вход по аккаунту


Towards Clean Fuels Molecular-Level Sulfur Reactivity in Heavy Oils.

код для вставкиСкачать
individual sulfur molecular groups. Moreover, we have
coupled this sulfur characterization tool with systematic
desulfurization experiments in pilot-scale reactors to gain
molecular-level insights into the heavy-oil sulfur reaction
chemistry. These studies provide a first-ever “reactivity scale”
for individual sulfur molecular groups in HO and reveal that
the reactivity scale is profoundly influenced by the reaction
temperature. This study, which represents a major advance in
unraveling the complex sulfur chemistry in HO, is expected to
assist the production of clean fuels by opening-up new
avenues for process improvements and superior catalyst
Our sulfur characterization method can be used to
quantitatively discriminate between different sulfur molecular groups present in HO in less than five hours. The content
of nonaromatic sulfur compounds is quantitatively determined by chemical analysis, while the remaining compounds
are identified chromatographically, as described in detail in
the Experimental Section. Figure 1, which shows chromatoHeavy-Oil Desulfurization
DOI: 10.1002/ange.200503660
Towards Clean Fuels: Molecular-Level Sulfur
Reactivity in Heavy Oils
Tushar V. Choudhary,* Jim Malandra, John Green,
Stephen Parrott, and Byron Johnson
Due to environmental concerns there is a continual worldwide thrust towards clean transportation fuels, with the
emphasis on minimization of sulfur content.[1–6] The threshold
limit for sulfur in gasoline and diesel is expected to be
regulated to 50 ppm of weight (ppmw) or less within the next
few years,[1, 5, 6] and several billion dollars are anticipated to be
spent to meet these stricter regulations.[1] Sulfur removal from
heavy oil (HO) is an important process for meeting sulfur
regulations in transportation fuels.[5, 7–11] From an economic
viewpoint, it is therefore extremely desirable to increase the
efficiency of the HO sulfur removal process. Unfortunately,
whereas there is a vast and fundamental understanding of
sulfur removal chemistry for gasoline/diesel,[3, 5, 8] the molecular-level understanding of the ultra-complex HO desulfurization chemistry is inadequate, thereby limiting process/
catalyst improvements. Because of its enormous complexity,
previous studies have been unable to provide the much
needed quantitative molecular-level structure–reactivity
information about HO sulfur chemistry.[12–18]
Herein, we present a method that can be used routinely to
quantitatively classify the total HO sulfur content into
[*] Dr. T. V. Choudhary, Dr. J. Malandra, Dr. J. Green, Dr. S. Parrott,
Dr. B. Johnson
ConocoPhillips Company
Bartlesville Technology Center
Bartlesville, OK 74004 (USA)
Fax: (+ 1) 918-661-8761
Angew. Chem. 2006, 118, 3377 –3381
Figure 1. Chromatograms of sulfur compounds in HO1. I = intensity,
tNP = retention time on a nonpolar column, tP = retention time on a
polar column; T = thiophenes, b = benzothiophenes, DBT = dibenzothiophenes, PhT = phenanthrothiophenes, BNT = benzonaphthothiophenes, 5R-c = compact five-rings thiophenes, 5R-e = extended fiverings thiophenes. The six-rings thiophenes, although not marked here
because of their very low content, are present just beyond the extended
five-rings thiophenes.
grams of sulfur compounds in heavy oil 1 (HO1), clearly
demonstrates the ability of the method to distinguish between
different sulfur compounds in complex heavy oils. Such
quantitative resolution over a wide range (carbon range: C10C44) has never been reported; in fact, prior to this, molecularlevel quantification of only around 20 % of the total HO
sulfur compounds has been reported.[12, 14]
Each series of peaks in Figure 1 represents a homologous
series of aromatic sulfur compounds in HO1. The parent
structure for each of the homologous series was identified by
mass spectrometric analysis. The corresponding quantitative
information for HO1 in a condensed form is summarized in
Table 1. Benzothiophenes, dibenzothiophenes, benzonaphthothiophenes, and nonaromatic sulfides are the dominating
compounds in HO1, while thiophenes, phenanthrothiophenes, thiophenes with a total of five annulated rings
either in compact (compact five-rings thiophenes) or
extended form (extended five-rings thiophenes), and thio-
2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Table 1: Sulfur compounds[a] in heavy oils and feed properties.
