You are on page 1of 8

TDLAS

Application Note

PROCESS INSTRUMENTS

Use of TDLAS Based Analyzers for Real-Time Process


Control in Ethylene Oxide Production
Ethylene oxide (EO), which is derived from
ethylene, is a high volume chemical feed stock
used to make a variety of other chemicals.
The most important in terms of volume is the
production of ethylene glycol which consumes
more than 60% of the ethylene oxide produced.
Ethylene glycol is primarily used as an antifreeze

Ethylene oxide is produced by the carefully


controlled oxidation of ethylene over a silver
catalyst. Controlling the concentrations of the
compounds in the reactor and the temperature/
pressure is critical. If the conditions are too mild,
little ethylene oxide is produced but if they are
too severe, the reaction runs away and carbon

agent in motor vehicle radiators and at airports to


remove ice from runways and aircraft.

dioxide and water are the only products. If silver


oxide is used as a catalyst the dominant reaction
is formation of ethylene oxide.

Ethylene oxide can also be converted by


thermal hydration into mixture of glycols, namely
monoethylene glycol, diethylene glycol and
triethylene glycol which are used in cosmetic and
pharmaceutical industries as a base material to
carry active ingredients. These glycols are also
used in the ink and printing industry as antifoam
agents in distillation processes. In addition
ethylene oxide is also used for production of
ethoxylates and polyether polyols which are used
in the manufacture of biodegradable detergents
and polyurethane resins and as a solvent for
paints and lacquers. Another use of ethylene
oxide is in the production of ethanolamine by
reaction with ammonia. Ethanolamine is used
as a feedstock in the production of detergents,
emulsifiers, polishes, pharmaceuticals, corrosion
inhibitors and other chemical intermediates.

The Production of Ethylene Oxide


To operate with maximum yield and safe
conditions ethylene oxide production requires
a sophisticated process control system. There
are a number of constituents in the process that
must be analytically measured in real time. This
application note will describe the use of Tunable
Diode Laser Absorption Spectroscopy (TDLAS) for
the measurement of oxygen, carbon dioxide and
ethylene in the ethylene oxide production process.

Some ethylene still ends up being further oxidized


(as much as 25% in some processes) resulting
in the production of carbon dioxide and water as
main by-products through this reaction.

Both reactions are exothermic, each gram mole


of ethylene converted to ethylene oxide produces
25 kcal of heat and each gram mole of ethylene
converted to carbon dioxide and water produces
316 kcal of heat. Temperature is a prime concern
in the process design. The temperature of the
oxidation is controlled by a heat exchanger built
into reactor. The reaction temperature is controlled
to 250290C and is under a pressure range of
1525 atm. The residence time of the ethylene/
oxygen feed in the reactor is only about one
second.

www.ametekpi.com

TDLAS

The ethylene oxide produced represents only a


small fraction of the total effluent stream leaving
the reactor with the majority being unconverted
ethylene. As a result, ethylene oxide manufacturing
requires recycle of ethylene back into the reactor.
The limit of the conversion of ethylene and oxygen
to ethylene oxide during each pass through the
reactor is determined by the oxygen concentration
in the reactor. Normally, ethylene concentrations in
the reactor are in the range 2040% by volume and
oxygen concentration is dictated by the flammable
limit. To prevent an explosive condition the oxygen
must be limited to less than ~10%. To increase the
flammable limit, diluents in the form of nitrogen,
argon, methane, ethane or natural gas are added to

Gas Component,
mole% or ppm
Ethylene

35

33

Oxygen

Ethylene Oxide

0.01

2.1

Carbon Dioxide

Water

0.2

0.5

Argon

Nitrogen

Methane

44.8

44.8

Ethane

0.2

0.2

Ethylene Dichloride

0.2 ppm

0.2 ppm

Vinyl Chloride

1.2 ppm

1.2 ppm

The ethylene oxide production process is shown in


Figure 1. The ethylene oxide product is recovered
from the reactor effluent gases by absorption in
aqueous absorbent. The ethylene oxide is then
stripped from the absorbent and separated from the
residual water, and light ends by fractionation. The
aqueous absorbent from the ethylene oxide stripper
is recycled back to ethylene oxide absorber. The
overhead stream containing ethylene and carbon
dioxide is sent to carbon dioxide absorber where
carbon dioxide is partly removed from the recycling
gas by absorption in a hot potassium carbonate
solution.

