CA2200946A1 - Removal of large molecules from a fluid - Google Patents

Abstract

A process for the removal of large molecules from waste or process streams using an M41S material as an adsorbent. The process is useful for the removal of trace amounts of polynuclear aromatics from complex hydrocarbon mixtures in the vapor phase, such as reformate.

Description

~00946 ~ WO96/10537 PCT~S95/11387 REMOVAL OF LARGE MOLECULES FROM A FLUID

This invention relates to a process for the removal of large molecules, such as polynuclear aromatics, from a fluid using a~ M41S adsorbent.
Porous i~organic solids have great utility as catalysts and separation media for industrial applications.
Catalytic and sorptive activity are enhanced by the extensive surface area provided by a readily accessible microstructure characteristic of these solids.
The porous materials in use today can be sorted into three broad categories using the details of their microstructure as a basis for classification. These categories are l) amorphous and paracrystalline supports, 2) crystalline molecular sieves and 3) modified layered materials.
Variatio~s in the microstructures of these materials manifest them$elves as important differences in the catalytic and sorptive behavior of the materials, as well as differences in various observable properties used to characterize them. For example, surface area, pore size and variability in pore sizes, the presence or absence of X-ray diffraction patterns, as well as the details in such patterns, and the appearance of the materials when their microstructure is studied by transmission electron microscopy and~ electron diffraction methods can be used to characterize porous inorganic solids.
Amorphous and paracrystalline materials represent an important class of porous inorganic solids which have been used for many years in industrial applications. Typical examples of these materials are the amorphous silicas commonly used in catalyst formulations and the paracrystalli~e transitional aluminas used as solid acid catalysts and petroleum reforming catalyst supports.
The amorphous materials are generally characterized as "amorphous" since they are substances having no long range order. Unfortunately, this can be somewhat misleading ~00946 ~
WO96/10537 PCT~S95/11387 since almost all materials are ordered to some degree, at least on the local scale. An alternate term which has been used to describe these materials is "X-ray indifferent".
The microstructure of the silicas consists of 100-2S0 A
(10-25 nm) particles of dense amorphous silica (Kirk-Othmer EncYclo~edia of Chemical Technology, 3rd Edition, Vol. 20, John Wiley & $ons, New York, p. 766-781, 1982), with the porosity resulting from voids between the particles. Since there is no long range order in these materials, the pore sizes tend to be distributed over a rather large range.
This lack of order also manifests itself in the X-ray diffraction pattern, which is usually featureless.
Paracrystalline materials such as the transitional aluminas also have a wide distribution of pore sizes, but exhibit better defined X-ray diffraction patterns usually consisting of a few broad peaks. The microstructure of these materials consists of tiny crystalline regions of condensed alumina phases and the porosity of the materials results from irregular voids between these regions (K.
Wefers and Ch~n~kya Misra, "Oxides and Hydroxides of Aluminum", Technical Paper No. 19 Revised, Alcoa Research Laboratories, p. 54-59, 1987).
Despite any differences arising between these paracrystalli~e or amorphous materials, neither substance has long range order controlling the sizes of pores in the material. Consequently, variability in pore size is typically quite high. The sizes of pores in these materials fall into what is known in the art as the "mesoporous range", which, for the purposes of this Application, is from about 13 to 200 A (1. 3-20 nm).
In sharp contrast to these structurally ill-defined solids are materials whose pore size distribution is narrow because it is controlled by the precisely repeating crystalline nature of the materials' microstructure. These materials are referred to as "molecular sieves", the most important examples of which are zeolites.

~=

~00946 WO96/10537 PCT~S95/11387 Zeolites, both natural and synthetic, have been demonstrated in the past to have catalytic properties for various types of hydrocarbon conversion. Certain zeolitic materials are ordered, porous crystalline aluminosilicates having a defi~ite crystalline structure as determined by X-ray diffraction. These crystalline structures contain a large number of small cavities which may be interconnected by a number of still smaller channels or pores. These cavities and pores are uniform in size within a specific zeolitic material. Since the dimensions of these pores provide access to molecules of certain dimensions while rejecting those of larger dimensions, these materials are known as "molecular sieves". These molecular sieves have been utilized in a variety of ways in order to take advantage of their properties.
Molecular sieves, both natural and synthetic, include a wide variety of positive ion-containing crystalline silicates. These silicates can be described as a rigid three-dimensional framework of Sio4 and Group IIIB element oxides, e.g. Al04, in which the tetrahedra are cross-linked by the sharing of oxygen atoms. The ratio of the total Group IIIB element, e.g. aluminum, and Group IVB element, e.g. silicon, atoms to oxygen atoms is 1:2. The electrovalence of the tetrahedra containing the Group IIIB
element, e.g. aluminum, is balanced by the inclusion of a cation in the crystal. Examples of such cations include alkali metal Qr alkaline earth metal cations. This can be expressed wherein the ratio of the Group IIIB element, e.g.
aluminum, to the number of various cations, such as Ca/2, Sr/2, Na, K or Li, is equal to unity. One type of cation may be exchan~ed either entirely or partially with another type of cation utilizing ion exchange techniques in a conventional manner. By means of such cation exchange, it has been possible to vary the properties of a given silicate by suitable selection of the cation. The spaces between the tetrahedra are occupied by molecules of water prior to dehydration.