Compound type[b]/Feed property HO1[c]
total sulfur content
32 290
11 518
dibenzothiophenes (total)
C0 dibenzothiophenes
C1 dibenzothiophenes
C2 dibenzothiophenes
C3 dibenzothiophenes
C4+ dibenzothiophenes
phenanthrothiophenes (total)
C0 phenanthrothiophenes
C1 phenanthrothiophenes
C2 phenanthrothiophenes
C3+ phenanthrothiophenes
benzonaphthothiophenes (total)
C0 benzonaphthothiophenes
C1 benzonaphthothiophenes
C2 benzonaphthothiophenes
C3+ benzonaphthothiophenes
compact 5-rings thiophenes
(total) (5R-c)
C0 5R-c
C1 5R-c
C2+ 5R-c
extended 5-rings thiophenes
(total) (5R-e)
thiophenes (total) (T)
benzothiophenes (total) (B)
C0 benzothiophenes
C1 benzothiophenes
C2 benzothiophenes
C3 benzothiophenes
C4 benzothiophenes
C5 benzothiophenes
C6+ benzothiophenes
6-rings thiophenes (total) (6R)
nonaromatic sulfides (NArS)
other properties[e]
total nitrogen content [ppmw]
total core aromatic compounds
[wt. %]
Parent structure
[a] Sulfur content given in ppmw. [b] Cn indicates the hydrocarbon group(s)
containing n carbon atoms that are attached to the aromatic ring system; Cn+ is
used for hydrocarbon group(s) containing more than n carbon atoms. [c] HO1
is a blend of atmospheric gas oil, vacuum gas oil, coker gas oil, and coker
distillate. [d] HO2 is a blend of atmospheric gas oil and vacuum gas oil.
[e] API = 141.5/(Specific Gravity at 288.55 K)131.5, IBP–FBP = initial boiling
point–final boiling point.
phenes with a total of six annulated rings (six-rings thiophenes) are present in much smaller amounts. The sulfur
distribution for heavy oil 2 (HO2) is also shown for
comparison. From these data, it is apparent that heavy oils
can differ appreciably in their total sulfur content and
HO1 and HO2 were hydrotreated in pilot reactors under
an industrially relevant range of process conditions (temperature: 622–655 K; H2 pressure: 7–14 MPa; catalyst: commercial NiMo). The hydrotreated products obtained from this
study were characterized for sulfur distribution. Based on
these data, the sulfur molecules were collated into the
following seven groups: group 1: thiophenes and benzothiophenes;
group 2:
C0/C1 dibenzothiophenes;
group 3:
C2+ dibenzothiophenes; group 4: phenanthrothiophenes;
group 5: benzonaphthothiophenes and compact five-rings
thiophenes; group 6: extended five-rings thiophenes and sixrings thiophenes; and group 7: nonaromatic sulfides.
The kinetic parameters for HO1 were determined assuming parallel first-order kinetics (described in the Experimental
Section) for the seven groups. Figure 2 shows parity plots for
experimental and calculated sulfur contents for individual
sulfur groups and for the total sulfur content obtained under
different process conditions. The kinetic parameters obtained
Figure 2. Comparison of calculated and experimental sulfur contents
for the individual sulfur groups given in Table 1 (a) and of the total
bulk sulfur content (b) for HO1 (filled symbols) and HO2 (open
symbols). Process conditions: temperature: 622–655 K; H2 pressure:
7–14 MPa; liquid hourly space velocity: 1 h1.
2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. 2006, 118, 3377 –3381
from the fitting to the experimental data for HO1 (& in
Figure 2 a and ~ in Figure 2 b) were used to predict the sulfur
distribution (&) and total sulfur content (~) for HO2. HO2
has an appreciably different sulfur content/distribution than
HO1; furthermore, the nature of the heavy oils is also very
different as only one of them contains coker liquids. The
excellent predictive ability, as seen from Figures 2 a and b,
lends substantial credence to the sulfur analysis and kinetic
The relative reactivities of different sulfur compounds at
three temperatures are shown in Figure 3. The relative
reactivity in this study is defined as the ratio of the rate
constant of a sulfur molecular group at a given temperature to
the rate constant of C0/C1 dibenzothiophene compounds. The
Figure 3. Relative reactivities (kS component/kC0/C1 dibenzothiophenes) of sulfur
molecular groups as a function of desulfurization reaction temperature. For the abbreviations, see Figure 1 and Table 1.
nonaromatic sulfides, thiophenes, benzothiophenes, benzonaphthothiophenes, extended and compact five-rings thiophenes, and six-rings thiophenes are easier to desulfurize than
C0/C1 dibenzothiophene, while the C2+ dibenzothiophenes
and phenanthrothiophenes are more difficult to desulfurize
at all the temperatures investigated in this study. Since the
distribution of the aromatic sulfur compounds (Table 1) is
dominated by compounds containing a large number of sidechains, considerable steric hindrance may be expected for
these molecules; this decreases the ability of the sulfur
compound to interact with the active sites of the catalyst.
