Green retarder technology for the

styrene industry
Increased environmental and safety awareness in the styrene industry has
led to the development of more acceptable alternatives to toxic nitrophenolic

Lisheng Xu, Javier Florencio, Vincent Lewis and Christopher Morrison Nalco
Ana Guzman, Carmen Monfort and Ana Olivares Repsol Química Tarragona

S
tyrene is an important petrochemical prod- polymer formed in a specific production unit is a
uct that is used as a starting material for a function not only of the process temperature, but
variety of polymer products, polystyrene in also of the residence time and styrene monomer
particular. There are two main industrial routes concentration in the distillation towers. Higher
for the production of styrene; namely, the EB/ process temperatures result in increased rates of
SM route and the PO/SM (sometimes referred to polymerisation. Furthermore, increased residence
as the SM/PO) route. The EB/SM route produces times result in a greater amount of polymer
styrene through the dehydrogenation of ethyl- forming.
benzene. The PO/SM route furnishes styrene Proper polymer control in the purification
through the oxidation of ethylbenzene and section of a styrene plant is important for two
subsequent addition of propylene to co-produce main reasons. First, styrene monomer that is
propylene oxide and styrene, after dehydration polymerised is an economic loss for the producer,
of the intermediate alcohol. Ethylbenzene, used as it can no longer be sold as styrene monomer.
in both cases, is produced through an alkylation Second, if polymerisation goes unchecked, high
reaction between benzene and ethylene. As of polymer levels in styrene monomer will result in
2009, total global styrene capacity was about 30 an increase in stream viscosity. While polysty-
million t/y, of which about 80% was produced rene is soluble in styrene monomer, very high
through the EB/SM process and 20% through levels of polystyrene can make the stream overly
the PO/SM process. viscous, causing difficulties in pumping the
Regardless of specific licensed technologies for material through the process. Occasionally, high
styrene production, crude styrene produced in levels of divinylbenzene (DVB) can also result in
the reaction section needs to be purified to yield the formation of insoluble crosslinked polymer
polymer-grade styrene monomer. This is typi- that will deposit inside the tower.
cally accomplished through a series of distillation
towers designed to separate styrene from unre- Technologies to control polymer
acted ethylbenzene as well as other reaction There are two categories of compounds that are
byproducts. commonly used for polymer control. The first
category is a slow-reacting compound, typically
Polymer control in a styrene unit known as a retarder. Today, the most commonly
Styrene is a very reactive monomer. If not prop- used retarders in the industry are nitrophenolic-
erly inhibited, it will polymerise rapidly via a free based products, and in particular DNBP
radical mechanism at elevated temperatures typi- (2,4-dinitro-6-sec-butylphenol). The second
cally encountered in the styrene purification category is much faster reacting and is some-
process. The polymerisation reaction is self-initi- times referred to as a true inhibitor. The true
ating; no initiators such as peroxides are required inhibitors available on the market are mostly
to initiate the reaction process. The amount of proprietary formulations that offer a faster poly-

