The Future Role of Hydrogen in Petrochemistry and Energy Supply DGMK Conference October 4-6, 2010, Berlin, Germany

Catalytic Ring Opening of Decalin – Bifunctional versus Hydrogenolytic Pathways J. Weitkamp*, S. Rabl*, A. Haas*, D. Santi*, M. Ferrari**, V. Calemma** *Institute of Chemical Technology, University of Stuttgart, Germany ** Eni R&M Division, San Donato Milanese, Italy

Abstract

Ir/silica, Pt/La-X and Rh/H-Beta were prepared and tested in the hydroconversion of cis- decalin at different temperatures. The catalytic tests were carried out under hydrogen in a high-pressure flow-type apparatus at 5.2 MPa. On the three catalysts open-chain decane yields up to 20 % were achieved, which is much higher than the yields reported so far in the literature. Pt/La-X and Rh/H-Beta behave as bifunctional catalysts with a high tendency for skeletal isomerization. On these catalysts the so-called paring reaction via carbenium ions occurs, leading to iso-butane and methylcyclopentane as main hydrocracked products. On Ir/SiO2, carbon-carbon bond cleavage occurs through hydrogenolysis on the noble metal without prior isomerization. As a consequence the product spectrum is less complex than on the bifunctional catalysts which makes the system particularly amenable to mechanistic studies.

Introduction

Polynuclear aromatics in diesel fuel bring about various undesirable properties, such as poor ignition characteristics and cetane numbers, an increased propensity for soot formation and unfavorable cold-flow properties. For these reasons, the content of polynuclear aromatics in diesel fuels is limited by legislation, and certain refinery streams which are notoriously rich in these undesired components can be blended into diesel fuels only to a limited extent. The selective ring opening of polynuclear aromatics into high-value diesel components, in particular mildly branched alkanes, without degradation of the carbon number, continues to be among the great challenges of catalysis. It is generally agreed upon that such a ring opening must be preceded by a complete ring hydrogenation, and the resulting multi-ring naphthenes would be the true precursors of breaking up the rings.

Prior work with model hydrocarbons, typically decalin, revealed that ring opening (or, synonymously, hydrodecyclization) can occur on acidic, bifunctional or metallic catalysts. Acidic zeolites were found to open one ring of decalin so that alkylnaphthenes with a single ring are formed, but at elevated conversions these catalysts tend to degrade the carbon number into C9- hydrocarbons and to deactivate rapidly [1, 2]. Bifunctional catalysts consisting of the acid form of a large-pore zeolite, mostly faujasite or Beta, and a noble metal were predominantly used in hydrodecyclization studies [3-9]. Monofunctional metallic catalysts on non-acidic supports have been scarcely used for ring opening of multi-ring naphthenes. However, among these investigations is a particularly thorough one devoted to the hydrogenolysis of various bicyclic naphthenes over iridium on non-acidic supports. Interestingly, while opening of one ring in bicyclic naphthenes is described in all above- mentioned reports, opening of both rings to the particularly desired alkanes with the carbon number of the feed is even not mentioned - with two exceptions: McVicker et al. [10] and Daage et al. [11, 12] observed traces of open-chain decanes (OCDs) and nonanes (OCNs) in the hydrogenolysis of decalin (bicyclo[4.4.0]decane) and perhydroindan

DGMK-Tagungsbericht 2010-3, ISBN 978-3-941721-07-4 77 The Future Role of Hydrogen in Petrochemistry and Energy Supply

(bicyclo[4.3.0]nonane), respectively, on 0.9 wt.-% Ir on Al2O3, and noticeable amounts of octanes in the hydroconversion of bicyclo[3.3.0]octane on the same catalyst. Very recently, Mouli et al. [13] reported on the formation of OCDs from decalin on a bifunctional catalyst, viz. Ir,Pt/H-Y zeolite, but their best OCD yields (4 %) and selectivities (5 %) were low.

