DIESEL

Custom Catalyst Systems for

Higher Yields of Diesel

Brian Watkins In those areas of the world which are experiencing low costs of natural gas, there has been a decrease in
Manager, the cost of hydrogen, and this, combined with the growth in global demand for middle distillates, has
Hydrotreating Pilot prompted refiners to look to improve profitability by increasing middle distillate yields. Options under consid-
Plant and Technical eration have included operating an FCC (Fluid Catalytic Cracker) pretreater in a mild hydrocracking mode,
Service Engineer switching to maximum LCO1 (Light Cycle Oil) mode or extending the endpoint of feed to a ULSD (Ultra Low
Sulfur Diesel) unit and converting the heavy fraction into diesel range material. The use of opportunity feed-
Charles Olsen
stocks and synthetic type feedstocks can also be considered2. These approaches require specialized cata-
Director, Distillate R&D
lyst systems capable of providing some cracking conversion or changes to traditional unit operation, and
and Technical Service
careful attention must be given to minimizing production of excess gas and naphtha while maximizing

Advanced diesel. Another seemingly simple option is to maximize the product volume swell from an existing ULSD
Refining unit through a change in catalyst and understanding the demand on operating conditions. This approach to
Technologies increasing diesel yields requires a detailed understanding of feed and operating conditions such that the hy-
Chicago, IL, USA drotreater can be operated at the maximum product volume swell for the majority of the unit cycle. In this
case, the benefits of increased diesel yield need to be balanced against the potential costs of increased hy-
drogen consumption and decreased cycle length.

A critical element in all the approaches to increasing diesel yield is the proper design and selection of a cat-
alyst system for the hydrotreater. This paper summarizes some of these various catalytic options and the
operating conditions that can be implemented to increase yields of middle distillate using existing assets
with minimal investment.

As a first step, it is useful to understand the chemistry involved in hydrotreating and, in particular, the chem-
istry required for maximizing product volume swell. Table I lists several different classes of hydrocarbon
compounds that can be found in diesel range feeds. The data shows that as hydrogen is added to a mole-
cule, the density of the compound decreases. This indicates that even some simple reactions involved in
hydrotreating result in a decrease in density of the product or put another way, result in an increase in prod-
uct volume. This is especially apparent with aromatics species.

Grace Catalysts Technologies Catalagram® 33

 Class Compound Formula Density, g/cc ˚API H/C Ratio
Iso Paraffin 2,3-dimethyl-octane C10H22 0.738 60.3 2.2
Paraffin n-decane C10H22 0.730 62.3 2.2
Olefin 1-decene C10H20 0.741 59.5 2.0
Naphthene Decalin C10H18 0.897 26.3 1.8
Mono Aromatic Tetralin C10H12 0.970 14.3 1.2
Poly Aromatic Naphthalene C10H8 0.738 -7.4 0.8

TABLE I: Selected Compounds Boiling in the Diesel Range

Aromatics Saturates
Rings Compound Formula Density, g/cc Boiling Point, ˚F Compound Formula Density, g/cc Boiling Point, ˚F
2 Naphthalyne C10H8 1.140 424 Decaline C10H8 0.897 374
3 Fluorene C13H10 1.202 563 Perhydro Fluorene C13H22 0.920 487
Perhydro
3 Phenanthrene C14H10 1.180 630 C14H24 0.944 518
Phenanthrene
4 Pyrene C16H10 1.271 759 Perhydro Pyrene C16H26 0.962 604