~i~00946 ~
WO96/10537 PCT~S95/11387 Prior art techniques have resulted in the formation of a great variety of synthetic zeolites. Many of these zeolites have come to be designated by letter or other convenient symbols, as illustrated by zeolite A (U.S.
Patent 2,882,243); zeolite X (U.S. Patent 2,882,244);
zeolite Y (U.S.. Patent 3,130,007); zeolite ZK-5 (U.S.
Patent 3,247,195); zeolite ZK-4 (U.S. Patent 3,314,752);
zeolite ZSM-5 (U.S. Patent 3,702,886); zeolite ZSM-ll (U.S.
Patent 3,709,979); zeolite ZSM-12 (U.S. Patent 3,832,449);
zeolite ZSM-2Q (U.S. Patent 3,972,983); ZSM-35 (U.S. Patent 4,016,245); and zeolite ZSM-23 (U.S. Patent 4,076,842), merely to name a few.
The SiO2/Al203 ratio of a given zeolite is often variable. For example, zeolite X can be synthesized with SiO2/Al203 ratios of from 2 to 3; zeolite Y, with ratios from 3 to about 6. In some zeolites, the upper limit of the SiOJAl203 ratio is unbounded. ZSM-5 is one such example wherein the SiO2/Al203 ratio is at least 5 and up, as measured within the limits of present analytical measurement techniques. U.S. Patent 3,941,871 (Re. 29,948) discloses a porous crystalline silicate made from a reaction mixture containing no deliberately added aluminum in the recipe and exhibiting the X-ray diffraction pattern characteristic of ZSM-5. U.S. Patents 4,061,724; 4,073,865 and 4,104,294 describe crystalline silicates of varying alumina and metal content.
Additiona~lly, aluminum phosphates are taught in the U.S. Patents 4,310,440 and 4,385,994, for example. These aluminum phosphate materials have essentially electroneutral lattices. U.S. Patent 3,801,704 teaches an aluminum phosphate treated in a certain way to impart acidity.
An early reference to a hydrated aluminum phosphate which is crystalline until heated at 110C, at which point it becomes amorphous or transforms, is the "H1" phase or hydrate of aluminum phosphate of F.d'Yvoire, Memoir Presented to the Chemical Society, No. 392, "Study of ~!00946 WO96/10537 PCT~S95/11387 Aluminum Phosphate and Trivalent Iron", July 6, 1961 (received), pp. 1762-1776. This material, when crystalline, is identified by the JCPDS International Center for Diffraction Data card number 15-274. Once heated at 110C, however, the d'Yvoire material becomes amorphous or transforms to the aluminophosphate form of tridymite.
Compositions comprising crystals having a framework topology after heating at 110C or higher and exhibiting an X-ray diffraction pattern consistent with a material having pore windows formed by 18 tetrahedral members of about 12-13 A ( 1. 2-1.3 nm) in diameter are taught in U.S. Patent 4,880,611.
A naturally occurring, highly hydrated basic ferric oxyphosphate mineral, cacoxenite, is reported by Moore and Shen, Nature, Vol. 306, No. 5941, pp. 356-358 (1983) to have a framework structure containing very large channels with a calculated free pore diameter of 14.2 A. R. Szostak et al., Zeolites: Facts Fiqures, Future, Elsevier Science Publishers B.y.~ 1989, present work showing cacoxenite as being very hydrophilic, i.e. adsorbing non-polar hydrocarbons, only with great difficulty. Their work also shows that thermal treatment of cacoxenite causes an overall decline in X-ray peak intensity.
Silicoaluminophosphates of various structures are taught in U.S. Patent 4,440,871. Aluminosilicates containing phosphorus, i.e., silicoaluminophosphates of particular structures are taught in U.S. Patents 3,355,246 (i.e. ZK-21) and 3,791,964 (i.e., ZK-22). Other teachings of silicoaluminophosphates and their synthesis include U.S.
Patents 4,673,559 (two-phase synthesis method); 4,623,527 (MCM-10); 4,639,358 (MCM-l); 4,647,442 (MCM-2); 4,664,897 (MCM-4); 4,638,357 (MCM-5); and 4,632,811 (MCM-3).
A method for synthesizing crystalline metallo-aluminophosphates is shown in U.S. Patent 4,713,227. An antimonophosphoaluminate and the method for its synthesis are taught in U.S. Patent 4,619,818. U.S. Patent 4,567,029 ~0 0 g 4 6 ~
W096/10537 PCT~S95/11387 teaches metalloaluminophosphates, and titaniumalumino-phosphate and the method for its synthesis are taught in U.S. Patent 4,S00,651.
The phosphorus-substituted zeolites of Canadian Patents 911,416; 911,417; and 911,418 are referred to as "aluminosilicQphosphate" zeolites. Some of the phosphorus therein appears to be occluded, not structural.
U.S. Patent 4,363,748 describes a combination of silica and al~minum-calcium-cerium phosphate as a low acid activity catalyst for oxidative dehydrogenation. Great Britain Patent 2,068,253 discloses a combination of silica and aluminum-calcium-tungsten phosphate as a low acid activity catalyst for oxidative dehydrogenation. U.S.
Patent 4,228,Q36 teaches an alumina-aluminum phosphate-silica matrix as an amorphous body to be mixed with zeolite for use as cracking catalyst. U.S. Patent 3,213,035 teaches improving hardness of aluminosilicate catalysts by treatment with phosphoric acid. The catalysts are amorphous.
Other patents teaching aluminum phosphates include U.S. Patents 4~,365,095; 4,361,705; 4,222,896; 4,210,560;
4,179,358; 4,158,621; 4,071,471; 4,014,945; 3,904,550; and 3,697,550.
The precise crystalline microstructure of most zeolites manifests itself in a well-defined X-ray diffraction pattern that usually contains many sharp ~; m~
and that serves to uniquely define the material.
Similarly, the dimensions of pores in these materials are very regular, due to the precise repetition of the crystalline microstructure. All molecular sieves discovered to date have pore sizes in the microporous range, which is usually quoted as 2 to 20 A (.2-2 nm), with the largest reported being about 12 A ( 1. 2 nm).
Certain layered materials, which contain layers capable of being spaced apart with a swelling agent, may be pillared to provide materials having a large degree of porosity. Examples of such layered materials include Og46 WO96/10537 PCT~S95/11387 clays. Such clays may be swollen with water, whereby the layers of the clay are spaced apart by water molecules.
Other layered materials are not swellable with water, but may be swolle~ with certain organic swelling agents such as amines and quaternary ammonium compounds. Examples of such non-water swellable layered materials are described in U.S.
Patent 4,859,648 and include layered silicates, magadiite, kenyaite, trititanates and perovskites. Another example of a non-water s~ellable layered material, which can be swollen with certain organic swelling agents, is a vacancy-containing titanometallate material, as described in U.S.
Patent 4,831,Q06.
Once a layered material is swollen, the material may be pillared by interposing a thermally stable substance, such as silica, between the spaced apart layers. The aforementioned U.S. Patents 4,831,006 and 4,859,648 describe methods for pillaring the non-water swellable layered materials described therein and are incorporated herein by reference for definition of pillaring and pillared materials.
Other patents teaching pillaring of layered materials and the pillared products include U.S. Patents 4,216,188;

4,248,739; 4,176,090; and 4,367,163; and European Patent Application 2Q5,711.
The X-ray diffraction patterns of pillared layered materials can vary considerably, depending on the degree that swelling and pillaring disrupt the otherwise usually well-ordered layered microstructure. The regularity of the microstructure in some pillared layered materials is so badly disrupted that only one peak in the low angle region on the X-ray diffraction pattern is observed, at a d-spacing corresponding to the interlayer repeat in the pillared material. Less disrupted materials may show several peaks in this region that are generally orders of this fundamental repeat. X-ray reflections from the crystalline structure of the layers are also sometimes observed. The pore size distribution in these pillared wos6/1~ ~ 9 4 6 J~ PCT~S95/11387 layered materials is narrower than those in amorphous and paracrystalli~e materials but broader than that in crystalline framework materials.
Indeed, X-raY diffraction patterns have come to play an important role in identification of various crystalline materials, especially pillared layered materials.
Nevertheless, it is the physical properties of these materials which render them valuable assets to the scientific and, industrial communities. These materials are not only valua~ble when employed in the petroleum industry, but they have also been found to exhibit properties useful for a variety of applications including such fields as nonlinear optics and the biological and chemical sciences.
One particular area of interest involves employing these porous crystalline materials in the fields of chemistry and biology in order to effect the separation of substances contained within a mixture.
As the en,vironmental requirements become stricter, removal of polynuclear aromatics from products and waste streams is de$irable and will be required. Polynuclear aromatics (PN~'s) are large petroleum molecules, many of which are carcinogens, and are generated as undesirable by-products in reactions such as reforming and regeneration of catalysts, e.~., MLDW catalysts. As they are formed, they are entrained in low concentrations in the vapor leaving the reactor or in the gaseous combustion products leaving the regeneration process. One simplest method for the separation of polynuclear aromatics from raw reformate is by distillation, but capital expense is high.
Therefore, it is an object of the present invention to separate large molecules from small molecules by selective adsorption at elevated temperatures using M41S material.
It is a further object of the present invention to remove traces of PNA's from the hydrocarbon vapor phase by selective adso,rption at elevated temperature using M41S
material.