www.digitalrefining.com/article/1000341 PTQ Q1 2012 1

 during both normal operations
and emergency shutdowns.
Untreated
During normal plant opera-
Rate of polymer formation Retarded
tions, a faster reacting true
Inhibited
inhibitor provides superior
polymer control versus a
retarder such as DNBP. In fact,
the faster the reaction rate of
an inhibitor, the more effective
it is in reducing polymer forma-
tion. Nalco offers Prism
inhibitors that are several
orders of magnitude faster than
DNBP for polymer control in a
Time
styrene unit. Some styrene
producers only use DNBP for
Figure 1 Polymer inhibition behaviour: retarder vs inhibitor polymer control, but with less
satisfactory results. However,
when emergency shutdowns
9 -1 -1 occur, a retarder is essential in
2R . RR ≅ 10 M s (rate limit)
protecting the unit from exces-
R. + PRISM™ Inhibitor ≅ 107 -108 M-1s-1 sive polymer build-up. When a
plant loses power, it loses, in
R. + Hydroxylamines ≅ 105 M-1s-1 most cases, its ability to pump
R. + AOH (retarders) ≅ 103-104 M-1s-1 the hot process stream out of
the distillation towers. It also
R. + ≅ 102 M-1s-1 loses in most cases its ability to
inject polymer inhibitor and/or
retarder. Due to tower insula-
Table 1 Relative reaction rates tion, the hot process fluid will
remain at elevated tempera-
mer inhibition rate than DNBP. Figure 1 shows tures for an extended period of time. This long
the difference between a retarder and an inhibi- residence time can cause excessive polymer
tor. In the graph, uninhibited styrene has an build-up if not properly controlled. In severe
inherent polymerisation rate, as represented by cases, high polymer levels eventually lead to
the line marked as untreated. A retarder would solidification and can turn the distillation tower
slow the polymerisation rate and is represented contents into a solid block of polymer — a
by the red line with reduced slope. A true inhibi- devastating outcome for any facility. A fast-
tor is faster reacting than a retarder; therefore, reacting inhibitor, while excellent for controlling
the polymerisation rate when an inhibitor is used polymer during normal operations, can be
would be less than a retarder, as represented by quickly consumed during an emergency shut-
the orange line. However, due to its rapid rate of down, thus losing its ability to protect the
reaction, inhibitor is consumed much faster than tower. On the other hand, a retarder — owing
a typical retarder. Once consumed, polymerisa- to its slower reaction rate — will remain effec-
tion resumes at the rate of an uninhibited tive for longer and offer prolonged protection
system, as represented by the second half of the for the tower. The ratio between inhibitor and
orange line. retarder should be considered carefully to result
in optimal cost performance while ensuring
Optimal approach to polymer inhibition safety during an emergency shutdown.
The optimal approach to polymer control in a A summary of the relative polymerisation rates
styrene unit is to use a combination of inhibitor of various inhibitors and retarders is shown in
and retarder. This ensures polymer control Table 1.

2 PTQ Q1 2012 www.digitalrefining.com/article/1000341

Green retarder true retarder. Since a retarder is utilised to offer
As an effective retarder, DNBP has been the longer polymer control times during an emer-
product of choice for the styrene industry. It is a gency shutdown, a higher reaction rate than
reliable and economic means of protecting the DNBP is not necessarily a desirable feature. Cost
tower during emergency shutdowns. But DNBP is also an important factor in the adaptation of
is also highly toxic. It causes reproductive and green retarder technology. Although the green
developmental system damages in mammals. It retarder concept has been around for some time,
is classified by European Regulations as a CMR the costs of such products have been prohibi-
(carcinogenic, mutagenic and reproductive toxin tively high until recently, which has prevented
substances) because it is a reproductive toxin their widespread application.
category 2 and 3, a chemical that will be subject The Prism Green Retarder technology was
to further reviews for its future use. Potential developed as part of Nalco’s Prism product line
exposure to DNBP poses severe health risks. It is with these requirements in mind. Specifically, it
also soluble in water at levels that can be highly has considerably lower toxicity and lower water
toxic and therefore a potential hazard in the solubility than DNBP. Lower water solubility
environment. Initially sold as a herbicide, its use reduces the risk of contaminating the water
in agriculture has long been banned in the US system, in the case of an accidental release. In
and many other countries. While the risk of Figure 2, the colour of water in the presence of
exposure to DNBP in an industrial setting is either Prism Green Retarder or DNBP can be
certainly lower than in an agricultural setting, visually compared. In the case of green retarder,
handling DNBP is by no means safe, since it has the water colour is unchanged, while in the case
a very low LD50. Industrial accidents do occur, of DNBP the water turned to a pale yellow
and operators and the environment in the imme- colour due to the high water solubility of DNBP.
diate vicinity of the plant are always at risk as The green retarder also has negligible NOx
long as DNBP is utilised. In recent years, the generation potential. Table 2 shows a compari-
potential risks of using DNBP in a styrene plant son between DNBP and the green retarder.
have been highlighted in many countries. From a performance standpoint, the green
DNBP is also a major generator of NOx when retarder is comparable to DNBP in its ability to
incinerated. In a typical styrene plant, DNBP control polymer. Figure 3 shows the polymer inhi-
exits the unit along with heavy tars, and the tar
stream is typically used as fuel. While recovering
fuel value, burning of tar containing DNBP
produces a significant amount of NOx. Up to
one-quarter of the DNBP (by weight) is converted
to NOx when burned. NOx is a known environ-
mental pollutant, and many countries have NOx
emission limits in place.
Future sustainable development of the chemi-
cal industry is largely dependent upon how
successful the industry is in improving the
health, environment and safety standards of its
operations. This requires finding innovative ways
by which highly toxic materials can be replaced
by more environmentally friendly alternatives.
The development of the Prism Green Retarder
technology is an effort aimed at offering the
industry a green alternative to DNBP.
To develop a green retarder, there are two
main criteria. First, it must be less toxic than
DNBP and generate less NOx when burned.
Second, it should have similar polymer control Figure 2 Colour comparison of the water phase in con-
properties (reaction rate) as DNBP. It must be a tact with DNBP and green retarder

