Feed molecular properties explain cracking results

Bruce A. Lerner, Frank L. Himpsl
Engelhard Corp.
Iselin, N.J.

Although feeds may have similar global properties, catalytic cracking conversion results may differ due to different feed molecular properties.

In a recent study, Engelhard determined that only by understanding the molecular characteristics of a resid was it possible to select the proper catalyst to maximize conversion to favorable products and minimize coke and gas. Engelhard examined two feeds with similar global characteristics but quite distinct molecular distributions.

In recent years, the refining industry has moved to processing residual feeds in the catalytic cracker. In fact, the majority of recent cracking expansions has been designed to process resid exclusively. In response to the demand for catalyst for these crackers, manufacturers are designing and developing specific catalysts that have good hydrothermal stability, high metal tolerance, and good bottoms upgrading.¹

Resid feed properties

Unfortunately, no single resid catalyst is proper for all feeds because resids, like gas oils, differ in nature. Feed-property parameters are necessary for proper selection of a cracking catalyst.

For gas oils, simple, established methods obtain properties such as API gravity, aniline point, Conradson or Ramsbottom carbon, refractive index, and K factor. Also, modified thin layer or column chromatography can provide global information about the quantity of paraffins and aromatics in a given gas oil feed. Attempts have been made to use these properties to correlate feed with performance.^2-5

Unlike gas oils, the heavy nature of resids prohibits accurate determination of some of these standard properties. Furthermore, many resid processors use a portion of resid from either distillation or hydrotreating operations and blend it with gas oil. These scenarios make performance correlation difficult with resid feeds.

Typically, resids are broadly characterized as highly paraffinic or highly aromatic, but these gas-oil properties can also be used to describe resids. In addition, the refractiveness of resid can be described by the percent boiling over 1,000° F. and by the concarbon level. The quantity of metals and other heteroatoms, such as sulfur and nitrogen, in the feed can characterize individual resids and can strongly influence catalyst performance.

Feeds A and B characteristics

• Global. The study used two resid feeds, Feeds A and B, with roughly the same distillation, concarbon, and paraffin/aromatic distributions as determined by modified column elution chromatography.
  Table 1 [6,942 bytes] shows the general properties of the two feeds used in this study. Inspection shows similarities in all common parameters, even exact duplication of the quantity of aromatics and paraffins in each feed.
• Molecular. The feeds were further characterized by a new High Performance Liquid Chromatography (HPLC) technique to determine the molecular distribution of each feed in more detail.

The HPLC consists of the liquid chromatograph and an interfaced Hewlett Packard Chemstation for data collection and analysis. The HPLC has a quaternary HPLC pump and an on-line solvent degasser which provide excellent peak-area reproducibility.

The instrument is also equipped with a liquid autosampler/injector and two detectors: a diode-array detector for aromatics and an evaporative light scattering detector for total mass.

The dual-column system is encased in an oven equipped with a thermostat for consistent and accurate temperature maintenance. HPLC feed samples are dissolved in an appropriate solvent prior to injection. The technique is applicable to oil fractions with an initial boiling point of 650° F. or higher.

Table 2 [5,450 bytes] shows the molecular characteristics of each feed determined from the HPLC method. Saturates contain all the paraffins and naphthenes. Aromatics are broken down into 1 through 4-ring aromatics and polars, which include 5+ ring aromatic condensates.

Fig. 1 [42,242 bytes] shows the distribution of aromatics by the number of aromatic rings, e.g., 25% of all the aromatics in Feed B are single rings. While the total quantities of the aromatic class compounds are almost identical (67-68%), the distributions of the aromatics types are different.

Special to this HPLC technique is not only that it allows for the quantification of the total aromatics broken down into number of aromatic rings, (e.g., 1-ring aromatics, 2-ring aromatics) but it quantifies the aromatic core content as well. For example, both molecules in Fig. 2 [48,795 bytes] are 2-ring fractions. Molecule 1, however, has two mass components: the two aromatic rings which constitute the aromatic core and the additional mass of the sidechain, a connecting alkyl group.

The detectors in the HPLC technique can separate these mass contributions to determine sidechain components. Fig. 3 [53,333 bytes] shows the percentage aromatic core in the total aromatic mass. The corresponding mass of aliphatic sidechains associated with each aromatic fraction, calculated by the difference between Fig. 1 and Fig. 3, is shown in Fig. 4 [47,337 bytes].

Fig. 1 shows that Feed B has more 1-ring aromatics than any other aromatic fraction. The level of 1-ring aromatics for Feed B is nearly double that of Feed A.