We recently developed experimental techniques tailored for the study of catalytic hydrodecyclization of two-ring naphthenes. The aim of this paper is two-fold: Using a non- acidic 2.6Ir/SiO2 catalyst and two bifunctional zeolite catalysts (1.0Pt/La-X and 5.0Rh/H- Beta), it will be demonstrated that significantly higher yields and selectivities of OCDs can be obtained from decalin than hitherto reported both by hydrogenolytic and bifunctional hydrodecyclization. Moreover, we will outline the very different reaction paths involved in the hydrogenolysis of decalin on iridium and its hydrodecyclization on the bifunctional zeolite catalysts.

Experimental Section

Preparation and Characterization of the Catalysts

Within the frame of this study, three catalysts were prepared, namely non-acidic Ir/SiO2 and two bifunctional zeolites Pt/La-X and Rh/H-Beta. Table 1 gives some important physico- chemical properties of the supports and the metal-loaded catalysts.

Table 1: Physico-chemical properties of the supports and catalysts.

Sample Specific Pore volume / Metal loading / Metal dispersion surface area / cm3 g-1 wt.-% m2 g-1 Silica 391 1.07 - - La-X 529 0.39 - - Na-Beta 513 0.96 - - Ir/silica - - 2.59 1.02 Pt/La-X - - 1.00 0.43 Rh/H-Beta - - 4.99 0.42

A Varian optical emission spectrometer with an inductively coupled plasma (ICP-OES) Vista- MPX CCD was used for chemical analysis of the samples. The metal loading of all samples is defined as the mass of the metal per mass of dry catalyst. To detect the mass of the dry catalyst, it was first stored in a desiccator over a saturated aqueous solution of calcium nitrate for at least 24 h. The precise water content of the resulting samples was then measured by means of a Setaram Thermogravimetric Analyzer (TGA) Setsys TG-16/18. In the TGA experiment, the sample was heated in a nitrogen flow from room temperature to 600 °C with a heating rate of 20 K min-1. For the determination of the noble-metal dispersion the amount of irreversibly adsorbed hydrogen was measured in a Quantachrome Autosorb-1- C instrument by static volumetry. The samples were reduced similar to the treatment prior to the catalytic experiments and evacuated. After cooling, two isotherms were measured at T = 313 K. The first isotherm was considered to be a combination of physi- and chemisorption, and the second isotherm, measured after evacuating the sample, was interpreted as physisorption only. The difference of these two isotherms originating from irreversibly and strongly adsorbed molecules was applied for calculating the noble-metal dispersion with an assumed adsorption stoichiometry of nH / nnoble metal = 1. Porous properties were measured by N2 adsorption at -196 °C in a Quantachrome Autosorb-1-C instrument after degassing the samples at 350 °C for 16 h. For the calculation of BET specific surface areas p / p0 values

78 DGMK-Tagungsbericht 2010-3 The Future Role of Hydrogen in Petrochemistry and Energy Supply

between 0.1 and 0.3 were applied.

Ir/SiO2 was prepared by electrostatic adsorption [14] of [Ir(NH3)5Cl]Cl2 onto silica (Aerosil 380, Degussa/Evonik). By addition of NH4OH to a suspension of silica in demineralized water a pH ≈ 10 was obtained in order to deprotonate the silanol groups. Subsequently, an ion exchange with an aqueous solution of [Ir(NH3)5Cl]Cl2 was possible, followed by filtration and washing with demineralized water. The Aerosil particles had an average particle size of 7 nm and were pressed without a binder and again crushed and sieved to a size between 0.20 and 0.32 mm, before they were used in the flow apparatus. The complex was then decomposed in situ in a flow of synthetic air at 150 °C for 3 h, whereupon the metal was reduced in flowing hydrogen at 400 °C for 2 h.

The starting material for Pt/La-X zeolite was Na-X (Strem Chemicals) with nSi / nAl = 1.21. It -1 was two times ion-exchanged with an aqueous solution of 0.076 mol l La(NO3)3 for 2 h at 80 °C. After each ion exchange, the zeolite was heated in an air flow at 450 °C in order to allow the La3+ ions to migrate into the small cages [15]. At this stage 90 % of the sodium ions were exchanged with lanthanum ions. The zeolite was suspended in demineralized water, and an aqueous solution of [Pt(NH3)4]Cl2 was added dropwise under vigorous stirring. Afterwards, the suspension was kept at 80 °C under stirring for 4 h, and the resulting solid was filtered off, washed with demineralized water and dried at 80 °C in air. Next, the complex was decomposed in an air flow at 300 °C. The noble metal was reduced in situ in flowing hydrogen at p = 5.2 MPa and T = 380 °C for 2 h. Before using this catalyst in the flow-type apparatus the 2 to 3 μm particles of zeolite X were pressed, crushed and sieved to a size fraction between 0.20 and 0.32 mm.

