Assessing Compositional Changes of Nitrogen ... - ACS Publications

Hydrotreating of Typical Diesel Range Gas Oils Using a Novel ... This paper describes the identification of nitrogen-containing compounds in a typical...
1 downloads 0 Views 81KB Size
Download PDF
Ind. Eng. Chem. Res. 2000, 39, 533-540

533

Assessing Compositional Changes of Nitrogen Compounds during Hydrotreating of Typical Diesel Range Gas Oils Using a Novel Preconcentration Technique Coupled with Gas Chromatography and Atomic Emission Detection Peter Wiwel, Kim Knudsen, Per Zeuthen,* and Duayne Whitehurst Haldor Topsøe Research Laboratories, DK-2800 Lyngby, Denmark

This paper describes the identification of nitrogen-containing compounds in a typical feed for diesel oil hydrotreating and how their individual concentrations change upon hydrotreating over a conventional sulfided CoMo/Al2O3 catalyst at commercial conditions. A preconcentration procedure followed by gas chromatographic (GC) analysis utilizing a highly sensitive nitrogenspecific detector (atomic emission detector) allowed the quantitative analysis of individual nitrogen-containing compounds (N compounds) at levels as low as 0.05 µg N/mL. The nitrogen compounds in the feed and products were identified by comparison with reference compounds as well as by high-resolution GC/mass spectrometric characterization. The relative reactivities of individual compounds in the diesel fuel feed were determined and the most refractory compounds identified. Alkyl-substituted carbazoles were found to be the major compound class in the feed and to be among the least reactive N compounds in the feed. Just as in the case of alkyldibenzothiophene hydrodesulfurization, carbazoles having alkyl substituents at positions adjacent to the nitrogen atom were found to be the least reactive N compounds in the diesel fuel feed for hydrodenitrogenation. Introduction Crude oil generally contains low levels of organic nitrogen compounds (0.1-2%), but the nitrogen content strongly increases with increasing boiling point of the particular oil fraction.1-3 Diesel fuels are prepared commercially from mixtures of straight run distillates and cracked products of heavier feedstocks. Typically, the nitrogen levels found in diesel fuel feeds range from 20 to 1000 µg N/mL. These N compounds can be divided into four different chemical classes: aliphatic amines, anilines, and two heterocyclic aromatic compound groups, viz., five-membered pyrrolic and six-membered pyridinic ring systems. Most of the nitrogen in heavier petroleum fractions is present as aromatic heterocycles with multiple rings (quinolines, acridines, indoles, carbazoles, and benzocarbazoles).2-5 Aliphatic amines, anilines, and pyridinic compounds constitute the basic nitrogen components, whereas indoles and carbazoles are the nonbasic components. However, only limited information is available on how the concentration of these different N compound classes varies in different feedstocks. Analysis of nitrogencontaining compounds has been difficult because of the low concentrations found in typical feedstocks and the often tedious cleanup and enrichment procedures that are necessary prior to gas chromatographic (GC) analysis.6-22 The greatest advances in identifying the N compounds present in petroleum streams came from geochemists interested in crude oil components and how these changed during oil migration, e.g., Smitter, Igniatiadis et al.,6-15 Yamamoto,16 and Larter, Li et al.17-22 This * To whom correspondence should be addressed. E-mail: [email protected]. Fax: +45 45 27 29 99. Phone: +45 45 27 2138.

earlier work utilized extensive preseparation schemes that were quite selective for isolation of the N compounds but were difficult to quantify. Following the joint development of a new GC/atomic emission detector (AED) system by David Grudowski of Chevron Research and Technology Company and Hewlett-Packard, we utilized this new instrument to develop new insight into the composition of N species in gas oils.23a In the area of hydrotreated stream compositions, prior work identified alkylcarbazoles as low-reactivity N compounds,9,11,15 but little information on the relative rates of individual compound conversions was gained because of the above-mentioned difficulties in quantifying the results. In this paper we report a novel simple method for obtaining quantitative results for the contents of individual N compounds in real gas oils and their hydrotreated products. The method consists of a preconcentration step to eliminate the hydrocarbon matrix that interferes with the nitrogen analysis by atomic emission detection, which is highly sensitive and provides a linear response with nitrogen concentration at low levels. The removal of organic nitrogen is essential to many different refinery processes. It is well-known that nitrogen molecules have a significant influence on the formation of coke, as many of these molecules are known to be coke precursors. Kinetically, organic nitrogen molecules have a pronounced inhibitive effect on hydrotreating reactions such as hydrodesulfurization (HDS), other hydrogenolysis reactions, and hydrogenation. Moreover, removal of nitrogen is important in order to lower the emission of nitrogen oxides when oil fractions are burned. Organic nitrogen is removed catalytically by the socalled hydrodenitrogenation process (HDN). Generally, HDN is the most difficult hydrotreatment reaction, and