Sulfur removal from petroleum fractions is possible by
a) direct extraction, which involves sulfur removal in a single
hydrogenolysis step, and b) hydrogenation, which involves
hydrogenation of the aromatic ring(s) of the sulfur compound
prior to the hydrogenolysis step.[19] Hydrogenation of the
aromatic ring(s) facilitates the hydrogenolysis step by diminishing the steric hindrance and increasing the electron density
on the sulfur atom.
Based on the reactor studies undertaken to examine the
influence of hydrogen pressure on the reactivity of the sulfur
compounds at 622 K, the following dependence on the
hydrogen pressure was observed: nonaromatic sulfides < C0/
C1 dibenzothiophenes < C2+ dibenzothiophenes < benzothiophenes < benzonaphthothiophenes phenanthrothiophenes compact five-rings thiophenes < extended five-rings
Angew. Chem. 2006, 118, 3377 –3381
thiophenes six-rings thiophenes. A higher dependence of
the reactivity on hydrogen pressure suggests a relatively
greater contribution from the hydrogenation mechanism.
The reactivity scale for the different sulfur compounds
(Figure 3) at 622 K is given by: nonaromatic sulfides > thiophenes benzothiophenes @ extended
thiophenes six-rings thiophenes > benzonaphthothiophenes compact five-rings thiophenes > C0/C1 dibenzothiophenes @
C2+ dibenzothiophenes > phenanthrothiophenes.
The low influence of hydrogen partial pressure on the
reactivity of the nonaromatic sulfides suggests that they react
predominantly by the direct hydrogenolysis route; the high
electron density expected on the sulfur atom for these
compounds is responsible for their rapid removal. Thiophenes
react rapidly due to their low steric hindrance. The high
dependence of the reactivity of the benzothiophenes on the
hydrogen pressure suggests that the hydrogenation pathway
also plays an important role in their removal: the large bond
orders for benzothiophenes at the unsaturated bond that
needs to be hydrogenated prior to sulfur removal are
expected to facilitate rapid hydrogenation,[20] which explains
the high reactivity of these compounds. The C0/
C1 dibenzothiophenes exhibit a weak dependence on hydrogen pressure and react mainly by the direct hydrogenolysis
route. The steric hindrance for these dibenzothiophenes is
expected to be considerably lower than that for
C2+ dibenzothiophenes, and hence their reactivity is higher.
The C2+ dibenzothiophenes show low sulfur reactivity on
account of their high steric hindrance and small bond orders
(slower hydrogenation).[20]
The multi-aromatic-rings sulfur molecules (more than
three rings), such as extended five-rings thiophenes, six-rings
thiophenes, benzonaphthothiophenes, and compact five-rings
thiophenes, show a higher dependence on hydrogen pressure,
thus indicating that the hydrogenation pathway is important
for them. Similar compounds are known to have high
hydrogenation rates for the first aromatic ring,[21, 22] therefore
sulfur compounds containing multiple aromatic rings are
expected to show higher reactivity (compared to dibenzothiophenes) via the hydrogenation route due to facile hydrogenation of their first aromatic ring.
Figure 3 shows that the reactivity scale is significantly
influenced by the reaction temperature. The relative reactivity of the multi-aromatic-rings sulfur compounds increases
upon increasing the reaction temperature from 622 to 655 K,
whereas the relative reactivity of the nonaromatic sulfides
(hydrogenolysis pathway) decreases. This suggests that an
increase in reaction temperature assists sulfur removal by the
hydrogenation mechanism (by increasing the rate of hydrogenation) to a greater extent than sulfur removal by the direct
hydrogenolysis mechanism. In line with this, we observed a
large increase in the total experimental hydrogen consumption upon increasing the temperature. The dependence of the
hydrogenation rate on temperature is not the same for
different compounds,[21] and hence the relative rates are
expected to vary to a different extent with temperature.
