This document gives an overview of the challenges and common pitfalls of calibration in trace and ultratrace gas analysis. Special attention is directed to the principle of permeation and the technical solutions that are currently available for reliable preparation of gas standards.

1. 1
CALIBRATION IN
GAS ANALYSIS
Practical Guide
For Everyone Involved in
Quantitative Gas Analysis
www.is-x.com
@ J. Vercammen

2. 2
SUMMARY
This document gives an overview of
the challenges and common pitfalls
of calibration in trace and ultratrace
gas analysis.
Special attention is directed to the
principle of permeation and the
technical solutions that are currently
available for reliable preparation of
gas standards.
Accurate and reproducible
analysis of permanent gases
and volatiles is challenging.
“

3. 3
+ Sampling Bags
+ Dynamic
Blending
+ Permeation
+ Schematics
+ Details
+ Specifications
+ Dilution Strategy
+ Standard
Addition
+ Tube
Configurations
+ Choosing the
Right Tube
02 03 04
Preparation of
Gas Standards
Pages 11-15
Hardware
Solution
Pages 16-20
Permeation
in Practice
Pages 21-28
+ Definition
+ Case Study
+ Key Performance
Denominators
Performance
Validation
Pages 5-10
01
TABLE OF CONTENTS
INTRO
Performance
Validation
Page 4

4. 4
THE MISPERCEPTION OF ANALYTICAL SIMPLICITY
The misperception of
analytical simplicity, which is
conveyed by popular TV
shows, such as CSI: and CSI:
Miami, is considered reality
by a majority of users.
Due to the introduction of ever
faster, more sensitive and more
selective gas analyzers, such as
specific sensors, standalone MS
devices and high performance
GC/MS systems, the analysis of
gaseous and/or volatile
components has become
commonplace in a wide variety of
application domains.
Most of these systems are black
boxes, which can be implemented
almost immediately and require no
specific analytical knowledge at all.
In order to guarantee reliable
measurement results and correct
conclusions in function of time, any
analytical technique, albeit one that
uses a basic sensor or a complex
analyzer, demands that these
requirements have been met:
 Method development by an
experienced user
 Comprehensive method
validation
 Traceable quality control
according international
standards
 Extensive training and support
If one of these requirements cannot be met or has been granted less importance,
method reliability is greatly compromised.
“

5. 55
Performance Validation
A comprehensive set of tests that are carried out to determine
whether an instrument is suitable for the intended use.
PART 1
“

6. 66
Validation results are measured against the minimal
requirements imposed by the reference method.
Several international guidelines are available that describe the minimum set of tests and procedures that
need to be carried out during the performance validation of a trace gas analyzer. Generally speaking, at
least the following parameters need to be determined:
 Response linearity Construction and statistical evaluation of calibration curves
 Sensitivity Limits-of-detection for individual contaminants
 Reproducibility At various concentration levels
 Recovery Spiking experiments using real matrix
+ LINK: ANALYTICAL VALIDATION
“

7. 7
CASE STUDY: CATALYST CONTAMINANTS
Catalysts are ubiquitous in the
petrochemical industry. They are
applied to control and accelerate
various chemical reactions, such as for
example the polymerization of
ethylene and propylene to
polyethylene and polypropylene,
respectively. New generation of
catalysts are highly effective, but at
the same time extremely prone to
contamination by traces of
contaminants that are present in the
feedstock.
In order to determine these impurities
at the ultratrace levels at which they
occur (viz. Table 1), substantial
measures need to be taken,
particularly with respect to the
calibration performance of the
instrument applied to measure them.
All of these chemicals are reactive and/or polar, and do not
store reliably as static mixtures in cylinders.
Name Structure Specification
Arsine AsH3 < 20 ppb
Phosphine PH3 < 20 ppb
Ammonia NH3 < 100 ppb
Hydrogen sulfide H2S < 20 ppb
Carbonyl sulfide COS < 20 ppb
Nitrogen dioxide NO2 < 50 ppb
Hydrogen cyanide HCN < 100 ppb
Hydrogen fluoride HF < 200 ppb
Hydrogen chloride HCl < 200 ppb
Phosgene COCl2 < 50 ppb
Sulfur dioxide SO2 < 50 ppb
“

8. 8
Several factors need to be addressed during the performance validation of a
trace gas analyzer. Generally speaking, these factors are either related to the
nature of target contaminant (compound-specific factors) or the relative
composition and complexity of the sample matrix (matrix-specific factors).
Compound-specific factors:
 Compound polarity
 Compound reactivity
Matrix-specific factors:
 Relative humidity
 Matrix contains the target component (unknown concentration)
 Matrix is highly reactive and unstable
 Components react with each other
KEY PERFORMANCE DENOMINATORS
Immediate impact on instrument
performance characteristics. De facto
determined by applied compound
delivery and dilution technique(s).
Recovery experiments with real matrix
are required to evaluate the true impact
of matrix composition on system
performance. Be ware of experimental
design and data interpretation.

