ISO/DIS 6974-4:2025(en)

ISO/TC 193/SC 1

Secretariat: NEN

Date: 2024-12-13

Natural Gas — Determination of composition and associated uncertainty by gas chromatography — Part 4: Guidance on gas analysis

Gaz naturel — Détermination de la composition et de l'incertitude associée par chromatographie en phase gazeuse — Partie 4: Conseils sur l'analyse des gaz

© ISO 2025

All rights reserved. Unless otherwise specified, or required in the context of its implementation, no part of this publication may be reproduced or utilized otherwise in any form or by any means, electronic or mechanical, including photocopying, or posting on the internet or an intranet, without prior written permission. Permission can be requested from either ISO at the address below or ISO’s member body in the country of the requester.

ISO copyright office

CP 401 • Ch. de Blandonnet 8

CH-1214 Vernier, Geneva

Phone: +41 22 749 01 11

Email: copyright@iso.org

Website: www.iso.org

Published in Switzerland

Contents Page

1 Scope 1

2 Normative references 1

3 Terms and definitions 1

4 Symbols 1

5 Overview 2

6 Sample 3

6.1 General 3

6.2 Gas origin 3

6.3 Phase of sample 4

6.4 Sample pressure 4

6.5 Sampling 4

7 Sample introduction 5

7.1 General 5

7.2 Sample loop 5

7.2.1 General 5

7.2.2 The temperature of the sample loop 5

7.2.3 The pressure within the loop 5

7.2.4 Sample shut-off before injection 6

7.2.5 The atmospheric pressure effect 6

7.3 Injection 6

7.4 Vacuum injection 7

8 Separation 8

8.1 General 8

8.2 Columns 9

8.3 Carrier gas 9

8.3.1 Types of gases 9

8.3.2 Carrier gas flowrate 9

8.3.3 Purity of the carrier and auxiliary gas 10

8.4 Temperature 10

8.5 Separation columns 11

8.6 Back-Flush 11

8.7 Maintenance related to column performance 12

8.8 Environmental conditions 12

8.9 General setup 12

8.10 Correction for the presence of oxygen and argon 13

8.10.1 General 13

8.10.2 Gas containing oxygen. 13

8.10.3 Gas containing argon 14

8.10.4 Air contamination correction for natural gas spot samples 14

8.10.5 Correction when the amount of argon has been determined 14

8.10.6 Correction when the amount of argon has not been determined 15

9 Detection 16

9.1 Typical detectors for natural gas analysis 16

9.2 Peak resolution 17

9.3 Detector 20

10 Data processing 20

10.1 Data 20

10.1.1 General 20

10.1.2 Conversion 21

10.1.3 Allocation or peak identification 21

10.1.4 Data file format 21

10.2 Peak integration 22

10.2.1 General 22

10.2.2 Principle 22

10.3 Chromatogram 23

10.3.1 General 23

10.3.2 File 23

10.3.3 A/D Conversion 23

10.3.4 Sampling frequency 23

11 Calibration 23

12 Optimization 23

12.1 General 23

12.2 Method 23

12.3 Repeatability 24

13 Precision and bias 24

14 Use of control charts (from ISO 6975:1997) 24

15 Test Report 27

Annex A (informative) Determination of hydrogen, helium, oxygen, nitrogen, carbon dioxide and hydrocarbons up to C8 using two packed columns 28

A.1 Application Ranges 28

A.2 Principle 28

A.3 Materials 29

A.3.1 For the determination of helium, hydrogen, oxygen and nitrogen 29

A.3.2 For the determination of nitrogen, carbon dioxide and hydrocarbons from C₁ to C₈ (separation on Porapak column), 29

A.4 Apparatus 30

A.4.1 Laboratory gas chromatographic (GC) system 30

A.5 Procedure 32

A.5.1 Gas chromatographic operating conditions 32

A.5.2 Performance requirements 34

A.5.3 Determination 34

A.6 Example: Single-oven gas-chromatographic system consisting of two columns 35

Annex B (informative) Determination of nitrogen, carbon dioxide and C1 to C5 and C6+ hydrocarbons for a laboratory and on-line measuring system using two columns 39

B.1 Application ranges 39

B.2 Principle 39

B.3 Materials 40

B.3.1 Helium carrier gas, 40

B.3.2 Working-reference gas mixtures (WRM), 40

B.4 Apparatus 40

B.4.1 Laboratory gas chromatographic (GC) system, 40

B.5 Procedure 41

B.5.1 Gas chromatographic operating conditions 41

B.5.2 Performance requirements — Peak resolution 45

B.5.3 Determination — Outline of the analysis 46

B.6 Expression of results 46

B.6.1 Calculation 46

B.6.2 Precision and accuracy 46

B.7 Procedure for setting valve timings and restriction setting 46

B.8 Final time settings 47

Annex C (informative) Isothermal method for nitrogen, carbon dioxide, C1 to C5 hydrocarbons and C6+ 48

C.1 Application ranges 48

C.2 Principle 48

C.3 Materials 49

C.3.1 Carrier gas, 49

C.3.2 Auxiliary gases, 49

C.3.3 Reference materials 49

C.3.4 Reference gases, 49

C.3.5 Gas mixture containing n-Pentane and 2,2-Di-Me-butane, 50

C.4 Apparatus 50

C.4.1 Gas chromatograph, 50

C.4.2 Column oven, 50

C.4.3 Valve oven, 50

C.4.4 Pressure regulator, 50

C.4.5 Injection device, 50

C.4.6 Backflush valve, 50

C.4.7 Column isolation valve, 50

C.4.8 Columns, 50

C.4.9 Tube and packing. 50

C.4.10 Method of packing, 52

C.4.11 Thermal Conductivity Detector (TCD), 52

C.4.12 Controller/Peak Measurement System, 52

C.4.13 Auxiliary valves, tubing and other accessories, 52

C.5 Scheme of the configuration 52

C.6 Procedure 54

C.6.1 Control of the apparatus 54

C.6.2 Column Conditioning 55

C.6.3 Operation of the apparatus 55

C.7 Expression of results 58

C.7.1 Uncertainty 58

C.8 Example of application 58

C.8.1 General considerations 58

C.8.2 Calculation of mole fractions 61

C.8.3 Calculation of uncertainties of mole fractions 68

C.8.4 Comparison of mean normalization and run-by-run approaches 69

C.8.5 Reporting of results 69

C.8.6 Excel spreadsheet 69

C.9 Procedure for Setting Valve timings and Restrictor Setting 70

C.9.1 Initial Flow Settings 70

C.9.2 Backflushing 70

C.9.3 V3 Timing 70

C.9.4 Final timings 71

Annex D (informative) Determination of hydrogen, helium, oxygen, nitrogen, carbon dioxide and C₁ to C₈ hydrocarbons using three capillary columns 72

D.1 Application ranges 72

D.2 Principle 74

D.2.1 Analysis of natural gas sample 74

D.2.2 Auxiliary gases 74

D.3 Materials 74

D.3.1 Carrier gases 74

D.3.1.1 Argon (Ar) 74

D.3.1.2 Nitrogen (N2) 74

D.3.1.3 Helium (He) 74

D.3.2 Auxiliary gases 74

D.3.2.1 For FID detection: 74

D.3.2.1.1 Nitrogen (N₂) 74

D.3.2.1.2 Air 74

D.3.2.1.3 Hydrogen (H₂) 75

D.3.2.2 For methanizer 75

D.3.2.2.1 Hydrogen, 75

D.3.2.2.2 Pressurized laboratory air 75

D.3.3 Reference materials 75

D.3.3.1 Working reference gas mixture (WRM) 75

D.3.3.2 Performance test gases. 75

D.3.3.2.1 For methanizer operation 75

D.3.3.2.2 Gas containing benzene and cyclohexane 75

D.3.3.2.3 Gas containing hydrogen and helium 76

D.4 Apparatus 76

D.4.1 Gas chromatograph system(s) 76

D.4.1.1 Two column ovens 76

D.4.1.1.1 Instrument 1 oven 76

D.4.1.1.2 Instrument 2 oven, 76

D.4.1.2 Flow regulators 78

D.4.1.3 Gas sampling valves (GSV) 78

D.4.1.4 Valveless or micro-valve column-switching system 78

D.4.1.5 Thermal conductivity detector (TCD) and flame ionization detector (FID) 78

D.4.1.5.1 Instrument 1 detectors 78

D.4.1.5.2 Instrument 2 detector 78

D.4.1.6 Data acquisition system 78

D.4.1.7 Methanizer 78

D.4.2 Capillary columns 79

D.4.2.1 PLOT fused silica capillary precolumn, 79

D.4.2.2 Molecular sieve PLOT fused silica capillary column 79

D.4.2.3 Non-polar WCOT fused silica capillary column, 79

D.5 Procedure 81

D.5.1 Operating conditions 81

D.5.1.1 Gas chromatograph 81

D.5.1.2 Column conditioning 81

D.5.1.3 Sample introduction 82

D.5.2 Performance requirements 83

D.5.2.1 Column performance evaluation 83

D.5.2.2 Relative response factors 83

D.5.2.3 Response characteristics 83

D.5.3 Determination 84

D.5.3.1 Components He, H2, O2, N2, CH4, CO, CO2, C2H2, C2H4 and C2H6 84

D.5.3.2 Hydrocarbons and higher 84

D.5.3.3 Detection 84

D.5.3.3.1 TCD 84

D.5.3.3.2 FID 84

D.5.3.3.3 Data acquisition 85

D.6 Calculation 88

Annex E (informative) Natural gas -Extended analysis - Gas-Chromatographic method 89

E.1 Introduction 89

E.2 Scope 89

E.3 Definitions 90

E.3.1 Resolution 90

E.3.2 Main components 91

E.3.3 Associated components 91

E.3.4 Trace components 91

E.3.5 Other components 91

E.3.6 Response 91

E.3.7 Reference component 91

E.3.8 Relative response factor (for an FID) 92

E.3.9 Concentration of a group of components 92

E.4 Principle 92

E.5 Analysis and analytical requirements 93

E.5.1 Apparatus and materials 93

E.5.1.1 Analytical system 93

E.5.1.2 Reference gas mixtures 93

E.5.1.2.1 Certified-reference gas mixtures (CRMs) 93

E.5.1.2.2 Working-reference gas mixtures (WRMsl 93

E.5.1.2.3 Control gas 94

E.5.2 Structure of the analysis 94

E.5.3 Resolution 95

E.5.3.1 Main Components 95

E.5.3.2 Trace components 95

E.5.3.3 Associated components 95

E.6 Procedures 95

E.6.1 Setting up the analytical system 96

E.6.2 lnjection 96

E.7 Calculations 96

E.8 Annex A (informative) Determination of the response curves for the main components in a sample 96

E.9 Annex B (informative) Further details on the matrices used in annex A 96

E.10 Annex C (informative) List of retention indices 96

Annex F (informative) Natural gas -Extended analysis - Gas-Chromatographic method 99

F.1 Introduction 99

F.2 Measurement of C6+ peak 99

Annex G (informative) GPA 2286-95 101

Bibliography 102

Foreword

ISO (the International Organization for Standardization) is a worldwide federation of national standards bodies (ISO member bodies). The work of preparing International Standards is normally carried out through ISO technical committees. Each member body interested in a subject for which a technical committee has been established has the right to be represented on that committee. International organizations, governmental and non-governmental, in liaison with ISO, also take part in the work. ISO collaborates closely with the International Electrotechnical Commission (IEC) on all matters of electrotechnical standardization.

The procedures used to develop this document and those intended for its further maintenance are described in the ISO/IEC Directives, Part 1. In particular the different approval criteria needed for the different types of ISO documents should be noted. This document was drafted in accordance with the editorial rules of the ISO/IEC Directives, Part 2 (see www.iso.org/directives).

Attention is drawn to the possibility that some of the elements of this document may be the subject of patent rights. ISO shall not be held responsible for identifying any or all such patent rights. Details of any patent rights identified during the development of the document will be in the Introduction and/or on the ISO list of patent declarations received (see www.iso.org/patents).

Any trade name used in this document is information given for the convenience of users and does not constitute an endorsement.