WO96/10537 ~ ~ 0 0 9 4 6 ~ PCT~S95/11387 The process of the present invention is useful for separating large molecules from small molecules in a fluid, i.e. liquid or gas phase, at elevated temperatures. The process of the present invention is particularly useful for removing large PNA's by adsorption from the hot vapor effluent of a reformer and from the combustion gases of catalyst rege~eration processes. The sorbent used in the process of the present invention comprises M4lS, and more particularly ~igh-silica MCM-41. The M41S sorbent has a pore diameter sufficiently large to admit PNA's into the pores, but sufficiently small to adsorb PNA's strongly.
The large pore volume provides a high capacity for the sorbate. The adsorbent has a high thermal and hydrothermal stability at the sorption conditions and vapor phase composition from which the PNA's are to be removed. The M4lS sorbent is also regenerable.
The invention therefore includes a process for the removal of sorbate molecules from a fluid comprising adsorbing said sorbate molecules on an adsorbent composition comprising an inorganic, porous, crystalline phase material and which exhibits, after calcination, an X-ray diffraction pattern with at least one peak at a d-spacing greater than 18 A Units (1.8 nm) with a relative intensity of 100, and a benzene sorption capacity of greater than 15 grams of benzene per 100 grams of the material at 50 torr (6.7 kPa) wherein the ratio of crystalline phase material pore diameter to kinetic pore diameter of said sorbate molecules is in the range of 1:1 to 4:1.
The inve~tion further includes a process for the removal of polynuclear aromatic molecules from a hydrocarbon vapor phase comprising adsorbing said polynuclear aromatic molecules on an adsorbent composition comprising an inorganic, porous, crystalline phase material which exhibits, after calcination, an X-ray diffraction pattern with at least one peak at a d-spacing greater than 18 A Units (1.8 nm) with a relative intensity of 100, and a ~!2~ 0 ~ 4 6 , WO96/10537 PCT~S95/11387 benzene sorption capacity of greater than 15 grams of benzene per lQ0 grams of the material at 50 torr (6.7 kPa) and 25C, wherein the crystalline phase material pore diameter is no~t larger than three times the polynuclear aromatic molecules kinetic pore diameter.
Large molecules, such as PNA's, can be separated from smaller molecules by adsorption at elevated temperatures using an M4lS sorbent. A large molecule that straddles the pore being attached, by sorption, to both opposite walls is more tightly sorbed than a small molecule that is adsorbed on one wall only. Since the pores of M41S are very large, the M41S material is useful for separation of large petroleum molecules, such as PNA's, as well as biological molecules, such as hormones, insulin Bovine Pancreatic Trypsin Inhibitor (BPTI) and metal complexes of proteins.
M41S material can be prepared with uniform pores over a wide range of pore sizes. In the process of the present invention, the pore size of the M4lS material is selected based on the kinetic diameter of the molecules to be sorbed. The ratio of the M4lS pore diameter to kinetic pore diameter of the molecules to be sorbed is in the range of from 1:1 to 4:1 and preferably in the range of from 2:1 to 3:1. For the separation of polynuclear aromatics, M41S
sorbent having uniform pores of a size not larger than three times the kinetic diameter of the PNA molecules is important.
Traces of polynuclear aromatics (PNA's) are removed from the hydrocarbon vapor phase by adsorbing the PNA's on an adsorbent having a pore size adequate to adsorb the PNA's at a temperature at which lighter hydrocarbons and other gases are not adsorbed by such adsorbent.
Light hydrocarbons, such as benzene, cyclohexane and n-hexane are not sorbed on MCM-41 at 100C at 40 Torr (5.3 kPa) vapor pressure while they are sorbed at 25C. Larger molecules are believed to be sorbed at higher temperatures at comparable vapor pressure. The process is therefore able to separate large molecules from smaller molecules at Z;~09946 .
~ WO96/10537 PCT~S95/11387 elevated temperatures. M4lS material may be used as a sorbent for large molecules, such as PNA's, at conditions where even a large excess of light hydrocarbons cannot compete successfully.
The polynuclear aromatics which are removed by the process of the present invention include three, four, five or even a higher number of aromatic rings. Naphthalenes may also be removed. The process of the present invention is useful for the removal of polynuclear aromatic compounds from refined petroleum fractions including catalytic reformate.
There is a great variety of PNA's present in reformate. PNA's are distinguished not only by the arrangement and number of condensed rings, but also by the number and type of alkyl side chains, as well as the positions of these side chains on the rings. Not all PNA's are yellow, but the yellow ones all appear to absorb at 432nm.
Typically, the heavy portion of the reformate comprises the following PNA's:
Fluorenes 0.049%
Phenanthrenes 0.049%
Fluoranthenes/Pyrenes 0.015%
Chrysenes 0.0019%
Total PNA's 0.115%
The percentages are based on the total reformate.
The kinetic diameter, also referred to as critical diameter and ~;n;mum cross-sectional diameter, determined through computer modeling, of several PNA's is shown below 30 in Angstrom:
Chrysene 7.5A (.75 nm) Pyrene 8.gA (.89 nm) Perylene 9. oA ( . g nm) Benzpyrene 9. lA ( . g1 nm) These molecules are too small to straddle the pores of the M4lS materials. Derivatives of the above named and other PNA's may be somewhat larger.