www.digitalrefining.com/article/1000341 PTQ Q1 2012 3

 Comparison of DNBP and green retarder (mechanical, operational and
chemical) approach should be
carried out to ensure maximum
Acute toxicity Water solubility, NOx potential,
Oral LD50 rats, mg/kg mg/L kg_NO/T
return on investment from the
DNBP 27 52 250 chemical treatment. This
Green retarder >5000 0.02 Negligible approach involves a detailed
plant survey from both a mechan-
* Comparison based on active component. Toxicity characteristics of ethylbenzene solvent
are not reflected in the comparison. ical and operational standpoint.
Critical parameters to take into
Table 2 consideration include tower
bottom temperatures, tower
internal design, stream composi-
0.8 tions, tower residence times, tar
Prism Green Retarder
recycle ratios, as well as others.
DNBP
Soluble polymer data are evalu-
Pol yme r con t e n t at

0.6 ated against chemical injection

st e ady st at e , %

dosages. Under the same operat-

ing conditions, the soluble
0.4
polymer level is inversely propor-
tional to the inhibitor reaction
0.2
rate and dosage. A faster reacting
inhibitor will result in a lower
soluble polymer make, as will a
0.0 higher inhibitor dosage of a
85ºC / 1 hr 85ºC / 4 hr 110ºC / 1 hr 110ºC / 2 hr 120ºC / 1 hr slower reacting inhibitor. It is
150 ppm 150 ppm 120 ppm 240 ppm 500 ppm
therefore advantageous to employ
the fastest reacting polymer
Figure 3 Polymer control performance: DNBP vs green retarder inhibitor available.
Baseline data are collected and
bition performance of both DNBP and the green benchmarked against a database that contains
retarder under various conditions in laboratory information from over 30 styrene plants. This
tests. database enables opportunities for improvement
to be identified when designing treatment details.
Programme implementation The programme design also takes into considera-
The accepted best practice for polymer inhibition tion the styrene market’s conditions to ensure
in a styrene unit is to combine the use of a maximum return on investment.
retarder and an inhibitor, providing strong poly- To ensure a smooth transition, it is advised that
mer inhibition during normal plant operations inhibitor application should be commenced while
and adequate protection in the event of an emer- leaving DNBP injection unchanged. To facilitate
gency shutdown. During the implementation of a a safe withdrawal of DNBP and the introduction
Prism programme using an inhibitor and green of green retarder, the inhibitor injection should
retarder, injection points and dosages are be increased prior to the withdrawal of DNBP. At
adjusted to optimise plant performance and elevated inhibitor levels, DNBP dosage is gradu-
economics. A key performance indicator for most ally reduced until it is completely removed from
styrene producers is the soluble polymer level in the system. Green retarder is immediately intro-
the bottom of the purification section distillation duced once DNBP is removed from the system.
tower. To accurately measure soluble polymer Dosage of both Prism Green Retarder and
levels, Nalco developed a proprietary test method Inhibitor should be optimised gradually until the
that is more accurate than the often-used ASTM design dosage is reached. Close monitoring of the
method D2121. transition process is critical. Soluble polymer
Before the implementation of a replacement levels should be monitored at least once a day,
programme, a comprehensive MOC analysis optimally once a shift.