For an explanation on the voluntary nature of standards, the meaning of ISO specific terms and expressions related to conformity assessment, as well as information about ISO's adherence to the World Trade Organization (WTO) principles in the Technical Barriers to Trade (TBT) see the following URL: www.iso.org/iso/foreword.html.

This document was prepared by Technical Committee ISO/TC 193, Natural gas, Subcommittee SC 1, Analysis of natural gas.

This edition cancels and replaces the previous editions of ISO 6974-3:2000, ISO 6974-4:2000, ISO 6974-5:2014, ISO 6974-6:2002 and ISO 6975:1997, which in part have been technically revised.

The main changes compared to the previous edition(s) are extensive, as this document is the compilation of the formentioned documents, added to a collection of knowledge on natural gas chromatography.

A list of all parts in the ISO 6974 series can be found on the ISO website.

Any feedback or questions on this document should be directed to the user’s national standards body. A complete listing of these bodies can be found at www.iso.org/members.html.

Introduction

The composition of natural gasses varies immensely, and the addition of biogas, biomethane, hydrogen, syngas and other natural gas substitutes only adds to chromatography spectrum. The gas chromatographic system should be designed not only to separate the components of economic (short-term) interest, but also for components considered as trace or of no-interest. These trace components could be potentially problematic for public health, safety or assets.

A precise and stable analysis of the main and trace components of the gas can be obtained by an analyser that is fit-for-purpose. Satisfactory performance of a natural gas analyser requires that the method has good precision and response characteristics which allow component concentrations over the range of interest to be accurately compared with the equivalent components in the reference mixture (calibration).

This document elaborates on all steps involved with gas chromatography, and is comprised of extracts from older parts of this standard, and other standards. Added to the extracts is guidance, originating from common sense and experience, which may help the user and manufacturer to ascertain that their analyser is fit-for-purpose and that optimum analysis results are obtained from the analyser.

Natural Gas — Determination of composition and associated uncertainty by gas chromatography — Part 4: Guidance on gas analysis

This document gives guidance for obtaining the best analysis results possible from a gas chromatograph (GC) when analysing natural gas and natural gas substitutes for combined use with the most recent versions of ISO 6974’s part 1, 2 and 3 (examples are given).

The following documents are referred to in the text in such a way that some or all of their content constitutes requirements of this document. For dated references, only the edition cited applies. For undated references, the latest edition of the referenced document (including any amendments) applies.

ISO 6974‑1, Natural gas — Determination of composition and associated uncertainty by gas chromatography — Part 1: General guidelines and calculation of composition

ISO 6974‑2, Natural gas — Determination of composition and associated uncertainty by gas chromatography — Part 2: Uncertainty calculations

ISO 6974‑3:2018, Natural gas — Determination of composition and associated uncertainty by gas chromatography — Part 3: Precision and bias

ISO 7504:2015, Gas analysis — Vocabulary

ISO 10715:2022, Natural gas — Gas sampling

ISO 14532:2014, Natural gas — Vocabulary

ISO 23219:2022, Natural gas — Format for data from gas chromatograph analysers for natural gas — XML file format

For the purposes of this document, the terms and definition used are taken from ISO 14532 and ISO 7504.

ISO and IEC maintain terminological databases for use in standardization at the following addresses:

— ISO Online browsing platform: available at https://www.iso.org/obp

— IEC Electropedia: available at https://www.electropedia.org/

The table below, Table 4‑1, lists the symbols used in the main text of this document.

Table 4‑1 — Symbols

This guidance document covers a wide range of subjects in relation to the analysis of natural gas.

Below, in Figure 1, a schematic overview is given, the main clauses are indicated according to the process-sequence.

Figure 1 — Gas Analysis Overview

All gas analysis starts with a sample. The origin and physical state of the sample are of importance to the subsequent steps in the analytical process. This clause describes the most common concerns in the first step of that process. More detailed methods of sampling can be found in ISO10715.

A good understanding of the gas production process and the consequences for possible trace components should be consulted for. Although a periodic in-depth investigation of these components seems like a wise way of operating, it has the innate possibility to miss the (near) out-of-spec occurrences.

A few examples from the field of biomethane:

— Periodic replacement of activated carbon filters is a clear and common way of operating, but the flooding of a filter is primarily caused by contaminant concentration times flow, and not by time alone. A small investment in an extra separation column or back-flush with detector option provides up-to-date information of the contaminant-level and leads to better prediction of the filter-pack efficiency and proper moment of replacement.

— Biomethane usually contains nitrogen, which, more often than not, originates from the air added to the fermentation process to decrease the production of hydrogen sulfide and other sulphur compounds. When analyzing for oxygen in the produced biomethane, argon should be sufficiently separated from oxygen because argon will be present in a 0,934 0 to 78,084 ratio to nitrogen, whereas the oxygen might be lower in concentration, because it is used in the process. For instance, at 5 mol% nitrogen this adds up to about 600 ppm argon.

— Water should be identified in the chromatogram to determine co-elution or even calibrated for, because a high level (> 50 ppm) is quite common and could get as high as 1 vol%. Measurement of the water concentration with a water dewpoint analyser provides a way to prevent a mismeasurement.

Syngas or other gasses from synthetic sources may contain unsaturated hydrocarbons or other volatile components that are low in concentration individually, but could amount to be significant when lumped together, either in total concentration or in calorific content.

Natural gas has a wide range of concentration for all occurring components. In some cases, the gas has quite a high amount of higher hydrocarbons and aromatic compounds, but could also be almost pure methane. This wide range of component concentrations has implications for the set-up of the gas analyser (and will be detailed further on in this document).

For most applications the sample phase will be gaseous (with low water dewpoint and low hydrocarbon dewpoint, i.e. dry gas), but it may very well be that the sample has a high hydrocarbon or water dewpoint or is liquified.

This document will not treat samples that are liquid, multiphase or have a propensity towards multiphase behavior, and for this document it is assumed that the introduced sample is entirely gaseous. There are various solutions available for such samples, ranging from heated sample container cabinets and heat-traced sample lines up to very advanced evaporation systems.

Although pressure (and temperature) of the sample is closely related to phase behavior of the gas, it also is of concern for the sampling part of the analyser:

— If the sample is at ambient, sub-atmospheric or high pressure it needs to be decided if the sample is to be compressed, pumped or reduced for sampling.

— The gas wetted valves and tubing used need to be assessed accordingly.

— The sample loop pressure needs to be able to equilibrate within the time between filling and injection.

— The sample flow needs to be enough to properly purge the sample line, from the sample container to the sample loop, so a representative sample is taken.

— The sample pressure and flow should not overload the vent system (e.g. of the lab or of other analyzers).

These items are treated in more detail further in this document.

The sampling of the gas of interest at its source is not treated in this document but is described extensively in ISO 10715, along with guidelines on the materials used for the sample container and the sample line.

In relation to the configuration of the gas analyser mainly two types of sampling occur:

— Single, from a sample container

— Continuous, from e.g. a gas supply line or process

Being closely related to the next step, sample introduction, this is treated in more detail in the next clause.

A consistent means of sample introduction is required so that equivalent amounts of sample and reference mixture are compared.

In most applications, a switching valve is used for sample introduction. The sample is purged through a loop. This defines the sample size and when the valve operates. The loop is switched into the carrier gas stream. The loop contents are swept by the carrier gas onto the separation system, see Figure 2. An alternative for micro-GCs uses pressure switching for a selected short time to achieve the effect. There are some viscosity influences, and it is not considered here.

The sample loop has a defined volume, but the size of injected sample is influenced by the temperature and pressure of the contained gas. The sample loop is sufficiently flushed until the volume of sample or reference gas has fully purged the previous loop contents.

We therefore need to consider the following.

The sample loop can be controlled separately or fitted in the column oven. The size is not usually critical, but the stability is. Temperature variations will cause different effective sample amounts and hence poorer repeatability.

Higher temperature stability can be achieved by adding masses of metal around the sample-loop and sample-valve. Combination with insulation and shielding the analyzer from airflows (e.g. air-conditioning vents and analyser exhausts) will further improve stability.

The sample line usually purges to atmosphere (possibly through a vent line extending outside the laboratory or analyzer housing). If the sample flowrate and the resistance of the vent line create a backpressure, the pressure in the sample loop will vary with flowrate. Usually the flows of calibration gas and sample gas are controlled by different means and may be set to different values, so there may be a difference in effective sample size between the two and hence the possibility of bias error.

When using pressure correction it is advisable to connect the (high-accuracy) barometer as closely as possible to the sample-loop, i.e. connect the barometer directly to the exit of sample shut-off valve outlet. This prevents measuring the fluctuating under- or overpressure of the laboratory.

Also, be aware of the possibility of automatic pressure correction, installed by the manufacturer of the GC. It should either be disabled, removed or have the capability to be externally calibrated.

This is a common procedure. Stopping the sample (or calibration gas) flow a few seconds before injection allows the loop pressure to decay to atmospheric. As a result, it will only vary according to atmospheric pressure variations, which can be significant.

The optimum sample shut-off time is influenced by the stability (or deterioration) of the sample. This is caused by sample composition, the sample-loop material and sample-loop temperature.

The atmospheric pressure effect can be measured at the exact time of injection and corrected for or ignored. If ignored and the method is a single operation one, normalization will correct it, since sample loop pressure has the same relative influence on all components. In fact, for single operation, atmospheric pressure correction followed by normalization (pressure correction alone never sets results to 100 %) gives the same result as normalization alone.

However, in multipoint calibration no component has the same calibration function. As a consequence, any pressure different from standard will give a slightly different correction on the raw analyzer output compared to the other components. This effect occurs both during calibration and analysis, so for multipoint calibration and subsequent analysis, pressure correction should be applied.

When only a small sample volume is available, or the GC is not equipped to handle atmospheric samples, a direct injection (e.g. by syringe) on the GC injector is a good alternative.

Most GC are equipped with a split/splitless injector, a small heated chamber that either flushes the sample on to the column (splitless mode) or flushes a portion of the sample into the column (split mode).

In split mode, a part of the mixture of sample and carrier gas in the injection chamber is exhausted through the split vent. Split injection is preferred when working with samples with high analyte concentrations.

Splitless injection, see Figure 2, is best suited for trace analysis with low amounts of analytes. In splitless mode the split valve opens for a pre-set amount of time to flush the sample on the separating column. This pre-set time should be optimized, a shorter time ensures less tailing but loss in response, a longer time increases tailing but also signal.

Figure 2 — Operation of a simple GC configuration using a switching valve; (a) sampling and (b) injection

Different types of Injection can be specified:

— Manual injection: if you don’t have to analyze too many samples, manual injection can be used. Be aware that the injection should be repeatable in order to have repeatable results.

— Auto-sampling: in case of many samples, an auto-sampler can be a good option and also to have more repeatable injections and results. For gas samples, the auto-sampler could be more complicated than for liquid samples.

— Sample-loop: the capacity of the sample loop should be chosen based on the detection limit wanted and on the column dimension. The split flow also plays a role in this case.

— Time-based valve injection: this kind of valve can be used for a better repeatability of the injection and so, for the analysis

Injection using a Deans type switch. This is an effluent switching device based on controlling flows of carrier gas instead of mechanical valves in the analytical flow path. This technique offers high inertness and a wear-free operation.

In process GCs a combination of an autosampler with a time-based valve injection or Deans type switch is most often seen.

If the sample pressure is very low, or the amount of sample available is very small it may be necessary to first evacuate the sampling system prior to injection. By first removing any previous sample, air or carrier gas from the sample loop using a vacuum pump prior to introducing the sample will greatly reduce the amount of sample required to get a representative sample of the gas. See Figure 3 for schematic overview.

Figure 3 — Vacuum injection

Procedure:

1. Connect the sample, vacuum pump and pressure gauge as per the above figure, with all three valves closed.

2. Evacuate the sampling system by opening valve V1.

3. Once a suitable pressure (vacuum) is achieved (< 10 mbar) isolate the vacuum pump using valve V1.

4. Monitor the pressure indicator to ensure that ambient air is not leaking into the system, if the pressure is steadily increasing, check all the fittings and return to step 2.

5. The sample should then be slowly introduced to the system; valve V1 should be opened first then valve V2 can be slowly opened and used to control the flow. Once the pressure within the system reaches atmospheric pressure, close valves V2 and V3.