10.1021/ie990554e CCC: $19.00 © 2000 American Chemical Society Published on Web 01/08/2000

534

Ind. Eng. Chem. Res., Vol. 39, No. 2, 2000

Table 1. Temperature Program for the GC Oven and Parameters for the AED a. Temperature Program for the GC Oven injection port temperature (°C) 275 AED transfer line/cavity 300 temperature (°C) oven temperature (°C) 80-300 °C at 7 °C/ min, hold 10 min column pressure (psi) 13.2 carrier gas helium split ratio 7:1 injection volume (µL) 1 b. Parameters for the AED data rate (Hz) 2.5 reagent gases (psi) oxygen/hydrogen/methane: 80:40:50 makeup flow (mL/min) 240

little is known about which N compounds are the most problematic, their mechanisms of conversion or the kinetics of their conversion during HDN processes. When the new analytical procedure described in this paper and the new instrument developed by Chevron/ Hewlett-Packard23a,b are utilized, it is now possible to follow the kinetics of individual N compounds in real feeds as these are converted to products in typical hydrotreating processes. This is the first of a series of papers intended to elucidate the limitations of HDN processes with today’s catalysts and processes. Experimental Section Gas Oil Samples. For this study, a typical feed and a hydrotreated product in diesel hydrotreating have been investigated. The feedstock was a blend of a straight run Kuwait gas oil and a light cycle oil (LCO) from a North Sea crude. The blend ratio was 33/67. The feedstock contained cracked materials and therefore had relatively high concentrations of nitrogen and aromatic compounds. The product was produced in an isothermal bench-scale downflow reactor at 350 °C, pressure ) 30 bar, and LHSV ) 1.0 h-1. The feed contained 302 µg N/mL of nitrogen and 0.627 wt % sulfur, whereas the product contained 10 ppm of sulfur and 70 µg N/mL. Analysis of sulfur compounds was conducted using gas chromatography with sulfur-specific chemiluminescence detection. The total nitrogen content was determined by oxidative combustion and chemiluminescence detection (ASTM D4629-91). A commercial catalyst, sulfided Co-Mo on alumina, TK-554, from Topsøe was used in this study. This catalyst has a very high HDS activity but, like other CoMo catalysts, a moderate HDN activity. GC and GC/MS Techniques. Reference N compounds were used to identify many of the methyl-, dimethyl-, trimethyl-, and tetramethylcarbazoles. Dr. M. Kuroki, Shibaura Institute of Technology, Saitama, Japan,24 kindly provided these. Carbazole and acridine were purchased from Fluka, Buchs, Switzerland. All gases used for the chromatography were of the highest purity available. In addition, helium and hydrogen were further cleaned by the use of an in-line gas purifier and an in-line palladium diffusion unit. GC was performed on a Hewlett-Packard model 6890 with a split/splitless injector port and an AED model G2350A. The chromatographic separation was carried out on a 30 m HP1MS 0.32 mm internal diameter capillary column with a film thickness of 1.0 µm (Hewlett-Packard). Details on the temperature program of the gas chromatographic oven are listed in Table 1a.