~00~46 WO96/10~37 PCT~S95/11387 The usual feedstock for catalytic reforming is a petroleum fraction known as naphtha and having an initial boiling point of 180F (82C) and an end boiling point of 400F (204C). Catalytic reforming is a vapor phase reaction effected at temperatures ranging from 500 to 1050F (260-~66C) and at pressures ranging from 50 to 1000 psig (446-6996 kPa), preferably from 85 to 350 psig (687-2515 kPa). The reaction is carried out in the presence of sufficient hydrogen to provide a hydrogen to hydrocarbon mole ratio of from 0.5:1 to 10:1. Further information on catalytic reforming processes may be found in, for example, U.S. Patent Nos. 4,119,526 (Peters et al.); 4,409,095 (Peters); and 4,440,626 (Winter et al.).
The temperature for adsorption is dependent on the sorption properties of the other components present in a fluid phase and on the pore diameter of the M41S sorbent.
Generally the temperature is in the range of from 100 to 500C (212 to 932F) or higher. Temperatures in the range of from 100 to 400C (302 to 702F) are important, particularily in the presence of water. Temperatures in the range of from 200C to 320C (392 to 608F) are employed for the removal of PNA's from reformate.
The sorption material used in the process of the present invention includes a novel synthetic composition of matter comprising an ultra-large pore size crystalline phase. This material may be an inorganic, porous, non-layered, crystalline phase material which can be characterized (in its calcined form) by an X-ray diffraction pattern with at least one peak at a d-spacing greater than 18 Angstrom with a relative intensity of 100 and a benzene sorption capacity of greater than 15 grams of benzene per 1~0 grams of the material at 50 torr (6.7 kPa) and 25C. This material and its preparation and properties are described in further detail in U.S. Patent 5,102,643.
An important form of the crystalline material is an inorganic, porous material having a hexagonal arrangement of uniformly sized pores with a maximum perpendicular z~on946 WO96/10537 PCT~S95/11387 cross-section pore diameter of at least 13 A Units (l.3 nm), and typically within the range of from 13 A Units (1.3 nm) to 200 A ~nits (20 nm), identified as MCM-41. This material exhibits a hexagonal electron diffraction pattern that can be i~dexed with a d1oo value greater than 18 Angstrom whic~ corresponds to at least one peak in the X-ray diffraction pattern. This material and its preparation and properties are described in further detail in U.S. Paten~ 5,098,684.
The inor~anic mesoporous crystalline material of this invention may have the following composition:
M~/q (Wa Xb Yc Zd h) wherein W is a divalent element, such as magnesium, and/or a divalent first row transition metal, e.g., manganese, cobalt and iron, preferably cobalt; X is a trivalent element, such as aluminum, boron, iron and/or gallium, preferably aluminum; Y is a tetravalent element such as silicon and/or germanium, particularly silicon; Z is a pentavalent element, such as phosphorus; M is one or more ions, such as, for example, ammonium, Group IA, IIA and VIIB ions, usually hydrogen, sodium and/or fluoride ions; n is the charge of the composition excluding M expressed as oxides; q is the weighted molar average valence of M; n/q is the number of moles or mole fraction of M; a, b, c, and d are mole fractions of W, X, Y and Z, respectively; h is a number of from l to 2.5; and (a+b+c+d) = l.
An important embodiment of the above crystalline material is when (a+b+c) is greater than d, and h = 2. A
further embodiment is when a and d = 0, and h = 2.
MCM-41 having a high silica/alumina ratio greater than 3000/l is particularly important.
In the as-synthesized form, the material may have a composition, on an anhydrous basis, expressed empirically as follows:

~i~O O ~ 4 6 WO96/10537 -14- PCT~S95/11387 rRMn/9 (Wa Xb YC Zd h) where R is the total organic material not included in M as an ion, and r is the coefficient for R, i.e., the number of moles or mole fraction of R.
The M an~ R components are associated with the material as a result of their presence during crystallizatiQn, and are easily removed or, in the case of M, replaced by post-crystallization methods described below.
To the e~tent desired, the original M, e.g., sodium, ions of the a~-synthesized support material can be replaced in accordance with conventional ion-exchange techniques.
Preferred replacing ions include metal ions, hydrogen ions, hydrogen precursor, e.g., ammonium, ions and mixtures of these ions. Replacing ions include hydrogen, rare earth metals and metals of Groups VIIA (e.g., Mn), VIIIA (e.g., Ni),IB (e.g., Cu), IVB (e.g., Sn) of the Periodic Table of the Elements and mixtures of these ions.
The crystalline (i.e., having sufficient order to provide a diffraction pattern such as, for example, by X-ray, electron or neutron diffraction, following calcination, ~ith at least one peak) mesoporous material may be characterized by its structure, which includes extremely large pore windows as well as by its high sorption capacity. The term "mesoporous" is used here to indicate crystals having uniform pores within the range of from 13 Angstrom (1.3 nm) to 200 Angstrom (20 nm). The mesoporous materials have uniform pores within the range of from 13 Angstrom (1.3 nm) to 200 Angstrom (20 nm), more usually from 15 Angstrom (1.5 nm) to 100 Angstrom (10 nm).
Since these pores are significantly larger than those of other crystalline materials, it is appropriate to refer to them as ultra-large pore size materials. For the purposes of this application, a working definition of "porous" is a material that adsorbs at least 1 gram of a small molecule, such as Ar, N2, n-hexane or cyclohexane, per 100 grams of ~OQ9~6 ~
WO96/10537 PCT~S95/11387 the solid, at appropriate temperature and atmospheric pressure.
The synthesis of the material is described in U.S.
Patent Nos. 5,108,725 and 5,0S7,296.
, The material can be distinguished from other porous inorganic solids by the regularity of its large open pores, whose pore siz,e more nearly resembles that of amorphous or paracrystalli~e materials, but whose regular arrangement and uniformity of size (pore size distribution within a single phase of, for example, + 25%, usually + 15% or less of the average pore size of that phase) resemble more those of crystalline framework materials such as zeolites. The preferred MCM-4l materials have a hexagonal arrangement of large open cha,nnels that can be synthesized with open internal diame,ters from 13 Angstrom (l.3 nm) to 200 Angstrom (20 nm). The term "hexagonal" is intended to encompass not only materials that exhibit mathematically perfect hexagonal symmetry within the limits of experimental measurement, but also those with significant observable deviations from that ideal state. A working definition as applied to the microstructure of MCM-41 would be that most channels in the material would be surrounded by six nearest neighbor channels at roughly the same distance. Defects and imperfections will cause significant numbers of channels to violate this criterion to varying degrees, depending on the quality of the material's preparation. Samples which exhibit as much as + 25% random deviation from the average repeat distance between adjacent channels still clearly give recognizable images of the present ultra-large pore materials. Comparable variations are also observed in the d1oo values from the electron - diffraction patterns.
The size of the pores in the present mesoporous materials is large enough that the spatiospecific selectivity with respect to transition state species in reactions such, as cracking is minimized (Chen et al., "Shape Selective Catalysis in Industrial Applications", ~ Q 9 46 z~
W096/l0537 PCT~S95/11387 Chemical Industries, 36, 41-61 (1989) to which reference is made for a di$cussion of the factors affecting shape selectivity). Diffusional limitations are also minimized as a result of the very large pores.
The most regular preparations of the present support material give an X-ray diffraction pattern with a few distinct maxima in the extreme low angle region. The positions of these peaks approximately fit the positions of the hkO reflections from a hexagonal lattice. The X-ray diffraction pattern, however, is not always a sufficient indicator of the presence of these materials, as the degree of regularity in the microstructure and the extent of repetition of the structure within individual particles affect the number of peaks that will be observed. Indeed, preparations with only one distinct peak in the low angle region of the X-ray diffraction pattern have been found to contain substantial amounts of MCM-41 in them. Other techniques to illustrate the microstructure of this material are transmission electron microscopy and electron diffraction. Properly oriented specimens of the MCM-41 material show a hexagonal arrangement of large channels and the corresponding electron diffraction pattern gives an approximately hexagonal arrangement of diffraction ~; r~ .
The dloo spacing of the electron diffraction patterns is the distance between adjacent spots on the hkO projection of the hexagonal lattice and is related to the repeat distance aO between channels observed in the electron micrographs through the formula dloo = aO~3/2. This dloo spacing observed in the electron diffraction patterns corresponds to the d-spacing of a low angle peak in the X-ray diffraction pattern of the material. The most highly ordered preparations of the MCM-41 material obtained so far have 20-40 distinct spots observable in the electron diffraction patterns. These patterns can be indexed with the hexagonal hkO subset of unique reflections of 100, 110, 200, 210, etc., and their symmetry-related reflections.