TABLE II: Aromatic Compounds Found in Diesel Range Feed

Table II lists several different aromatic and fully saturated com- sure is around 500 psi, so the data cover the range of H2 pressures
pounds which occur in diesel range feedstock along with some se- typically encountered in ULSD. The total aromatics conversion
lected properties. It is apparent that dramatic shifts in boiling point nearly doubles with a 2.5 times increase in H2 partial pressure.
and density can be realized by hydrogenating aromatic compounds.
The density decreases by 20-25% with boiling points shifts any- Figure 2 shows how the aromatics conversion changes with temper-
where from 50-150°F upon saturation of the aromatic rings. ature in a typical ULSD unit. The figure compares the conversion
observed for both a NiMo and a CoMo catalyst. The data clearly in-
This suggests that in order to achieve a high degree of product vol- dicates that the NiMo catalyst has the greater aromatic saturation
ume swell in ULSD, a detailed understanding of aromatic and activity of the two catalysts shown. The product aromatics concen-
polynuclear aromatic (PNA) hydrogenation is required. It is well un- tration is over 4% (absolute) lower for the NiMo catalyst compared to
derstood that hydrogenation of aromatic compounds is a reversible the CoMo catalyst. This difference in aromatics conversion ac-
reaction, and that the equilibrium conversion is less than 100% counts for the higher H2 consumption typically seen for a NiMo com-
under typical conditions. The equilibrium conversion is highly de- pared to a CoMo catalyst. The chart also shows the influence of
pendent on temperature and hydrogen partial pressure. Figure 1 equilibrium on aromatics conversion. As the temperature increases
shows how the saturation of aromatics in diesel changes with H2 beyond about 670-680°F the conversion actually begins to decrease
partial pressure at a typical temperature for ULSD. The base pres- as the rate of the dehydrogenation reactions has increased enough

80.0 29.0

75.0 Constant Temperature 27.0

Aromatics Conversion, wt.%

Total Aromatics, vol.%

70.0
25.0
65.0

60.0 23.0

55.0
21.0
50.0
19.0
45.0

40.0 17.0

35.0
15.0
30.0 600 620 640 660 680 700 720 740
0.5 1.0 1.5 2.0 2.5 3.0
Reactor Outlet Temperature, ˚F
H2 Pressure/Base H2 Pressure
CoMo NiMo

FIGURE 1: Effect of Pressure on Aromatics

Hydrogenation FIGURE 2: Aromatic Reduction in ULSD

34 Issue No. 113 / 2013

 1.00
k1-10*k2
0.90
~+2 H2
0.80

0.70

Concentration
Poly-Nuclear Aromatic ~+2 H2 Tri-Aromatic 0.60

0.50

0.40
~+2 H2
0.30

0.20

Di-Aromatic Mono-Aromatic 0.10

+3 H2
0.00
Residence Time

Monorings Polyrings Sat’d

“Naphthene”

FIGURE 4: 1st Order Reversible Reactions in Series

Concentration Profile

FIGURE 3: Stepwise Saturation of a Poly Aromatic genation reaction. There is a rapid decrease in the concentration of
Compound
the 2-ringed aromatic species at short residence times and a corre-
sponding increase in the mono-ringed species. As contact time in-
to compete with saturation reactions. At high enough temperatures creases however, the mono-ring aromatic concentration begins to
both catalysts give the same conversion since they are operating in decrease and the fully saturated species begin to build up. This
an equilibrium-controlled regime. type of concentration profile suggests that there is a range of resi-
dence times in the unit corresponding to a maximum in the mono-
One significant consequence of achieving a high level of saturation ringed aromatic concentration.
of multi-ring and mono-aromatic ring compounds is higher hydrogen
consumption. However, not all aromatic species are created equal A variety of substituted naphthalene’s have also been shown to fol-
when it comes to hydrogen consumption. Figure 3 shows a simple low a similar reaction network with the rate of hydrogenation of the
schematic of the reaction pathway for saturating a 4-ring poly aro- first aromatic ring approximately equal to that observed for naphtha-
matic compound. The hydrogenation occurs in a stepwise fashion lene. The hydrogenation of biphenyl occurs in a stepwise fashion as
where one aromatic ring at a time is being saturated, with each step well, with the rate of hydrogenation of the first aromatic ring about an
along the pathway being subject to equilibrium constraints. The rate order of magnitude faster than that of the mono ring compound. An
limiting step to the fully saturated species is hydrogenation of the interesting difference is that the rate of the first hydrogenation reac-
last aromatic ring (the mono aromatic), and this step consumes the tion in naphthalene is approximately an order of magnitude faster
most hydrogen of the reactions shown in the reaction pathway.
Three moles of hydrogen are required to hydrogenate the mono-
ringed compound compared to two moles of hydrogen to hydro-
genate the rings in the poly aromatic compounds.

