached the end of the column at the time of backflushing will have experienced the full acceleration It now finds itself in the high pres- sure, low velocity region, and has to do it all over again Its time to elute in backflush will be almost identical to its forward flow time By contrast, a com- ponent which had travelled only a short distance from the injection point by the time of backflushing will have spent all the forward flow time in the low veloc- ity region After backfl ushing, it is in the high velocity region, and will be eluted
as a backflushed peak in a much shorter time than it spent travelling forward Hence there is a mechanism of separation, in reversed order, during back- flushing The fact that this is not evident when backflushing in applications such
as that illustrated in Figs 1.10 and 1.12 is because in these, the backflushed sec- tion is short Hence, while it is subject to high pressure, low velocity on forward flow, and vice versa on backflush, the pressure drop across the section itself is small, and backflush separation does not occur Figure 1.18 is a chromatogram showing the application [9] The advantages of this approach are that it requires
less hardware (a single ten-port valve would cover both injection and backflush- ing) and the column can be long enough to give good N,, CH,, CO, separation without a consequent time penalty for the heavier components The main disad- vantage is to do with the nominal C,+ group Since the backflush separation is good enough to separate C, from C, and C, from C,, it is almost certainly per- forming some separation of C,, C, and C, This being so, the C,+ group, which is
a sharp, distinct peak in Fig 1.11, becomes relatively “smeared out” and more difficult to quantify
I 2.2.7 Rapid analysis
Process natural gas analysers have analysis cycle times typically in the range