AI ChE 174855

Feedstock Contaminants: A Case Study

Patricio Herrera
Research Scientist
NOVA Chemicals Research and Technology Centre, Calgary, AB, Canada

Vasily Simanzhenkov
Research Engineer
NOVA Chemicals Research and Technology Centre, Calgary, AB, Canada

Chi Wong
Research Engineer
NOVA Chemicals Research and Technology Centre, Calgary, AB, Canada

Les Benum
Team Leader
NOVA Chemicals, J offre, AB, Canada


Prepared for Presentation at the 2010 Spring National Meeting
San Antonio, TX, March 22
nd
, 2010


AIChE and EPC shall not be responsible for statements or opinions contained in
papers or printed in its publications








Feedstock Contaminants: A Case Study
Patricio Herrera
Research Scientist
NOVA Chemicals Research and Technology Centre, Calgary, AB, Canada
Vasily Simanzhenkov
Research Engineer
NOVA Chemicals Research and Technology Centre, Calgary, AB, Canada
Chi Wong
Research Engineer
NOVA Chemicals Research and Technology Centre, Calgary, AB, Canada
Les Benum
Team Leader
NOVA Chemicals, J offre, AB, Canada


Abstract: An advanced characterization program was designed and
implemented to identify the chemical nature of contaminants present in the feed
to a flexi-cracker. The solid contaminants removed from the naphtha feed had
similar amounts of inorganic and organic materials. Iron and silicone were the
most abundant elements present in the inorganic fraction. Iron oxides and
hydroxides, iron sulfide, and quartz made up the crystalline portion of the solids.
In a unique finding, particles of iron-based compounds were found encapsulated
in the silicon species. Most significantly, the size of almost half of the particles
removed from the naphtha feed was under 1 micron, and about two-thirds were
below 2 micron. Analysis of the organic fraction identified the presence of sulfur-
containing chemicals, paraffins, aromatics, olefins, oxygenates, and dienes.











INTRODUCTION

Different contaminants are present in almost all the feed streams to steam
cracking plants. Anything other than hydrocarbons that enters the plant has the
potential to cause unintended side effects, resulting in problems such as fouling,
corrosion, product contamination, etc. Upsets in the content of certain
contaminants in the feedstock are usually caused by upsets of upstream
operating units. However, contaminants can enter the production site not only
with the feed stream from an upstream producer, but from a number of different
sources. All the contaminants can be classified according to the way they enter
the process site. There are, generally, five possible ways in which contaminants
can be introduced into the steam cracking process. These are (1):
1. With feedstocks.
2. By chemical reactions during the thermal cracking process.
3. During the olefins recovery processes.
4. During shipping and storage.
5. Injected on purpose or self-inflicted contaminants.
Understanding the nature and effect of contaminants on the steam
cracking process is crucial by setting the limits or specifications for process
streams, which are being processed in a plant. Some contaminants can cause an
immediate effect on the product quality or equipment performance; whereby,
some of them can result in long-term effects such as corrosion or fouling.
The general trend for the industry for the past few decades is to reduce
the contaminants level to the acceptable levels for major ethylene unit streams.
This trend is directly related to:
1. Introduction of new generations of polymerization catalysts, which have
higher activity but reduced poisons or contaminants tolerance levels.
2. Introduction of new high efficiency equipment (i.e. new low coke coils,
anti-coking additives, etc.). This sets new, lower limits for acceptable
contaminants levels.
New specifications for a number of different contaminant groups are one
order of magnitude lower compared to the same values in the past. Technologies
for contaminants removal are becoming more and more advanced, following the
general industry trend. However, more advanced technologies are often also
more cost extensive. This is why it is becoming more and more important to
understand the role and nature of different process contaminants in order to
make technologically and economically sound decisions with regards to setting
specifications for contaminants levels.
This paper provides analytical protocols and results of detailed
characterization of selected process streams from NOVA Chemicals
manufacturing sites.