Zeolite Beta with nSi / nAl = 14.0 was synthesized from colloidal silica (Ludox HS-40), aluminum sulfate 18 hydrate and tetraethylammonium hydroxide solution as a template via the dry-gel conversion method [16]. To remove the template, the as-synthesized zeolite was heated in a nitrogen flow from room temperature to 450 °C with a rate of 1 K min-1, holding at 450 °C for 24 h, then switching the gas flow to synthetic air and holding at 450 °C for another 24 h. Subsequently, a two-fold ion exchange was carried out at 80 °C with a 1 mol l-1 solution of NaNO3. Zeolite Na-Beta was washed nitrate-free and suspended in demineralized water at 80 °C, and an aqueous solution of [Rh(NH3)5Cl]Cl2 was added dropwise under vigorous stirring. Thereafter, stirring was continued for 24 h at 80 °C. Next, the zeolite was filtered off, washed with demineralized water and dried in air at 80 °C. The complex was decomposed in an air flow at 300 °C, and the noble metal was reduced in situ in flowing hydrogen at p = 5.2 MPa and T = 360 °C for 2 h. Before using this catalyst in the flow-type apparatus the 5 to 30 μm particles of zeolite Beta were pressed, crushed and sieved to a size fraction between 0.20 and 0.32 mm.

Hydroconversion of Decalin

All catalytic experiments were conducted in a high-pressure flow-type apparatus equipped with a saturator and a fixed-bed reactor. The saturator was filled with an inert solid (Chromosorb P/AW) and liquid cis-decalin (Merck, ≥ 98 %), and it was kept at 135 °C corresponding to a decalin partial pressure of 25 kPa. The total pressure amounted to 5.2 MPa in all experiments. The mass of dry catalyst was 0.17 g, 0.28 g and 0.19 g for Ir/SiO2, Pt/La-X and Rh/H-Beta, respectively. The hydrogen flow rate and the LHSV were 130 cm3 min-1 referenced to 0.1 MPa and room temperature and 0.4 h-1, respectively. With each catalyst, the reaction temperature was varied. After a time-on-stream of 4 h, a gaseous product sample was introduced into an on-line capillary gas chromatograph (Agilent 7890 A or Hewlett-Packard HP 6890 N) for analysis equipped with a flame ionization detector and a Supelco Petrocol DH150 capillary column. Furthermore, an integral liquid product sample was collected for off-line analyses in a cooling trap held at -10 °C.

DGMK-Tagungsbericht 2010-3 79 The Future Role of Hydrogen in Petrochemistry and Energy Supply

Product Analysis and Evaluation of the Data

Catalytic hydroconversion of decalin tends to result in very complex product mixtures consisting of up to ca. 200 or more hydrocarbons. Assignment of the peaks in a chromatogram was mainly based on GC/MS analyses of the liquid products in the cooling trap. A separate gas chromatograph (Agilent 6890 N) equipped with a Supelco Petrocol DH150 capillary column and coupled to a mass spectrometer (Agilent 5876 B inert XL MSD) was used for this purpose. In addition, commercially available pure hydrocarbons were coinjected with the liquid product samples from the catalytic experiments. Furthermore, for a safe identification of the OCDs, n-decane was isomerized in separate experiments on a Pt/H,Na-Y zeolite catalyst at 262 °C, and the mixtures of iso-decanes thereby generated (a reliable analysis of which was possible on the basis of previous studies of catalytic n-alkane isomerization [17]) were co-injected with the liquid products from decalin hydroconversion.