Surprisingly, the phenanthrothiophenes, despite containing multiple aromatic rings, are the least reactive compounds
at 622 K. The hydrogenation of the first ring of phenanthrene
2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
(hydrocarbon analogue of phenanthrothiophene) is known to
be unusually slow,[21, 22] in fact it is even slower than the
hydrogenation of the first aromatic ring of compounds
containing only two aromatic rings. Therefore, at low temperatures (622 K) the phenanthrothiophenes react very slowly due
to lower hydrogenation rates. However, increasing the temperature makes the hydrogenation more facile and leads to a
considerable increase in reactivity. In contrast, the reactivity of
C2+ dibenzothiophenes does not increase as much with the
reaction temperature as the hydrogenation pathway (based on
its relatively lower hydrogen dependence) is less important for
these compounds. Therefore, although the reactivity of
C2+ dibenzothiophenes is higher than that of phenanthrothiophenes at 622 K, their reactivities are similar at 655 K.
We have developed a method for the molecular-level
classification of HO sulfur compounds and have demonstrated that the sulfur compounds present in HO exhibit
distinctly different reactivity, depending on their structure.
Furthermore, their reactivity is a complex function of the
process conditions. The nature and distribution of the different sulfur compounds is expected to define the ability to
remove sulfur from these feeds. For example, feeds dominated by phenanthrothiophenes and C2+ dibenzothiophenes
will be more difficult to desulfurize. In the case of the lowerboiling petroleum fraction (distillate range), the nature and
distribution of sulfur compounds are known to dictate the
selection of the catalyst and the processing conditions to
achieve the desired product sulfur.[19] Information about the
nature and distribution of the sulfur compounds in HO may
similarly be used to maximize the HO desulfurization
efficiency. Process innovations and the development of
advanced desulfurization catalysts require a deep understanding of the fundamentals of the process under consideration.[19, 23, 24] The detailed molecular-level information apropos HO sulfur chemistry provided by this study is expected to
help in process innovations and desulfurization catalyst
(“conventional” and “biocatalyst”) improvements.
Experimental Section
Determination of the sulfur content: In the first step, nonaromatic
sulfides were selectively oxidized to sulfoxides, which were separated
by adsorption on silica gel, thereby avoiding interference in the
quantitative analysis of other sulfur compounds. ffs a first step, the
sample was heated in an oven at 363 K for 0.5 h. One gram of the
heated sample was treated with 0.15 g of tetrabutylammonium
periodate in 12 mL of toluene/methanol (5:1, v/v) at 358 K for 1 h,
washed with water, and separated into two fractions on a silica gel
column. The fraction containing sulfoxides was quantitatively analyzed for the total sulfur content. The presence of sulfoxide groups
was confirmed by infrared spectroscopy. The remaining fraction
(devoid of sulfoxides) was introduced onto a diphenyl-/dimethylpolysiloxane (5/95) column (nonpolar) and then subsequently onto a
diphenyl-/dimethylpolysiloxane (50/50) column (polar). The nonpolar
column separates according to boiling point, and a modulator
reintroduces the sample onto the second column, which separates
according to polarity; a sulfur-chemiluminescence detector (SCD)
and flame-ionization detector (FID) were used for sulfur and
hydrocarbon detection, respectively. The amounts of sulfur compounds determined by this analysis were combined with the amount
of nonaromatic sulfide and normalized to the total sulfur of the
sample. The sulfur recovery ranged from 75 to 95 %; the average
sulfur recovery for feed samples was about 85 %, whereas the average
recovery for hydrotreated product samples was about 80 %. The
sulfur loss is due to formation of emulsions during the water-wash.
However, the relative concentrations are not altered by this process:
Control studies at different recoveries showed that the sulfur analysis
is not significantly different for two analyses with considerably
different sulfur recovery. This method is applicable to coker gas oils,
vacuum gas oils, heavy atmospheric gas oils, heavy cycle oil, and other
deashphalted heavy oils.
Hydrotreating experiments: The hydrotreating experiments with
a commercial NiMo catalyst were undertaken in pilot units at the
Bartlesville Technology Center, ConocoPhillips (COP). The HO
feeds used in this study were obtained from COP refineries.
Presulfiding was achieved using a distillate feed. Experiments were
undertaken at varying reactor temperatures (622–655 K) and pressures (7–14 MPa); the liquid hourly space velocity was maintained
constant at 1 h1 and H2/feed ratio at 356 m3/m3. Four pilot units were
used to obtain the data in this study. An excellent reproducibility for a
range of product properties was obtained between different pilot
units for a given HO under identical process conditions.
The parallel first-order kinetics are described by the Equations (1) and (2), where Sf = sulfur content of the feed, x = 1–7 (group
total Sproduct ¼
½SfgroupðxÞ expðkgroup x Pagroup
kgroup x ¼ Agroup x expðEgroup x =RTÞ
numbering as given in the text), a = pressure dependence term,
LHSV = liquid hourly space velocity, k = rate constant, A = preexponential factor, and E = activation energy.