BACKGROUND

In the year 2007, Manufacturing East Corunna Site and Manufacturing
West J offre Site experienced shorter than expected run lengths on a number of
heaters, in which different coil metallurgies were installed. The impacted coils
were taken out of service and detailed surface morphology analysis was
performed on the feed site of these coils. This analysis revealed that coil inlet
surfaces were severely damaged / degraded by iron based contaminants.
Whereby, no contamination or damage was identified at the coil outlets. All iron
contaminated surfaces did show the presence of new surface morphology:
globule like particles or globule like surface morphology.
As a result of these findings, it was decided to perform a detailed study of
contaminants for all feed streams, which are entering the ethane cracking
heaters.
Analytical protocols and data obtained for naphtha feed to cracking
heaters will be reviewed in detail in the following sections of this paper.

ANALYTI CAL PROTOCOLS

All the naphtha samples were analyzed for total solids content via vacuum
filtration. Solid and liquid phases obtained during the filtration were separately
analyzed via two different analytical protocols, which are presented in Table 1.
The analytical protocol was developed for detailed investigation of the possible
source of iron contamination as well as identification of the nature of iron
present in the feed stream. One can see that only proton induced x-ray emission
(PIXE) analysis was performed on the liquid phase of the naphtha samples. This
analysis was seen to be sufficient to give initial data on dissolved metals content.
There were no concerning levels of metals identified in the liquid phase of
naphtha, this is why this protocol was not further developed for more detailed
characterization of the samples.

Table 1. Analysis of Naphtha

Vacuum filtration of a naphtha sample (2 m filter)
Liquid phase Solid phase (particulates)
PIXE PIXE
TGA-FTIR
Elemental analysis (CHNOS)
XRD
Particle Size Distribution by SEM
SEM/EDX
Solvent extraction
GC/MS Headspace analysis
Transmission Electron Microscopy


RESULTS AND DISCUSSION

The following sections of this paper will present the analytical data
obtained from different naphtha samples analyzed within this case study via the
analytical protocols described in the previous section.

LIQUID PHASE ANALYSIS
PIXE
Analytical results of dissolved metals analysis in different naphtha samples
are shown in Table 2.

Table 2. PIXE Analysis of Filtered Naphtha (wppm)
Metal
Naphtha
Tank A
Naphtha
Tank B
Naphtha to
Tanks A & B
Naphtha Upstream
of cracking heater
Na 0.376 0.709 0.560 0.968
Si 0.611 0.159 2.16 5.43
P 1.28 0.499 0.363 3.50
S 7.60 5.93 3.00 227
K 0.1000 0.0500 0.270 1.83
Ca 0.392 0.101 0.174 1.27
Fe 0.0372 0.0278 0.0500 0.614

One can see from Table 2 that phosphorous and sulfur were identified as
the only major contaminants in these four different naphtha samples. All other
analyzed elements, including iron, are present at a very low concentration.



SOLID PHASE ANALYSIS

A 2 m filter was installed on a slip stream of one of the naphtha feed
pipelines at NOVA Chemicals. This filter collected a significant amount of solids
after two weeks in service. After this period of time, it was taken out of service,
particulate matter from the filter was carefully removed and analyzed via the
analytical protocols described previously in this paper. A photograph of the
particulate matter is shown in Figure 1.


Figure 1. Digital photograph of Naphtha Filter Solids (NFS).

TGA-FTIR

A TGA Thermogram is shown in the Figure 2. One can see from Figure 2
that approximately 55 wt.-% of the sample was thermally decomposed in a
nitrogen environment, which can be considered as a total content of thermally
crackable organic matter. One can also see a strong heat flow peak at
approximately 800C. Nitrogen flow through the TGA analyzer was switched from
nitrogen to air at this point. No significant weight loss was observed at the same
time with significant heat release, this is mainly due to the fact that the major
part of the organics was removed from the sample in the first half of the TGA
analysis due to thermal cracking. As a result, the strong heat flow at this analysis
stage can be explained by oxidation of the metals species, which formed mainly
non-volatile oxides.
TGA-FTIR analysis shown in Table 3, qualitatively shows compounds
which were identified in the evolved gas from TGA at different temperatures.
One can see from the table that the main compounds evolved from the sample
are water, carbon dioxide, carbon monoxide and sulfur dioxide. Noticeable
amounts of hydrocarbons were identified at lower temperatures (225C and
390C). Whereby, only oxides were evolved from the sample at temperatures
higher that 733C. Additionally, halogenides and traces of mercury were
identified at 390C. All these compounds are naturally occurring in most
hydrocarbon feeds, as a result, it was not surprising to find traces of these
compounds in the particulate matter collected from the naphtha sample.

