3 The Expanding Library of Zeolite Structures: Novel Structures, Novel Features

186



7 Structural Chemistry of Zeolites



(b)



(a)



(c)

Figure 7.13 The tsch¨ rtnerite mineral’s framework structure

o

(a) contains the tsch¨ rtnerite supercage (b), a combination

o

of D6Rs, β-cages, α-cages, and the supercage (c).



alongside some of the structures they have produced. In parallel with this, strategies

of using combinations of organic bases as potential co-template mixtures have also

had important results in the synthesis of zeolites and related solids, particularly as

shown via recent high-throughput studies by Blackwell et al. at UOP [23].

In addition to the use of designed potential SDAs, modification of the inorganic

components of the gel has been found to play a key role in phase selectivity.

The work of Camblor at the ITQ on the introduction of fluoride ions into low

water content synthesis gels gave many new porous silica polymorphs of zeolites

with large pore volumes, where the fluoride ions have a dual role of assisting

silicate condensation and crystallization and balancing the positive charge on the

framework [24]. Early examples of the success of this model for new structures

include the highly porous small-pore zeolite ITQ-3 (with 2D 8-ring pore system),

the large-pore 1D channel structure ITQ-4 (there are now many 1D 12-ring channel

structures known), and the 3D 12-ring ITQ-7 [25]. This approach has subsequently

been used very productively by many researchers in the field. The structural role of

the fluoride is discussed in Section 7.3.3.

Changing the composition and elemental ratios of framework-forming cations

has also been found to exert a large influence on the phase to form. Zones has

investigated the effect of variation of Si/Al and Si/B ratios in the gel as additional



7.3 The Expanding Library of Zeolite Structures: Novel Structures, Novel Features



parameters in the syntheses with new potential SDAs [26], and Corma (and the

Mulhouse and Stockholm groups) found that the inclusion of germanium had a

strong structure-directing influence, because of its propensity to favor the formation

of D4Rs [11, 27, 28] . Furthermore, the addition of inorganic cations and variation

of alkalinity (OH− /T) have been shown to have an important influence on the

zeolite structure to form [29].

A combination of these innovative synthetic strategies has been responsible

for the upsurge in reported novel synthetic zeolite types since 1990 (Figure 7.2).

The need for state-of-the-art structural characterization, combined with specialist

organic synthesis and high-throughput screening of potential SDAs and mixtures

of SDAs under a very wide range of synthetic variables, means that the synthesis of

ever more complex structures is a specialized enterprise. This approach has brought

many significant structural highlights in the last decade or so, including larger

pore sizes, new pore connectivities, increased structural complexity, a broadening

of the compositional range of known structure types, and chiral structures. In

the second main part of this chapter we summarize some of the most important

developments.

7.3.2

Novel Structures and Pore Geometries



One of the most obvious advances in new structural chemistry has been the preparation of zeolites with extra-large pores, larger than the 12-rings found in faujasites

and Beta (Table 7.1). The first 14-ring zeolite, UTD-1 [30] (University of Texas, Dallas), was prepared using the permethylated cobalticenium ion (Figure 7.14). Other

14-ring window pure silicate and germanosilicates were prepared at Caltech (CIT-5)

[31], at Chevron, (SSZ-53 and SSZ-59) [32], and at Mulhouse (IM-12) [33], and a

beryllosilicate (OSB-1) with 14-ring openings in the framework has also been synthesized [34]. Furthermore, two silicates with channels bounded by 18-rings have

been reported, the gallosilicate ECR-34 [35] and most recently the germanosilicate

ITQ-33 [36]. The dimensions of the pore openings as defined crystallographically

are given in Table 7.1. The large-pore nature of ECR-34 has been demonstrated

by the adsorption of large hydrocarbons such as perfluorotri-n-butylamine. The

gallium and germanium contents have important roles in directing these structures, for example, in favoring D4Rs in ITQ-33, and tend to reduce their overall

hydrolytic stability. Nevertheless, these structures signpost the way to thermally

regenerable extra-large-pore acid catalysts. Finally, a germanosilicate ITQ-37 with

channels linked by highly noncircular 30-rings has recently been reported [37]. This

is described in more detail in Section 7.3.5.

Many new structure types with 3D pore connectivity have resulted from these

structures, including those with connectivity in all dimensions via openings of

10-ring or greater (Table 7.2). These are of great interest as adsorbents and catalysts

because of their enhanced molecular transport properties and their resistance

to blocking in catalytic reactions. TNU-9 [38] and SSZ-74 [39], for example, add

to the important class of zeolites with 3D 10-ring channel systems, previously



187



188



7 Structural Chemistry of Zeolites

Table 7.1



Window sizes of extra-large pore silicates.