To facilitate the subsequent discussion, the reaction products will be classified into the following groups: (i) Skeletal isomers (sk-Isos) of decalin with two naphthenic rings and a molar mass of 138 g mol-1; (ii) ring opening products (ROPs) with one remaining naphthenic ring and a molar mass of 140 g mol-1; (iii) open-chain decanes (OCDs) with a molar mass of 142 g mol-1; (iv) dehydrogenated products (DHPs), mainly tetralin and naphthalene and (v) hydrocracked products (C9-) with less than ten carbon atoms. On all three catalysts, trans- decalin was formed in a very fast reaction, but this isomer of the feed was treated as unconverted decalin.

Particularly desired products are the ROPs and OCDs. Conversely, the hydrocracked products (C9-) are particularly undesired. We nevertheless placed much emphasis on their detailed analysis hoping that the nature of the C9- hydrocarbons formed could furnish valuable mechanistic information. For a quantitative discussion, the modified hydrocracking selectivity will be used which is defined as the molar amount of hydrocarbons with j carbon atoms formed (j = 1 to 9) divided by the molar amount of decalin converted into hydrocracked products:

 n j  S *  formed (1) j  nDecalin converted to C 9-

Results and Discussion

Description of the Products Formed from Decalin

In Figure 1 the conversion of decalin and the selectivities of different groups of products on the non-acidic Ir/silica and the two bifunctional zeolite catalysts are depicted. Clearly, the two bifunctional zeolite catalysts are significantly more active and bring about a complete or nearly complete decalin conversion at ca. 250 °C where the conversion on Ir/silica is still below 20 %.

80 DGMK-Tagungsbericht 2010-3 The Future Role of Hydrogen in Petrochemistry and Energy Supply

100 100 100 Ir/silica Pt/La-X Rh/H-Beta

80 80 80

60 60 60

40 40 40

20 20 20

Conversion or selectivity / % selectivity or Conversion

0 0 0 260 280 300 220 240 260 200 220 240 Temperature / °C Temperature / °C Temperature / °C Figure 1: Conversion of decalin ( ) and selectivities of skeletal isomers ( ), ring opening products ( ), open-chain decanes ( ), hydrocracked products ( ) and dehydrogenated products ( ) on the Ir/silica, Pt/La-X and Rh/H-Beta catalysts at different temperatures.

Very drastic differences between Ir/silica and the two bifunctional zeolite catalysts are also observed in the product selectivities: Essentially no skeletal isomers of decalin are formed on Ir/silica, regardless of the reaction temperature and conversion. Instead, ring opening products strongly predominate at low conversion on Ir/silica. With increasing temperature and conversion, the selectivity of ROPs decreases, while correspondingly, the selectivities of open-chain decanes and C9- hydrocarbons increase. The selectivity of OCDs passes through a maximum of 27 % at T = 290 °C. At still higher temperatures, the undesired degradation into hydrocarbons with less than 10 carbon atoms becomes rapidly dominant.

On the two bifunctional zeolite catalysts the prevailing products at low conversions are skeletal isomers of decalin. With increasing conversion, their selectivities decrease, and those of the ring opening products and open-chain decanes increase until they reach maxima. The maximal selectivities of ROPs and OCDs are 24 % and 13 %, respectively, for zeolite Pt/La-X and 17 % and 19 %, respectively for Rh/H-Beta. Increasing the temperatures and conversions further again leads to a rapid increase of the selectivities of the undesired C9- hydrocarbons. On all three catalysts used in this study, the formation of dehydrogenated products is negligible.

The occurrence of open-chain decanes on both the non-acidic Ir/silica and the two bifunctional zeolite catalysts is remarkable in view of the very scarce information concerning this group of particularly interesting products in the prior literature. Note that the maximum selectivities of OCDs observed in the present study for Ir/silica, Pt/La-X and Rh/H-Beta are, respectively, about 5-, 2.5- and 4-times higher than the most favorable value reported so far in the literature [13]. The catalysts used in this study seem to be excellent starting materials for further increasing the selectivities of open-chain alkanes in the hydroconversion of bicyclic naphthenes.

DGMK-Tagungsbericht 2010-3 81 The Future Role of Hydrogen in Petrochemistry and Energy Supply

Table 2: Maximum yields of open-chain decanes and selectivities of open-chain nonanes and hydrocracked products obtained on all catalysts.