General information on the AED technique can be found in ref 23a,b. The initial instrument optimization was done according to suggestions by Grudowski et al.23a Nitrogen emission at 388 nm was used with the AED because of its high selectivity over carbon. Details on the parameters used with the AED are listed in Table 1b. A Hewlett-Packard model 6890 GC coupled to a Hewlett-Packard model 5973 MS was used to confirm the molecular masses of the reference compounds and to identify additional N compounds in the samples. Sample Preparation. The nitrogen-containing compounds were concentrated by solid-phase extraction (SPE) prior to analysis using a procedure similar to that described by Larter, Li et al.17-22 The polar compounds, including all of the nitrogen compounds in this study, were trapped on a pure silica SPE column (Isolute-Si from International Sorbent Technology) and then recovered for analysis by flushing with acetone. For quantification of the nitrogen compounds, an internal standard was added to the oil prior to the SPE treatments. Acridine was found to be useful for this purpose because this compound was not found in the feedstock or in the product gas oil. The procedure developed for the present studies was considerably simpler than that of Larter et al.17 It consisted of diluting the standard into about 0.5 mL of the oil sample and then diluting the mixture to 10 cm3 with heptane. The diluted oil was injected onto the SPE column and the column flushed with an additional 10 cm3 of heptane to remove the hydrocarbons. The N compounds and other polar materials were then recovered by flushing the SPE column with 10 cm3 of acetone. Prior to analysis of the acetone effluent by GC/AED, the sample was concentrated to about 0.1-1 cm3 by evaporation. Analysis of the total nitrogen content in the nonretained fraction from the SPE extraction showed that more than 99% of the N compounds were retained by the SPE column. Using this procedure, it was found that 90-108% of the nitrogen determined by ASTM D4629-91 could be detected by the GC/AED system. The detection limit for individual compounds using this SPE extraction procedure was found to be as low as 10 pg N/mL in the sample being analyzed, under ideal conditions. Results and Discussion Identification of organic nitrogen compounds in various oil fractions most often involves complicated chromatographic techniques in order to isolate the basic as well as the nonbasic compounds. Derivation of the isolated N compounds is often needed in the identification of the individual compounds. The GC analysis of the SPE extract of the feedstock using the nitrogen-specific detector (AED) is shown in Figure 1. Visually, the nitrogen-specific chromatogram can be divided into two fractions, one before and one after a retention time of 23 min. The fraction eluting after 23 min accounts for a significant amount of the total nitrogen in the sample. Analysis of carbazole and methyl-substituted carbazole standards and verification of the masses by GC/MS showed that compounds that eluted after 23 min were predominantly methylsubstituted carbazoles. The compounds that eluted earlier than 23 min in Figure 1 were identified as indole, quinoline, aniline, and their methyl-substituted derivatives. Positive identifications of 64 of the N components shown in Figure 1 were made and are summarized in Table 2. The exact positions of the alkyl groups were identified for most, but not all, of these compounds.

Ind. Eng. Chem. Res., Vol. 39, No. 2, 2000 535

Figure 1.

Part a and b of Figure 2 show the expanded feed chromatogram. Figure 2a represents compounds eluting earlier than 23 min, whereas Figure 2b shows compounds eluting after 23 min. Table 2 lists the identified peaks in their absolute amounts in the feedstock and the hydrotreated product. The ID numbers refer to peaks in Figure 2a,b. It is seen that most of the basic nitrogen compounds that were identified in reasonable amounts were present in the fraction eluting earlier than that of the carbazoles and that these were quite reactive in the hydrotreating process. However, as will be discussed later, the carbazoles were quite refractory, and the reactivity of the various carbazole derivatives varied substantially with the position and number of alkyl substituents on the carbazole nucleus. Alkyldibenzothiophenes also behave in this way, as will be discussed later. For reference, the chemical structures and the nomenclature of carbazole and its sulfur analogue dibenzothiophene are as follows:

The total nitrogen of the feed and product was determined to be 302 and 70 µg N/mL, respectively (ASTM D4629-91). Although it is difficult to use the GC/AED as a total nitrogen detector, a simple way of checking the efficiency of the SPE extraction is to add up all of the signals from the detector and compare this to the known total amount of nitrogen. It should be noted that there is, in addition to the sharp individual peaks, a substantial broad hump of unresolved N species that must be considered. The identification of all of the minor peaks constituting the broad hump will be the subject of future research work. When the whole spectrum including the broad hump is integrated, the total nitrogen detected by GC/AED was found to be in the order of 90-110% of that determined by ASTM D462991. Table 2 lists only the major species observed, and their concentrations were determined by integration of the areas of the peaks above the broad hump. The total areas of the identified peaks accounted for about 62%