6. The sample can then be injected.

7. Repeat steps 2-6 until the required number of injections is achieved.

If there is sufficient sample it may be beneficial to omit step 6 after the first evacuation and re-pressurization cycle to ensure the sampling system is fully purged before injecting.

A separation system, consisting of one or more chromatographic columns, which allows consistent retention times and adequately separates all components of interest, is required to obtain reliable results.

This involves at least one column, usually more, in a temperature-controlled oven with a supply of carrier gas controlled either by pressure or by flow. Where multiple columns are used, the configuration is changed through the analytical cycle. This can be accomplished either by switching valves or by a Deans type pressure control. Either system uses timed events within the analyzer controller to create the changes.

To obtain optimum separation and reliable results the user has the possibility to select from a wide range of separation options, and with a wide range of combinations it is possible to have an analysis that produces the components of interest at adequate accuracy or analyses to give results as detailed as possible. The following sub-clauses highlight some of the system parts available for a gas analysis system configuration.

The gas chromatographic columns are a vital part of the analyser and can be chosen according to specifications:

— type of column (packed, capillary)

— stationary phase specification:

— stationary phase thickness (capillary column)

— characterization of solid phase (packed column)

— column dimensions

All these have an influence on the separation, retention time (speed of analysis) and sensitivity.

The carrier gas is essential to the gas chromatographic separation of the injected sample. The choice of carrier gas depends on the analytes that are to be separated. In multi-column setup, different carrier gasses can be selected in order to be able to analyze the complementary components, i.e. use helium as carrier gas to analyze for nitrogen and use nitrogen as carrier gas to analyze for helium.

Gases, supplied from cylinder or generator, used in GC analyses are divided by their use for the separation and detection. The most common carrier gases are:

— Helium

— Hydrogen

— Argon

— Nitrogen

— Air

The choice of gas can depend on the available gas supply, the detection method and safety concerns.

Component retention times are a direct function of carrier gas flowrate, and so flow variations will have a possible effect on peak identification. With the thermal conductivity detector, which is concentration sensitive, increased flowrate will give smaller peak areas (but similar peak heights) and vice versa. With the flame ionization detector, which is mass sensitive, flowrate change will have a secondary effect on peak area measurement due to the fact that detector sensitivity varies with the carrier gas/hydrogen ratio.

Most common configurations of the carrier gas flowrate are:

— Constant velocity

— Constant pressure

— Programmed

The purity of the carrier and auxiliary gas has an influence on the (long-term) performance of a separation column:

— A carrier gas purity shall be at least 99,99 % (4,0), but 99,999 % (5,0) or even 99,999 9 % (6,0) purity is preferred. The addition of a carrier gas filter can even further increase the purity of the carrier gas and, more importantly, shield the separation column from impurities like water, oxygen and hydrocarbons, inherent to the (lack of) purity or originating for instance from a carrier gas cylinder change. Be aware that the carrier gas filter should be preferably ordered, pre-filled or thoroughly rinsed (before use) with the same gas as the carrier gas to prevent transient effects. Maintenance of these filters is important and most filters have indicators to show when the filter needs replacement. Upon installation all connections should be checked for leakage.

— Carrier and auxiliary gases can be supplied from high pressure cylinders, compressors (air) or from generators. Especially for the use of hydrogen, a hydrogen generator can be a safer option. Generators for hydrogen, nitrogen and air can produce high purity gases from deionized water (hydrogen) or compressed air (nitrogen and air).

— A column that has been set-up for separation of “lighter” components will accumulate “heavier” components, either present in the sample-gas or in the carrier gas, especially when operating in isothermal mode. Periodically increasing the temperature to the maximum allowable temperature for some time is a good way to remove accumulated components, restoring the separating properties of the column, and should be part of the maintenance procedure.

— Components that are separated on the column but present in the carrier gas could appear as negative peaks, or ghost peaks in the next chromatogram. A series of blank runs, both without injection and with injection of the carrier gas, can provide useful insight to the baseline performance of the analyser and the actual purity of the carrier gas.

NOTE In this case, attention should be paid to the fact that some columns can be damaged if they are run without carrier gas

— For all supply lines for carrier gasses, fuel gasses and support gasses, impermeable tubing like copper or stainless steel, should be used to prevent impurities for permeating in or out. Special attention in this regard should be given to the use of hydrogen.

Component retention times remain consistent if the column performance does not vary, and the oven temperature and carrier gas flowrate remain constant. This applies to both isothermal and temperature programmed operation, see Figure 4, to each stage of the program in the latter case. Exact accuracy of the oven temperature is not critical, but consistency is. Changes in column temperature will have little or no effect on the peak size, but will cause alterations in retention times, with possible effects on peak identification.

The oven temperature can be set in different configurations:

— Isothermal; isothermal temperature programming is used generally to have more repeatable results and when the components are not too heavy and so the analysis is not too long. See top.

— Step-Programmed; programmed temperature generally is used when the components to be analyzed have a very wide, but clustered range of boiling points and to shorten the analysis total time in case of heavy compounds. In the temperature transition the baseline is more unstable and is less suited for separation purposes.

— Ramping; a continue rising temperature program, see Figure 5. This type of programming is used when a sample contains a significant amount of higher boiling components. See bottom part of Figure 4.

Figure 4 — Isothermal vs temperature ramp

Figure 5 — Temperature programming

The separation column is the functional part of the gas chromatograph. Through interactions between the individual compounds in the sample and the stationary phase within the column separation occurs. Several types can be discerned:

— Packed or capillary phase. Packed columns are more common in process GCs or for the separation of bigger quantities of gas. Capillary columns give a better separation of the components since have a higher HETP but can cope with lower quantities of gas. Capillary columns can accept lower quantities of sample, so, a split flow should be regulated in order not to saturate the column.

— Stationary phase type. The choice of the stationary phase of a column should be based on the components of the mixture that are to be separated. It can be chosen to separate by polarity, functional groups, boiling point, etc.

— Dimensions of the column, diameter and length. It should be chosen based on what is the wanted separation(resolution) of the components of the mixture, and taken into account the quantities or the detection limit that should be reached.

Back-Flush is a useful part of a chromatographic system and consists of a short column of which the flow after a determined time is reversed. Key advances are:

— Protection of the separation column against slow-eluting components.

— Maintaining a more stable baseline, by preventing ghost peaks from the previous injection.

— Gives an indication that the gas contains components that are beyond the GC’s designed range but are in the gas nevertheless, if equipped with a detector.

It might be therefore prudent to equip every column with a Back-Flush, even the columns used for the “heavier” components.

An in-depth perspective on the use of the backflush technique on the C6+ components is given in Annex F.

A column that has been set-up for separation of “lighter” components will accumulate “heavier” components, either present in the sample-gas or in the carrier gas, especially when operating in isothermal mode, during an analysis period. Periodically increasing the temperature (and carrier flow) to the maximum allowable temperature for some time is a good way to remove accumulated components and restore the separating properties of the column, and should be part of the maintenance procedure, either during a non-analysis or a maintenance period.

The available GC-columns typically used, have very stable characteristics. The exception is the Molecular Sieve type, which is only used for natural gas if oxygen/argon/nitrogen separation is required. Molecular Sieve (aluminum oxide) is a desiccant and adsorbs water at trace levels in carrier and sample gases. This reduces its activity and causes retention times to reduce with consequential loss of resolution. If a molecular sieve column is used in the system, this column should be conditioned at high temperature (above 200 °C) for some hours (see the manufacturer’s instructions) every 5-6 months or every time the resolution between peaks decreases.

The environmental conditions surrounding the GC influence its results and stability. This influence maybe gradual of nature or more abrupt. Examples are:

— Ambient pressure. When a GC sample loop is filled with the gas to be analyzed, the pre-injection stabilization of the sample loop pressure to ambient pressure causes the injected amount to be proportional to the ambient pressure. If the GC does not use pressure compensation for the sample loop pressure, the results are to be corrected for the pressure measured in the vent line as close to the sample loop as possible with a high-accuracy barometer, at the moment of injection.

— Temperature. Although fluctuations in the normal temperature range will influence the operation of a GC, the sample too is influenced by the surrounding temperature. A temperature conditioned room or cabinet will increase the stability of the produced analytical results. The reduction of the sample pressure to the injection pressure should also take place in that room or cabinet.

— Pressure variations. Detectors like, for instance, FID or TCD are also influenced by pressure pulses (caused, for instance, by the opening and closing of a door to the analyser cabinet or to a laboratory operating in under- or over-pressure). These pressure peaks cause false peaks in the chromatogram, which may lead prematurely starting or ending the analytes’ peak, resulting in a misreading.

— Unless the humidity is beyond the permissible range, stated by the manufacturer, no influence is to be expected from the ambient humidity.

The configuration of the GC determines the range of possibilities a GC has. A proper setup gives the user enough options to optimize the analysis method for varying compositions. Most seen option are:

— Single or multiple columns: the choice for a single or multiple columns depends on the analysis to be performed. If the mixture to be analyzed is complex with different kinds of compounds (e.g. hydrocarbons and permanent gases and sulfur compounds), it could be advisable to use multiple columns in order to have good separation and short times of analysis.

— Single or multiple carriers: usually it is more practical to use a single carrier, but in some cases where the compound to be analyzed has the same or very similar properties as the carrier gas, it is advisable to use multiple carriers (e.g. hydrogen determination with helium carrier)

— With or without back-flush: The back-flush is useful in combination with a pre-column, when there are compounds in the mixture that have longer retention times in the column and there is no interest in them, so to clean the column and have a shorter time for the analysis, back-flush can be a good option. Back-flush can be used also for the determination of heavier compounds that are useful as one peak (as for example, the determination of C6+ in natural gas analysis)

— With or without pre-column: a pre-column can be used in combination with back-flush to do a pre-separation of a family of compounds in order to shorten the time of analysis, but also to protect the column against contamination.

Normally, natural gas does not contain oxygen or argon. If, however, natural gas samples are found to contain oxygen or argon, this may be due to contamination by improper sampling of the gas. In such a case, the mole fraction of nitrogen and all other components shall be corrected according to the following procedures.

NOTE A historic benchmark for deciding how to correct for oxygen is 0,02 % oxygen. Depending on the sampling method and circumstances, it can be decided whether or not to correct for air.

If the gas contains oxygen, correct the mole fraction of nitrogen according to Formula (1):

(1)

where

In Formula 1 it is assumed that the TCD responses for N₂ and O₂ are equal. To minimise the error for this correction, the respective response factors (of the nearest calibration points) should be taken into account .

This correction applies in most cases to spot sampling.

The mole fraction, , expressed as a percentage, of component j in the sample corrected for the presence of oxygen is normalized to 100 % according to Formula 3:

(3)

where

In certain biomethane producing processes, i.e. air is added to reduce the production of gaseous sulphur components during fermentation, or nitrogen is added to adjust the physical properties of final gas. In the latter case, the origin of this nitrogen will in most cases specifically remove only oxygen, so the produced nitrogen gas will still contain the remaining air components, like the noble gasses.

The amount of argon, present in air, accounts for 0,934 volume percent. This means that for every 100 00 mol ppm of nitrogen in the final gas, also 120 mol ppm of argon is introduced.

When analysing for oxygen in such a process, a separation column must be set up in such a way that argon is sufficiently separated from oxygen. Otherwise, argon’s peak area will significantly add to oxygen’s peak area, resulting in an overcorrected nitrogen content, and should therefore be calculated for based on the nitrogen content and deducted from the oxygen peak area, before calculating and correcting for oxygen.

Often oxygen is found in samples of natural gas from sources where no oxygen should be present, this is primarily due to contamination of atmospheric air during the sampling process. If oxygen is found the best possible course of action would be to acquire a new sample free from contamination; however this is often not possible due to remote analysis or samples taken at a particular time being of interest. This annex will provide a method for the correction of air contamination for natural gas spot samples which come from sources which are known to be oxygen free. Particular care should be taken in this regard since differing fuel gas sources are often comingled with natural gas and some of these sources may contain oxygen.