S Catalyst Tr / % XDec / % SOCDs / % YOCDs,max. / % SOCNs / % C9- / % Ir/silica 290 73 27 20 9 35 Pt/La-X 250 95 13 12 0 30 Rh/H-Beta 228 97 14 14 0 51

In Table 2, the maximal yields of OCDs for the three catalysts are compared. Using this criterion, again the non-acidic Ir/silica catalyst shows the best performance of the three catalysts used in this study: At T = 290 °C, YOCDs reaches its maximal value of 20 %. Very interestingly, at this reaction temperature, open-chain nonanes (OCNs) are formed on Ir/silica with a selectivity of 9 %, whereas no such products occur at all in the hydroconversion of decalin on Pt/La-X or Rh/H-Beta. This result will be re-addressed in the discussion of the reaction paths on the non-acidic iridium and the bifunctional zeolite catalysts.

100 Ir/silica Pt/La-X Rh/H-Beta 80

60

* / % */

j

S

40

20

0 1 2 3 4 5 6 7 8 9 Number of C-atoms Figure 2: Modified hydrocracking selectivities Sj* on the three catalysts:

Catalyst Ir/silica: Tr = 270 °C; XDec = 37 %; YC9- = 7 %;  Sj* = 216 %.

Catalyst Pt/La-X: Tr = 250 °C; XDec = 95 %; YC9- = 28 %;  Sj* = 185 %.

Catalyst Rh/H-Beta: Tr = 228 °C; XDec = 97 %; YC9- = 49 %;  Sj* = 196 %.

The pronounced differences between Ir/silica on the one hand and the two bifunctional zeolite catalysts on the other hand become also obvious from the carbon number distributions of the C9- products observed at yields of hydrocracked products between 7 % and 49 % (Figure 2). Plotted against the carbon number are the modified hydrocracking selectivities as defined in Eqn. (1). To the best of our knowledge, this is the first time that such distribution curves are shown for the catalytic hydroconversion of decalin.

All three distribution curves are nearly symmetrical around C5, i. e., equal molar amounts of Cj and C(10-j) are formed (except for C1 and C9 on the two bifunctional zeolite catalysts). Overall, the distribution curves for Pt/La-X and Rh/H-Beta are very similar: They are M- shaped with pronounced maxima at C4 and C6. It is noteworthy that the C4 fractions consist mainly (Pt/La-X: 96 mol-%, Rh/H-Beta: 96 mol-%) of iso-butane, and the C6 fractions contain large amounts (Pt/La-X: 81 mol-%, Rh/H-Beta: 39 mol-%) of methylcyclopentane. Presumably, a large portion of the methylcyclopentane originally formed on Rh/H-Beta was ring-opened to 2-methylpentane (28 mol-%), 3-methylpentane (19 mol-%) and n-hexane

82 DGMK-Tagungsbericht 2010-3 The Future Role of Hydrogen in Petrochemistry and Energy Supply

(5 mol-%).

By contrast, the distribution curve for the non-acidic Ir/silica catalyst resembles a hammock with a strongly preferred formation of C1 and C9 fragments and very little C4, C5 and C6. We believe that such a distribution curve can only be interpreted in terms of a mechanism of formation of the C9- hydrocarbons via hydrogenolysis on the noble metal.

Proposed Pathway of Decalin Conversion on the Pt/La-X and Rh/H-Beta Zeolite Catalysts

As one can clearly see from Figure 1, a striking difference between the bifunctional catalysts and Ir/SiO2 is the very high tendency of the former to form skeletal isomers of decalin at low conversion, whereas negligible amounts of sk-Isos are formed on Ir/SiO2 over the whole temperature range. Skeletal isomerization of decalin occurring as a first step on bifunctional catalysts is in accordance with a mechanism proposed by Kubička et al. for decalin transformation on Pt/H-Beta [18]. A comprehensive identification of the numerous skeletal isomers of decalin is difficult, but hints can be collected by GC/MS. Suggestions as to which kind of isomers are likely to be formed were made in Ref. [19]. The large number of skeletal isomers of decalin can all undergo ring opening and hydrocracking leading to very complex product mixtures. Tracing back the precise pathway of formation of each of these product hydrocarbons is probably a too ambitious task.