Zeolite

(code)



Framework

composition



UTD-1

(DON)

CIT-5

(CFI)

SSZ-53

(SFH)

SSZ-59

(SFN)

IM-12

(UTL)

OSB-1

(OSO)

ECR-34

(ETR)

ITQ-33



SiO2



ITQ-37



Connectivity

(MRs)



Window dimensions of

˚

the largest pore (A)



Reference

(year)



14



(8.2 × 8.1)



[30] (1999)



SiO2



14



(7.5 × 7.2)



[31] (1998)



Si0.97 B0.03 O2



14



(8.7 × 6.4)



[32] (2003)



Si0.98 B0.02 O2



14



(8.5 × 6.2)



[32] (2003)



Si0.82 Ge0.18 O2



14 × 12



(9.5 × 7.1) × (8.5 × 5.5)



[33] (2004)



Si0.66 Be0.33 O2



14 × 8 × 8



[34] (2001)



Si0.75 Ga0.24 Al0.01 O2



18 × 8 × 8



(7.3 × 5.4) × (3.3 × 2.8)

× (3.3 × 2.8)

(10.1) × (6.0 × 2.5)



Si0.66 Al0.04 Ge0.30 O2



18 × 10 × 10



Si0.58 Ge0.42 O2



30 × 30 × 30



(12.2) × (6.1 × 4.3) ×

(6.1 × 4.3)

(19.3 × 4.9) ×

(19.3 × 4.9) ×

(19.3 × 4.9)



[35] (2003)

[36] (2006)

[37] (2009)



exemplified by ZSM-5 and ZSM-11. They are the most complex structures yet

observed and are discussed further in Section 7.3.4, along with the 2D 10-ring

zeolite IM-5 [40]. In addition, new 3D-connected 12-ring silicates ITQ-17 [25] and

germanosilicates ITQ-17 [41], ITQ-21 [42], and ITQ-26 [43] add to the previously

known structures with this connectivity, the faujasite and Beta structures (see, for

example, Figure 7.15). ITQ-17 is related to the disordered Beta structures originally

prepared, having the same framework layers, but these are stacked differently, in an

ordered tetragonal arrangement and including D4Rs. In this structure, originally

hypothesized and named Beta polymorph C (Beta C), the three perpendicular

12-ring channel systems intersect at the same place. Originally prepared as a

germanate FOS-5 [44], this has more recently been prepared as the germanosilicate

ITQ-17. ITQ-21 and -26 also possess 3D 12-ring channel systems, and here again

the D4Rs typical of germanium-containing silicates are a crucial structural element.

One of the most important novel classes of 3D-interconnected channel structures that has been prepared contains 12-ring channels intersecting with 10-ring

channels. The zeolites CIT-1 [45], ITQ-24 [46], and MCM-68 [47] are examples of

this type of framework, each prepared with a complex SDA. There is considerable

interest in investigating possible novel shape-selective catalytic performances in

this type of structure. Other novel solids with 3D connectivity include ITQ-33

(18 × 10 × 10). In addition to these new structures with 3D-connected porosity,



7.3 The Expanding Library of Zeolite Structures: Novel Structures, Novel Features



Table 7.2 Recent (post-1990) tetrahedrally connected zeolite

structures with 3D connectivity via at least 10MR openings,

compared with ZSM-5, Y, and Beta zeolites.



Zeolite

(code)



Framework

composition



Space

group



Pore

system (MR)



Dimensions

˚

(A)



Reference

(year)



Faujasite

(FAU)

Beta

(*BEA)



Si1−x Alx O2

(x = 0–0.4)

Si1−x Alx O2

(x = 0–0.05)



Fd3m



12 × 12 × 12



P4122



12 × 12 × 12



[12]

(1958)

[16]

(1988)



ZSM-5

(MFI)

ITQ-7

(ISV)



Si1−x Alx O2

(x = 0–0.05)

SiO2



Pnma



10 × 10*



P42 /mmc



12 × 12 × 12



CIT-1

(CON)



Si0.96 B0.04 O2



C2/n



12 × 12 × 10



ITQ-17

(BEC)



Si0.64 Ge0.36 O2



P42 /mmc



12 × 12 × 12



ITQ-21

(no code)

ITQ-24

(IWR)



Si0.66 Ge0.34 O2



Fm-3c



12 × 12 × 12



Si0.89 Ge0.09 Al0.06 O2



Cmmm



12 × 10 × 10



MCM-68

(MSE)



Si0.90 Al0.10 O2



P42 /mnm



12 × 10 × 10



ITQ-33

(no code)

TNU-9

(TUN)

SSZ-74

(-SVR)

ITQ-26

(IWS)

ITQ-37

(no code)



Si0.66 Al0.04 Ge0.30 O2



P6/mmm



18 × 10 × 10



Si0.95 Al0.05 O2



C2/m



102 × 10*



Si0.96



Cc



10 × 10*



Si0.8 Ge0.2 O2



I4/mmm



12 × 12 × 12



Si0.58 Ge0.42 O2



P41 32 or

P43 32



30 × 30 × 30



(7.4) × (7.4) ×

(7.4)