As mentioned previously, carbon dioxide is formed in


the reaction. This carbon dioxide is mostly removed
in the processing of the recycle stream but it can be
partially left in the recycle gas to act as a diluent. To
optimize the ethylene oxide yield, catalyst selectivity
can be increase by adding moderators such as
1,2-dichloroethane or vinyl chloride. Typical gas
stream compositions for the inlet and outlet of the
reactor are shown in Table 1.

Methane %
Ethylene %
Oxygen %
Carbon Dioxide %

Ethylene
Recycle

Ethylene
Oxygen
Diluent
Moderator
Methane %
Ethylene %
Oxygen %

Reactor
Outlet

Table 1. Typical Stream Composition for the Reactor with



Methane Ballast

the reactor mixture.

Reactor
Inlet

Methane %
Ethylene %
Oxygen %
Carbon Dioxide %
EO %

4
CO2
Absorber

Carbon
Dioxide

CO2
Stripper

Off Gas
5

2
Reactor

Methane ppm
Ethylene ppm
EO ppm

EO
Absorber

EO Desorber

Figure 1. Typical ethylene oxide process with key analysis points.

Methane ppm
Ethylene ppm
EO ppm

Ethylene
Oxide

experimental protocol is simply a matter of uploading


a file. The Model 5100 HD offers significant
advantages over prior analytical approaches for
customers monitoring ethylene, oxygen and carbon
dioxide in ethylene oxide manufacturing processes.

The process is controlled to optimize the conversion


of ethylene to ethylene oxide but at the same time
must assure that there are safe levels of oxygen in
the reactor. Carbon dioxide in the reactor outlet is
an undesirable side reaction product. The carbon
dioxide creation consumes valuable ethylene and
needs to be minimized through process control.
Optimizing the relative concentrations of oxygen and
ethylene also requires real-time process analyzers.
Historically ethylene oxide plant analyzers have
included paramagnetic oxygen analyzers for oxygen
measurements at the inlet and outlet of the reactor,
infrared filter photometers for inlet ethylene feed
control and Process Gas Chromatographs (PGCs) to
measure all of the components of the gas at reactor
outlet. Recent advances in Tunable Laser Diode
Laser Absorption Spectroscopy (TDLAS) based
analyzers can provide measurements for closed
loop process control for product yield optimization,
plant monitoring and alarms, quality control and
environmental safety.

The purpose of oxygen measurement in the


ethylene oxide reactor is to maximize the oxygen
concentration while controlling to a limit of around
8% to eliminate the possibility of an explosive
condition. The typical oxygen concentration
measurement range for this application is 012 %
by volume. There are two primary locations for the
measurement of oxygen in ethylene oxide plants.
These locations are inlet and outlet of the ethylene
oxide reactor.

Oxygen Measurements

Paramagnetic oxygen analyzers have been


traditionally used in the ethylene oxide
manufacturing process. According to Siemens, a
leading manufacturer of paramagnetic analyzers,
the common practice is to use three paramagnetic
analyzers per sample point to assure a 99% up time.
TDLAS based oxygen analyzers can be successfully
used for this application without the need for this
level of redundancy. TDLAS is non-contact analysis
technique with long term stability, high specificity,
selectivity and reliability. TDLAS analyzers offer the