After isomerization and ring opening, the carbon number distributions of the C9- products observed on both catalysts (Figure 2) may furnish valuable hints concerning the next steps in the complex reaction network: Very similar M-shaped carbon number distributions with pronounced maxima at C4 and C6 and iso-butane and methylcyclopentane as the main hydrocracked products have been reported previously for hydrocracking of C10 one-ring naphthenes on both non-zeolitic [20] and zeolitic [21] bifunctional catalysts. The term “paring reaction” has been coined for this surprisingly selective reaction. The chemistry of the paring reaction is rather well understood: Starting from any arbitrary one-ring naphthene with 10 carbon atoms, carbocations are formed on the acid sites of the catalyst. These are very reluctant to show endocyclic carbon-carbon bond rupture, but rather undergo several skeletal isomerizations in the alkyl side chain (see Figure 3) until one out of three possible isomers with branchings in the α,γ,γ-positions (relative to the positively charged carbon atom) is formed. Carbenium ions with such a carbon skeleton can undergo very rapid (so-called type A [22]) β-scissions in the alkyl side chain. As Figure 3 shows, these very strongly favored β- scissions all lead to iso-butane and methylcyclopentane as hydrocracked products.

Figure 3: The three possible C10 cycloalkylcarbenium ions that can undergo exocyclic type A β-scission. M-CPn: methylcyclopentane; iso-Bu: iso-butane.

While invoking the occurrence of the paring reaction satisfactorily explains the large amounts

DGMK-Tagungsbericht 2010-3 83 The Future Role of Hydrogen in Petrochemistry and Energy Supply

of hydrocracked products with 4 and 6 carbon atoms on the two bifunctional catalysts, it does not account for the formation of small amounts of hydrocarbons with other carbon numbers. In particular the formation of methane and C9 hydrocarbons can hardly be interpreted by a mechanism via carbocations. Nor is the catalytic pathway for the noticeable amounts of open-chain decanes on the two bifunctional zeolite catalysts clear at this time. It cannot be ruled out that a small mechanistic contribution of hydrogenolysis on the noble metals plays a role in the formation of both C1 and C9 moieties and of open-chain decanes from one-ring C10 naphthenes.

Proposed Pathway of Decalin Conversion on the Ir/SiO2 Catalyst

On the non-acidic Ir/SiO2 catalyst the hydrodecyclization pathway appears to be much simpler, and less product hydrocarbons are formed than on bifunctional catalysts. The main reason is that iridium lacks the capability of isomerizing decalin, so the reaction starts with a direct opening of one six-membered ring by hydrogenolysis on the noble metal. Relatively high temperatures are required for this direct ring opening, since carbon-carbon bonds in a six-membered ring are difficult to cleave [10].

Decalin Hydrogenolysis on iridium on non-acidic or very weakly acidic supports

+ H2 + H 2 Butylcyclohexane n-Nonane - CH4 4-Methyloctane + H2 3-Ethylheptane Propylcyclohexane + H2 + H2 OCDs 4-Ethylheptane - CH4 3,4-Dimethylheptane

1-Methyl- + H2 2-propylcyclohexane 3-Ethyl-2-methylhexane + H2 3-Ethyl-4-methylhexane + H 1-Ethyl-2-methyl- 2 - CH4 cyclohexane

1,2-Diethylcyclohexane

direct ROPs direct C9-ROPs direct OCNs

OCD pathway OCN pathway

Figure 4: Two main pathways for the hydrogenolytic steps on iridium on non-acidic or very weakly acidic supports.