of the total nitrogen in the feedstock and about 65% of the total nitrogen in the product. Most of the minor compounds were identified as quinolines, isoquinolines, benzoquinolines, naphthylamines, and diphenylamines. Carbazole and monomethyl-substituted carbazoles represented about 27% of the total amount of the carbazole fraction in this feed. The data also showed that 1-methylcarbazole was the most predominant compound in the feedstock. Of the di- and trimethylcarbazoles, substitution at the 1 position was also observed to be the most predominant. The high relative concentrations of 1-substituted carbazoles in natural crudes have been attributed to preferential migration of nitrogen compounds having sterically shielded N centers, which have low adsorptive interactions with the clay minerals through which the crude has migrated.17-19 The high relative concentrations of 1-substituted carbazoles in the products are attributed to lower reactivity of these species, as will be discussed later. The compositions shown in Table 2 compare well with past reports on the compositions of N compounds in gas oils;7-22 however, this new procedure allows a more quantitative assessment of the contents of individual species and how they change during processing. The main advantages of this new procedure are simplification of the concentration procedures, elimination of interfering carbon AED signals through hydrocarbon removal, and use of a new high-sensitivity AED detector that provides linear response with nitrogen concentration at very low levels. It was found that individual N components could easily be detected and quantified at concentrations of less than 0.05 µg N/mL. GC analysis of the hydrotreated product is shown in Figure 3. Parts a and b of Figure 4 show the expanded product chromatogram. The conditions used for the hydrotreatment of the feedstock were quite severe. Supplemental analyses of the sulfur species remaining in the hydrotreated product showed that the sulfur content had been reduced from 0.627 wt % S to about 10 ppm S, or 99.8% conversion. All of the aliphatic, monoaromatic, and diaromatic S compounds were completely removed, and even dibenzothiophene and 4-methyldibenzothiophene were completely converted. The 10 ppm S remaining consisted primarily of more highly substituted dibenzothiophenes, and specifically only 1.2 ppm S was detected as 4,6dimethyldibenzothiophene. The feedstock was found to

536

Ind. Eng. Chem. Res., Vol. 39, No. 2, 2000

Table 2. Feed and HDT Product Compositions and Relative Rate Constants peak ID no. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64

compound name aniline C1-aniline C1-aniline ethylaniline C2-aniline C2-aniline C2-aniline C2-aniline C2-aniline quinoline propylaniline C3-aniline C3-aniline indole C3-aniline C3-aniline C3-aniline butylaniline C4-aniline 7-methylindole 2- and 3-methylindole 4-, 5-, and 6-dimethylindole C2-indole C2-indole C2-indole C2-indole C2-indole C2-indole C3-indole C3-indole C3-indole C3-indole C3-indole C4-indole C4-indole carbazole 1-methylcarbazole 3-methylcarbazole 2-methylcarbazole 4-methylcarbazole 1,8-dimethylcarbazole 1,3-dimethylcarbazole 1,6-dimethylcarbazole 1,7-dimethylcarbazole 1,4- and 1,5-dimethylcarbazole 3,6-dimethylcarbazole 2,6-, 3,5-, 2,7-dimethylcarbazole 2,4- + 1,2-dimethylcarbazole 2,5-dimethylcarbazole 2,3-dimethylcarbazole 1,4,8-trimethylcarbazole 3,4-dimethylcarbazole 1,3,5-trimethylcarbazole 1,5,7-trimethylcarbazole 2,4,6-trimethylcarbazole 1,3,4- and 2,4,7-trimethylcarzole 1,4,5-, 2,3,6- and 2,3,5-trimethylcarbazole 3,4,6-trimethylcarbazole C3-carbazole C4-carbazole C4-carbazole C4-carbazole C4-carbazole C4-carbazole sum of identified peaks total N (by ASTM D4629-91)

concentration feed product (µg N/mL) (µg N/mL)

conversion (%)

0.5 1.6 1.2 0.2 1.7 1.7 0.9 0.5 0.7 0.3 0.1 0.7 0.7 2.1 0.6 1.0 0.4 0.2 0.2 2.0 1.8 4.3 3.8 2.2 2.5 2.3 0.3 0.4 1.9 2.1 0.8 2.6 0.8 1.0 0.5 7.6 11.1 6.0 7.7 7.5 5.7 4.5 5.0 6.0 10.6 2.9 7.9 5.8 3.8 5.2 7.2 2.2 8.8 5.1 1.4 4.4 2.8 2.5 1.2 1.4 1.3 1.2 1.1 3.2 185.8

0.9 4.0 3.1 0.6 1.9 1.3 1.5 0.7 1.0 0.7 0.7 0.6 2.7 45.4

100 100 100 100 100 100 100 100 100 100 0 100 100 100 100 100 100 63 100 98 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 97 84 96 93 86 59 74 73 69 56 79 81 75 68 60 52 61 54 40 58 57 54 41 40 28 45 42 50 17 76

302.0

70.0

77

0.1

0.1 0.0

0.2 1.8 0.2 0.5 1.1 2.3 1.2 1.3 1.9 4.7 0.6 1.5 1.4 1.2 2.1 3.5

relative first-order constanta

100 G 49 G 87 G 73 G 53 G 24 G 37 P 36 P 32 R 22 M 43 P 45 M 38 M 31 R 25 G 20 G 25 P 21 R 14 G 24 P 23 P 21 P

All rate constants are set relative to carbazole k ) 100. G ) good quality, P ) poor quality, R ) reasonably good, and M ) rate for mixture reasonably good. a

contain 52.2 ppm S as 4,6-dimethyldibenzothiophene. Thus, even this highly refractory S compound was 97.7% converted. By comparison, only about 77% of the nitrogen compounds were converted. Thus, many of the N compounds present in this feed were considerably less reactive than the hindered dialkyldibenzothiophenes.