A number of poly aromatic species have been studied over the 1

years leading to a good understanding of the chemistry involved in
PNA saturation.3 In the case of naphthalene, the reaction begins
1
with the hydrogenation of one of the aromatic rings to form tetralin, a
mono-ring aromatic. The next reaction is hydrogenation of the re-
20
maining aromatic ring to produce decalin, the fully saturated
species. The reactions occur sequentially with the rate of hydro-
genation of the final aromatic ring an order of magnitude lower than 16

saturation of the first aromatic ring. The reactions can be modeled

as a series of first order reversible reactions. Figure 4 shows the
species concentration profiles as a function of residence time for a
hydrogenation reaction sequence such as that for naphthalene just
discussed. The rate of the first hydrogenation reaction in the series FIGURE 5: Relative Rate Constants for Saturating
Aromatics
is an order of magnitude faster than the rate of the second hydro-

Grace Catalysts Technologies Catalagram® 35

 18 55%

Aromatic Conversion, vol.%

40
16 Sulfur Total
35

Product Sulfur, wppm

50%
14
Aromatic Content

30
12 45%
25 10
40%
20 8
15 6 35%
4
10 30%
2
5
0 25%
0 640 660 680 700 720 740
0 0.2 0.4 0.6 0.8 1
Temperature, ˚F
1/LHSV, hrs

Mono Rings Poly Rings Saturates

FIGURE 7: Comparison of HDS and Aromatic
Saturation Using CoMo
FIGURE 6: Aromatic Concentrations in ULSD

than the rate of hydrogenation of the first ring in biphenyl. Figure 54 18 55%

Aromatic Conversion, vol.%

compares relative reactions rates for selected aromatics species. 16 Sulfur Total
50%

Product Sulfur, wppm

14
Figure 6 summarizes pilot plant data demonstrating how the aro-
12 45%
matic species change in ULSD product as a function of the resi- 10
40%
dence time (i.e. 1/LHSV) (Liquid Hourly Space Velocity) in the 8
6 35%
reactor. Notice how the curves look very similar to the simple exam-
4
30%
ple discussed in Figure 4. For PNA saturation, the 2-ringed aro- 2
matic going to the mono ring aromatic, there is a fairly steep decline 0 25%
620 640 660 680 700 720 740
in concentration as a function of residence time below about 0.5 hr.
Temperature, ˚F
Above that point, which represents space velocities of 2 hr-1 or less,
there is very little change due to equilibrium constraints. For mono-
ringed aromatic saturation there is a steady increase in conversion FIGURE 8: Comparison of HDS and Aromatic
Saturation Using NiMo
as the residence time is increased, and eventually the mono-ringed
concentration begins to decrease indicating that mono-ring satura-
tion gets a lot more favorable as the LHSV is decreased. These The maximum aromatic saturation in this case is about 42% at just
data show that PNA saturation occurs fairly readily under typical hy- under 700°F as shown on the chart.
drotreating conditions, but saturation of mono rings aromatics is
much more difficult and is aided by lower LHSV. Figure 8 shows results for a NiMo catalyst on the same feed and
conditions. In this case just over 640°F is required to achieve 10
Hydrotreaters with very short residence time (high LHSV) will have ppm product sulfur and, at that temperature, about 38% aromatics
difficulty achieving higher volume swells due to the much slower rate conversion is achieved. Comparing with the data in Figure 7 it is ap-
of saturating the final aromatic ring. These units will require a higher parent that the NiMo catalyst is significantly more active for HDS
temperature in order to drive the kinetic saturation portion of the re- than the CoMo catalyst, and it achieves slightly higher aromatics sat-
action. This can have some negative effects on catalyst perform- uration when running to make 10 ppm product sulfur despite running
ance by decreasing the expected cycle time due to the higher start at a lower temperature.
of run temperature and the increased fouling rate associated with it.
ART (Advanced Refining Technologies) was interested in exploring Comparing the catalysts in maximum aromatic saturation mode re-
aromatics saturation and the impact of product volume further, and veals significantly larger differences between catalysts. Maximum
completed some pilot plant work for a refiner. The feedstock used aromatics conversion occurs at 685°F for the NiMo catalyst and, at
for this case study contained 50% cracked material, and the operat- that temperature, the aromatics conversion is 52%. The NiMo cata-
ing conditions included 850 psi hydrogen pressure and a H2/Oil ratio lyst is achieving over 10 numbers higher aromatics conversion than
over four times the hydrogen consumption. the CoMo catalyst.