As shown in Figure 4 direct ring hydrogenolysis is expected to give the direct ROPs butylcyclohexane (B-CHx), 1-methyl-2-propylcyclohexane (1-M-2-P-CHx) and 1,2- diethylcyclohexane (1,2-DE-CHx), the last two naphthenes occurring as cis- / trans-isomers. Indeed, B-CHx, cis- and trans-1-M-2-P-CHx and cis- and trans-1,2-DE-CHx are the predominating products formed on Ir/SiO2 at low conversion of XDec = 17 %, they account for 96 % of all ring opening products. Increasing the conversion leads to the appearance of two new classes of products, namely (i) open-chain decanes (Figure 4, OCD pathway) formed by endocyclic hydrogenolysis of the direct ROPs, and (ii) to propylcyclohexane (P-CHx) and 1- ethyl-2-methylcyclohexane (1-E-2-M-CHx) formed by exocyclic abstraction of methane from

84 DGMK-Tagungsbericht 2010-3 The Future Role of Hydrogen in Petrochemistry and Energy Supply

the five direct ROPs by hydrogenolysis. We refer to these latter products as direct C9-ROPs. The important role of exocyclic methane abstraction is reflected in the carbon number distribution of the hydrocracked products for Ir/SiO2 (Figure 2): The hammock-shaped curve originates mainly from the preferred formation of C1 and C9 fragments. The C9 fraction consists mainly (82 %) of the direct C9-ROPs 1-E-2-M-CHx and P-CHx. Almost the entire remaining C9 hydrocarbons (16 %) are open-chain nonanes (OCNs) formed by hydrogenolytic ring opening of the direct C9-ROPs (Figure 4, OCN pathway). Again, this step proceeds in a very selective manner: at moderate conversion more than 90 % of all OCNs are products of the ring opening of the direct C9-ROPs. The seven OCN isomers formed in this manner are listed in Figure 4. The individual OCD isomers formed via the OCD pathway are not listed individually in Figure 4, because some OCDs to be expected by endocyclic hydrogenolysis of the direct ROPs could not be safely identified in the chromatograms.

As an alternative pathway for OCN formation, hydrogenolysis of OCDs could also be invoked, but this has not been shown in Figure 4 to keep it sufficiently simple.

The formation of open-chain nonanes from decalin in noticeable amounts seems to be a characteristic feature of hydrogenolytic ring opening on non-acidic iridium catalysts (see Table 2). No open-chain nonanes are formed on the two bifunctional catalysts, while open- chain decanes are formed on all three catalysts used in this study.

Conclusions

The catalysts used in this study can be divided into two groups. The bifunctional catalysts consisting of platinum or rhodium supported on acidic La-X or Beta zeolites show a high tendency for isomerization of decalin. After ring opening of the first ring, the paring reaction leading to iso-butane and methylcyclopentane occurs, but a competitive reaction exists leading to the highly desired open-chain decanes with a maximal yield between 10 and 15 %. These yields are about three times higher than the best results reported so far in the literature.

Iridium supported on non-acidic silica is a pure hydrogenolysis catalyst with negligible activity for skeletal isomerization. Ring opening proceeds on the noble metal. Iridium opens directly one naphthenic six-membered ring of decalin. The five direct ring opening products react further into one of the following directions: (i) Hydrogenolysis of an endocyclic carbon-carbon bond leads to open-chain decanes or (ii) methane is split off from the alkyl side chain whereby C9-ROPs are formed. A consecutive hydrogenolytic ring opening of the latter gives open-chain nonanes. Yields of open-chain decanes and open-chain nonanes as high as 20 and 6 %, respectively, could be attained on Ir/SiO2. The pronounced tendency of this catalyst for methane abstraction is reflected by hammock-shaped carbon number distributions of the hydrocracked products.

Many details of the bifunctional and the hydrogenolytic pathways of decalin hydrodecyclization remain to be elucidated in further studies. Nevertheless, it can be stated that the three catalysts looked at in this investigation are excellent starting materials for the development of further improved ring opening catalysts.