Figure 2. TGA Thermogram of Naphtha Filter Solids (NFS).

Table 3. TGA-FTIR analysis of Naphtha Filter Solids
Compounds present in the evolved gases at selected temperatures (C)
225 390 733 805
Water, carbon
dioxide, sulfur
dioxide, carbonyl
sulfide,
unidentified
hydrocarbon
Water, methane,
carbon dioxide,
carbon monoxide,
chloroethyl mercury,
octachloro
cyclopentene, and
triiodomethane
Water, methane,
carbon dioxide,
and carbon
monoxide
Water, carbon
dioxide, carbon
monoxide and
sulfur dioxide



Elemental analysis of NFS

Results of elemental analysis of NFS are shown in Table 4. Elemental
analysis was performed via sample combustion and post-combustion product
analysis. One can see from Table 4 that NFS contain significant amounts of
heteroatoms (S, N and O) as well as a very low C:H ratio. This can be seen as an
indication that the sample was mainly heavy, aromatic, organic matter as well as
very rich in carbonates and organic oxygenates.


Table 4. Elemental analysis of NFS
C H N O S Total C:H
20.66 3.46 1.54 21.31 3.99 50.96 0.50

PIXE

Analytical results of metals analysis in the solids particulate matter are
shown in Table 5. One can see that the major elements in the analyzed NFS
sample are iron, silicon, sulfur an zinc. Iron and sulfur are the most abundant
elements of corrosion products in a refinery, as a result, it is not surprising to see
these two elements among the most abundant elements in the NFS sample.
Silicon is one of the most popular naturally occurring contaminants of
hydrocarbon feeds. Also, silicon containing additives are widely used in many
refineries and petrochemical plants. This results in two sources of silicon in the
hydrocarbon feed: natural content in the feed and self-inflicted contamination,
which are introduced with the recycle streams.

Table 5. Elemental analysis of NFS.
Element Mass concentration
Iron 30.86%
Silicon 12.44%
Sulfur 4.82%
Zinc 0.11%
Chlorine 0.11%
Manganese 485ppm
Calcium 229ppm
Mercury 644ppm
Potassium 127ppm
Chromium 640ppm
Copper 194ppm
Nickel 102ppm

XRD

Table 6 shows XRD data obtained for the NFS sample. One can see from
Table 6 that the main compounds identified in NFS via XRD are typical corrosion
products, as well as quartz. Quartz is one of the most abundant compounds on
earth. It is also one of the most abundant contaminants in the hydrocarbon
feeds. As a result, the most probable source of quartz in the sample is from the
feed to the plant.




Table 6. XRD analysis of NFS
Components Concentration %
Quartz SiO
2
2.8
Lepidocrocite FeO(OH) 22.8
Pyrite FeS
2
5.6
Magnetite Fe
3
O
4
8.7
Goethite FeO(OH) 60.1
Total 100

SEM/ EDX

Particle size distribution of NFS was analyzed via Scanning Electron
Microscopy (SEM). The results of this analysis is shown in Figure 3.


Figure 3. Particle size distribution of NFS.