Obviously, control of the ethylene oxide plant


performance cannot be solely limited to TDLAS
analyzers. However, there are applications
where they can be used very effectively including
measurements of ethylene, oxygen and carbon
dioxide.
AMETEKs Model 5100 HD shown in Figure 2 is
an extractive type gas analyzer designed for hot/
wet sample analysis. There is no requirement for
complicated sample conditioning although a heated
transfer line may be required to keep the sample in
the gas phase. The 5100 HD offers high specificity
and sensitivity and uses a sealed reference cell
for laser line locking and continuous analyzer
performance verification. The analyzer uses a
digital implementation of Wavelength Modulation
Spectroscopy (WMS), where changing the

Figure 2. 5100 HD TDLAS analyzer

TDLAS

From the carbon dioxide absorber overhead the


recycle ethylene gas is sent back to the reactor for
a new cycle. The absorbent from the carbon dioxide
absorber is sent to carbon dioxide stripper where
carbon dioxide is separated and then vented to
atmosphere. The potassium carbonate is returned to
the carbon dioxide absorber.

TDLAS

advantages of faster response time and low cost


of installation and ownership in comparison with
conventional techniques such as paramagnetic
devices and PGCs.

sample compartment oven. The laser, reference cell


and photo diodes are located in the main electronics
compartment, isolated from the heated sample
compartment.

The AMETEK 5100 HD TDLAS oxygen analyzer


uses a vertical cavity surface emitting diode
(VCSEL) laser based source. Like all AMETEK
TDLAS analyzers, it employs an all-digital protocol
for the modulation of the laser drive current and
the demodulation of the detector response and a
reference cell. The reference cell provides laser
line lock and allows continuous verification of the
performance of the analyzer.

The analytical performance of the TDLAS oxygen


analyzer has been evaluated at various customer
sites within the required specification ranges of
pressure and temperature. There are no spectral
interferences from other compounds in the TDLAS
measurement of oxygen in ethylene oxide gas
production. The data shown in Figure 4 represent
the response of the instrument to a series of oxygen
challenges in the concentration range of interest.
The duration of each of the challenges was 60

The optical set up for the TDLAS measurements


is shown in Figure 3. The beam from the laser
diode is collimated with a plano-convex lens, which
is used as the window of the sample cell. Since
the absorption signal is weak, an extended path
length sample cell is used. Multiple reflections of
the laser light through the sample cell is achieved
using a spherical mirror with a flat mirror at the focal
point. Another long focus lens is used as an optical
output window of the sample cell and operates as
a condenser. On the output of the sample cell, the
beam is divided in to two parts by a beam splitter.
Signals from sample and reference cell silicon
photodiode detectors are input to separate channels
of the main electronics unit.

minutes or more with return to the 0% gas baseline


value between challenges. The speed of the T90
response time was 20 seconds and was determined
using a flow rate of 2L/min. The data acquisition

Figure 3. Optical Set Up for AMETEK 5100 HD TDLAS Oxygen Analyzer.

The sample cell temperature is controlled with


accuracy of +/- 0.1C and can be set in the range
of 40C150C by setting the temperature of the

Figure 4. Oxygen in Ethylene Methane Gas Stream.


Response to a Series of Concentration Challenges (A),
Speed of the Response (B).

The AMETEK 5100 HD oxygen analyzer


demonstrates a high level of repeatability.
Instrumental drift during 15 hours is less than 0.05%
of the oxygen concentration. These data are shown
in Figure 5.

0.06
Average = -0.006%, STD = 0.013%

0.04

sensitivity of PGCs the relatively long measurement


time (2-3 minutes) is a serious disadvantage.
TDLAS-based measurements provide real time
monitoring with data acquisition rate of typically 2
seconds. As was mentioned above the residence
time of the feed in the reactor is only about 1
second. It is known that during regular plant startups
with a new catalyst the concentration of carbon
dioxide can change suddenly. To maximize the yield
of ethylene oxide it is important to have a real time
monitoring of the carbon dioxide concentration.

0.02
0.00

Oxygen, %

-0.02
-0.04
-0.06

10

12

14

Time, hours
Figure 5. AMETEK 5100 HD Oxygen Analyzer Drift.