Figure 7 summarizes the HDS and aromatics conversion observed Figures 9 and 10 summarize the same data, but now show the im-
for the CoMo catalyst in that test. A temperature of 665°F was re- pact on product volume. The yields for the CoMo catalyst system
quired to achieve 10 ppm sulfur in this case. At that temperature are shown in Figure 9 along with the product sulfur. In this case the
about 36% aromatics hydrogenation was achieved which is less difference in yields from operating in ULSD mode versus a maxi-
than the maximum possible aromatic saturation for these conditions. mum volume swell mode is very low. The difference in aromatics

36 Issue No. 113 / 2013

 18 102.8
740
16 Sulfur Yield, vol.% 102.6 EOR for ULSD
Product Sulfur, wppm
14 102.4
720

Yields, vol.%
12
102.2
10
102.0
8 700

Temperature, ˚F
101.8

End of ASAT Mode

6
4 101.6
680
2 101.4
0 101.2 660
640 660 680 700 720 740
Temperature, ˚F
640 HDS - NiMo
HDPNA - NiMo
FIGURE 9: Distillate Yields Using a CoMo Catalyst 620
System
600
0 10 20 30 40 50 60 70
Cycle Length, Months

18 105.0
16 FIGURE 11: Maximum Saturation Versus ULSD Mode
104.5
Product Sulfur, wppm

14 Comparison

Yields, vol.%
12 104.0
10
103.5
8 switching to maximum saturation mode the reactor temperature is
6 103.0 ramped up to the conditions resulting in maximum PNA/HDA aro-
4
102.5 matic conversion. The temperature is then adjusted to maintain a
2
0 102.0 constant saturation level. PNA/HDA saturation activity deactivates
620 640 660 680 700 720 740
at a slower rate relative to HDS activity, so the rate of temperature
Temperature, ˚F
increase in PNA mode is much slower than for the HDS mode. The
EOR for the PNA mode of operation is determined by the required
FIGURE 10: Distillate Yields Using a NiMo Catalyst
System sulfur level of 10 wppm, at which point the ULSD unit switches to
maintaining the 10 wppm product sulfur until the EOR temperatures
conversion for ULSD and maximum aromatics is not large enough to are met.
result in any significant change in product volume. There is no eco-
nomic incentive to run for maximum volume swell with this system. The fact that saturation activity deactivates at a slower rate than
HDS activity is validated by comparing commercial operating data.
The situation is different for the NiMo catalyst as shown in Figure 10. API upgrade is often used as a simple measure of aromatics satura-
Operating in ULSD mode results in estimated distillate yields that tion and can be tracked through the cycle. Figure 12 summarizes
are about 1% higher compared to the CoMo catalyst. Of course this commercial data for API upgrade from eight different units ranging in
comes at the cost of additional hydrogen consumption with the NiMo operating pressure from 615-1900 Psig and 0.77-3.7 LHSV. The
catalyst and assumes that the extra hydrogen required for stable op- data indicate that the API upgrade is maintained throughout the
eration is readily available at a reasonable cost. The figure also cycle in these cases.
highlights the yields for running to maximum aromatic saturation.
Running the unit for maximum volume swell requires an increase in ART next examined the value of operating in a maximum volume
temperature to around 670°F. At this temperature there is over 1.0% swell mode vs. ULSD mode. For this model it was assumed that the
additional volume gain which also results in 40-60 SCFB (standard unit processes 50,000 barrels per day, and the data from Figure 10
cubic feet per barrel) additional hydrogen consumption. is used to estimate the yields. Data, like that from Figure 12, is used
to determine the cycle length expected for operating in either the
When estimating the benefits of operating a hydrotreater in maxi- HDS (ULSD mode) or maximum yield mode. Based on the under-
mum saturation mode vs. simply maintaining ULSD it is also impor- standing from Figure 11 that the expected run length will be the
tant to realize that the entire cycle is not expected to produce the same for either mode, costs such as turnaround costs and operating
additional volume swell. Figure 11 shows the results of modeling the costs will be equal and will not need to be applied in determining the
differences in ULSD temperature profiles during the cycle for the two financial impact of operating in ASAT mode. For this example the
operating strategies. In ULSD mode the reactor temperature is in- catalyst systems are also identical, so that cost of the catalyst would
creased to maintain a constant product sulfur of 10 wppm. The end not need to be included in the financial evaluation, but when consid-
of run (EOR) is typically determined by a maximum outlet tempera- ering catalyst changes this cost would be included. Therefore, the
ture and often this is the point when the product color is out of speci- only difference between these two modes of operation are the addi-
fication. In this case, EOR is reached in about 59 months. In tional barrels of product produced by operating in ASAT mode, and

