DEVELOPMENT AND COMMERCIALIZATION OF, A HIGH ACTIVITY, LOW COST PARAFFIN ISOMERIZATION CATALYST W.S. Graeme, M.N.T. van der Laan Akzo Nobel Catalysts ABSTRACT Akzo Nobel s high activity paraffin isomerization catalysts have been used successfully since 1995 in both C 4 and C 5 /C 6 applications. The introduction of AT-2(G) marked only the beginning of a continuous development and commercialization effort focused on improving Akzo Nobel's products. In 1999 AT-10, a butane isomerization catalyst with the same high activity level of AT-2 but at only half the platinum content, was commercialized. In the same year was introduced, even now it is still the most active catalyst for isomerization of C 5 /C 6. To improve liquid distribution characteristics, the catalyst shape was change from trilobe to cylindrical in 2001. The cylindrical shape contributes to optimized performance in hydrogen once through units. This continuous catalyst improvement effort has resulted in a new class of catalyst,, a catalyst Akzo Nobel jointly developed with Axens. is a low-density catalyst that has not only the high volume activity of but also lower fill costs and platinum requirements associated with low-density. In this paper the development concept of is presented. Pilot plant test results for both and are offered for comparison. The test-work results show that its lower density does not hinder the performance of. 1
INTRODUCTION In 1995 Akzo Nobel entered the market of paraffin isomerization catalysts with the introduction of AT-2 for butane isomerization. Since 1995, Akzo Nobel has continuously been working to improve its isomerization catalyst portfolio. The first result of this work was the introduction of AT-2G in 1996. In 1999 Akzo Nobel introduced, a catalyst with the same platinum content as AT-2G, but higher activity due to improved platinum dispersion. is still the most active catalyst available in today s market. Not only has improving activity been the focus of Akzo Nobel but also there has been a focus on developing cost effective catalysts. The first was the introduction of AT-10 for butane isomerization in 1999. AT-10 has the same high activity level of AT-2, but at only half the platinum content. The reduction in platinum content is possible because the platinum function in the catalytic reaction does not limit the rate of isomerization. Since 2001 all Akzo Nobel paraffin isomerization catalysts have been produced with a cylindrical shape. Our extensive testing has shown that this shape contributes to optimized hydraulic behavior. This is important in view of the increased use of hydrogen once-through (HOT) flow-schemes that operate at low gas velocities with low gas phase fractions making the process more vulnerable to maldistribution. The cylindrical shape will protect the high activity and improve the utilization of the catalyst. Our customer's need to reduce catalyst expenses has continued to drive the Akzo Nobel development of cost-effective catalysts for paraffin isomerization. Considerable effort was expended to develop the alternative, a catalyst Akzo Nobel jointly developed with Axens. The challenge was clear, bring a catalyst to market with the high activity typical of Akzo Nobel s AT catalysts but with lower catalyst costs and platinum requirements. This paper describes the development concept of the low density catalyst and presents pilot plant testing data showing that even with lower density and lower platinum requirements, high volumetric activity is maintained. Since 1995 more than 20 customers throughout the world have chosen a catalysts from the AT-family. All the AT-catalysts are platinum containing chlorinated alumina catalysts. The unique chlorination step uses metal alkyls, instead of aluminum trichloride. This results in a higher acid site density and thus better stability through increased water and water precursor tolerance. I. DEVELOPMENT CONCEPT The refining and petrochemical industry is under constant pressure to reduce the operating expenses and by extension catalyst costs. Akzo Nobel's prior catalyst developments initially concentrated on introducing a more active and more stable catalyst into the market. But with many units operating close to their thermodynamic equilibrium, catalyst activity is only one element of many decision criteria. Short loading a reactor, one of the advantages high activity offers, is not a consideration for most operating companies. As the target for development of Akzo Nobel chose to combine both a low fill cost and a high activity. The target of the development efforts was defined as: 1. Produce a catalyst with at least the same volumetric activity of the most widely used third party catalyst. 2. Decrease the catalyst and platinum requirements by about 20%. 2