9. 9
Next to the factors described in the previous section,
there’s one other important factor to take into account
during the validation of a trace gas analyzer. This factor,
which exhibits an exponential effect on the factors, largely
determines the outcome of the validation scheme. This
factor is the concentration level at which the target
contaminants need to be determined.
It’s quite clear that percentage levels need an entirely
different strategy compared to ppm’s, but do you really
understand the difference between ppm, ppb and ppt?
Ppm’s are still quite tangible for most analysts but ppb’s
and ppt’s that’s an entirely different ballgame. Did you
know, for example, that 1 ppb corresponds to 1 second in
32 years? And 1 ppt to 1 second in 32,000 years?!
The concentration level of the target contaminant largely
determines the outcome of the validation scheme.
“

10. 10
One ppt = one secondin 32,000 years
The requirements imposed by
legislation are often extremely strict.
Just try to imagine that you are
responsible to find that single grain
of sand on a sandy beach and that
you can’t make any mistakes.
Quite demanding isn’t it?
Let’s give it a try:
Can you find the orange dot?

11. 1111
Preparation of Gas Standards
Inaccuracies due to miscalibration are prime sources
for errors in quantitative gas analysis.
PART 2
“

12. 12
SAMPLING BAGS
Stepwise dilution by means of
inflatable plastic bags, such as Tedlar
or ALTEF, is a very popular approach
to prepare gas standards for
quantitative analysis.
Unfortunately, this procedure is
highly unsuitable for accurate analysis
of components at trace and ultratrace
levels. Moreover, more reactive
components, such as aldehydes,
amines and sulfur-containing
components might be entirely lost.
A detailed study with respect to
the difficulties related to the use
of sample bas was carried out by
Restek scientists in 2011.
+ LINK: GAS SAMPLING BAGS

13. 13
DYNAMIC BLENDING
Unfortunately, dynamic blending is characterized by its own
set of particular problems.
Calibration mixtures with a complexity as required for the analysis of catalyst
contaminants, are typically dynamically blended and used immediately.
Unfortunately, dynamic blending is characterized by its own set of particular
problems.
To create a 1 ppm-by-volume (ppmv) mixture by direct blending in a single
stage, for example, it is required to measure and control 1 mL/min of
component vapour and blend it into 1,000 L/min of matrix gas. Obviously,
this is not a practical solution. Generally, multiple dilution stages are required
and each stage adds additional error. But even then the problem remains of
accurately controlling a small flow of component vapour.
“

14. 14
Permeation tubes are devices that act as a flow control mechanism for
dispensing a very small flow of permeate vapour through a polymeric
membrane, usually Teflon.
The component is held in contact with one side of the polymeric membrane.
Component vapour passes through the membrane at a rate determined by the
permeability of the membrane and the vapour pressure of the component.
The membrane permeability is function of membrane temperature, so when
the component is a liquid sealed inside a permeation tube, the component
vapour flow, i.e. emission rate, is set by the operating temperature of the tube.
PERMEATION
Seal
Seal
Gas Phase
Permeable
Liquid Phase
Component vapour passes through the membrane at a rate
determined by the permeability of the membrane and the
vapour pressure of the component.
“

15. 15
Matrix gas flowing over the outer side of the membrane, i.e. around the
permeation tube, mixes with the emitted vapour to form a trace concentration
mixture. For a stable component, the emission rate is extremely steady as long
as there is visible liquid in the tube.
To measure the emission rate, the tube is held at operating temperature in a
steady flow of carrier gas and weighed periodically over carefully timed
intervals. The rate of weight loss is the emission rate of the tube. Since
temperature and weight loss measurements are traceable to NIST, the
emission rate is traceable, too.
PERMEATION
The emission rate is extremely steady as long as there is visble
liquid in the tube.
“
Dilution Flow

16. 1616
Hardware Solution
Essentially one needs only to control tube temperature and
matrix gas flow to create a known, reproducible mixture
PART 3
“