This product contains the most refractive nitrogencontaining compounds found in this feed. Dominant species were identified as methylated carbazoles. This result supports the results previously reported by Schmitter, Igniatiadis et al.9,11,14,15 This work indicated the general resistance of carbazole derivatives toward hydrotreating.

Ind. Eng. Chem. Res., Vol. 39, No. 2, 2000 537

Figure 2.

Most of the lower boiling compounds identified in the feedstock are converted under these rather severe hydrotreating conditions into ammonia and hydrocarbons (see Table 2). This means that indole, quinoline, aniline, and their methylated derivatives have a higher reactivity than the carbazoles. Essentially, no basic compounds were observed in this severely hydrotreated product. Most of the compounds in the product that did elute earlier than 23 min were new compounds that were not present in the feed. These were identified as aniline derivatives and are believed to be reaction intermediates in the HDN of the carbazoles. Basic organonitrogen compounds have often been described as the strongest inhibitors of the HDS reaction and have therefore been the most studied model nitrogen compounds.25-27 However, only a few quantitative inhibition effects have been reported. In studies by Nagai28 and La Vopa,29 the inhibition effect of the nonbasic carbazoles appears unexpectedly strong. Carbazole’s adsorption parameter is comparable to that of much more basic species. This unusual inhibition is

most likely due to the consequence of its conversion under reaction conditions into basic nitrogen compounds, which are stronger inhibitors. The ease of hydrodenitrogenation is often debated to be dependent upon the nitrogen compound type and its base strength. Holmes reported that nitrogen removal is independent of the compound’s basicity and/or that the nitrogen compounds are converted to more basic compounds during hydrogenation.30 Thus, competitive adsorption by intermediates complicates kinetic interpretation considerably. The only nonbasic nitrogen compound, for which quantitative reaction networks have been reported, is indole.31,32 The hydrogenation of indole to give indoline was proposed to be a rapid reaction that reached virtual equilibrium. No organonitrogen products were detected in which the benzenoid ring was saturated, which implies the preferential hydrogenation of the nitrogencontaining ring as for other organonitrogen compounds. Very little information has been reported on the reaction pathway for carbazoles or methyl-substituted

538

Ind. Eng. Chem. Res., Vol. 39, No. 2, 2000

Figure 3.

carbazoles. We have recently investigated the HDN of several such compounds and found that both carbazole and 1,4-dimethylcarbazole are in rapid equilibrium with their tetrahydro analogues.33 This supports the previous conclusions for indole31,32 and the suggested flatwise adsorption of nitrogen compounds on the active catalytic sites.34 However, we have found, from model compound studies on carbazoles, that the HDN products of carbazole and its methyl derivatives always include large quantities of benzene-ring-opened and methylcyclopentane structures.33 Such products are difficult to explain using the conventional reaction pathways that have been proposed in the literature.25,26 In the present study we have investigated the reactivity of the individual carbazole derivatives by comparing the amounts present in the feed with those present in the product. Assuming simple first-order kinetics, we have estimated the relative rate constants for the individual compounds using carbazole as the reference (k ) 100). It should be noted that the accuracy of these calculations is somewhat in question for many of these compounds because the peaks were not fully resolved (see Figures 2b and 4b). These estimated rate constants are presented in Table 2 along with a qualitative judgment as to the accuracy of the numbers. When the chromatograms of feed and product (Figures 2b and 4b) as well as the results shown in Table 2 are compared, it can be seen that the relative abundance of the various compounds changed considerably during hydrotreatment. In the feed, carbazole and monomethylcarbazoles were the most predominant species, whereas in the product, the di- and trimethylcarbazoles were the dominant species. This shows that, as the degree of substitution increased, the reactivity decreased. In the feedstock, carbazole and monomethylcarbazoles accounted for 27% of the nitrogen, whereas in the product, they only amounted to 8%. The reason for the specific increased resistance of these nonbasic N compounds to HDN appears to be related to the position of the alkyl substituents on the aromatic nuclei. This is quite similar to the behavior of alkyldibenzothiophenes in that substitution adjacent to the heteroatom lowers the reactivity substantially.25,33,34 Previous authors have noted that hydrotreated products contain high amounts of 1-substituted carba-