4 PTQ Q1 2012 www.digitalrefining.com/article/1000341

Emergency shutdown risk
assessment 105
5% polymer
During an emergency shutdown,
95 10% polymer
polymer continues to form in
15% polymer

Typical pumping limit

hot distillation towers. If not

Temperature, ºC
85 20% polymer
acted upon expeditiously, all
existing inhibitor — and ulti- 75
mately retarder — will be
65
consumed, causing catastrophic
tower failure as a result of poly- 55
mer solidification. While a
retarder does provide prolonged 45
protection due to its lower reac-
35
tivity, its protection is finite, and 0 100 200 300 400
it too will be consumed should Viscosity, cP
the situation not be rectified
swiftly. Most plants have emer-
gency shutdown response Figure 4 Viscosity profile of styrene with differing polymer content
procedures in place to mitigate
potential problems. Some
Lights
common measures include grav- Retarder
ity-fed or pumping of cold injection
ethylbenzene to dilute and cool
the tower contents, and emer-
Inhibitor
gency generator-powered pumps SM product
injection
to remove the tower contents or
Lights
to add extra retarder or tower
inhibitor. Inhibitor
Based on the emergency injection
procedure of the plant, one can Styrene
estimate the time required to tower
carry out the procedure to the
point where the tower is consid-
ered safe from polymer build-up.
This time can vary widely from Heavies
plant to plant due to different
emergency response mecha- Figure 5 Styrene purification for a PO/SM unit
nisms, specific column design
and operating conditions. lation train of a particular styrene plant, the
Ideally, the protection time a polymer control inhibitor and retarder dosages, and experimen-
programme (retarder and inhibitor) offers should tally determines the rate of polymer accumulation
be greater than the amount of time required to in the system.
implement the aforementioned emergency The available protection time is defined as the
procedure(s). Thus the critical question is: does time elapsed after the emergency shutdown until
the polymer control programme offer sufficient the polymer level reaches a critical level whereby
tower protection time in the event of an emer- the process stream is too viscous to pump. To
gency shutdown? determine at what polymer level the viscosity
To answer this question, Nalco developed a becomes excessively high, a series of viscosity
methodology to help determine the available measurements were taken using styrene mono-
protection time of a polymer control programme mer containing various levels of polymer (see
in the event of an emergency shutdown. This Figure 4). From the graph, one can see that at
tool examines operating conditions in the distil- above 20% polymer content, a significant