Since atmospheric air has a well-defined composition it is possible to determine the amount of atmospheric nitrogen and argon in a sample using the ratio of oxygen to nitrogen and argon. The correct nitrogen value can be determined as per Formula 4 below

(4)

where

Trace argon is sometimes present in natural gas, if argon is also measured and nitrogen is corrected for air contamination the argon must also be corrected via Formula 5.

(5)

where

NOTE It is possible that if all of the argon in the mixture is from air contamination that could possibly be negative due to slight biases within the measurement system. If this occurs then should be considered to be zero. If is significantly negative it would suggest strongly that air contamination is not the source of oxygen present in the gas.

Once the corrected values have been determined the mixture must be re-normalised according to Formula 6 for all components except for oxygen, nitrogen and argon

(6)

where

To determine the corrected normalised amount of nitrogen and argon we can use Formula 7

(7)

where is the un-normalized mole fraction, expressed as a percentage of component j in the sample after air contamination correction.

The corrected normalised oxygen value = 0.

When the argon is not separated from the oxygen and the two components co-elute the composite peak is typically assumed to be oxygen.

The composite peak will contain the oxygen and argon, this method assumes all of the argon in the sample is from air contamination, and that the bulk material the sample was taken from contains no argon; however if the bulk material does contain argon this will introduce a bias to the measurement.

The method also assumes that argon and oxygen have the same response factor, this will also introduce a bias to the measurement. These biases should typically be small but can be eliminated by separating and measuring the oxygen and argon.

The corrected nitrogen value can be determined as per Formula 8 below

(8)

where

Once the corrected values have been determined the mixture must be re-normalised according to Formula 9 for all components except for nitrogen and oxygen

(9)

where

To determine the corrected normalised amount of nitrogen we can use Formula 10

(10)

The corrected normalised oxygen value = 0

The choice of the detector depends on the purpose of the gas analysis results. Most common detectors are:

— Thermal conductivity (TCD): It is a universal detector, as long as there is enough difference of the conductivity between the carrier gas and the components of the mixture. Being universal, it is not very specific, and the detection limit can be insufficient in case of low concentrations.

— Flame ionization (FID): it is a detector suitable for the analysis of hydrocarbons, since it responds proportionally to the carbon number in the components.

The most used detector is the TCD, as it does not require a support gas and is a very compact detector.

Other types of detection are available that give a better detection of certain types of components and for lower concentrations. For example, mass spectrometry (MS), which can be used specially to identify unknown compounds. It is a universal detector, but it is expensive and more complicated to use compared to the TCD.

It is important that all components are measured without interference from others, see Figure 7. For each component to be quantified independently, the resolution between neighboring peaks shall exceed 1,5. This can be assessed by measuring peak resolution.

Gas-chromatographic resolution is a characteristic of the separation of two adjacent peaks and is measured as twice the distance between the maxima of the named peaks divided by the sum of the intercepts on the baseline made by tangents drawn to the peaks at half their height (see Figure 6). The resolution R_AB may be expressed by following Formula 11:

(11)

where

Figure 6 — Resolution of two adjacent peaks

Although the resolutions of all peaks are important, there are particular pairs of peaks which are critical: their satisfactory resolution ensures that of other pairs. A reference for peaks resolution that allows obtaining the limit precision values, is given in Annex A.5.2.1.1 and Annex C.6.3.4 of this document.

Ideally, component peaks should start from and finish at the baseline. The data system evaluates the area bounded by the peak profile and the calculated position of the baseline if the peak had not been there. This baseline calculation is more reliable if the data system can assess the baseline position both before and after the peak and hence project a straight line between the two. This is not always possible, particularly if there are time constraints on the analytical cycle, and the signal between adjacent peaks does not return to the baseline (or where it would have been) so that there is some overlap between peaks. See Figure 6.

Where overlap occurs, there is inevitably some misallocation of the areas (quantitative measurements) of the two peaks. If the peaks are symmetrical (approximately Gaussian) and of comparable size, the loss from one balances the gain from the other, and the area allocations are close to being correct. If, as would be the case with iso and normal butane, the responses are very similar and the properties almost identical, then the errors involved in calculating their contributions to calorific value, for example, would be small. If the peaks are of substantially different sizes and the respective substances have different physical properties, then more significant errors will arise.

A good example is nitrogen and methane, where the peak resolution is complicated by the asymmetry of the methane peak in particular. See Figure 8.

Figure 7 — Example of overlapping peaks

Figure 8 — Good separation, but asymmetry for the methane peak (right)

Figure 9 — Figure 8, zoomed in

This is illustrated in the chromatograms shown, which uses the data and recreates the chromatogram from a typical C6+ analyzer. The full chromatogram (Figure 7) is shown at the top, and the nitrogen/methane pair below it (Figure 8), at different scales (see Figure 8 and Figure 9) to allow the tangents to each peak to be plotted. The calculated peak resolution is 0,98. Conventional wisdom would say that this is unacceptably low, and a value of 1,5 of higher should be aimed for. It would need considerable change to the chromatography to achieve a value of 1,5, and almost certainly a longer analysis time. It would also be debatable whether a resolution of 1,5 would give a better result than the existing 0,98.

If the resolution is less than desirable or resolving would require an unproportional amount of time or money, then an (off-line) optimization of the integrating parameters (for the components range of interest) combined with an investigation for linearity could be the most effective way to go.

The detector(s) should be chosen in order to be suitable for all the separated components of interest and has a stable and repeatable response characteristic for the separated components.

The thermal conductivity detector (TCD) is the most widely used for natural gas analysis. Its ability to measure non-flammable gases (principally nitrogen and carbon dioxide) means that is the only choice for single-operation analysis. It can also measure sufficiently low amounts of C4 and C5 hydrocarbons (and the pseudo-component C6+) to be satisfactory for calculation of the more common properties, such as CV, SG and WI.

The thermal conductivity detector responds to conductive heat loss (hence its flow sensitivity). However, in GC applications thermal conductivity is the main effect. The most important parameter is the temperature difference between the sensing element (filament or thermistor) and the detector body.

If a TCD is chosen for the analysis, it should be taken into account that the thermal conductivity of the carrier and of the components to be determined should be sufficiently different to have a sensible signal (for example, in the case of the determination of hydrogen, helium as carrier is not a good choice since they have almost the same thermal conductivity; in this case argon or nitrogen are good choices for carrier gas).

The flame ionization detector (FID) can be used for hydrocarbons, but only in the context of multi-operation analysis. An FID cannot detect inert gases, (high levels of) methane or hydrogen. It is more sensitive to higher hydrocarbons than a TCD. In conjunction with TCD an FID is more suited for CV determination where the gas contains higher amounts of higher hydrocarbons.

The most complete GC for Natural Gas analysis is a setup with both a TCD and FID. However, it requires (much) more support gas and has more safety requirements. In combination with a Boiling Point separating column is also suitable for Hydrocarbon Dew Point calculations.

Hence, we need to be aware of:

1) The sensing element temperature, which is a function of the current flowing through it.

2) The detector block temperature. The detector can be mounted in the column oven, which sets it to the column temperature (only practical for isothermal analysis) or, more commonly, in a separate heated enclosure. In this latter case the constraint on the temperature value is that it should not be lower than the minimum column temperature (or the maximum temperature programmed value) to avoid condensation of column bleed (effluent/analyte) in the detector. The temperature stability is important for baseline stability and consistency of response.

3) The detector pressure. The amount of analyte is directly influenced by the pressure in the detector. Although it would be difficult to actively correct the detector output for the pressure, it is recommended to shield the detector from sudden pressure shifts as much as possible. For instance, this can be achieved by using soft closing doors to the laboratory and installing a “delay line” (tubing-volume-tubing) between the detector and the vent-line. Sudden pressure shifts appear to have a particular propensity for the starting or ending zone of a peak.

The data of the chromatograph uses a data handling system which converts the analogue detector output into qualitative (what is the component?) and quantitative (how much of it?) data.

The data system converts the analogue detector signal to a series of digital sums for each component. If necessary, it converts these sums to molar quantities according to the calibration. From these quantities it calculates physical properties from the resulting composition. It also provides time-based commands to control valve or flow switching during the chromatogram. In a fully automated system, it controls repeated analyses on a regular basis with calibration carried out at selected intervals.

The analogue to digital conversion is trouble-free. The correct allocation of peak area depends upon the system recognizing the start and finish of each peak. The system should allow for baseline drift without confusing it with the start or finish of peaks and not confusing electrical spikes or baseline noise with real signals. It must also handle peaks which are not fully resolved. In this case the signal between the peak maxima does not reach the ‘true’ baseline. It creates a ‘valley’ point between the peaks, and defines a perpendicular line from the valley to the baseline. This represents the end of the first peak and the beginning of the second. In particular circumstances, the system will create a tangent skim whereby the second peak is deemed to be sitting on the tail of the first one.

How these decisions are made can be checked by inspection of the chromatogram generated by the data system. The location of the peak start and finish is usually identified on the chromatogram and the projected baseline under the peak is shown. If the user does not agree with the decisions made, changing the integration parameters, such as peak width, slope sensitivity and turning integration on and off should lead to more acceptable results. The subtlety of these allocations and adjustments are such that specification of how the system should make decisions can hardly be more instructive than “ensure correct operation”. The operator’s experience is the critical input.

A convenient way to transfer data between systems is the XML file format, as described in ISO 23219.

The XML file format is useful for output from ISO 6974-1 for composition, and ISO 6974-2 for uncertainty, and for input for ISO 10723 for performance evaluation. Typically, these would be the gas composition as provided on an analysis certificate, or results from a performance evaluation that would be read into an Excel spreadsheet for data processing.

XML is defined by the tags – keywords enclosed in angular bracket, e.g. <gas>, with a terminating tag with / in front of the keyword, e.g. </gas>

The schema is short but defines the file format that has flexibility to be of use for all natural gas data.

The schema has (up to) four levels:

Within each level the layout is basically: keyword, value, units and uncertainty.

The XML tags defined in the schema are case sensitive (they are all lowercase) and have no leading or trailing spaces. The tags have no attributes, i.e. no name=”value” within the tag.

The content of the tags, e.g. keywords, are case insensitive, and leading and/or trailing spaces are ignored. In the examples (see annexes B to E), the contents are in uppercase; often with trailing spaces in order to improve readability.

The only likely need for lowercase is to distinguish the S.I. prefixes of m (milli) and M (mega). Keywords should not contain spaces.

For more detailed information, see ISO 23219, Natural gas – XML Data File Format.

Peak integration is basically the numerical processing of the detector signal during or after sample analysis to obtain a component response proportional to the component content in the sample being analyzed.

The principle and challenge of integrating the peaks in a chromatogram is finding the balance between baseline noise and peak resolution. The simplest way to integrate a peak is to start the integration when the signal has risen significantly above the noise of the baseline and stop integration after the signal has peaked vice versa, when the signal can no longer be discriminated from the baseline noise.

Figure 10 — Example of the Peakwidth-Threshold box

Above, Figure 10 is an example of peak detection, set by peakwidth and threshold (box); the signal is no longer within the threshold range during the peakwidth time range. The moment this occurs, is defined as the start of the peak. Vice versa, the same logic is used for detecting the end of the peak.

Most GC manufacturers are providing a wide range of peak detection types and parameters, and, also,provide all information necessary to obtain optimum solution for the component range of interest.

A chromatogram is a visual representation of the detector signal of the analysis run, and is often combined with per-peak information, i.e. peak name and retention time, either during analysis or afterwards.

Chromatogram files are commonly stored in a GC producers’ proprietary file type, a binary file or another type of compressed file type.

The analogue signal from the detector needs to be digitized to be processed digitally. The bit depth of this step needs to be in agreement with the signal level to produce an optimal signal to noise ratio. An automatic bit depth-level, based on the signal level, may also be available.

The sampling frequency is set such to allow the detection of the peak of interest with smallest peak width with sufficient resolution, so reliable peak detection can be performed. Setting the sampling frequency to high will not lead to better peak detection, but only increase the file size of the chromatogram data.

Calibration is described in ISO 6974 part 1, the associated uncertainty in ISO 6974 part 2.

The correct way to integrate peaks in a chromatogram is to set integration parameters that lead to the best calibration fit with the lowest polynome for the composition range of interest.