zoles.9,11,14,15 Our results confirm those results and provide additional kinetic data on the relative reactivity of carbazoles substituted in different positions. The difference in reactivity of the different methyl-substituted carbazoles is, however, not quite as pronounced as that of the dibenzothiophene analogues. It is wellknown that 4-methyldibenzothiophene and 4,6-dimethyldibenzothiophene are among the most refractory sulfur compounds (please note the different nomenclature for dibenzothiophenes and carbazoles described above). Relative differences in reactivity according to Gates et al.34 and Mochida et al.35 are as follows: DBT/ 4-MDBT/4,6DMDBT is 10:3:1. This difference in reactivity was proposed to be due to steric inhibition of a transition state in the rate-limiting step of the overall reaction mechanism.35 Fortunately, we were able to estimate the first-order rate constant for the HDS of 4,6-dimethyldibenzothiophene in our present HDN studies. The rate constant for HDS of 4,6-DMDBT was observed to be almost the same as the rate constant for the HDN of carbazole. Thus, alkyl-substituted carbazoles appear to react at rates about 1/10 as fast as those of alkyldibenzothiophenes of comparable structures. Surprisingly, it was noted that alkyl substitution at the 4 position of carbazole lowers the reactivity almost as much as substitution at the 1 position (adjacent to the N heteroatom). No literature data are available for the relative reactivity of alkyldibenzothiophenes of similar structures, but it has been inferred that substitution at positions other than those adjacent to the S heteroatom has little effect on reactivity.25 This may imply that HDS and HDN proceed by different mechanisms. However, studies of additional alkyldibenzothiophenes are needed to clarify this issue. It is well-known that there are two reaction pathways for HDS of such alkyl-substituted dibenzothiophenes.25 One route involves prior hydrogenation of one aromatic ring prior to removal of the sulfur atom (hydrogenative route), and the other route involves direct extraction of the sulfur atom without ring hydrogenation (direct route). It has been shown that the rates of the hydrogenative routes are not substantially affected by the position or number of alkyl substituents on dibenzothiophene, whereas the direct route is strongly reduced by substituents adjacent to the sulfur atom.34,35

Ind. Eng. Chem. Res., Vol. 39, No. 2, 2000 539

Figure 4.

CoMo-based catalysts operate primarily via the direct HDS route; thus, for the sulfided CoMo catalyst used in the present studies, it is not unreasonable that 4,6DMDBT reacts more slowly. However, the observed relative rates for HDN of different alkyl-substituted carbazoles seem rather surprising and quite similar to the relative rates for HDS of alkyldibenzothiophenes. We have found in other studies that the rate-limiting step in HDN is the cleavage of the first C-N bond,33 and it is curious to see very similar relative rates for HDN and HDS for compounds of similar structures. Because it is known that the HDS rates with the catalysts used in this study primarily correspond to the differences in the rates of the direct HDS as influenced by alkyl groups adjacent to the sulfur atom, the observed lower reactivities for similarly substituted alkylcarbazoles may also imply that the HDN rates are also subject to some steric restriction on an intermediate, as is the case for alkyldibenzothiophenes. Work is in progress to clarify this issue.

Conclusions Using a novel analytical procedure, including preconcentration of N compounds and a GC analysis coupled with a highly sensitive and selective AED, we have been able to identify most of the refractive organic nitrogen compounds present in a typical gas oil and its hydrotreated product. Alkyl-substituted carbazoles were found in large amounts in the feed, and carbazole compounds substituted at position 1 were found to be the most abundant. 1-Methylcarbazole was the single most predominant species. Carbazoles were found to be the most refractory organic N compounds in the feed toward HDN. Generally, the more methyl substituents on the carbazole, the lower the reactivity. However, there appeared to be extraordinarily low reactivity for carbazoles having substituents at the 1 and 8 positions, just as is known for comparable dibenzothiophene HDS reactivities. Surprisingly, alkyl substitution at the 4 and 5 positions also appeared to lower the reactivity substantially, which is