The kinetic parameters were obtained from fitting the kinetic
model to the experimental data obtained for HO1 under a range of
process conditions. The scatter in the parity plot is mostly within
experimental error (arising from pilot reactor operations and sulfur
analysis). In order to validate the kinetic analysis, the parameters
obtained from the HO1 data were used to predict the desulfurization
performance of HO2 under the same process conditions. HO2 was
chosen because it has significantly different properties (different
source, sulfur content and distribution). Figure 2 a provides the data
for individual sulfur groups, while Figure 2 b provides the corresponding total sulfur content data. Figure 2 b has significantly fewer points
as each data point represents the sum of the data of individual sulfur
groups under a given process condition. In this particular study, the
relative inhibition effect from nitrogen compounds is not expected to
be important in terms of predictability as both HO1 and HO2 have
similar total nitrogen content (Table 1). In a future study, the effect of
individual nitrogen molecular groups (molecular composition), if any,
on the desulfurization reaction should be determined.
Received: October 15, 2005
Revised: February 20, 2006
Published online: April 5, 2006
Keywords: analytical methods · desulfurization ·
fuel technology · green chemistry · sulfur
[1] S. F. Venner, Hydrocarbon Process. 2000, 79, 51 – 60.
[2] R. T. Yang, A. J. Hernandez-Maldonado, F. H. Yang, Science
2003, 301, 79 – 81.
[3] S. Brunet, D. Mey, G. Perot, C. Bouchy, F. Diehl, Appl. Catal. A
2005, 278, 143 – 172.
[4] N. Gupta, P. K. Roychoudhary, J. K. Deb, Appl. Microbiol.
Biotechnol. 2005, 66, 356 – 366.
[5] I. V. Babich, J. A. Moulijn, Fuel 2003, 82, 607 – 631.
[6] C. Song, X. L. Ma, Appl. Catal. B 2003, 41, 207 – 238.
2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. 2006, 118, 3377 –3381
[7] S. W. Shorey, D. A. Lomas, W. H. Keesom, Hydrocarbon Process. 1999, 78, 43 – 51.
[8] C. Song, Catal. Today 2003, 86, 211 – 263.
[9] V. Bavaro, World Refrig. 2000, 10, 34 – 37.
[10] C. Botchwey, A. K. Dalai, J. Adjaye, Energy Fuels 2003, 17,
1372 – 1381.
[11] T. A. Reid, Akzo Nobel Catalyst Courier 2002, no. 42.
[12] X. L. Ma, K. Sakanashi, T. Isoda, I. Mochida, Fuel 1997, 76, 329 –
[13] N. Ghaloum, G. Michael, Z. Khan, J. Liq. Chromatogr. Relat.
Technol. 2002, 25, 1409 – 1420.
[14] X. L. Ma, K. Sakanashi, I. Mochida, Ind. Eng. Chem. Res. 1996,
35, 2487 – 2494.
[15] T. M. Schaub, R. P. Rodgers, A. G. Marshall, Energy Fuels 2005,
19, 1566 – 1573.
[16] J. T. Andersson, S. K. Panda, S. Wolfgang, Abstr. Pap. ACS, 230th
ACS National Meeting, September 2005.
[17] F. Ali, Z. H. Khan, N. Ghaloum, Energy Fuels 2004, 18, 1798 –
[18] H. Behbehani, M. K. Andari, Pet. Sci. Technol. 2000, 18, 51 – 61.
[19] K. G. Knudsen, B. H. Cooper, H. Topsoe, Appl. Catal. 1999, 189,
205 – 215.
[20] X. L. Ma, K. Sakanashi, T. Isoda, I. Mochida, Energy Fuels 1995,
9, 33 – 37.
[21] A. Stanislaus, B. H. Cooper, Catal. Rev. Sci. Eng. 1994, 36, 75 –
[22] S. C. Korre, M. T. Klein, R. J. Quann, Ind. Eng. Chem. Res. 1995,
34, 101 – 117.
[23] M. Breysse, G. Djega-Mariadassou, S. Pessayre, C. Geantet, M.
Vrinat, G. Perot, M. Lemaire, Catal. Today 2003, 84, 129 – 138.
[24] Y. Ishii, S. Kozaki, T. Furuya, K. Kino, K. Kirimura, Curr.
Microbiol. 2005, 50, 63 – 70.
Angew. Chem. 2006, 118, 3377 –3381
2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Без категории
Размер файла
211 Кб
fuel, level, molecular, towards, sulfur, clean, reactivity, heavy, oils
Пожаловаться на содержимое документа