The following steps, using the two-factorial or Design of Experiments approach, for the optimization of the integration parameters are recommended:

1) Unless proper integration demands it, do not start integration directly after injection. The pressure difference caused by injection or backflush can result in a substantial baseline swing. Start integration a few seconds before the first peak, 10 % of its retention time (RT) is a good guideline.

2) Assuming a components determination has already been performed, optimization of integration parameters starts at the first peak in the chromatogram.

3) Inject the lowest concentration of reference gas for the current peak in the calibration range once.

4) In the chromatogram, determine the peak width at half-height (PW) of the current peak. This is the lowest value in the PW-range to be investigated, PWmin.

5) Inject the highest concentration of reference gas for the current peak in the calibration range once. Determine the peak width at half-height (PW). This is the highest value in the PW-range to be investigated, PWmax.

6) Determine the noise level at the current RT by injecting a zero-gas or by performing a run without injection. Within the time period of PWmin, find the difference between the lowest and highest baseline level. This is the minimal threshold for integration, THmin.

7) Analyse all reference gasses at least 10 times.

8) Recalculate following a fractional or DoE approach to obtain the optimal integration parameters, in the following range:

9) PWmin/2 to 2*PWmax, in an odd amount of steps

10) THmin*{1,2,5}*{1,10,100}

11) Determine, either for the lowest and highest content level the integration parameters that combined give the lowest relative standard deviation, or the integration parameters that give the simplest and best fit for all reference gasses.

12) Restart at 3. for the next peak.

This formula is not the only solution to the optimal integration parameters. However, choosing the integration parameters based on a single chromatogram or the most expected concentration will lead to either integration failure or unnecessary high uncertainty in parts of the concentration range of interest.

To obtain a correct estimate of the repeatability, multiple analysis over a longer period of time are required. Additionally the results of the (components) control charts can be used.

Typical values for precision and bias can be found in ISO 6974-3:2018.

Construct, for each component in the control gas, a control chart with the mean value for the component and concentrations representing the mean ±2 standard deviations and the mean ±3 standard deviations marked on the y-axis. Draw lines parallel to the x-axis from these points. Each time the control gas is analysed, plot the value using the x-axis as the time scale.

If it is known that the standard deviation for a component varies with concentration, and that the range of concentrations likely to be encountered is sufficiently wide for these variations to be significant, it may be advisable to have two control charts for such a component, representing the system behaviour towards test gases of different compositions.

Compare the plotted values from each analysis of the control gas with the mean value and the ±2 standard deviation lines and ±3 standard deviation lines. Figure 11 shows a typical example for nitrogen at around 2,5 %. This shows relatively little scatter around the mean value, and gives reassurance that the measurement of this component is satisfactory.

If individual results fall outside the warning limits (±2 standard deviations) more than just occasionally, this can indicate that:

— Either there is a systematic tendency for results to be too high or too low (provided that only the upper or the lower warning limit has been crossed),

— Or the random error for measurement of that component has increased (if both limits are crossed randomly).

Figure 12 shows a control chart for carbon dioxide. Results for the first few days remain close to the mean, but the plot then shows a clear drift downwards. Although the ±3 standard deviation limit has not been exceeded, this clearly suggests that some systematic error is present, allowing underestimation of the carbon dioxide.

Figure 11 — Example of a control chart for nitrogen

Figure 12 — Example of a control chart for carbon dioxide

Figure 13 illustrates an increased random error. Days 5 to 8 appear normal, but from days 9 to 13 much greater but non-systematic variations occur. Once again, no single result exceeds ±3 standard deviations, but some attention to the method is clearly required.

The initial control limits selected are the result of a single repeatability measurement made before the chart is drawn. More information becomes available as the chart is used, and it is reasonable to redraw these limits after 25 or 50 data points have been collected. This assumes, of course, that the method has remained stable, as in Figure 12. Do not use data which clearly indicate some fault to revise the control limits.

Figure 13 — Example of a control chart for ethane

Report the results in accordance with ISO 6974-1.

1. 
   (informative)
   
   Determination of hydrogen, helium, oxygen, nitrogen, carbon dioxide and hydrocarbons up to C8 using two packed columns

NOTE This informative annex A contains an extract of the text of ISO 6974-3:2000 with editorial corrections.

• 1. Application Ranges

This method is applicable to the analysis of gases containing constituents within the mole fraction ranges given in Table A.1.

Table A.1 — Application Ranges

• 1. Principle

Determination of nitrogen, carbon dioxide and hydrocarbons from C₁ to C₈ by gas chromatography using two chromatographic columns. A molecular sieve 13X column coupled with a thermal conductivity detector (TCD) is used for the separation and detection of hydrogen, helium, oxygen and nitrogen, and a PR^[1] column coupled with a TCD and a flame ionization detector (FID) in series is used for the separation and detection of nitrogen, carbon dioxide and hydrocarbons from C₁ to C₈. The two analyses are carried out independently and the results are combined.

The correction for oxygen is described in clause 8.10.

Quantitative results are achieved by determining the response of the TCD detector with reference-gas mixtures and using relative response factors of the FID detector.

The resulting composition of the natural gas is normalized to 100 %.

• 1. Materials
     1. For the determination of helium, hydrogen, oxygen and nitrogen

Separation on molecular sieve 13X column consisting of the following.

A.3.1.1 Argon carrier gas

> 99,99 % pure, free from oxygen and water.

If the purity of the gas is less than that specified, it is essential to check that the type of impurity present does not interfere with the analysis. Also, even if the carrier gases argon and/or helium fall within the specification, some of the impurities present in the gases can nevertheless interfere with the analysis. Under these circumstances, appropriate purification is essential.

A.3.1.2 Working-reference gas mixtures (WRM), consisting of:

A.3.1.2.1 Gas mixtures containing helium and hydrogen with nitrogen or argon as the matrix gas.

A.3.1.2.2 Gas mixtures containing oxygen and nitrogen with argon as the matrix gas.

NOTE 1 Take care to prevent explosion of gas mixtures.

NOTE 2 In the case of analysis using only one instrument, the WRM with oxygen and nitrogen as components and argon as the matrix gas can be replaced by oxygen with nitrogen as matrix gas. By addition of helium to the WRM this gas could also be used for the daily calibration.

A.3.2 For the determination of nitrogen, carbon dioxide and hydrocarbons from C₁ to C₈ (separation on Porapak column),

Consisting of the following:

A.3.2.1 Helium carrier gas, > 99,99 % pure, free from oxygen and water.

A.3.2.2 Working-reference gas mixtures (WRM),

Consisting of multi-component gas mixtures containing: nitrogen, carbon dioxide and hydrocarbons from C₁ to C₃ (optional to C₄).

An example of the composition of the working-reference gas mixture is given in Table A.2.

Table A.2 — Example of the composition of the working-reference gas mixture.

A.3.2.3 FID gases, consisting of

• 1. Apparatus
     1. Laboratory gas chromatographic (GC) system

Consisting of two columns, a molecular sieve 13X column and a PR column, which are contained in two column-ovens or can be placed in the same column-oven.

The gas sample is injected on each column by means of a 6-way sample valve. Signal responses of components in the gas sample are detected using TCD and/or FID detectors.

NOTE The gas sample can be injected into the PR and molecular sieve column in series using a column isolation technique.

A.4.1.1 For the determination of helium, hydrogen, oxygen and nitrogen,

Equipped with the following specific components and characteristics.

A.4.1.1.1 Gas chromatograph,

Capable of temperature-programmed operation and equipped with a TCD and the following specific equipment:

A.4.1.1.2 Injection device,

Consisting of a by-pass-type injector (gas-sampling valve) having an injection capacity of 1 ml and capable of being heated to a temperature setting of 110 °C.

The sample volume shall be reproducible such that successive runs agree within 1 % for each component.

A.4.1.1.3 Columns, two with the same type of packing and with the same dimensions.

The second column is normally used for drift compensation during the temperature programme. If drift is compensated by means of an electronic integrator, the second column is not necessary.

Columns shall satisfy the following requirements:

NOTE Some injection devices are unable to deal with temperatures above 250 °C and can cause conditioning problems.

A.4.1.1.4 Thermal conductivity detector (TCD).

A.4.1.2 For the determination of nitrogen, carbon dioxide and hydrocarbons from C₁ to C₈,

Equipped with the following specific components and characteristics.

A.4.1.2.1 Gas chromatograph,

Suitable for dual-column application and equipped in series with a TCD and an FID.

A.4.1.2.2 Injection device,

Consisting of a by-pass-type injector (gas-sampling valve) having an injection capacity of 1 ml and capable of being heated to a temperature setting of 110 °C.

A.4.1.2.3 Columns,

Two of the same type of packing and with the same dimensions.

The second column is normally used for drift compensation during the temperature programme. If drift is compensated by means of an electronic integrator, the second column is not necessary.

A.4.1.2.4 Detectors,

Having the following characteristics:

— for components including hydrocarbons up to C3: thermal conductivity detector (TCD)

— for hydrocarbons from C4 to C8: flame ionization detector (FID)

— Ethane and propane can be detected by an FlD if the mole fraction is less than 1 %. In either case, the time constant shall not be greater than 0,1 s. If C3 is used as reference component, it shall be detected by an FID.

— the TCD and FID detectors shall be connected in series

NOTE If the mole fraction of oxygen is less than 0,02 %, the nitrogen value can be taken from the PR analysis, assuming that hydrogen is not present in the gas sample.

• 1. Procedure
     1. Gas chromatographic operating conditions
        1. For the determination of helium, hydrogen, oxygen and nitrogen

Set the operating conditions for the apparatus (A.4.1.1.) as follows.

— initial temperature: 35 °C for 7 min

— temperature rate: 30 °C/min to 250 °C

— final temperature: maintain at 250 °C for 10 min

NOTE Alternative procedures for analysis on the molecular sieve 13X column are described in Annex A.6. Variations in the programming give a better separation in some cases.

— set according to manufacturer's instructions

— temperature: between 140 °C and 160 °C

— carrier gas: argon

• • • 1. For the determination of nitrogen, carbon dioxide and hydrocarbons from C₁ to C₈
         1. GC conditions

Set the operating conditions for the apparatus (A.4.1.2.) as follows:

• • • • 1. Column stability check

Check the baseline stability of the column by means of blank runs.

No individual peak shall originate from a constituent having a mole fraction exceeding 0,04 %. If larger peaks are seen, repeat blank runs until satisfactory. If necessary, prepare new columns, preferably from a different batch of PR.

NOTE 1 Different batches of PR often show variation in performance. For example, the retention sequence of benzene and cyclohexane can be reversed. It is therefore recommended that the retention time of benzene and cyclohexane be determined from time to time, and certainly after new columns have been installed.

NOTE 2 Baseline stability can be checked as follows:

• • 1. Performance requirements
       1. Resolution efficiency
          1. Molecular sieve 13X column

The height of the valley between the peaks above the baseline shall be no greater than 10 % of the height of the larger peak under the operating conditions following injection of a sample containing equivalent amounts (a mole fraction of about 0,4 %) of hydrogen and helium (see Table A.3). If this criterion is not met, condition the packing for a longer period or prepare a new column.

Assess the peak resolution in accordance with ISO 7504.

Table A.3 — Required peak resolution

• • • • 1. PR column

The height of the valley between the 2‑methylbutane and pentane peaks above the baseline shall be no greater than 10 % of the height of the larger peak under the operating conditions following injection of a sample. If this criterion is not met, condition the packing for a longer period or prepare a new column.

• • • 1. Response

Determine the response characteristics for each of the gases determined in accordance with ISO 6974-2 at least once a year.

• • • 1. Relative response factor

Determine the relative response factors in accordance with ISO 6974-2.

• • 1. Determination
       1. Outline of the analysis

The analysis is outlined as follows.

Examples of typical chromatograms of this analysis are given for information in Figure A.1 and Figure A.2 of Annex A.6.

• • • 1. Estimation of other components

Estimate the fraction of other components in accordance with ISO 6974-1.

Backflushing shall not be carried out.

• 1. Example: Single-oven gas-chromatographic system consisting of two columns

Both analytical columns are placed in a single column oven and provided with a linear temperature programmer capable of obtaining a rate of temperature increase of 30 °C/min over the specified range.