Since the cost of the platinum can be more than the costs of the catalyst (depending on the market price for platinum) this affect alone substantially lowers the catalyst cost when compared to third party catalysts. In cases where platinum from an existing catalyst will be reclaimed the excess will decrease the capital employed in the unit. A series of pilot plant experiments were designed to compare the volumetric activity of with and the current widely used catalyst. II. EVALUATION OF DEVELOPMENT CONCEPT The pilot plant used to compare the different catalysts consists of two reactors in series. Like most commercial units the order of the reactors can be changed allowing a direct comparison of the activity of two catalyst types at the same operating conditions. In the experiments the catalysts were compared on a weight basis. Based on model calculations the results were corrected for its lower density to facilitate the comparison of activities on a volume basis. Two experiments were carried out. Experiment 1: Reactor one (lead reactor) is loaded with and reactor two (lag reactor) is loaded with. The experiment starts with in lead position; is move to the lead position later in the test. Experiment 2: Both reactors (lead and lag reactor) are loaded with the third party catalyst. The conditions chosen are a simulation of a commercial unit s operation, including the low hydrogen to hydrocarbon ratio (H 2 /HC) common to hydrogen once through operation. Due to test phenomena the test results may not scale up directly to a commercial operation and tend to be conservative. To demonstrate that the ranking of the catalysts is H 2 /HC ratio independent a higher than normal H 2 /HC ratio has been applied in the pilot plant test. The feedstock used for these tests is a combination of fresh feed and a recycle stream from a de-isohexanizer column. Table 1 gives the combined feedstock characterization. The same feedstock was used for both experiments. Table 1: Composition of combined feedstock Component wt% C 5 - components 0.8 Pentane 19.3 Iso-pentane 7.8 Cyclopentane 1.5 Hexane 21.1 2,2 di-methyl-butane 0.5 2,3 di-methyl-butane 2.8 2 methyl-pentane 15.9 3 methyl-pentane 12.5 Cyclohexane 4.2 Methyl-cyclopentane 9.1 C 7 + components 2.7 Benzene 1.8 3
The sum of C 6 naphthenes, benzene and C 7 + (the X factor) for this combined feed is 17.8 %wt, which is a typical value for today s C 5 /C 6 isomerization unit feeds. In table 2, the test conditions are summarized. The hydrogen to hydrocarbon ratio (H 2 /HC) is measured at the outlet of the lag reactor. All tests were carried out at a pressure of 30 barg. Table 2: Test conditions Condition WHSV H 2 /HC ratio Catalyst type Temperatures Lead Lag Lead Lag h -1 mol/mol C C 1 1.9 0.05 165 140 2 1.9 0.05 165 140 3 1.9 0.05 165 140 4 1.8 0.20 165 140 5 1.8 0.20 165 140 For the pilot plant test two different hydrogen to hydrocarbon ratios were used. During the tests, and were alternately operated in lead position at these different ratios. (See conditions 1,2 and 4,5 above.) Since the reactor temperatures of the test remained unchanged, a direct activity comparison between these two catalysts can be made. The test on the reference catalyst was carried out at the same conditions for benchmarking with Akzo Nobel s catalysts. In this test the number of conditions could be limited to changing the hydrogen to hydrocarbon ratio since both reactors were loaded with the same catalyst (see conditions 2 and 5). 1. Pilot plant test results In figure 1 the results from the test on and are presented. In this chart the paraffin isomerization number (PIN) of the product from the lead and lag reactors are plotted as function of time. The PIN represents the degree of paraffin isomerization and is the accepted indicator for comparing the isomerization activity of different catalysts. The PIN is defined as follows: isopentane in product PIN = sum of C paraffins in product 5 + 2,2 dimethylbutane and 2,3 dimethylbutane in product sum of C paraffins in product 6 4
Condition 1 2 3 4 5 120 110 lead: lag: 100 PIN 90 80 70 PIN ex lead reactor PIN ex lag reactor 60 0 312 Time on stream Figure 1. PIN results of / test In figure 2 the results from the test on the reference catalyst are presented. Condition 2 5 120 110 100 PIN 90 80 70 60 PIN ex lead reactor PIN ex lag reactor 0 312 Time on stream Figure 2. PIN results on reference catalyst 5