Z~00946 WO96/10537 PCT~S95/11387 -In its calcined form, the crystalline material may be further characterized by an X-ray diffraction pattern with at least one peak at a position greater than 18 Angstrom (1.8 nm) d-sp~cing (4.909 2~ for Cu K-alpha radiation) which corresponds to the d1oo value of the electron diffraction pa~ttern of the material, and an equilibrium benzene adsorption capacity of greater than 15 grams benzene/100 grams crystal at 50 torr (6.7 kPa) and 25C
(basis: crystal material having been treated in an attempt to insure no pore blockage by incidental contaminants, if necessary).
The equilibrium benzene adsorption capacity characteristic of this material is measured on the basis of no pore blockage by incidental contaminants. For instance, the sorption test will be conducted on the crystalline material phase having any pore blockage contaminants and water removed by ordinary methods. Water may be removed by dehydration techniques, e.g., thermal treatment. Pore blocking inorganic amorphous materials, e.g., silica, and organics may be removed by contact with acid or base or other chemical agents such that the detrital material will be remov.ed without detrimental effect on the crystal.
More particularly, the calcined crystalline material may be characterized by an X-ray diffraction pattern with at least two peaks at positions greater than 10 Angstrom d-spacing (8.842~ for Cu K-alpha radiation), at least one of which is at a position greater than 18 Angstrom (1.8 nm) d-spacing, and no peaks at positions less than 10 Angstrom (1 nm) d-spacing with relative intensity greater than 20% of the strongest peak. Still more particularly, the X-ray diffraction pattern of the calcined support material will have no peaks at positions less than 10 Angstrom (1 nm) d-spacing with relative intensity greater than 10% of the strongest peak. In any event, at least one peak in the X-ray diffraction pattern will have a d-spacing that corresponds to the dloo value of the electron diffraction pattern of the material.

~!~0 0 9 4 6 WO96110537 PCT~S95/11387 The calcined inorganic crystalline material may also be characteri~ed as having a pore size of 13 Angstrom (l.3 nm) or greater as measured by physisorption measurements, described below. Pore size is defined by the m~;rllm perpendicular pore diameter of the crystal.
X-ray diffraction data were collected on a Scintag PAD
X automated diffraction system employing theta-theta geometry, Cu K-alpha radiation, and an energy dispersive X-ray detector. Use of the energy dispersive X-ray detector eliminated the need for incident or diffracted beam monochromators. Both the incident and diffracted X-ray beams were collimated by double slit incident and diffracted collimation systems. The slit sizes used, starting from the X-ray tube source, were 0.5, l.0, 0.3 and 0.2 mm, respectively. Different slit systems may produce differing intensities for the peaks. The support materials that have the largest pore sizes may require more highly collimated incident X-ray beams in order to resolve the low angle peak from the transmitted incident X-ray beam.
The diffraction data were recorded by step-scanning at 0.04 degrees of 2~, where ~ is the Bragg angle, and a counting time of lO seconds for each step. The interplanar spacings, d's, were calculated in Angstrom, and the relative intensities of the lines, I/Io, where Io is one-hundredth of the intensity of the strongest line, above background, were derived with the use of a profile fitting routine. The intensities were uncorrected for Lorentz and polarization effects. The relative intensities are given in terms of the symbols vs = very strong (75-lO0), s =
strong (50-74), m = medium (25-49) and w = weak (0-24).
The diffraction data listed as single lines may consist of multiple overlapping lines which under certain conditions, such as very high experimental resolution or crystallographic changes, may appear as resolved or partially resolved lines. Typically, crystallographic changes can include minor changes in unit cell parameters and/or a chan~e in crystal symmetry, without a substantial ~ WO96/10537 ~ Z 0 0 9 4 6 PCT~S95/11387 change in stru~cture. These minor effects, including changes in relative intensities, can also occur as a result s of difference$ in cation content, framework composition, nature and degree of pore filling, thermal and/or 5 hydrothermal history, and peak width/shape variations due to particle size/shape effects, structural disorder or other factors known to those skilled in the art of X-ray diffraction.
The equilibrium benzene adsorption capacity is 10 determined by contacting the support material, after dehydration or calcination at, for example, 540C for at least one hour and other treatment, if necessary, in an attempt to remove any pore blocking contaminants, at 25C
and 50 torr (~.7 kPa) benzene until equilibrium is reached.
15 The weight of benzene sorbed is then determined as described below.
The above crystalline material, especially in its metal, hydrogen and ammonium forms, may be readily converted to another form by thermal treatment 20 (calcination). This thermal treatment is generally performed by heating one of these forms at a temperature of at least 400C for at least 1 minute and generally not longer than 2~ hours, preferably from 1 to 10 hours. While subatmospheric pressure can be employed for the thermal 25 treatment, atmospheric pressure is desired for reasons of convenience, $uch as in air, nitrogen, ammonia, etc. The thermal treatment can be performed at a temperature up to 750C. The thermally treated product is particularly useful in the catalysis of certain hydrocarbon conversion 30 reactions.
MCM-41 can be prepared by one of several methods, each with particula~r limitations.
A first method involves a reaction mixture having an X2O3/Y02 mole ratio of from 0 to 0.5, but an Al2O3/SiO2 mole 35 ratio of from 0 to 0.01, a crystallization temperature of from 25C to 250C, particularly from 50C to 175C, and an organic directing agent, hereinafter more particularly -~og46 t WO96/1053 PCT~S95/11387 described, or, particularly a combination of that organic directing age~t plus an additional organic directing agent, described below. This first method comprises preparing a reaction mixt~re containing sources of, for example, alkali or alkaline earth metal (M), e.g., sodium or potassium, cation if desired, one or a combination of oxides selected from the grou~ consisting of divalent element W, e.g., cobalt, trivalent element X, e.g., aluminum, tetravalent element Y, e.~., silicon, and pentavalent element Z, e.g., phosphorus, a~ organic (R) directing agent, described below, and a solvent or solvent mixture, such as, for example, C1-C6 alcohols, C,-C6 diols and/or water, especially water. The reaction mixture has a composition, in terms of mole ratios of oxides, within the following ranges:

Reactants Useful Preferred x2o3/Yo2 o to 0.5 O.OOl to 0.5 Al203/siO2 0 to O.Ol O.OOl to O.Ol 23/ (Y2+Z25)O. l to lO0 O.l to 20 20 X203/(Y02+W0+Z205)O.l to lO0 o.l to 20 Solvent/l to 15005 to lO00 ( YO2+WO+Z205+X203 ) OH-/Y02 0 to lO 0 to 5 (M2je+P~2/fO)/0.01 to 20 û.û5 ~o 5 25 (Y02+WO+Z205+X203) M2/e/ O to lO 0 to 5 YO2+WO+Z205+X203 ) R2~fO/O.Ol to 2.00.03 to l.0 ( YO2+WO+Z205+X203 ) where e and f are the weighted average valences of M and R, respectively.
In this first method, when no Z and/or W oxides are added to the reaction mixture, the pH is important and must be maintained at from 9 to 14. When Z and/or W oxides are present in the reaction mixture, the pH is not narrowly Z~00946 ~
WO96/10537 PCT~S95111387 -2l-important for synthesis of MCM-41. In this, as well as the following methods for synthesis of MCM-41, the R2~fO/(Y02+WO+z~05+x2o3) ratio is important. When this ratio is less than ~.Ol or greater than 2.0, impurity products tend to be synthesized at the expense of the desired crystalline ma~terial.
A second method for synthesis of MCM-41 involves a reaction mixture having an X203/Y02 mole ratio of from 0 to 0.5, a crystallization temperature of from 25C to 250C, particularly from 50C to 175C, and two separate organic directing agents, i.e., the organic and additional organic directing agents, described below. This second method comprises preparing a reaction mixture containing sources of, for example, alkali or alkaline earth metal (M), e.g., sodium or potassium, cation if desired, one or a combination of oxides selected from the group consisting of divalent element W, e.g., cobalt, trivalent element X, e.g., aluminum, tetravalent element Y, e.g., silicon, and pentavalent element Z, e.g., phosphorus, a combination of organic directing agent and additional organic directing agent (R), each described below, and a solvent or solvent mixture, such as, for example, C1-C6 alcohols, C1-C6 diols and/or water, especially water. The reaction mixture has a composition, in terms of mole ratios of oxides, within the following ran~es:

~o ~ 9 4 6 zl W096/10537 PCT~S95/11387 Reactants Useful Preferred X203/YO2 0 to 0.5O.OOl to 0.5 ~' X203/(YO2+Z20s~ O.l to lOOO.l to 20 X203/(YO2+WO+z205) O.l to lOOO.l to 20 Solvent/ l to 15005 to lOOO
( Y2+W+Z20s+X203 ) OH-/YO2 0 to lO O to 5 (M2,eO+R2/fO)/ O.Ol to 200.05 to 5 ( YO2+WO+Z205+X203 ) M2/c/ O to lO O to 5 ( YO2+WO+Z205+X203 ) R2/fO/ O.l to 2.00.12 to l.O
(YO2+WO+Z205+X203 ) where e and f are the weighted average valences of M and R, respectively.
In this second method, when no Z and/or W oxides are added to the reaction mixture, the pH is important and must be maintained at from 9 to 14. When Z and/or W oxides are present in the reaction mixture, the precise value of the pH is not important for crystallization.
A third method for synthesis of MCM-4l is where X
comprises aluminum and Y comprises silicon, the crystallization temperature must be from 25C to 175C, particularly from 50C to 150C, and an organic directing agent, described below, or, preferably a combination of that organic directing agent plus an additional organic agent, described below, is used. This third method comprises preparing a reaction mixture containing sources of, for example, alkali or alkaline earth metal (M), e.g., sodium or potassium, cation if desired, one or more sources of aluminum and/or silicon, an organic (R) directing agent, hereinafter more particularly described, and a solvent or solvent mixture, such as, for example C1-C6 alcohols, C1-C6 diols and/or water, especially water. The reaction mixture has a composition, in terms of mole ratios of oxides, within the following ranges:

WO9G/10537 Z ~ ~ ~ 6 ' PCT~S95/11387 Reactants Useful Preferred Al203/siO2 0 to 0.5 O.OOl to 0.5 Solvent/SiO2 l to 1500 5 to lO00 oH-/SiO2 0 to lO 0 to 5 5 (M2/eO+R2/fO)/ O.Ol to 20 0.05 to 5 ( S io2+Al2o3 ) M2~eo/(sio2+Al2o3) 0 to 5 0 to 3 R2~f0/(siO2+Al2o3)O.Ol to 2 0.03 to l where e and f are the weighted average valences of M and R, respectively.
In this third method, the pH is important and must be maintained at from 9 to 14. This method involves the following steps:

(l) Mix the organic (R) directing agent with the solvent or solvent mixture such that the mole ratio of solvent/R2/~0 is within the range of from 50 to 800, particula,rly from 50 to 500. This mixture constitutes the "primary template" for the synthesis method.

(2) To the primary template mixture of step (l) add the sources of oxides, e.g., silica and/or alumina such that the ratio of R2/fO/(SiO2+Al203) is within the range of from O.Ol to 2Ø

(3) Agitate the mixture resulting from step (2) at a temperatu,re of from 20C to 40C, particularly for from 5 minutes to 3 hours.

(4) Allow the mixture to stand with or without agitatio~, preferably at a temperature of from 20C to lO0C, and preferably for from lO minutes to 24 hours.

(5) Cry$tallize the product from step (4) at a temperature of from 50C to 175C, particularly for ~00946 ~
wos6llos37 PCT~S95/11387 from 1 hour to 72 hours. Crystallization temperatures higher in the given ranges are most preferred.

A fourth method for the synthesis of MCM-41 involves the reaction mixture used for the third method, but the following specific procedure with tetraethylorthosilicate the source of silicon oxide:

(1) Mix the organic (R) directing agent with the solvent or solvent mixture such that the mole ratio of solvent/R2~fO is within the range of from 50 to 800, particularly from 50 to 500. This mixture constitutes the "primary template" for the synthesis method.

(2) Mix the primary template mixture o~ step (1) with tetraethylorthosilicate and a source of aluminum oxide, if desired, such that the R2~fO/SiO2 mole ratio is in the range of from 0.5 to 2Ø

(3) Agitate the mixture resulting from step (2) for from 10 minutes to 6 hours, particularly from 30 minutes to 2 hours, at a temperature of from 0C to 25C, and a pH of less than 12. This step permits hydrolysis/polymerization to take place and the resultant mixture will appear cloudy.

(4) Crystallize the product from step (3) at a temperature of from 25C to 150C, particularly from 95C to 110C, for from 4 to 72 hours, particularly from 16 to 48 hours.

In each of the above methods, batch crystallization of the crystalline material can be carried out under either static or agitated, e.g., stirred, conditions in a suitable reactor vessel, such as for example, polypropylene jars or teflon lined or stainless steel autoclaves.

~00946 WO96/10537 PCT~S95/11387 Crystallization may also be conducted continuously in suitable equipment. The total useful range of temperatures for crystallization is noted above for each method for a time sufficient for crystallization to occur at the temperature used, e.g., from 5 minutes to 14 days. The crystals are then separated from the liquid and recovered.
Following the synthesis,the crystalline material should be subjected to treatment to remove part or all of any organic constituent.
When a source of silicon is used in the synthesis method, it is preferred to use at least in part an organic silicate, such as, for example, a quaternary ammonium silicate. Non-limiting examples of such a silicate include tetramethylammonium silicate and tetraethylorthosilicate.
By adjusting conditions of the synthesis reaction for each method, like temperature, pH and time of reaction, etc., within the above limits, various embodiments of the MCM-41 crystalline material with a desired average pore size may be prepared. In particular, changing the pH, the temperature or the reaction time may promote formation of product crystals with different average pore size.
Non-limiting examples of various combinations of W, X, Y and Z contemplated for the first and second synthesis methods include:

W X Y Z
-- Al Si ---- Al -- P
-- Al Si P
Co Al -- P
Co Al Si P
~ ---- ---- S i including the combinations of W being Mg, or an element selected from the divalent first row transition metals, e.g., Mn, Co and Fe; X being B, Ga or Fe; and Y being Ge.