DGMK-Tagungsbericht 2010-3 85 The Future Role of Hydrogen in Petrochemistry and Energy Supply

References

[1] D. Kubička, N. Kumar, P. Mäki-Arvela, M. Tiitta, V. Niemi, T. Salmi, D.Y. Murzin, J. Catal. 222 (2004) 65-79. [2] M. Santikunaporn, J.E. Herrera, S. Jongpatiwut, D.E. Resasco, W.E. Alvarez, E.L. Sughrue, J. Catal. 228 (2004) 100-113. [3] D. Kubička, N. Kumar, P. Mäki-Arvela, M. Tiitta, V. Niemi, H. Karhu, T. Salmi, D.Y. Murzin, J. Catal. 227 (2004) 313-327. [4] M.A. Arribas, A. Martínez, Appl. Catal., A 230 (2002) 203-217. [5] M.A. Arribas, P. Concepción, A. Martínez, Appl. Catal., A 267 (2004) 111-119. [6] M.A. Arribas, A. Corma, M.J. Díaz-Cabañas, A. Martínez, Appl. Catal., A 273 (2004) 277-286. [7] H. Ma, X. Yang, G. Wen, G. Tian, L. Wang, Y. Xu, B. Wang, Z. Tian, L. Lin, Catal. Lett. 116 (2007) 149-154. [8] M.A. Arribas, J.J. Mahiques, A. Martínez, in: “Zeolites and Mesoporous Materials at the Dawn of the 21st century”, A. Galarneau, F. Di Renzo, F. Fajula, J. Vedrine (Eds.), Studies in Surface Science and Catalysis, Vol. 135, Elsevier, Amsterdam, 2001, p. 303. [9] M.A. Arribas, A. Martínez, G. Sastre, in: “Impact of Zeolites and Other Porous Materials on the New Technologies at the Beginning of the New Millennium”, R. Aiello, G. Giordano, F. Testa (Eds.), Studies in Surface Science and Catalysis, Vol. 142, Part B, Elsevier, Amsterdam, 2002, p. 1015-1022. [10] G.B. McVicker, M. Daage, M.S. Touvelle, C.W. Hudson, D.P. Klein, W.C. Baird, Jr., B.R. Cook, J.G. Chen, S. Hantzer, D.E.W. Vaughan, E.S. Ellis, O.C. Feeley, J. Catal. 210 (2002) 137-148. [11] M. Daage, G.B. McVicker, M.S. Touvelle, C.W. Hudson, D.P. Klein, B.R. Cook, J.G. Chen, S. Hantzer, D.E.W. Vaughan, E.S. Ellis, in: “Zeolites and Mesoporous Materials at the Dawn of the 21st Century”, A. Galarneau, F. Di Renzo, F. Fajula, J. Vedrine (Eds.), Studies in Surface Science and Catalysis, Vol. 135, Elsevier, Amsterdam, 2001, p. 159. [12] M. Daage, G.B. McVicker, M.S. Touvelle, C.W. Hudson, D.P. Klein, B.R. Cook, J.G. Chen, S. Hantzer, D.E.W. Vaughan, E.S. Ellis, Prepr. – Am. Chem. Soc., Div. Pet. Chem. 47 (2002) 38-40. [13] K.C. Mouli, V. Sundaramurthy, A.K. Dalai, J. Mol. Catal., A 304 (2009) 77-84. [14] J.T. Miller, M. Schreier, A.J. Kropf, J.R. Regalbuto, J. Catal. 225 (2004) 203-212. [15] E.F.T. Lee, L.C. Rees, Zeolites 7 (1987) 143-147. [16] P.R.H.P. Rao, C.A. Leon y Leon, K. Uyama, M. Matsukata, Microporous Mesoporous Mater. 21 (1998) 305-313. [17] J. Weitkamp, Ind. Eng. Chem. Prod. Res. Dev. 21 (1982) 550-558. [18] D. Kubička, T. Salmi, M. Tiitta, D.Y. Murzin, Fuel 88 (2009) 366-373. [19] D. Kubička, M. Rönnholm, S.-P. Reinikainen, T. Salmi, D.Y. Murzin, Anal. Chim. Acta 537 (2005) 339-348. [20] R.F. Sullivan, C.J. Egan, G.E. Langlois, R.P. Sieg, J. Am. Chem. Soc. 83 (1961) 1156-1160. [21] S. Ernst, J. Weitkamp, in: Proc. Intern. Symp. Zeolite Catalysis, Siófok, Hungary, May 13 to 16, 1985, p. 457-466. [22] J. Weitkamp, P.A. Jacobs, J.A. Martens, Appl. Catal. 8 (1983) 123-141.

86 DGMK-Tagungsbericht 2010-3