540

Ind. Eng. Chem. Res., Vol. 39, No. 2, 2000

inconsistent with steric restrictions in the immediate vicinity of the heteroatom. The cause of this lower reactivity is presently unknown. Acknowledgment The authors thank Dr. M. Kuroki, Shibaura Institute of Technology, Saitama, Japan, for kindly providing all of the methyl-substituted carbazoles used as reference compounds. More details of the syntheses of the model compounds are given in ref 21. We are also deeply indebted to Dr. David Grudowski of Chevron Research and Technology Co. for his help with optimization of the GC/AED system and for continuing cooperation in the development of this and other analytical procedures.23a Literature Cited (1) Minderhoud, J. K.; Van Veen, J. A. R. First-Stage Hydrocracking: Process and Catalytic Aspects. Fuel Process. Technol. 1993, 35, 87. (2) Boduszynski, M. M. Characterization of “Heavy” Crude Components. Prepr. Pap.sAm. Chem. Soc., Div. Fuel Chem. 1985, 365-382. (3) Boduszynski, M. M. Composition of Heavy Petroleums. I. Molecular Weight, Hydrogen Deficiency, and Heteroatom Concentration as a Function of Atmospheric Equivalent Boiling Point up to 1400 °F (760 °C). Energy Fuels 1987, 1, 2-11. (4) Snyder, L. R. The Nitrogen and Oxygen Compounds in Petroleum. Prepr. Pap. Am. Chem. Soc., Div. Pet. Chem. 1970, 15 (2), C44-C62. (5) McKay, J. F. K.; Weber, J. M.; Latham, D. R. Characterization of Nitrogen Bases in High-Boiling Petroleum Distillates. Anal. Chem. 1976, 48, 891. (6) Schmitter, J. M.; Vajta, Z.; Arpino, P. Investigation of Nitrogen Bases from Petroleum. In Advances in Organic Geochemistry; Douglas, A. G., Maxwell, J. R., Eds.; Pergamon Press: New York, 1980. (7) Schmitter, J. M.; Igniatiadis, I.; Arpino, P. Distribution of Diaromatic Bases in Crude Oils. Geochim. Cosmochim. Acta 1983, 47, 1975-1984. (8) Dorbon, M.; Schmitter, J. M.; Garrigues, P.; Igniatiadis, I.; Ewald, M.; Arpino, P.; Guiochon, G. Distribution of Carbazole Derivatives in Petroleum. Org. Geochem. 1984, 7, 111-120. (9) Dorbon, M.; Igniatiadis, I.; Schmitter, J. M.; Arpino, P.; Guiochon, G.; Toulhout, H.; Huc, A. Identification of Carbazoles and Benzocarbazoles in a Coker Gas Oil and Influence of Catalytic Hydrotreatment on Their Distribution. Fuel 1984, 63, 565. (10) Schmitter, J. M.; Garrigues, P.; Igniatiadis, I.; De Vazelhes, R.; Perin, F.; Ewald, M.; Arpino, P. Occurrence of Tetra-Aromatic Aza-Arenes in Petroleum. Org. Geochem. 1984, 6, 579-586. (11) Schmitter, J. M.; Igniatiadis, I.; Dorbon, M.; Arpino, P.; Guiochon, G.; Toulhoat, H.; Huc, A. Identification of Nitrogen Bases in Coker Gas Oil and Influence of Catalytic Hydrotreatment on Their Composition. Fuel 1984, 63, 557-564. (12) Schmitter, J. M.; Arpino, P. Azaarenes on Fuels. Mass Spec. Rev. 1985, 4, 87-121. (13) Igniatiadis, I.; Dorbon, M.; Arpino, P. Analyse comparative et identification des de´rive´s du carbazole extraits de deux pe´troles bruts du Congo. Analusis 1985, 13, 406-414. (14) Igniatiadis, I.; Schmitter, J. M.; Arpino, P. Separation and Identification through GC and GC/MS of Nitrogen Compounds of a Heavy Deasphalted Oil. Evaluation of Their Distribution after Catalytic Hydrotreatment. J. Chromatogr. 1985, 324, 87-111. (15) Igniatiadis, I.; Kuroki, M.; Arpino, P. Identification of Carbazole Derivatives in a Hydrotreated Coker Gas Oil by Gas Chromatography and Gas Chromatography-Mass Spectrometry. J. Chromatogr. 1986, 366, 251-260. (16) Yamamoto, M. Fraction of Azaarenes during Oil Migration. Adv. Org. Geochem. 1991, 19, 389-402. (17) Li, M.; Larter, S. R.; Stoddart, D. Liquid Chromatographic Separation Schemes for Pyrrole and Pyridine Nitrogen Aromatic Heterocycle Fractions from Crude Oils Suitable for Rapid Characterization of Geochemical Samples. Anal. Chem. 1992, 64, 13371344.