17. 17
SCHEMATICS
Careful attention to a few instrumental features will greatly
improve its overall applicability and user-friendliness.
Permeation tubes are extremely
straightforward devices, and are very
easy to use in practice. Essentially one
needs only to accurately control tube
temperature and matrix gas flow to
create a known, reproducible mixture.
Careful attention to a few
instrumental features will,
however, greatly improve its
overall applicability and user-
friendliness.
A state-of-the-art gas calibration
system is micro-processor controlled,
modular and can be used in a variety
of applications. The system shown
here can be equipped with up to four
separate flow modules, all software-
controlled, as well as seven gas inlets.
MFCDilution
flow
Standby flow
MFC
MFC
MFCGas inlet (0 to 5)
Mixing chamber
To waste
To analyzer
Permeation
oven
“

18. 18
The flow path provides a continuous, uninterruptible,
fixed carrier flow over the permeation tube. With fixed
carrier flow, the component concentration is steady and
surfaces in the system equilibrate assuring rapid response
while minimizing system bias system.
Carrier flow is continuous, in order to avoid quick
contamination of the permeation chamber and its
surrounding elements with high concentrations of
components. Moreover, the carrier flow can be switched
to vent hence isolating the permeation oven from the rest
of the system. In this case, matrix flow is “zero gas”,
allowing evaluation of system blank levels. This feature is
particularly suitable when working with non-zero matrix
gas to prepare standards.
The temperature of the permeation oven is carefully
controlled. The emission rate from a liquid fed
permeation tube typically varies about 10% per degree
Celsius, so a stable operating temperature is essential for
mixture accuracy and stability. Control to  0.1°C assures
emission accuracy within 1%. Higher accuracy
temperature control has limited value because other error
sources then dominate overall accuracy of the final
mixture.
DETAILS

19. 19
Measurement and control of matrix gas flow is often the
most significant error source, especially when the matrix is
some gas other than nitrogen.
Common practice is that flow meters are actually
calibrated for nitrogen (or air), and accuracy is specified as
a percent-of-full-scale reading. For example, a 5 L/min
flow meter may show 5 L/min (1% = 50 mL/min) at full
scale, but at 0.25 L/min the accuracy stays at 50 mL/min,
which results in a substantially larger error of 20%.
Even more troublesome is using the flow meter for a non-
nitrogen matrix. Common practice then is to apply a
“flow factor,” i.e. a correction factor to compensate for
the physical characteristics of the matrix gas on the flow
meter. The flow factor correction is based on the ratio of
physical factors to those of nitrogen. The result is that the
error band is no longer well defined.
Flow meters used for permeation systems are calibrated at
5 or more points to assure  1% flow measurement
accuracy of the full meter range. When used with a non-
nitrogen matrix, the meter should be calibrated with the
actual matrix gas.
The great strength of the permeation tube
method is that the permeation tube itself
solves the most difficult flow control
problem, i.e. controlling the miniscule flow of
a wide range of vapours used as component
gases.
“
DETAILS

20. 20
SPECIFICATIONS
Accuracy MFC  1% full scale
Repeatability  0.25%
Operating temperatures 0 – 40°C
Power consumption Typ. 15W / 50W max
Dimensions 19”; LCD display, keyboard on front panel
Weight 12 to 18 kg
Permeation temperature Standard 30 to 110°C
Accuracy  0.05%
Positions 6 tubes
Standby flow 200 mL/min
Dilution flow Max 5 L/min
Data transmission RS232 (8 outputs, 24 inputs)
Gas inlets/outlets Up to 7 gas inlets via solenoid valves
Typical specifications of a state-of-the-art digital dilution unit with permeation capabilities
are presented in the table below.
Please contact us at info@is-x.com if further details are required.

21. 2121
Permeation in Practice
Essentially one needs only to control tube temperature and
matrix gas flow to create a known, reproducible mixture
PART 4
“

22. 22
DILUTION STRATEGY
Permeation tubes can be used to
prepare concentrations that are quite
low with a single step dilution. For
example, a tube emitting 50 nL/min
mixed into a 1 L/min flow creates a
50 ppbv mixture. Diluted into 5
L/min, the same emission gives a
concentration of 10 ppbv.
Theoretically, one could choose a 1
nL/min tube and dilute it in 10
L/min to get a 100 pptv mixture
(101:1 dilution) in a single dilution
step. But, as with all great schemes,
there are practical limits. With
permeation tubes, the basic limit is
the minimum emission rate that can
be actually measured. As mentioned
above, emission rate is usually
determined by measuring the rate of
weight loss.
A typical disposable permeation tube
weighs 10 – 15 mg, so the static
weight is well within the range of a
semi-micro balance. The minimum
readable weight change is 0.01 mg,
and a weight change of 1 mg is
needed to assure the possibility of
measuring weight loss at  1%
accuracy. For a 100 ng/min emission
rate, the minimum interval between
weightings is 7 days. At 10 ng/min
the minimum interval is over 2
months! The minimum time for
emission rate certification is then
about 8 months.
Emission rate is usually
determined by measuring the
rate of weight loss.
“