A schematic representation of the AMETEK 5100


HD analyzer for measurements of carbon dioxide
is shown in Figure 6. Examples of the 5100 HD
carbon dioxide analyzers performance for the 05%
is shown in Figure 7. The data shown in Figure 7
are the responses of the instrument to a series of
carbon dioxide challenges over the concentration
ranges of interest. The duration for each challenge
was several minutes. The response time (T90)
was 30 seconds at a flow rate of 2L/min. The data
acquisition rate was 2 seconds per measurement.

During one year of field trials in a similar control


application under the harsh conditions of a refinery
waste water treatment plant oil/water separator head
space, the analyzer had zero failures versus nine
failures of the paramagnetic analyzer running in
parallel.

Carbon Dioxide Measurements


There are several process control applications for
carbon dioxide measurements at ethylene oxide
plants:
Reactor inlet (0-5% CO2 volume)
Reactor outlet (0-5% CO2 volume)
Ethylene oxide stripper overhead (030%
CO2 volume)
Carbon dioxide stripper overhead (030%
CO2 volume)
5

TDLAS

It is known that at high concentrations carbon


dioxide adversely affects the selectivity of the
catalyst used in ethylene oxide production. Despite
the fact that the concentration of carbon dioxide
in the reactor is relatively low in comparison with
ethylene and methane, it is important to monitor
the concentration to maximize the yield of ethylene
oxide. As undesired product of the reaction, carbon
dioxide must be continuously removed from the
ethylene recycle stream by the carbon dioxide
absorber stripper system in order to control
the carbon dioxide levels in the reactor. Process
Gas Chromatographs (PGCs) have been used
historically to monitor carbon dioxide in the ethylene
oxide production process. However, despite the high

rate was 2 seconds/measurement. The standard


deviation of the reading on each of the challenges
was between 0.01% and 0.02% of the oxygen
concentration. The measured accuracy over the
concentration range was better than 0.1%.

TDLAS

Ethylene Measurements
The optimal conditions for the ethylene oxide
production process may be determined by real time
monitoring of ethylene concentrations at the different
locations of the process. Fast and accurate analysis
of feedstock and effluent streams are needed
to maximize conversion efficiency and catalyst
selectivity ratios. Historically, filter IR photometers
have been used to provide ethylene monitoring
in the ethylene oxide production process. Filter
photomers require frequent zero measuremens
which is undesirable for real-time process control.
TDLAS-based measurements provide real time
monitoring with a data acquisition and analysis
rate of a few seconds. As was mentioned above,
the residence time of the feed in the reactor is only
about 1 second. Also during normal plant startups
with new catalyst and unusual process events, the
concentration of ethylene can change suddenly. To
optimize the EO yield it is important to have real
time monitoring of the ethylene concentration in the
process.

Figure 6. 5100 HD Optical Set Up Used for Carbon Dioxide


Measurements.

The standard deviation of the readings was 25ppmv


for measurements in the 05% range. The accuracy
in the range 05% was 100ppmv.

Taking into account the need to measure ethylene


in the inlet and outlet of the reactor AMETEK offers
a TDLAS analyzer with dual sample cells. A photo
of the 5100 HD analyzer with two sample cells
and two sets of inlet and outlet valves is shown in
Figure 8. This configuration allows simultaneous
measurements of the ethylene concentration in two
separate gas streams with results for each stream
reported every two seconds.

Figure 7. Response to a series of concentration CO2 challenges in the


range 0 5%.

Figure 8. AMETEK 5100 HD analyzer with dual sample cells.