A molecular sieve 13X column is used for the determination of helium, hydrogen, oxygen. The detection of these components is carried out by TCD. The gas sample is injected using a by-pass-type injector with an injection capacity of 1 ml. Flow regulators are used to give suitable argon gas-flow rates.

A PR column is used for the determination of nitrogen, carbon dioxide, methane to normal octane. The detection is carried out by TCD in series with an FID. The gas sample is injected using a by-pass-type injector (gas sampling valve) with an injection capacity of 1 ml. Flow regulators are used to give suitable helium gas flow rates.

The configuration of such a chromatographic system is given in Table A.4.

Table A.4 — Configuration of the chromatographic system

Figure A.1 — Typical chromatogram of helium, hydrogen, oxygen and nitrogen using a Molecular sieve 13X column (with indication of the absolute retention time in minutes)

Figure A.2 — Typical chromatogram of nitrogen (oxygen), carbon dioxide and hydrocarbons from C1 to C8 using a PR column

1. 
   (informative)
   
   Determination of nitrogen, carbon dioxide and C1 to C5 and C6+ hydrocarbons for a laboratory and on-line measuring system using two columns

NOTE This informative annex B contains an extract of the text of ISO 06974-4:2000 with editorial corrections.

• 1. Application ranges

This method is applicable to the analysis of gases containing constituents within the mole fraction ranges given in Table B.1.

Table B.1 — Application ranges

• 1. Principle

Determination of nitrogen, carbon dioxide, methane, ethane, propane, butanes and pentanes by gas chromatography using two columns (a short one and a long one) of silicone oil DC-200 on CP^[2] column in a backflush arrangement. The short column retains hydrocarbons heavier than normal pentane which are eluted after backflushing as a C₆₊ composite peak. The long column is used for the determination of nitrogen, carbon dioxide, methane up to normal pentane. Detection is carried out by a thermal conductivity detector (TCD). Oxygen, argon, hydrogen and helium are not measured by this method.

• 1. Materials
     1. Helium carrier gas,

> 99,99 % pure.

• • 1. Working-reference gas mixtures (WRM),

The composition of which shall be chosen to be similar to that of the anticipated sample.

Prepare a cylinder of a working-reference gas mixture, by a gravimetric method, in accordance with ISO 6142, and/or certified and validated in accordance with ISO 6143. The working-reference gas mixture shall contain at least nitrogen, carbon dioxide, methane, ethane, propane, n-butane, iso -butane and possibly neo-pentane, iso -pentane and n-pentane.

• 1. Apparatus
     1. Laboratory gas chromatographic (GC) system,

Consisting of the following components.

• • • 1. Gas chromatograph (GC),

Capable of isothermal operation and equipped as follows:

column oven, capable of maintaining the temperature of the column temperature to within ±0,1 °C over the temperature range from 70 °C to 150 °C;

valve oven: capable of being maintained over the temperature range from 70 °C to 150 °C or alternatively having the capacity to fit the valves in the column oven;

flow regulators: capable of regulating the carrier gas flow rates.

• • • 1. Injection device,

Consisting of a ten-port sample-injection valve V1 and also used for backflushing C₆₊ components (two six-port valves may be used for these duties).

• • • 1. Metal columns packed with 28 % DC-200 on CP,

Satisfying the performance requirements given in clause B.6.2 and consisting of the following packing materials and column dimensions, given as examples, for use with conventional and readily available injection valves and TCD.

Columns shall satisfy the following requirements:

NOTE The following packing method is suitable: Close the column outlet with a sintered disc or glass wool plug. Connect a reservoir containing rather more packing than is needed to fill the column to the inlet and apply a pressure of 0,4 MPa of nitrogen to this reservoir. The flow of packing into the column is assisted by vibration. When the column is full, allow the pressure to decay slowly before disconnecting the reservoir.

• • • 1. Thermal conductivity detector (TCD),

With a time constant no greater than 0,1 s, and internal volume appropriate for the column sizes and flow rate used.

• • • 1. Controller/peak-measurement system,

Having a wide range of sensitivity (0 V to 1 V), capable of measuring peaks on a sloping baseline and able to control automatic operation of the valves according to a sequence selected by the operator.

• • • 1. Auxiliary equipment,

Consisting of valves, tubing and any other accessories, to control the flow of sample gas to the chromatograph and for shutting off this flow for a defined period of time before injection.

• 1. Procedure
     1. Gas chromatographic operating conditions

Set the operating conditions for the apparatus (B.4) as follows.

An example of the configuration is shown in Figure B.1 and Figure B.2. The measuring system comprises a ten-port sample injection/backflush valve. In configuration 1, see Figure B.1, the sample loop is flushed by the sample gas. When the valve is switched in configuration 2, see Figure B.2, injection is performed. The valve is returned to configuration 1 after all the n‑pentane leaves column 1 but before the lowest C₆ isomer leaves column 1 on its way to column 2.

A typical chromatogram is shown in Figure B.3.

The gas chromatographic conditions are summarized in Table B.2.

The procedure for setting valve timings and restriction setting is described in section 5.7.

Figure B.1 — Configuration 1

Figure B.2 — Configuration 2

Description of configuration 1 and 2 parts:

a. Gas sample

b. Vent

c. Sample loop (1 ml)

d. Carrier gas

e. TCD (reference channel)

f. TCD (analysis channel)

g. Column 2 (length: 9 m)

h. Column 1 (length: 0,45 m)

Table B.2 — Gas chromatographic conditions

Figure B.3 — Example of a typical chromatogram

• • 1. Performance requirements — Peak resolution

It is important that all components are measured with as less interference from others as possible. Possible interference can be assessed by measuring peak resolution in accordance with clause B.6.2 of this standard. Although the resolution of all peaks is important, there are no particular pairs of peaks which are critical. However, acceptable resolution of one pair of peaks can ensure acceptable resolution of other peaks.

Furthermore, the resolution required is likely to vary with respect to component uncertainty although it may be deemed acceptable for particular applications. If the procedure is implemented correctly, the values of acceptable peak resolution indicated in Table B.3 shall be expected. Higher resolution may require modification of column dimensions, temperature and flowrate, and would likely require longer analysis time.

Each value of resolution shall be tested as part of the normal analytical cycle, not by some alternative procedure designed only to measure these parameters.

Table B.3 — Acceptable peak resolution

NOTE The above values are outdated and originate from the described setup. The required resolution is described in clause 9.2.

• • 1. Determination — Outline of the analysis

The analysis is outlined as follows.

• 1. Expression of results
     1. Calculation

Refer to ISO 6974-1.

• • 1. Precision and accuracy

Refer to ISO 6974-3:2018.

• 1. Procedure for setting valve timings and restriction setting

1. The backflush operation allows n-pentane (n-C₅) to be measured by forward elution and all of the lightest C₆ (2,2, dimethyl butane) to be backflushed. Use a gas mixture containing n-C₅ and 2,2-dimethyl butane preferably with no other C₆ or heavier component present.

2. Set the initial time to 1,5 min (or as recommended by the manufacturer) after injection at which time the valve returns to configuration 1 (see Figure 1). Inject the gas mixture and record the chromatogram. 2,2-Dimethyl butane should appear as a backflushed component (C₆₊) shortly after the valve returns to configuration 1, and n-C₅ should appear as a normally eluted peak. If no C₆₊ peak is seen, reduce the initial time setting and repeat this operation.

3. Continue to inject the gas mixture, increasing the backflush time (valve to configuration 1) by 0,05 min for each successive injection until the backflushed C₆₊ peak area (in fact 2,2-dimethyl butane) starts to diminish.

4. Continue injecting the mixture, now reducing the backflush time setting by 0,05 min for successive injection. Note the time at which the area of the C₆₊ peak first becomes constant (designated as t_back,high).

5. Continue injections with further incremental reductions in the backflush time until the size of the n-pentane peak starts to diminish, with a corresponding increase in the size of the C₆₊ peak. Note the latest backflush time at which the areas of both peaks are still constant (designated as t_back,low).

6. Determine the value of t_back using Formula (B.1):

tback = (tback,low + tback,high) / 2 (B.1)

• 1. Final time settings

Implement the method given in Table B.4 (see Figure B.1 and Figure B.2).

Table B.4 — Valve configuration timing

1. 
   (informative)
   
   Isothermal method for nitrogen, carbon dioxide, C1 to C5 hydrocarbons and C6+

NOTE This informative annex C contains an extract of the text of ISO 06974-5:2014 with editorial corrections.

• 1. Application ranges

This method is applicable to the analysis of gases containing constituents within the working ranges given in Table C.1.

Table C.1 — Component working ranges

• 1. Principle

This chromatographic method uses a column switching/backflush arrangement, configured as shown in Figure C.1. The sample is injected onto a boiling-point column which is divided into short and long sections (columns 1 and 2). The long section (column 2) provides separation of C3 to C5 hydrocarbons, while C6 and heavier hydrocarbons are retained on the short section (column 1), from which they are backflushed and measured by the detector as a single peak. Two six-port valves can handle the sample injection and backflushing operations, or they may be dealt with together by a single 10-port valve.

Figure C.1 — A typical chromatogram

As Figure C.1 shows, nitrogen, carbon dioxide, methane and ethane pass rapidly and unresolved through the boiling-point column onto a porous polymer bead column (column 3), suitable for their separation. A six-port valve either connects this column or by-passes it during measurement of C₃ to C₅ components.

The separations that occur in the columns are as follows:

Column 1   Retains C₆+ components ready for backflushing as one composite peak.

Column 2   Separates Propane, iso-Butane, n-Butane, neo-Pentane, iso-Pentane and n-Pentane, (which elute after C₆+ has left column 1).

Column 3   Stores and separates Nitrogen, Methane, Carbon Dioxide and Ethane which elute after n-Pentane has left column 2.

• 1. Materials
     1. Carrier gas,

Helium (He), ≥99,995 % pure, free from oxygen and water.

• • 1. Auxiliary gases,

Compressed air, for valve actuation (If consumption is low, carrier gas may be used as an alternative for valve actuation).

• • 1. Reference materials
    2. Reference gases,

According to ISO 6974-1.

• • 1. Gas mixture containing n-Pentane and 2,2-Di-Me-butane,

Used to check valve timings (see clause C.9).

• 1. Apparatus
     1. Gas chromatograph,

Capable of isothermal operation and equipped with TCD.

• • 1. Column oven,

Temperature range 70 °C to 105 °C, capable of being maintained to within ±0,1 °C.

• • 1. Valve oven,

Controlled over the temperature range 70 °C to 105 °C, or valves fitted in the column oven.

• • 1. Pressure regulator,

To give suitable carrier gas flow rates

• • 1. Injection device,

V1, six-port sample injection valve

• • 1. Backflush valve,

V2, six-port, to allow rapid backflush of C₆+ components. As described in Clause C.2, a single 10-port valve may be used for both these tasks. The operating principle is the same.

• • 1. Column isolation valve,

V3, six-port. This directs the carrier gas through the porous polymer bead column (column 3) or by-passes it.

• • 1. Columns,

The columns must satisfy the performance requirements given in clause C.6.3.4. The following packing materials and column dimensions, given as examples, should be satisfactory, for use with conventional and readily available injection valves and TCDs. Any alternative combination of columns which provide similar separations and satisfy the performance requirements may be used. Micro-packed or even capillary columns can be chosen, with appropriately sized injection and detector systems, in which case packing or coating details would be different.

• • 1. Tube and packing.
       1. Configuration 1

Examples of the operating conditions for configuration 1 are given in Tabl C.2.

• • • • 1. Column 1,

A diatomite packed column with 28 % loading of silicone oil DC200/500 on 45/60 mesh particle size: 0,75 m (2,5 foot) long, 2 mm internal diameter and 1/8 in outer diameter.

• • • • 1. Column 2,

A diatomite packed column with 28 % loading of silicone oil DC200/500 on 45/60 mesh particle size: 5,2 m (17 foot) long, 2 mm internal diameter and 1/8 in outer diameter.

• • • • 1. Column 3,

A column packed with spherical beads of porous copolymers of polydivinylbenzene (DVB) with 15 % loading of Silicone Oil DC200/500 on 50/80 mesh particle size: 2,4 m (8 ft) long, 2 mm internal diameter and 1/8 in outer diameter.