Grace Catalysts Technologies Catalagram® 37

 14

12

10
Delta API
8

0
0 200 400 600 800 1000 1200 1400
Days on Stream
Refiner A Refiner B Refiner C Refiner D
Refiner E Refiner F Refiner G Refiner H

FIGURE 11: Commercial Delta API over a ULSD Cycle

80.0
45,000
79.8 44,000
Million Barrels

MMSCF Hydrogen

79.6 43,000
79.4 42,000
41,000
79.2
40,000
79.0
39,000
78.8 38,000
78.6 37,000
78.4 36,000
HDS Mode ASAT Mode 35,000
HDS Mode ASAT Mode

FIGURE 12: Commercial Delta API Over a ULSD FIGURE 13: Barrels Produced in HDS or ASAT Mode

the incremental hydrogen required to do so. For this financial analy- uct barrels. There is almost 80 SCFB standard cubic feet per barrel
sis the cost of the feed to the ULSD unit is assumed to be $5 lower in additional hydrogen consumption to produce those barrels which,
than the cost of the product being sold. over the 50-month cycle, amounts to over 5 million cubic feet of in-
cremental hydrogen consumed as shown in Figure 14. Using a hy-
With a detailed look at only the first 50 months of the cycle where drogen value of $3.00 per 1000 scf, the incremental hydrogen
the two modes of operation have different yield structures and hy- consumed amounts to a cost of just over $16 million dollars, or
drogen usages, Figure 13 shows that the HDS mode produces about $0.22 per barrel more than operating in HDS mode.
nearly 79 million barrels of product, while the ASAT mode produces
almost 80 million barrels of product. The barrels produced in the last However, the revenue from sale of the additional product barrels
9 months of the run are not included in this total, as both operating produced in ASAT mode is more than sufficient to cover the cost of
modes will need to finish in HDS operation in order to simply meet incremental hydrogen consumed. The net impact for this mode of
the ULSD product targets and maximum run length. operation is a $1.20 per barrel premium for operating in ASAT mode
versus HDS mode for the first 50 months of the cycle.
As was stated earlier, the difference between the two modes of oper-
ation is the use of hydrogen in order to produce the additional prod- ART has the ability to conduct detailed customer specific pilot plant
testing to provide the refiner the confidence and understanding of

38 Issue No. 113 / 2013

the various options available when considering a catalyst change. References
Numerous refiners have chosen to place ART catalyst into their
1. Watkins, B., Olsen, C., Hunt, D., NPRA Annual Meeting, Paper
ULSD hydrotreater in order to achieve the optimization between
AM 11-21, Balancing the Need for Low Sulfur FCC Products and In-
ULSD and maximum yield ULSD.
creasing FCC LCO Yields by Applying Advanced Technology for Cat
Feed Hydrotreating
Both the hydrotreating catalyst system and the operating strategy for
the ULSD unit are critical to providing the highest quality products.
2. Olsen, C., Watkins, B., 2009 NPRA Annual Meeting, Paper AM
Driving the hydrotreater to remove sulfur and PNA's improves prod-
09-78, Distillate Pool Maximization by Exploiting the use of Opportu-
uct value, but this needs to be balanced against the increased costs
nity Feedstock’s Such as LCO and Synthetic Crudes.
of higher hydrogen consumption. Use of tailored catalyst systems
can optimize the ULSD hydrotreater in order to produce higher vol-
3. Olsen, C., D’Angelo, G., 2006 NPRA Annual Meeting, Paper
umes of high quality products while balancing the refiners available
AM-60-06, No Need to Trade ULSD Catalyst Performance for Hy-
hydrogen.
drogen Limits: SmART Approaches

The complex relationship between hydrotreater operation and cata-

4. Girgis, M.J., Gates, B.C., Ind. Eng. Chem. Res., 30, 1991, p
lyst kinetics underscores the importance of working with a catalyst
2021
technology supplier that can tailor product offerings for each refiner’s
unique operating conditions. This knowledge enables ART to meet
the refiner’s objectives and maximize revenue.

Grace Catalysts Technologies Catalagram® 39