TDLAS

A schematic representation of the dual cell 5100 HD


analyzer is shown in Figure 9. The measurement of
ethylene is performed with two Distributed FeedBack
(DFB) lasers. The outputs of both lasers are coupled
into single-mode optical fibers, which in turn are
connected to fiber-optic beam splitters. The splitters
are used to divide the optical power in a 50/50 ratio
for use in the sample and reference measurements,
respectively. Reference cell contains a known
concentration of ethylene in a non-absorbing
matrix. The reference cell is used to lock the output
radiation wavelength of the laser for both sample
channels.
Figure 10. Validation of the ethylene analyzer performance.
Response to a series of ethylene challenges.

Figure 9. Schematic diagram of the dual cell 5100 HD analyzer for ethylene measurements.

rate of 2L/min. No significant trends or correlations


with the environmental temperature or sample
pressure were observed in the data. Over the 24
hour period, a mean value of 0.5% of ethylene, with
a standard deviation of 0.01%, was recorded; the
value of the drift was less than 0.03% ethylene.

The data shown in Figure 10 represents the


response of the instrument to a series of ethylene
challenges in a concentration range of 0-50%.
The duration of each of the challenges was
approximately 30 minutes with return to the 0%
gas baseline, which was represented by methane.
The T90 response time was 20 seconds at a flow
rate of 2L/min. A measurement for each stream is
completed every two seconds.
The standard deviation of the ethylene readings
on each of the challenges was between 0.1 % and
0.3% of the ethylene concentration. The value of the
accuracy evaluated at the levels of ethylene from 0
to 50% was in the range 0.30.5% ethylene.
Instrumental drift during 24 hours is shown in Figure
18. During the drift test, a sample with 0.5% ethylene
in nitrogen was run through the sample cell at a flow

Figure 11. Ethylene Analyzer 24 Hours Drift.

USE OF TDLAS BASED ANALYZERS FOR REAL-TIME PROCESS


CONTROL IN ETHYLENE OXIDE PRODUCTION

SALES, SERVICE AND MANUFACTURING:


USA - Pennsylvania

150 Freeport Road, Pittsburgh PA 15238 Tel: +1-412-828-9040, Fax: +1-412-826-0399

USA - Delaware

455 Corporate Blvd., Newark DE 19702 Tel: +1-302-456-4400, Fax: +1-302-456-4444

150 Freeport Road, Pittsburgh, PA 15238


Ph. +1-412-828-9040, Fax +1-422-826-0399

www.ametekpi.com

Canada - Alberta

2876 Sunridge Way N.E., Calgary AB T1Y 7H2 Tel: +1-403-235-8400, Fax: +1-403-248-3550

WORLDWIDE SALES AND SERVICE LOCATIONS:


USA - Houston, Texas
Tel: +1-713-466-4900, Fax: +1-713-849-1924
BRAZIL
Tel: +55 19 2107 4100

2014, by AMETEK, Inc.


All rights reserved. Printed in the U.S.A.
F-0438 Rev. 3 (1115)

CHINA
Beijing / Tel: +86 10 8526 2111, Fax: +86 10 8526 2141
Chengdu / Tel: +86 28 8675 8111, Fax: +86 28-8675 8141
Shanghai / Tel: +86 21 6426 8111, Fax: +86 21 6426 7818

One of a family of innovative process analyzer solutions from AMETEK Process Instruments.
Specifications subject to change without notice.

FRANCE
Tel: +33 1 30 68 89 20, Fax: +33 1 30 68 89 99
GERMANY
Tel: +49 21 59 91 36 0, Fax: +49 21 59 91 3639
INDIA
Tel: +91 80 6782 3200, Fax: +91 80 6782 3232
SINGAPORE
Tel: +65 6484 2388, Fax: +65 6481 6588

TDLAS

The monitoring of oxygen, carbon dioxide and ethylene in the ethylene oxide production process are
critical applications that demand an analyzer with a near 100% availability to maximize the production
of ethylene oxide and assure plant safety. The built-in verification feature of the 5100 HD provides a
real-time check on the analyzer performance. Fast response, no sample conditioning, and immunity
to interference from moisture make laser based AMETEK model 5100 HD a superior and dependable
choice.

Application Note

Conclusion

You might also like