Table C.2 — Example of instrument conditions, configuration 1

• • • 1. Configuration 2.

Examples of the operating conditions for configuration 2 are given in Table C.3.

• • • • 1. Column 1

A silica packed column with oxydiproprionitrile stationary phase: 0,3 m (1 foot) long, 0,75 mm internal diameter and 1/16 inch outer diameter.

• • • • 1. Column 2

A diatomite packed column with 20 % loading of silicone oil SF-96 on 80/100 mesh particle size: 2,1 m (7 foot) long, 0,75 mm internal diameter and 1/16 in outer diameter.

• • • • 1. Column 3

A column packed with spherical beads of porous copolymers of poly(divinylbenzene-co-ethylenedimethacrylate). with 15 % loading of silicone oil DC200/500 on 80/100 mesh particle size: 2,1 m (7 foot) long, 0,75 mm internal diameter and 1/16 in outer diameter.

Table C.3 — Example of instrument conditions, configuration 2

• • 1. Method of packing,

Any method which results in uniform column packing may be used.

To obtain a uniform column pack, the following method is suitable. The column outlet is closed with a sintered disc or glass wool plug. A reservoir containing rather more packing than is needed to fill the column is connected to the inlet and a pressure of 0,4 MPa of nitrogen is applied to this reservoir. The flow of packing into the column is assisted by vibration. When the column is full, allow the pressure to decay slowly before disconnecting the reservoir.

• • 1. Thermal Conductivity Detector (TCD),

With a time constant of not greater than 0,1 s, and internal volume appropriate for the column sizes and flow rate used.

• • 1. Controller/Peak Measurement System,

Wide range (0 V to 1 V), capable of measuring peaks on a sloping baseline. Be enabled to control automatic operation of the valves according to a sequence selected by the operator.

• • 1. Auxiliary valves, tubing and other accessories,

To control the flow of sample gas to the chromatograph and for shutting off this flow for a defined period of time before injection.

• 1. Scheme of the configuration

The valve positions during the separation process are shown in Figure C.2, Figure C.3, Figure C.4, Figure C.5. and Figure C.6 below.

Figure C.2 — a) Initial configuration: all valves in position 1

Figure C.3— b) Sample injection: V1 to position 2

Figure C.4 — c) Backflush C6+: V2 to position 2

Figure C.5 — d) Isolate N2, C1, CO2, C2; measure C3 to C5: V3 to position 2

Figure C.6 — e) Reconnect column 3 - measure N2, C1, CO2, C2:V3 to position 1

• 1. Procedure
     1. Control of the apparatus

Set up the gas chromatograph according to the manufacturer’s instructions.

When the method is first set up, establish the repeatability of measurement by repetitive analysis of a cylinder of test gas, commonly a typical natural gas. For each component, draw up a control chart showing the mean value (e.g. mole fraction or area), and the bounds representing 2 and 3 standard deviations. Analyse the test gas after each calibration of the analyser and compare the results with the data in the control charts.

NOTE The performance of the analyser is assessed by this procedure.

Any change in the method setup can give rise to differences in component responses and hence (where applied) to calculated uncertainties. In these circumstances do not attempt to fit data to an existing control chart; rather, repeat the operations that were undertaken when the method was first set up.

• • 1. Column Conditioning

The columns described in clause C.4.8 do not need conditioning or activation and are generally being used well within their temperature limits. However, a small amount of column bleed due to lower-boiling impurities may be evident on first use and result in unstable baselines. Operation of the analyser overnight with carrier gas flowing but no sample injections, at a temperature 20 °C to 40 °C above the recommended operating temperature should eliminate this effect.

Residual adsorbed moisture in the lines supplying carrier gas or sample gas can give rise to unexplained peaks over and above those expected. Operation overnight under the recommended conditions with sample injection should eliminate these effects.

• • 1. Operation of the apparatus
       1. Analytical method

Examples of the operating conditions for configurations 1, see clauses C.4.9.1 and are given in Table C.4 and Table C.5.

Table C.4 — Example of instrument conditions, configuration 1

Table C.5 — Example of instrument conditions, configuration 2

• • • 1. Sample introduction

Purge the sample valve with the gas to be analysed, using at least 20 times the volume of the valve and associated pipe work.

Stop the purge to enable the gas to reach the temperature of the valve and ambient pressure, and then start the analytical cycle, injecting the sample and switching the valves as required.

If this volume of sample is not enough to purge the valve, then contamination by air or by the previous sample will be evident. If either occurs, then use a larger volume of sample for purging.

NOTE The sample loop should be purged with gas for a precise time, at a defined rate, and the sample should then be allowed to equilibrate to ambient pressure before injection. In the absence of equipment which can confirm the latter, there should be a defined time between sample valve shut off and injection.

• • • 1. Analysis

The analytical system shown in Figure C.2 through Figure C.6 consists of one six-port sample injection valve, V1, one six-port backflush valve, V2, and one six-port bypass valve V3. Restrictor A maintains the pneumatic balance of the system when column 3 is isolated. The detailed setting-up procedure is given in see clause C.9. (One ten‑port valve may be used in place of the six-port valves V1 and V2, controlling both sample injection and backflushing of column 1.)

The timings of the valve switching operations shall ensure that:

a. V2 is returned to the backflush position (position 2) after all the n-pentane leaves column 1 but before the lowest C₆ isomer leaves column 1 on its way to column 2;

b. V3 is switched to isolate column 3 (position 2) before any propane leaves column 2 (on its way to column 3) and after all the ethane has left column 2 and entered column 3;

c. V3 is not returned to reconnect column 3 (position 1) until all the n-pentane has been detected, after having emerged from column 2 via column 1.

A typical chromatogram is shown in Figure C.7.

Figure C.7 — A typical chromatogram

• • • 1. Peak resolution

As described in clause 8.2, it is important that all components are measured without interference from others. The resolution between neighbouring peaks can be assessed according to ISO 7504:2001, 3.3.4.4. Although the resolutions of all peaks are important, there are particular pairs of peaks which are critical: their satisfactory resolution ensures that of other pairs (see Table C.6).

The resolution required is likely to vary with the component uncertainties which are deemed to be acceptable for particular applications. Two values are quoted below:

— medium resolution, which should be available if the procedure is implemented normally, and

— high resolution, which may require modifications to column sizes, temperature and flow rate, and is likely to involve a longer analysis time.

NOTE A resolution of 1,5 or higher indicates baseline separation between symmetrical peaks. A resolution of 1,0 is taken to be the minimum value for quantitative measurement.

Table C.6 — Peak resolution

NOTE The above values are outdated and originate from the described setup. The required resolution is described in clause 9.2.

• 1. Expression of results

Refer to ISO 6974-1.

• • 1. Uncertainty

Refer to ISO 6974-2.

• 1. Example of application
     1. General considerations

In this example, the analysis is considered to be of Type 2. Instrument response for all components is assumed to be first order with zero intercept. All components are measured directly against the same component in the Working Measurement Standard (WMS). No Other Components were determined. No pressure correction was employed, either during calibration or analysis of sample. Multiple operation methods (with or without bridging) were not employed.

Performance evaluation of the instrument according to ISO 10723 was carried out prior to calibration and analysis using seven test gases, each containing 11 components. From prior knowledge of the intended application and likely composition of gases to be presented to the analyser, the Working Range of the instrument is decided to be that given in Table C.7.

Table C.7 — Working range of the analyser

Performance evaluation resulted in the outputs shown in Table C.8. The mean errors, , shown in the second column of Table C.8, are deemed to be sufficiently close to zero that correction is unnecessary (see ISO 6974 1:2001, 6.9.4).

Table C.8 — Output from performance evaluation of the analyser

In the example given the mole fractions and their (standard) uncertainties are expressed to a large number of significant figures purely to aid checking of calculations and software.

Reporting of results should follow the guidelines indicated in clause 14 of ISO 6974-1:2001.

Table C.9 — Calibration of analyser with WMS

For each component the assumed analysis function was calculated from the mean of the 10 responses of the instrument to that component according to Formula (6) of ISO 6974‑1:2001. The coefficients of the calibration function b_1,j and their uncertainties are shown in Table C.10.

Table C.10 — Mean responses, coefficients of the assumed analysis function and their uncertainties

• • 1. Calculation of mole fractions
       1. Mean normalization method (see ISO 6974‑1:2001, 6.9.2)

Analysis was performed using 10 injections of unknown sample and the responses are shown in C.11.

Table C.11 — Analysis of the unknown sample — Responses

The mean responses, calculated according to ISO 6974‑1:2001, Formula (7) are shown in Table C.12.

Table C.12 — Analysis of the unknown sample — Mean responses and uncertainties

The raw mole fractions were calculated according to ISO 6974‑1:2001, Formula (9) and are shown in Table C.13.

The mole fractions were calculated according to ISO 6974‑1:2001, Formula (11) (note that x_oc is zero in this instance) and are shown in Table C.13.

Table C.13 — Raw mole fractions, mole fractions and their uncertainties

• • • 1. Run-by-run normalization method (see ISO 6974‑1:2001, 6.9.3)

For each injection of unknown the raw mole fractions were calculated according to ISO 6974‑1:2001, Formula (13) and are shown in Table C.14.

For each injection the mole fractions were calculated according to ISO 6974‑1:2001, Formula (15). Note that x_oc is zero in this instance and are shown in Table C.15.

Table C.14 — Raw mole fractions and their uncertainties for each run

Table C.15 — Mole fractions and their uncertainties for each run

For each component the mean of the mole fractions are computed from the mole fractions obtained from each injection using ISO 6974‑1:2001, Formula (16) and are shown in Table C.16.

Table C.16 — Mean mole fractions and their uncertainties

• • 1. Calculation of uncertainties of mole fractions
       1. Mean normalization method

(see ISO 6974-2, 5.3.2)

The uncertainties of raw mole fractions were first calculated using ISO 6974-2, Formula (3) and are shown in Table C.13. Input data to Formula (3) were obtained as follows:

— The uncertainties of mean responses to the unknown were calculated from ISO 6974-2, Formula (6) and are shown in Table C.12.

— The uncertainties of the mean coefficients of the assumed response were calculated from ISO 6974-2, Formula (7) and are shown in Table C.10.

— The additional terms in Formula (3) associated with nonlinearity errors are all included because no correction of the raw mole fractions was carried out. The nonlinearity terms are shown in Table C.8.

The uncertainties of mole fractions were calculated from ISO 6974-2, Formula (5) and are shown in Table C.13. Note that terms in xoc and u(xoc) are zero in this instance.

• • • 1. Run-by-run normalization method

(see ISO 6974-2, 5.3.3)

For each run the uncertainties of raw mole fractions were first calculated using Formula (14) of ISO 6974-2 and are shown in Table C.14. Input data to Formula (3) were obtained as follows:

— The uncertainties of responses to the unknown were calculated as the standard deviation of the responses to the unknown and are shown in Table C.16.

— The uncertainties of the coefficients of the assumed response were calculated from ISO 6974-2, Formula (17) and are shown in Table C.10.

— The additional terms in Formula (14) associated with nonlinearity errors are all included because no correction of the raw mole fractions was carried out. The nonlinearity terms are shown in Table C.8.

For each run the uncertainties of mole fractions were calculated from the numerator of ISO 6974-2, Formula (16) and are shown in Table C.15. Note that terms in x_oc and u(x_oc) are zero in this instance.

The uncertainties of the mean mole fractions were calculated from ISO 6974-2, Formula (16) and are shown in Table C.16. Note that terms in x_oc and u(x_oc) are zero in this instance.

• • 1. Comparison of mean normalization and run-by-run approaches

Strictly-speaking, the mole fractions calculated by the two approaches are slightly different and the degree to which they (and their uncertainties) will differ will depend upon the magnitude of the input uncertainties (principally the repeatability of the response of the instrument and the uncertainty in the composition of the WMS). For this example the input uncertainties are small and the differences between the two approaches are negligible.

• • 1. Reporting of results

For laboratory analyses in which expanded uncertainty is reported, the expanded uncertainty of mole fraction should be rounded to two significant figures using the normal rules of rounding, as described in ISO/IEC Guide 98‑3:2008^[1]. The numerical value of mole fraction should be rounded to the least significant figure in the expanded uncertainty.