www.digitalrefining.com/article/1000341 PTQ Q1 2012 5

 polymerisation risk assessment
Compared shutdown conditions vs normal operation resulting from
column detailed modelisation (SPRA) study was carried out to
ensure sufficient polymerisation
protection in the event of an
Lights column SM column emergency shutdown.
Normal Shutdown Normal Shutdown
SM concentration relative to SM concentration in The risk assessment study
lights column bottom during normal operation 1.00 1.02 0.04 0.37 took into account the tower
Temperature relative to temperature in lights
tower column during normal operation 1.00 0.98 1.29 1.10
internal design, actual operating
conditions and chemical
Note: Conditions calculated as relative to lights tower bottom normal operating conditions to protect
confidential plant data
dosages. The target was to
model conditions of the result-
Table 3 ing liquid hold-up mixture in
the tower when the emergency
increase in viscosity threatens the pump handling shutdown starts, and rate the polymerisation
ability of the fluid. severity for each tower in case of an emergency
shutdown.
Case study: Repsol Tarragona Additionally, bottom pumps design was
Introduction reviewed to confirm that they could pump a solu-
Repsol’s PO/SM styrene production unit in tion containing up to 20% polystyrene.
Tarragona, Spain, recently completed a field trial A quick evaluation would suggest the product
using Prism Green Retarder and Inhibitor. tower to be the most critical unit in case of a
Figure 5 shows a simple unit schematic of the shutdown because it typically operates with high
styrene purification section for the PO/SM unit. temperatures in the bottom and high-purity
Raw styrene from the reaction section enters the styrene. But after applying the SPRA methodol-
lights tower of the purification section first. This ogy, it was discovered that the lights tower was
tower is designed to remove ethylbenzene as well the most critical due to the combination of
as water on the overhead, leaving styrene and styrene concentration and temperature, resulting
other heavy components in the bottom. The in the liquid mixture accumulated in the bottom
lights tower feeds the product tower where after the shutdown. This demonstrates the value
styrene is distilled overhead, leaving heavier of applying a thorough analysis of the system
compounds to exit from the tower bottom. prior to lab simulations and the value of the
Prior to the green retarder trial, the polymer SPRA methodology.
control programme involved the use of Prism Table 3 shows a comparison between the
Inhibitor and DNBP retarder. Both products normal operating conditions in the bottom of
were injected into the reflux line of the lights each tower compared with a shutdown estima-
tower. Additionally, an inhibitor injection was tion by the model. It can be seen from the data
applied to the reflux line of the styrene product that the lights tower maintains roughly the same
tower to provide extra polymer control to that conditions, while the product tower bottom,
tower. despite increasing styrene concentration during
shutdown, remains less severe due to the very
Goals low styrene concentration in the bottom during
The goals of the trial were to replace toxic DNBP normal operation. This lower severity in the
with green retarder, while maintaining adequate product tower was confirmed later through labo-
polymer control both during normal operation ratory simulation.
and in the event of an unexpected emergency The model of the system revealed the lights
shutdown. tower to be the most severely affected unit, but
the polymerisation conditions during the shut-
Programme design down for that column could be classified as mild
The dosages of inhibitor and green retarder were when compared to other systems surveyed. This
carefully selected to ensure polymer control allowed for a substantial reduction in the
while remaining cost competitive relative to the retarder dosage. The reduced dosage showed in
existing DNBP-based programme. A styrene the industrial test proved to be enough to ensure

6 PTQ Q1 2012 www.digitalrefining.com/article/1000341

stable polymer during process
fluctuations during plant 12000
operation. 11000
Blank
Repsol’s operational proce- Blank (repeat)
10000
dures in case of emergency Current DNBP + inhibitor

Sol u bl e pol yme r, ppm

9000 Current DNBP + inhibitor
shutdown were studied in detail
to understand how polymerisa- 8000 Prism GR + inhibitor
Prism GR + inhibitor (repeat)
tion could be controlled by 7000
means other than the existing 6000
chemical residuals. Repsol’s 5000
design and procedures allow a
4000
rapid decrease in polymerisation
severity. A requested polymer 3000
control time of three hours was 2000
established as the requirement 1000
for the chemical treatment 0
programme in the simulation. 0.0 0.5 1.0 1.5 2.0 2.5 3.0
Figure 6 shows the polymer Time, hours
level over time for both unin-
hibited and inhibited (with Figure 6 Emergency shutdown risk assessment for lights column
proposed dosages of inhibitor
and retarder) situations for the
lights tower. The polymer level 1.00
was found to be below 1% after Styrene tower Prism
Pol yme r make – n or mal i se d