In table 3 the average PIN level observed at each test condition is summarized Table 3: Average PIN at outlet lead and lag reactor for / test Condition WHSV H 2 /HC ratio Catalyst type PIN Lead Lag Lead Lag h -1 mol/mol 1 1.9 0.05 77.7 93.8 2 1.9 0.05 83.7 95.1 3 1.9 0.05 78.2 93.8 4 1.8 0.20 93.2 107.8 5 1.8 0.20 98.1 108.1 In table 4 the results on the reference catalyst are presented. Table 4: Average PIN at outlet lead and lag reactor for reference catalyst Condition WHSV H 2 /HC ratio Catalyst type PIN Lead and lag Lead Lag h -1 mol/mol 2 1.9 0.05 Reference catalyst 72.5 86.9 5 1.8 0.20 Reference catalyst 84.9 100.1 2. Discussion of test results In the test that compares and, the PIN achieved with in lead position is 5-6 points higher than when is operated under the same conditions in lead position. Even when is operated in lag position a slightly better performance for the combined system is observed compared to operating in lag position. If the higher PIN is translated into activity, is 20% more active compared to on weight basis. However, in a commercial unit activity should be compared on volume basis rather than weight basis. Because the density difference between these catalysts is 20%, the activity of both and will be similar on volumetric basis. If the activity of the reference catalyst is compared with at the outlet of the lead reactor the delta PIN is even higher at 11-13 points. When this difference is translated into isomerization activity, is 40% more active on weight basis and 20% on volumetric basis. The conclusion from the pilot plant test is that and have the same volumetric activity and thus will show similar performance in a commercial unit. Based on the test conditions used for this comparison, and are 20% more active, on volumetric basis, than the widely used reference catalyst. This means that 20% more of the reference catalyst is required to achieve a similar performance. 6
3. Third party test results Several major oil company research centers have received samples of and to evaluate the performance claims. In close consultation with the research centers the pilot plant test formats were defined such that equal volumes of catalyst were loaded. By testing with equal volumes a direct comparison of the volumetric activity of both catalysts is possible. We received responses from two of these research centers both independently confirming that the activity of is comparable to. One of these companies is using in a C 5 /C 6 isomerization unit. The research center ran a one-month stability test on and found that the performance remained stable. In table 5 the conditions of the stability test are summarized. Table 5: Test conditions stability test Pressure 30 barg H 2 /HC at inlet of lead reactor 0.5 mol/mol LHSV 2 h -1 X factor feed 10.1 wt% Lead reactor temperature 155 C Lag reactor temperature 135 C Just before the stability test a pilot test on was conducted. This test was carried out to obtain a baseline for activity comparison between and at the test format defined. In figure 3 a comparison of the initial activity of and is presented. In figure 4 the results from the one month stability test are plotted. Since at reactor conditions the 2,3 dimethylbutane concentration in the isomerate product is at equilibrium conversion level, instead of PIN the customer uses the so called TIN number for activity evaluation: isopentane in product TIN = sum of C paraffins in product 5 2,2 dimethylbutane + sum of C paraffins in product 6 7
110 100 TIN 90 LHSV=2 h-1 LHSV=3 h-1 LHSV=2 h-1 80 70 60 0 50 100 150 200 250 300 350 Time on stream (hrs) Figure 3. TIN results of / test as carried out by customer 120 110 TIN 100 90 80 70 LHSV=3 h -1 T optimization 60 0 100 200 300 400 500 600 700 Time on stream (hrs) Figure 4. TIN results stability test Figure 3 shows that the activity levels of and are the same. Since both tests were carried out at the same LHSV, this confirms equal volume activity for both catalyst grades. The one month stability test underlines that the high initial activity of is maintained and even does not suffer from a temporary increase of the severity by 50% (LHSV from 2 to 3 h -1 ). 8
III. Conclusions With Akzo Nobel and Axens have introduced a catalyst that combines the high activity level of with lower fill costs and platinum requirements. During verification of the development concept, not only at Akzo Nobel, but also two major research centers, it was shown that the performance of is not hindered by its lower density. The first commercial unit in which will be installed is expected to come on stream early 2003. REFERENCES 1. Boer, M. de; Nat, P.J.; Broekhoven, E.H. van; New high performance catalysts for the isomerization of n-c 4 and n-c 5 /C 6 ; NPRA, 1997. 2. Boer, M. de; Johnson, C.; Commercial Experience with AT-2 and AT-2G Paraffin Isomerization Catalysts; Akzo Nobel Catalysts Symposium, 1998. 3. Decker, S.; Le Gall R.; New Improvement in Paraffin Isomerization Catalysts; ERTC paper 1999. 4. Nieuwland, J.J.; Laan, M.N.T. van der; Bruijn, A. de; Le Gall, R.; Catalyst development for paraffin isomerization: introduction of AT-10 and cylindrical AT-2G; Akzo Nobel Catalysts Symposium, 2001. 9