• • 1. Excel spreadsheet

An Excel spreadsheet implementing this example is available on request. It contains a user-defined function that implements Formula (5) and the numerator ISO 6974-2, Formula (16).

Although the spreadsheet will be made available in good faith, there is no implied warranty for its use in contractual, commercial or other applications and no guarantee that it is error-free. However, it has been tested by several experts and contains no known error at the time of going to press.

• 1. Procedure for Setting Valve timings and Restrictor Setting
     1. Initial Flow Settings

1) Set all the valves to position 1, see Figure C.10 so that the flow pattern is: column 1 (short boiling point) → column 2 (long boiling-point) → column 3 (porous polymer) → detector. Set the column temperature and carrier gas flow rate through the sensing side of the TCD to the manufacturer's values. In the absence of manufacturer's data, use 95 °C and 28 ml/min, for a system using 2 mm i.d. (1/8 in o.d.) columns.

2) Switch valve 3 to position 2, see Figure C.11 so that column 3 is by-passed. Allow the carrier flow to stabilize and then adjust restrictor A so that the flow through the sensing side of the TCD is identical to that measured in 1).

3) Set the TCD reference flow to the value measured in 1).

4) With valve 3 in position 2, inject a sample of natural gas by switching valve 1 to position 2. Record the chromatogram as components elute from column 2. The retention time for n-pentane should be about 2/3rds of the anticipated analysis cycle time. If it is significantly different, return to 1) and adjust the flow accordingly. Then repeat 2) to 4).

5) If no switching time is provided by the manufacturer, measure the time from injection to the valley minimum between ethane and propane (T_FIRSTCUT). This will be the initial time used for storing the lighter components on column 3.

• • 1. Backflushing

1) The backflush operation must allow all of the latest-eluting C₅ (n-C₅) to be measured by forward elution. and all of the lightest C₆ (2,2, dimethyl-butane) to be backflushed. Use a gas mixture containing n-C₅ and 2,2-di-methyl butane with no other C₆ or heavier components present.

2) Set an initial timing of 1 min (or as recommended by the manufacturer) after injection at which to switch valve V2 to position 2. Switch V3 to position 2 to isolate column 3. Inject the gas mixture and record the chromatogram. 2,2-di-methyl butane should appear as a backflushed component (C₆+) shortly after valve V2 switches from position 1 to position 2, and n-C₅ should appear as a normally eluted peak with a slightly longer retention time than that measured in C.9.3 item 4). (It has to travel through column 1 twice). If no C₆+ peak is seen, reduce the initial timing and repeat this section.

3) Continue to inject the gas mixture, increasing the backflush time (V2 to position 2) by 0,05 min on successive injections until the backflushed C₆+ peak (in fact 2,2-dimethyl butane) starts to diminish in area.

4) Continue injections of the mixture, now reducing the backflush time by 0,05 min on successive injections. Note the time at which the area of the C₆+ peak first becomes constant (T_{BACK HIGH}).

5) Continue injections with further incremental reductions in the backflush time until the size of the n-pentane peak starts to diminish, with a corresponding increase in the size of the C₆+ peak. Note the latest backflush time at which the areas of both peaks are still constant (T_{BACK LOW}).

6) Adopt the value of (T_{BACK LOW} + T_{BACK HIGH})/2 = T_BACK for backflush of column 1 (V2 → position 2).

• • 1. V3 Timing

1) In the absence of manufacturers data, set the timings of T_BACK for backflush (V2 → position 2) and T_{FIRST CUT} for isolation of column 3 (V3 → position 2). Switch all valves initially to position 1. Inject a sample of natural gas and, after the elution of n-pentane, switch valve 3 to position 1. Note this time (T_{V3 OFF}) and use it for these operations for the remainder of this section. Measure the peak area for the propane peak eluted from column 2 (via column 1) and that for the ethane peak eluted from column 3 (also via column 1).

2) Repeat the analysis, reducing T_{FIRST CUT} successively in increments of 0,05 min until the ethane peak eluted from column 3 reduces in size.

3) Continue with repeated analysis, now increasing the T_{FIRST CUT} in 0,05 min increments until a steady figure is obtained for ethane eluted from column 3. Note the lowest timing value at which this occurs as T_{FIRST LOW}.

4) Continue this process until the value for propane eluted from column 2 starts to reduce. Note the timing at which this starts to occur as T_{FIRST HIGH}.

5) Adopt the value of (T_{FIRST LOW} + T_{FIRST HIGH}) / 2 = T_V3ON as the time for the initial isolation of column 3.

• • 1. Final timings

Implement the method with the timings according to those in Table C.17 below:

Table C.17 — Timing table

1. 
   (informative)
   
   Determination of hydrogen, helium, oxygen, nitrogen, carbon dioxide and C₁ to C₈ hydrocarbons using three capillary columns

NOTE This informative annex D contains an extract of the text of ISO 06974-6:2002 with editorial corrections.

• 1. Application ranges

This method is applicable to the analysis of gases containing constituents within the working ranges given in D.1.

Table D.1 — Component working ranges

• 1. Principle
     1. Analysis of natural gas sample

Determination of hydrogen, helium, oxygen, nitrogen, carbon dioxide and hydrocarbons from C1 to C8 by gas chromatography using three capillary columns. A PLOT ) precolumn is used for the separation of carbon dioxide (CO2) and ethane (C2H6).

A molecular sieve PLOT column is used for the separation of the permanent gases helium (He), hydrogen (H2), oxygen (O2), nitrogen (N2) and methane (CH4).

A thick film WCOT ) column coated with an apolar phase is used for the separation of the C3 to C8 (and heavier) hydrocarbons.

The permanent gases helium (He), hydrogen (H2), oxygen (O2), nitrogen (N2) and methane (CH4) are detected with a thermal conductivity detector (TCD). The C2 to C8 hydrocarbons are detected with a flame ionization detector (FID).

• • 1. Auxiliary gases

Carbon monoxide (CO) and carbon dioxide (CO2) are detected using an FID after reduction of the components to CH4 by a methanizer. Use of a methanizer, makes it possible to detect CO and CO2 at a mole fractions greater than 0,001 %. If the samples do not include CO or CO2 or if the CO and/or the CO2 mole fraction exceeds 0,02 %, a methanizer is not required. CO and CO2 may then alternatively be detected with the TCD.

When analyzing natural gas substitutes, the PLOT column described in 3.1 can also be used for the separation of ethyne (C2H2) and ethene (C2H4) and the molecular sieve PLOT column can also be used for the analysis of carbon monoxide (CO).

• 1. Materials
     1. Carrier gases
        1. Argon (Ar)

99,999 % pure, free from oxygen and water.

• • • 1. Nitrogen (N2)

99,999 % pure

• • • 1. Helium (He)

99,999 % pure

• • 1. Auxiliary gases
       1. For FID detection:
          1. Nitrogen (N₂)

Or helium (He), W 99,996 % pure.

• • • • 1. Air

Free from hydrocarbon impurities, i.e. the mole fraction of hydrocarbons < 1 × 10^-4 %.

• • • • 1. Hydrogen (H₂)

99,999 % pure, free from corrosive gases and organic compounds.

• • • 1. For methanizer

Optional, when analysing natural gas substitutes:

• • • • 1. Hydrogen,

99,999 % pure (may also be used as make up gas).

• • • • 1. Pressurized laboratory air

For the operation of pneumatically actuated valves.

• • 1. Reference materials
       1. Working reference gas mixture (WRM)

The composition of which shall be chosen to be similar to the anticipated composition of the sample.

Mole fractions of the components shall not differ by more than the relative deviations stated in Table D.2.

A cylinder of distributed natural gas, containing all the components measured by this method may also be used as the WRM. Prepare the WRM in accordance with ISO 6142 and/or certify it in accordance with ISO 6143. The WRM shall contain at least nitrogen, carbon dioxide, methane, ethane, propane, i-butane, n-butane. In the case of an indirect determination, the working reference gas mixture shall contain the reference component with a concentration in agreement with the expected concentration range. Consequently, it may be necessary to use more than one WRM.

Table D.2 — Relative deviation between sample and WRM

• • • 1. Performance test gases.
         1. For methanizer operation

Optional, consisting of a volume fraction of 0,001 % to 0,02 % each of CH₄, CO and CO₂ in helium, for use when analysing natural gas substitutes.

• • • • 1. Gas containing benzene and cyclohexane

For use in verifying peak resolution.

• • • • 1. Gas containing hydrogen and helium

For use in verifying peak resolution.

• 1. Apparatus
     1. Gas chromatograph system(s)

Consisting of the following components:

• • • 1. Two column ovens

For temperature-programmed operation, capable of following a given linear temperature gradient (see Table D.3).

The columns may either be installed in a dual-oven gas chromatograph or in two separate instruments. The analyser should be capable of independently controlling the temperatures of both column ovens.

• • • • 1. Instrument 1 oven

Containing the PLOT precolumn and the molecular sieve column (see Figure D.1, Figure D.2 and Figure D.3).

Instrument 1 may alternatively be equipped with a column oven for isothermal operation for a temperature range from 40 °C to 140 °C and capable of maintaining the temperature to within ±0,1 °C at any point inside the oven chamber.

• • • • 1. Instrument 2 oven,

Containing the WCOT column.

Key

Figure D.1 — Schematic diagram of the column configuration at the time of sample injection

Key

Figure D.2 — Schematic diagram of the column configuration for the determination of CO2 and C2

Key

Figure D.3 — Schematic diagram of the column configuration for the determination

• • • 1. Flow regulators

Providing suitable flow rates for capillary columns.

• • • 1. Gas sampling valves (GSV)

Maintained at a constant temperature to within ±0,5 °C.

Sample loop volumes of about 0,25 ml may be used in conjunction with capillary split devices. Alternatively, microvalves with internal sample loops may be used without split devices.

• • • 1. Valveless or micro-valve column-switching system

Suitable for backflushing.

Examples of possible configurations of the column switching system are shown in Figure D.4 and Figure D.5.

• • • 1. Thermal conductivity detector (TCD) and flame ionization detector (FID)

Having a time constant and internal volume appropriate for operation with capillary columns. For analyses performed on two separate gas chromatographs, the instruments shall be equipped as follows:

• • • • 1. Instrument 1 detectors

A TCD and an FID.

• • • • 1. Instrument 2 detector

An additional FID.

• • • 1. Data acquisition system

Of suitable resolution and time constant, capable of automatic registration of analyses.

• • • 1. Methanizer

Optional, to catalytically reduce on-line carbon monoxide CO and carbon dioxide CO₂ to methane CH₄ when analysing natural gas substitutes.

These components can then be detected sensitively by the FID. A methanizer is not required if the samples do not contain CO.

If a methanizer is not installed, CO₂ and C₂H₆ are detected using the reference cell of the TCD, see Figure D.5, thus giving reversed (negative) peaks for these components. All other components are backflushed from the precolumn.

Determine the conversion efficiency of the methanizer (nickel catalyst) by injecting a test sample containing a known amount (volume fraction of 0,001 % to 0,02 %) of methane, carbon monoxide and carbon dioxide. If necessary, adjust the catalyst temperature to optimize conversion efficiency and peak symmetry. Also adjust the H₂ flow to optimize sensitivity. Under optimum operating conditions the methanizer has a conversion efficiency of nearly 100 %. It is recommended to determine the stability of the methanizer.

Small amounts of H₂S and probably other sulfur compounds reaching the methanizer cause immediate deactivation of the catalyst bed. For this reason H₂S shall be cut off by suitable column switching. A poisoned catalyst is identified by the onset of tailing on both CO and CO₂ peaks.

A periodic verification of the absence of leak is recommended. This can be performed by verifying that the injection of the carrier gas does not lead to any nitrogen or oxygen peak.

• • 1. Capillary columns

Consisting of the following:

• • • 1. PLOT fused silica capillary precolumn,

For the separation of air, CO₂, C₂H₂, C₂H₄, C₂H₆ and C₃H₈.

Sufficient resolution between CH₄ and CO₂ is required to enable column switching.

A 25 m × 0,53 mm i.d., with a 20 µm phase thickness, PoraPLOT U^[3] column is recommended because it provides good separation of these components.

• • • 1. Molecular sieve PLOT fused silica capillary column