three hours when inhibited Lights tower DNBP Green Retarder

0.8 Feed programme programme
either with DNBP or green
retarder programmes. This
provided confidence that the 0.6
designed and industrially tested
dosage combination of inhibitor 0.4
and retarder offers adequate
protection in the event of an
0.2
emergency shutdown based on
their emergency procedures.
The results of the SPRA study 0.0
can be summarised as: 1 11 21 31 41 51 61 71 81
• Lights tower is the most Days
severe column, but still this has
low severity when compared to Figure 7 Polymer make levels before and after green retarder implementation
other plants
• Green retarder programme dosages were under control both before and after the switch.
reduced to a minimum to ensure the required Polymer generation remained at the same average
three hours of polymer control time while having level and with the same variability with process
stable polymer during normal operation. fluctuations. It is interesting to note that when the
plant was using the old retarder-only programme
Results polymer generation was nine times higher than
When implementing the switch from DNBP to the average shown in Figure 7.
Prism Green Retarder, soluble polymer levels were Figure 8 shows the evolution of the costs of
closely monitored. Polymer levels from the feed to polymerisation in the plant over the years with
the lights tower and bottoms of both the lights the implementation of Prism Inhibitor and Green
tower and the product tower were plotted (see Retarder. Costs of polymerisation are those costs
Figure 7). It is clear that polymer levels were identified by the plant as being related to

www.digitalrefining.com/article/1000341 PTQ Q1 2012 7

 efficacy has been validated
1.00 through industrial testing.
Over the years Nalco Prism
programme has reduced Lisheng Xu is a Global Industry
Re l at i ve spe ci f i c cost

0.75 polymerisation costs by 40% while Development Manager with Nalco Company
phasing out very toxic retarders
in Sugar Land, Texas, and is responsible
for downstream petrochemicals market
0.50 development. He has over 20 years’
experience in the chemicals and
petrochemicals industry.
0.25 Javier Florencio is an Industry Technical
Consultant with Nalco Europe and is
responsible for technical support and new
0 projects development in the downstream
Retarder only Prism retarder Prism inhibitor petrochemicals European market. He has
+ DNBP + GR over 10 years’ experience in the chemicals
and petrochemicals industry.
Figure 8 Evolution of costs of polymerisation with the Prism programme Christopher Morrison is the Research and
Development Manager for Downstream
Petrochemicals with Nalco Company and has over 10 years’
polymer control: antipolymerants costs and the
experience in the chemicals and petrochemicals industry.
value of lost styrene production in the form of
Vincent Lewis is Research Scientist in the Downstream
polystyrene. It can be seen that DNBP replace- Petrochemicals Group with Nalco Company and has 23 years’
ment by green retarder has been achieved at experience in the chemicals and petrochemicals industry.
almost neutral cost, while the replacement Ana Guzmán is a Chemical Engineer with Repsol Química
programme has reduced the cost of polymerisa- Tarragona and has eight years’ combined experience in the
tion by 40% compared with the old retarder-only chemical and petrochemical industries.
programme. Carmen Monfort is the Plant Manager for the PO/SM Plant with
Repsol Química Tarragona and has over seven years’ experience
Conclusion in the chemicals and petrochemicals industry.
DNBP is a commonly used polymer retarder in Ana Olivares is the Product Engineer for Intermediate Products
with Repsol Química in Tarragona. She has 14 years’ experience
the styrene industry. However, the high toxicity
in the chemicals and petrochemicals industry.
associated with DNBP has prompted the devel-
opment of a green retarder as a safer and more
environmentally friendly replacement for DNBP.
Government regulations restricting the use of
DNBP in some countries have also created LINKS
demand for a green retarder, a trend that is
expected to develop worldwide.
More articles from the following categories:
Green retarder technology provides the benefits
Petrochemicals
of low toxicity, negligible NOx and polymer control Process Modelling & Simulation
performance comparable to DNBP. Programme

8 PTQ Q1 2012 www.digitalrefining.com/article/1000341