(18) Li, M.; Larter, S. R.; Frolov, Y. B. Adsorptive Interaction Between Nitrogen Compounds and Organic and/or Mineral Phases in Subsurface Rocks. J. High Res. Chromatogr. 1994, 17, 230236. (19) Larter, S. R.; Aplin, A. C. Reservoir Geochemistry: Methods, Applications and Opportunities. In The Geochemistry of Reservoirs; Cubitt, J. M., England, W. A., Eds.; Geological Society Special Publication No. 86; Geological Society of America: Boulder, CO, 1995. (20) Li, M.; Larter, S. R.; Stoddar, D.; Bjorøy, M. Fractionation of Pyrrolic Nitrogen Compounds in Petroleum during Migration: Derivation of Migration-Related Geochemical Parameters. In The Geochemistry of Reservoirs; Cubitt, J. M., England, W. A., Eds.; Geological Society Special Publication No. 86; Geological Society of America: Boulder, CO, 1995. (21) Larter, S. R.; Bowler, B. F. R.; Li, M.; Chen, M.; Brincat, D.; Bennett, B.; Noke, K.; Donohoe, P.; Simmons, D.; Kohenen, M.; Allan, J.; Telnaes, N.; Horstad, I. Molecular Indicators of Secondary Oil Migration Distances. Nature 1996, 383, 593-597. (22) Li, M.; Yao, V.; Stasiuk, L. D.; Fowler, M. G.; Larter, S. R. Effect of Maturity and Petroleum Expulsion on Pyrrolic Nitrogen Compound Yields and Distributions in Duvernay Formation Petroleum Source Rocks in Central Alberta, Canada. Org. Geochem. 1997, 26, 731-744. (23) (a) Grudowski, D. Presented at Advances in Desulfurization Technologysthe Design of New Desulfurization Catalysts and Process Schemes, Fukuoka, Japan, Nov 1997, and personal communication. (b) Webster, C.; Cooke, M. Capillary Gas Chromatography with Atomic-Emission Detection. Chromatogr. Anal. 1994, 35, 9-11. (24) Kuroki, M.; Tsunashima, Y. J. The Chemistry of Carbazoles. VII. Syntheses of Methyl Carbazoles. Heterocycl. Chem. 1981, 18, 709. (25) Girgis, M. J.; Gates, B. C. Reactivities, Reaction Networks, and Kinetics in High-Pressure Catalytic Hydroprocessing. Ind. Eng. Chem. Res. 1991, 30 (9), 2021-2058. (26) Ho, T. C. Hydrodenitrogenation Catalysis. Catal. Rev.s Sci. Eng. 1988, 30 (1), 117-160. (27) van Looij, F.; van der Laan, P.; Stork, W. H. J.; DiCamillo, D. J.; Swain, J. Key Parameters in Deep Hydrodesulfurization of Diesel Fuel. Appl. Catal. A 1998, 170, 1-12. (28) Nagai, M.; Kabe, T. Selectivity of Molybdenum Catalyst in Hydrodesulfurization, Hydrodenitrogenation, and Hydrodeoxygenation: Effect of Additivies on Dibenzothiophene Hydrodesulfurization. J. Catal. 1983, 81, 440. (29) La Vopa, V.; Satterfield, C. N. Poisoning of Thiophene Hydrodesulfurization by Nitrogen Compounds. J. Catal. 1988, 110, 375. (30) Holmes, S. A.; Thompson, L. F. Nitrogen Compound Distributions in Hydrotreated Shale Oil Products from Commercial-Scale Refining. Fuel 1983, 62, 709. (31) Olive`, J.-L.; Biyoks, S.; Moulinas, C.; Geneste, P. Hydroprocessing of Indole and o-ethyl Aniline over Sulfided CoMo, NiMo and NiW Catalysts. Appl. Catal. 1985, 19, 165. (32) Stern, E. W. Reaction Networks in Catalytic Hydrodenitrogenation. J. Catal. 1979, 57, 390. (33) Whitehurst, D. D.; Ovesen, C. H.; Nielsen, I. V. The Influence of Trace Amounts of Nitrogen Compounds on the Achievement of Future Sulphur Specifications in Diesel Fuel Part I: HDS Inhibition Studies with Model N-Compounds. To be published. (34) Gates, B. C.; Topsøe, H. Reactivities in Deep Catalytic Hydrodesulfurization: Challenges, Opportunities and the Importance of 4-methyldibenzothiophene and 4,6-dimethyldibenzothiophene. Polyhedron 1997, 16 (18), 3213. (35) Whitehurst, D. D.; Isoda, T.; Mochida, I. Present State of the Art and Future Challenges in the Hydrodesulfurization of Polyaromatic Sulfur Compounds. Adv. Catal. 1998, 42, 345-471 and references therein.

Received for review July 26, 1999 Revised manuscript received November 8, 1999 Accepted November 8, 1999 IE990554E