The most significant finding from the size distribution analysis, shown in
Figure 3, is that the size of almost half of the particles removed from the
naphtha feed was under 1 micron, and about two-thirds were below 2 micron.
Results of EDX analysis of ashe from TGA analysis, which was discussed
previously in this paper, are shown in Figure 4. One can see from Figure 4 that
the ashes contain only two major metals: iron and silicon.
Not treated NFS were separated into two fractions via filtration: particles
with effective diameter lesser than 3 m and particles with effective diameter
greater than 3 m. Both fractions were analyzed via SEM/EDX. Results of this
analysis are shown in Table 7. One can see from Table 7 that the major
difference between the two fractions of NFS is iron and silicon content. It is
shown that larger particles contain mainly iron and relatively low amounts of
silicon. On the other hand smaller particles contain mainly silicon and relatively
low amounts of iron.


Figure 4. SEM/EDX analysis of NFS ashes

Table 7. EDX analysis of NFS fractions
NFS Composition determined by EDX (wt.-%)
C O Si S Mn Fe
<3 m 32.83 42.82 13.78 1.08 9.49
>3 m 22.32 30.32 3.94 2.73 0.24 40.36


HEAD SPACE GC/ MS

The NFS sample was placed into a sampling container and kept under a
nitrogen environment at 250C for 10 minutes before the head space of the
sample was injected into a GC/MS analyzer. The GC/MS chromatogram, which
was obtained from this analysis, is shown in Figure 5. One can see from the
figure that nearly a hundred compounds were identified during this analysis.
Significant numbers of different sulfur containing compounds, such as hydrogen
sulfide, carbonyl sulfide, sulfur dioxide, mercaptans as well as aromatic sulfur,
were identified. Other compounds are represented by different oxygenates,
organic nitrogen containing compounds, aliphatic, and aromatic hydrocarbons.





Figure 5. Head space GC/MS of NFS

TEM

Transmission electron microscopy analysis was performed on the NFS
sample. Results of this analysis can be summarized as:
o Both typical corrosion and non typical corrosion products were
identified.
o Typical corrosion products are represented by large particles, which
contain large amounts of iron and very low amounts of silicon.
o Some of the small particles were iron compounds encapsulated in
silicon compounds.

An example TEM digital photograph of an iron rich particle, which is
encapsulated in a silicon / silicon oxide shell is shown in Figure 6. If one
compares Figure 6 and Figure 4, which shows SEM photograph, then it can be
seen that the encapsulated particles have very similar morphology with NFS ashe
particles. The mechanism of formation of these particles is not entirely clear as
well as their effect on the run length of ethane cracking heaters. This is the
topic of a separate ongoing study, which will not be reviewed within this paper.


Figure 6. TEM photograph of typical NFS

Conclusions

1. Similar amounts of organic and inorganic fractions were identified
in NFS.
2. Iron & Silicon were identified as the main elements in the inorganic
fraction of NFS.
3. Sulfur, Zinc, and Chlorine were also present in significant amounts
in NFS.
4. Typical corrosion products are most abundant in crystalline form in
NFS.
5. Very small particles of iron sulfides/oxides were identified.
6. Particles of iron compounds encapsulated in silicon species were
identified in NFS.
7. Homogeneous particles in NFS contain both iron and silicon
8. Size of almost half of the particles retained by naphtha filter is less
than 1 m.
9. About 64% of particles from NFS are smaller than 2 micron m.
10. More than 90 organic compounds identified by headspace analysis.

ACKNOWLEDGMENTS

This work was supported by NOVA Chemicals Corporation. Work and
contributions to this study of a number of manufacturing and research groups
from NOVA Chemicals Corporation is greatly appreciated.




NOMENCLATURE

EDX Energy Dispersive X-ray analysis
GC/MS Gas Chromatography / Mass Spectrometry
NFS Naphtha Filter Solids
PIXE Proton Induced X-ray Emission Spectroscopy
SEM Scanning Electron Microscopy
TEM Transmission Electron Microscopy
TGA-
FTIR
Thermogravimetric Analysis - Fourier Transform Infrared
Spectroscopy
XRD X-Ray Diffraction


References
1. Reid, J . and McPhaul, D., Contaminant Rejection Technology Update,
Volume 5, 8th Annual Ethylene Producers' Conference, 1996, New
Orleans, Louisiana`