23. 23
Measuring extremely low weight loss
rates is difficult even using a micro-
balance. The problem then shifts to
cleanliness and gas purity. During the
certification process the Teflon
membrane develops a static charge,
which attracts any particles in the
sweep gas flowing over the tube
during the emission measurement
process. This introduces additional
error. The practical limit for single
stage dilution using permeation tubes
is about 10:1. Best practice is to use a
secondary dilution step for lower
concentrations.
As concentrations go ever downward,
the ultimate problem is often the
purity of the matrix gas. Unknown
background contaminations can
introduce large random errors.
DILUTION STRATEGY

24. 24
During the course of performance validation, recovery experiments need to be carried out that use real sample matrix.
In order to avoid problems related to non-zero matrix gas, the method of standard addition has to be used. To use this
method, first observe the detection system response to the matrix gas alone, i.e., the “zero addition response.” Then
add the permeation tube output to the matrix gas and observe the response due to the added component. Since the
concentration in the matrix is constant, several “standard addition responses” can be obtained by varying the matrix gas
flow.
STANDARD ADDITION

25. 25
TUBE CONFIGURATIONS
Permeation tubes are available for approximately 500
chemicals with concentrations ranging from the sub-ppb
range to over 1000 ppm. Tubes can be provided in
different configurations wide variable lengths, depending
on compound chemistry and application.
The basic configuration uses disposable PTFE tubes,
filled with the appropriate chemical. Standard rate
disposable permeation tubes (SRT) generate moderate
concentrations of high vapour pressure compounds.
These compounds include SO2, H2S, NH3, Cl2, etc.
Additionally, SRT permeation tubes can generate very low
concentrations of low vapour pressure compounds such
as carbon disulfide, methylene chloride and nitric acid.
Typical emission rates yield concentrations of 1 to 10
ppm depending on the compound, operating temperature
and dilution flow rate.
Other disposable configurations include high rate
disposable permeation tubes (HRT) and extra life
disposable permeation tubes (ELF). The latter are fitted
with a 3 inch stainless steel reservoir for storing extra
component. Due to this impermeable reservoir, tube
lifetime is more than doubled.

26. 26
TUBE CONFIGURATIONS
Emission rate Accuracy
> 500 ng/min  2%
100 – 500 ng/min  2% or 5 ng/min
20 – 100 ng/min  5% or 2 ng/min
In addition, it is also possible to use refillable permeation
tubes (either gas-filled or liquid-filled).
Irrespective of the type applied, each tube is numbered
and laboratory certified at the temperatures specified. A
certificate listing a description of the tube and its
emission rate is supplied with each certified tube.
Emission rate data for each tube is kept on file at the
factory for future reference.
The accuracy of certification depends on the type of
tube, the component and the emission rate level. For most
components, the accuracy of disposable tubes is as
follows:

27. 27
CHOOSING THE RIGHT TUBE
Choosing the right tube or set of tubes depends on the concentration levels one wants to achieve. This is determined
by compound chemistry and reactivity, permeation rate in function of temperature, maximal tube length and maximal
dilution flow. Since a typical permeation oven can accommodate up to six tubes, careful calculations have to be carried
in order to determine optimum conditions for each of the tubes.
We have a compiled a dedicated spreadsheet that contains all the factors to take into when deciding on the most
suitable permeation tube conditions.
Sent an e-mail to info@is-x.com for a free copy of our permeation spreadsheet.

28. 28
COMPOUND LIST
Nowadays, a wide range of chemicals are applied in the permeation format.
Next to these standard tubes, it is also possible to fill your own tubes.
Please visit www.kin-tek.com/chemlist.html for the
most recent list of available permeation tubes.

29. 2929
Conclusion
Essentially one needs only to control tube temperature and
matrix gas flow to create a known, reproducible mixture
“

30. 30
It is worthless evaluating the performance of a trace gas analyzer without proper control of the applied calibration
procedure. In this document we have described and illustrated the use of permeation tubes as a simple and
straightforward means to prepare standards with great accuracy and precision.
Permeation tubes are available that contain a large variety of chemicals and different concentration ranges.
Please do not hesitate to contact us for further reference.
This document was prepared by Joeri Vercammen.
ISX