~2 0 0 9 4 6 ~
WO96/10537 PCT~S95/11387 An organic directing agent for use in each of the above methods for synthesizing MCM-41 from the respective reaction mixtures is an ammonium or phosphonium ion of the formula R1R2R3R4Q+, i.e., Rl1 R4 -IQ ~ R2 where Q is nitrogen or phosphorus and wherein at least one of R1, R2, R3 and R4 is aryl or alkyl of from 6 to 36 carbon atoms, e.g., -C6H13 , -C1oH21 , -C16H33 and -C18H3" or combinations thereof, the remainder of R1, R2, R3 and R4 being selected from hydrogen, alkyl of from l to 5 carbon atoms and combinations of these. The compound from which the above ammonium or phosphonium ion is derived may be, for example, the hydroxide, halide, silicate, or mixtures of these.
In the first and third methods above it is important to have an additional organic directing agent and in the second method it is required to have a combination of the above organic directing agent and an additional organic directing agent. That additional organic directing agent is the ammonium or phosphonium ion of the above directing agent formula wherein R1, R2, R3 and R4 together or separately are selected from the group consisting of hydrogen and alkyl of l to 5 carbon atoms and combinations thereof. Any such combination of organic directing agents go to make up "R" and will be in molar ratio of lO0/l to O.Ol/l, first above listed organic directing agent/additional organic directing agent.
The particular effectiveness of the required directing agent, when compared with other such agents known to direct synthesis of one or more other crystal structures, is believed due to its ability to form micelles which function as a template in the above reaction mixture in the nucleation and growth of the desired ultra-large pore crystals with the limitations discussed above. Non-~ O O 9 ~ 6 WO96/10537 PCT~S95/11387 limiting examples of these directing agents include cetyltrimethylammonium, cetyltrimethylphosphonium, cetylpyridinium, myristyltrimethylammonium, decyltrimethylammonium, dodecyltrimethylammonium and dimethyldidodecylammonium.
The reaction mixture components can be supplied by more than one source. The reaction mixture can be prepared either batchwise or continuously. Crystal size and crystallization time of M41S will vary with the nature of the reaction mixture employed and the crystallization conditions.
The crystals prepared by the synthesis procedure can be shaped into a wide variety of particle sizes. Generally speaking, the particles can be in the form of a powder, a granule, or a molded product, such as an extrudate having particle size sufficient to pass through a 2 mesh (Tyler) screen and be retained on a 400 mesh (Tyler) screen. In cases where the sorbent is molded, such as by extrusion, the crystals can be extruded before drying or partially dried and then extruded. The crystals of the mesoporous material may be composited with a matrix material to form the finished sorbent and for this purpose conventional matrix materials such as alumina, silica-alumina and silica are suitable with preference given to silica as a non-acidic binder. Other binder materials may be used, forexample, titania, zirconia and other metal oxides or clays.
The mesoporous material is usually composited with the matrix in amounts from 80:20 to 20:80 by weight, typically from 80:20 to 50:50 mesoporous material:matrix.
Compositing may be done by conventional means including mulling the materials together followed by extrusion of pelletizing i~to the desired finished catalyst particles.
A preferred method for extrusion with silica as a binder is disclosed in ~.S. 4,582,815.
The following examples illustrate the process of the present invention.

~2009~6 ~t WO96/10537 PCT~S95/11387 Example 1 Sorption is carried out using a commercial silica gel and MCM-41 of varying pore size. The physical properties of the sorbents are shown in Table 1 as follows:
Table 1 Nominal Average Pore Surface Pore Pore Diameter A Area Volume Diameter m2/g cm3/g A
Silica Gel Grade NA 720-760 0.43 23 (2.3 nm) MCM-41, d=43A 40 (4 nm) 940 1.0645 (4.5 nm) MCM-41, d=33~ 30 (3 nm) 967 0.9037 (3.7 nm) The average pore diameter, as calculated from the BET
pore volume and surface area, does not fully represent the pores of M41S, as it is influenced by the large external surface area of the small MCM-41 particles and also includes any non-structured material present. It is believed that the actual pore diameter of MCM-41 is slightly smaller than the d-spacing measured by XRD (x-ray diffraction) because the latter is the repeat spacing and includes the thickness of the walls.
The raw reformate used as a starting material has the 20 following properties:
Reformate Olefins 1 vol%
Aromatics 66 vol%
Saturates 33 vol%

25 D86 Distillation 10% 160F (71C) 50% 255F (124C) 90% 320F (160C) Sulfur = 0 ppm 30 Octane value ~0094~
WO96/10537 PCT~S95/11387 The raw reformate is yellow corresponding to a band at 432 nm (nanometers) in the visible spectrum. The band correlates with the color observed and is used for colorimetric ~etermination of the colored PNA in the reformate. The identity of the particular PNA responsible for the color and its absolute percentage is unknown. The absorbance of the raw reformate at 432 nm is arbitrarily set at 100%, $o that the data obtained after sorption allows the percentages of the initially present PNA removed and retained to be calculated. Only PNA's that sorb at this wavelength are monitored. However, it is believed that all PNA's are removed at the same rate.
The colored PNA's are separated from the desirable hydrocarbons by distillation. The sorption process of the present invention is useful for treating vapor emerging from a reactor. However, for testing of the adsorbents, it is necessary to vaporize the PNA's along with the bulk of the reformate. High temperatures are employed for such a flash vaporization and a temperature of 700F (371C) is used because colored materials are generated at 750F
(399C) and higher.
The raw reformate is vaporized by dripping it from a syringe pump needle onto 700F (371C) bed of tabular alpha-alumina. For a blank run, the sorption zone in the lower part of the reactor also contains the tabular alumina. The temperature of the lower part of the reactor is held at 45~F (232C). The run with alumina yields a yellow product, which contains 50% of the intially present PNA's after 1 hour on stream, gradually increasing to 95%
after 5 hours on stream. A temperature of 700F (371C) in the sorption zone filled with tabular alumina permits 9S%
of the yellow PNA's to pass.
The tabular alumina in the sorption zone is removed - and replaced with 2"(=6.6 cc) of MCM-41 (d= 33~) sized 30x60 mesh, and held at 450F (232C). The feed rate of the raw reformate is 11 cc/hour. The example is repeated ~ 2 U O 9 4 6 !~
WO96/10537 PCT~S95/11387 using MCM-41 ~d=43A). The results are shown in Table 2 below.
Table 2 Time onAbsorbance %
Stream, Removed Hours Raw Reformate -Al203 - 0.499 0.257 _0 2 0.320 -0 3 0.339 _0 4 0.410 _0 0.473 _0 Silica Gel Grade 40 (4 nm), 4.2 g 1 0.001 99.6 2 0.000 100 MCM-41, d=43A (4.3 nm), 1.71g 1 0.006 97.7 2 0.034 89.4*
.