Similarly, Feed B has fewer 3 and 4-ring aromatics and significantly fewer polars (which include more highly conjugated species) than Feed A. Fig. 3 shows that, except for the 1-ring aromatic, the aromatic core content of each aromatic fraction is significantly more for Feed A.

Therefore, Feed B contains more alkyl sidechains. These sidechains are nothing more than connected paraffins and olefins lumped into the aromatic category due to the parent ring on which they are attached.

Catalyst characteristics

A commercial Engelhard FCC resid catalyst was chosen based on the global properties of the feeds. Table 3 [5,705 bytes] gives the typical chemical and physical properties of the catalyst used in the study.

The catalyst contains an active matrix for bottoms cracking and special metals-trapping functions to inhibit coke and gas produced by contaminants nickel and vanadium. The catalyst uses a special USY (unstabilized Y zeolite) formed from the proprietary Pyrochem process which displays outstanding hydrothermal stability even in the presence of high levels of vanadium and under full burn regenerator operation.

Cracking results

Each of the feeds was cracked in a fixed fluid-bed reactor designed to process resids. ⁶ For each feed, a set of runs was made at 1,035° F. over a series of catalyst/oil ratios, and the selectivities regressed to constant conversion.

Table 4 [7,668 bytes] gives the catalytic results for each feed cracked over the commercial catalyst under the same reaction conditions. Feed B is significantly easier to crack. The activity of the catalyst at constant catalyst-to-oil ratio is more than 1.⁷ times greater in Feed B. Likewise, coke production is 36% lower and dry gas production is 30% lower than that of Feed A. Feed B generates 5 wt % less LPG.

In return for these low values, gasoline is 10 wt % higher, and LCO is 1.9 wt % higher for Feed B. These results are surprising considering the feeds had similar global properties. HPLC technique revealed, however, that the feeds' molecular characteristics were different.

The catalytic results may be rationalized from the molecular data presented in Figs. 1 and 3.

Fig. 1 shows that the percentage of aromatics which mechanistically cannot easily crack^7-9 are equal for the two feeds. The distribution of aromatics among the degrees of condensation, however, is different. As the number of rings in a polynuclear aromatic molecule increases, the rate of cracking decreases.¹⁰

Alkyl substituents (sidechains), however, can easily crack quite readily, yielding lower molecular weight fragments.

Feed B has most of its aromatics in the smaller ring groupings (1 and 2-ring aromatics), and these rings contain a large proportion of sidechains. Feed B also contains a significantly higher amount of sidechains across all of the aromatic fractions. The sidechains are easily catalytically cleaved off of the parent molecule9 leaving light and non-aromatic centers which become part of the gasoline and distillate range material.

Acknowledgment

The authors would like to acknowledge the efforts of Lianne Boot, AC Analytical Controls Inc., Rotterdam, for her assistance in analyzing the feed samples by HPLC.

References

1. Woltermann, G., Dodwell, G., and Lerner, B., "Modern Cracking Catalyst and Residue Processing Challenges," Paper No. AM-96-46, NPRA Annual Meeting, San Antonio, Mar. 17-19, 1996.
2. Gary, J.H., and Handwerk, G.E., Petroleum Refining Technology and Economics, Marcell Dekker, 1984.
3. Nelson, W.L., "Yields in cracking paraffinic feeds," OGJ, Sept. 3, 1979, p. 107.
4. White, Paul J., "Effect of feed composition on catalytic-cracking yields," OGJ, May 20, 1968, p. 112.
5. Castiglioni, B.P., "How to Predict FCC Yields," Hydrocarbon Processing, Vol. 62, No. 2, 1983, p. 35.
6. Himpsl, F.L., and McClung, R., in preparation.
7. Brouwer, D.M., Chemistry and Chemical Engineering of Catalytic Processes, Prins, R. and Schuitt, G.C.A. Eds., Sitjhoff and Noordhoff, 1980.
8. Lerner, B.A., Carvill, B.T., and Sachtler, W.M.H., "Enhanced Benzene Formation on Pt/H-Mordenite and Pb/H-Mordenite," Catalysis Letters 18, 227, 1993.
9. Greensfelder, B.S., The Chemistry of Petroleum Hydrocarbons, vol. 2, Reinhold Publishing, 1955.
10. Letzsch, W., and Ashton, A., Fluid Catalytic Cracking: Science and Technology, Magee, J.S., and Mitchell, M.M., Eds., Stud. Surf. Sci. Catal. V. 76, Elsevier, Netherlands, 1993.

The Authors