MCM-41, d=33A (3.3 nm), 1.64g 1 0.005 98.1 2 0.006 98.1 3 0.007 97.9 4 0.008 98.0 * The actual removal was even less, as a yellow deposit appeared at the reactor outlet.
Following adsorption, all the sorbents have a dark-brown zone at the vapor inlet side, indicating a high concentration of sorbate. The silica gel has a trail of color which changes gradually from yellow brown to light yellow with gradually decreasing intensity. MCM-41 (d=43 Angstrom (4.3 nm)) has a uniform tan color below the initial dark-brown zone. MCM-41 (d=33 Angstrom (3.3 nm)) has a uniform tan colored zone of 1 cm. length below the dark-brown zone, with the remainder of the bed white, z z o 0 9 4 6 ~t WO96/10537 -31- PCT~S95/11387 indicating that the PNA('s) are more readily sorbed by the smaller pore MCM-41.
The colors observed after sorption of PNA's may be explained as follows. A dark brown top layer represents a high concentration of yellow PNA's. Grey was observed only with sorbents of smaller pore size and is believed to be due to an electronic effect associated with sorbate bridging of pores, similar to the black color of benzene sorbed on fauiasite. The tan color is thought to be associated with the sorption on a pore wall without bridging. A blend of all these colors is observed on silica gel having a wide pore size distribution.
Example 2 Sorption is carried out using a commercial silica gel, a high silica zeolite Y having a silica:alumina ratio of 3300:1 and MCM-41 of varying pore size. The physical properties of the sorbents are shown in Table 3 as follows:

Table 3 Nominal Pore Average Pore Diameter A Surface Pore Diameter Area Volume A
m~/g cm3/g High Silica 7.2 (.72 nm) NA 0.307.2 (.?2 nm) 2 o Zeolite Y
Silica Gel Grade 40 NA 720-760 0.43 23 (2.3 nm) MCM-41, d=43~ 40 (4 nm) 940 1.06 45 (4.5 nm) MCM-41, d=33A 30 (3 nm) 967 0.90 37 (3.7 nm) MCM-41, D=27A 25 (2.5 nm) 975 0.78 31 (3.1 nm) The raw reformate, as set forth in Example 1, is vaporized by dripping it from a syringe pump needle onto 700F (371C) bed of tabular alpha-alumina. The sorption zone is filled~ with 1" (=3 . 3 cc) of sorbent and held at 4S0F (232OC). The feed rate of the raw reformate is 11 cc/hour. The results are shown in Table 4 below.

~ ~ O 9 4 ~ ~

~, ~ ~ ~ o ~ ~o t~ ,1 ., ~n ~, o ~ ~ cn 1~ --. o ~ _ ~ _ C)--:3 \ 6 ~ o ~ ~ ~ ~ o o ~ ~ I~

~1 D o o O O o o o U~

~ o ., ~

U~ o , z~ O O 9 4 6 ~s WO96/10537 PCT~S95/11387 Based on the weight of sorbent, the MCM-41 (d=27 Angstrom (2.7 nm)) performed better than the silica gel.
Following adsQrption, the silica gel has a brown-black color throughQut the bed. MCM-41 (d=43 Angstrom (4.3 nm)) has uniform tan colored below an initial dark-brown layer.
MCM-41 (d=33 Angstrom) has a dark-brown top layer, followed by a tan layer below to a pale yellow layer at the bottom of the bed. MCM-41 (d=27 Angstrom (2.7 nm)) has a dark-grey color throughout the bed with a slight yellow tint at the top of the bed. High silica Zeolite Y has a yellow top layer with a light-grey layer below.
The colors observed after sorption of PNA's may be explained as follows. A dark brown top layer represents a high concentration of yellow PNA's. Grey was observed only with sorbents of smaller pore size and is believed to be due to an electronic effect associated with sorbate bridging of pores, similar to the black color of benzene sorbed on fauiasite. The tan color is thought to be associated with the sorption on a pore wall without bridging. A blend of all these colors is observed on silica gel having a wide pore size distribution.
Silica gel appears to be more effective than MCM-41 (d=27A (2.7 nm)). However, it has to be taken into account that silica gel is denser, and three times as much mass is required to fill the same space. Silica gel has a wide pore size distribution, but the average pore size and the pore volume are smaller in comparison to MCM-41. The smaller pore volume silica gel is expected to show a more rapid breakthrough of the PNA's than MCM-41.
As Table 4 shows, smaller po~e size MCM-41 (d=27A (2.7 nm)) is more effective than MCM-41 (d=33A (3.3 nm)) and MCM-41 (d=43A (4.3 nm)). The effect may be explained by bridging of the pores, when molecules are adsorbed on opposite wall$, thus further decreasing the remaining pore openings. Thus, a pore diameter of 18A (1 . 8 nm) may be most effective for sorbing molecules of PNA size. Two PNA
molecules adsorbed on opposite walls of the pore may be ~00946 ~
WO96/10537 PCT~S95/11387 bridged by a third PNA molecule, as may occur with the MCM-41 (d=27A (2.7 nm)). The second molecule may also be sorbed on the wall not exactly opposite the first, but somehwhat nearer so that the bridging distance is smaller.
, It is further believed that a substance boiling at a high temperatu,re is still sorbed at a high temperature while a small molecule is found in the vapor phase at the same high temperature thus reducing the competition for sorption sites, of the large excess of small molecules in reformate with, PNA molecules.
Changes a,nd modifications in the specifically described embodiments can be carried out without departing from the scope of the invention which is intended to be limited only by the scope of the appended claims.

Claims (10)

CLAIMS:

1. A process for the removal of sorbate molecules from a fluid comprising adsorbing said sorbate molecules on an adsorbent composition comprising an inorganic, porous, crystalline phase material and which exhibits, after calcination, an X-ray diffraction pattern with at least one peak at a d-spacing greater than 18 A Units (1.8 nm) with a relative intensity of 100 and a benzene sorption capacity of greater than 15 grams of benzene per 100 grams of the material at 50 torr and 25°C, wherein the ratio of crystalline phase material pore diameter to kinetic pore diameter of said sorbate molecules is in the range of 1:1 to 4:1.

2. The process of claim 1, wherein said ratio of crystalline phase material pore diameter to kinetic pore diameter of said sorbate molecules is in the range of 2:1 to 3:1.

3. The process of claim 1, wherein said sorbate molecules are polynuclear aromatics.

4. The process of claim 3, wherein the crystalline phase material pore diameter is not larger than three times the polynuclear aromatic molecules kinetic pore diameter.

5. The process of claim 1, wherein said fluid is a hydrocarbon.

6. The process of claims 1 or 3, wherein said adsorbent composition comprises MCM-41.

7. The process of claims 1 or 3, wherein the adsorption is at a temperature in the range of from 100° to 500°C.

8. The process of claims 1 or 3, wherein the adsorption is carried out in the presence of water.

9. The process of claims 1 or 3, wherein the adsorption is carried out at a temperature in the range of from 150 to ~00C.

10. The process of claims 1 or 3, wherein said fluid is in the vapor phase.