(6.7 × 6.6) × (6.7

× 6.6) × (5.6 ×

5.6)

(5.1 × 5.5) × (5.3

× 5.6)

(6.5 × 6.1) × (6.5

× 6.1) × (6.6 ×

5.9)

(7.0 × 6.4) × (7.0

× 5.9) × (5.1 ×

4.5)

(6.7 × 6.6) × (6.7

× 6.6) × (5.6 ×

5.6)

(7.5) × (7.5) ×

(7.5)

(5.8 × 6.8) × (4.6

× 5.3) × (4.6 ×

5.3)

(6.8 × 6.4) × (5.8

× 5.2) × (5.2 ×

5.2)

(12.2) × (6.1 ×

4.3) × (6.1 × 4.3)

(5.6 × 5.5),(5.5 ×

5.1) × (5.5 × 5.4)

(5.9 × 5.5) × (5.6

× 5.6)

(7.05) × (7.3 ×

7.0) × (7.3 × 7.0)

(19.3 × 4.9) ×

(19.3 × 4.9) ×

(19.3 × 4.9)



* indicates



0.04 O2



that 3D connectivity is achieved via the intersection of two channel systems.

indicates a tetrahedral cation vacancy.



[17]

(1978)

[25]

(1999)

[45]

(1995)

[41]

(2001)

[42]

(2002)

[46]

(2003)

[47]

(2006)

[36]

(2006)

[38]

(2006)

[39]

(2008)

[43]

(2008)

[37]

(2009)



189



190



7 Structural Chemistry of Zeolites



(a)



(b)



(c)



(d)



(e)



(g)



(f)



Figure 7.14 Projections down the extra-large pore channels

of the 14MR zeolites (a) UTD-1, (b) CIT-5, (c) SSZ-53, (d)

SSZ-59, and (e) IM-12 and of the 18MR zeolites (f) ECR-34,

and (g) ITQ-33.



several new 2D-connected materials have been prepared and their structures solved

(MCM-22, ITQ-3, ITQ-13, ITQ-22, SSZ-56, etc.) [4]. Besides possessing interesting

gas adsorption and catalytic properties in their calcined forms, post-synthetic treatment is a possible route to secondary mesoporosity to increase the dimensionality

of molecular transport. Finally, new structures that show chirality have been prepared and are described in Section 7.3.5: the chiral and mesoporous ITQ-37 is a

remarkable example.

Another major result that has been achieved by these synthetic studies is a

widening of the available compositional range of zeolites with known structures.

The pure silica version of zeolite A (ITQ-29) [48] has been prepared by using organic

species (with low charge densities) in the syntheses, rather than Na+ cations. This

silica shows much higher hydrolytic stabilities than zeolite A. Similar results have



7.3 The Expanding Library of Zeolite Structures: Novel Structures, Novel Features



(a)



(b)



Figure 7.15 Projections down 12MR channels of (a) ITQ-17

(BEC) and (b) ITQ-26, both of which possess 3D-connected

12MR channel systems.



been achieved in UZM-4 (BPH) [23], which is a higher silica and more stable

version (Si/Al > 1.5) of Linde Q (BPH, Si/Al = 1.1) [4]. In a similar way, templating

studies have given silicate versions of structures initially prepared as germanates

or aluminophosphates. Among the latter, SSZ-16, SSZ-24, SSZ-55, and SSZ-73 are

the silica versions of AlPO4 -56, AlPO4 -5, AlPO4 -36, and STA-6, respectively [4]. It

is likely that silicate versions of other aluminophosphates without existing zeolite

analogs will also be obtained: the silicate versions of the larger pore, 3D-connected

DAF-1(DFO) and STA-14(SAO) [4] would be of particular interest.

7.3.3

Expansion of the Coordination Sphere of Framework Atoms



Although zeolites are defined by their tetrahedrally connected frameworks, there

are some examples where their framework cations adopt higher coordination.

The most important of these are when Si expands its coordination by bonding

to fluoride during synthesis and when framework Ti expands its coordination,

for example, with water or upon uptake of hydrogen peroxide. Whereas the first

observation is important in understanding fluoride synthesis [24], the second has

important catalytic consequences [49].

The fluoride ion behaves as an efficient mineralizing agent in the synthesis of

pure silica zeolites [24], where it catalyzes the hydrolysis of silica and enables the

formation of silicate frameworks at pH values of 7–9, where no reaction would

occur in its absence. A series of crystallographic and NMR studies have shown that

in silicas prepared in fluoride media, F− ions are present in the as-prepared solids,

coordinated to lattice silicon atoms, where they raise the coordination number of

Si to 5 (SiO4 F) [50]. The fluoride ion is often found to occur within small cages,

distributed over a number of different silicon sites with partial occupancy. For

example, it is found in [46 ] cages (LTA, AST) and also in cages in nonasil [41 54 62 ],

EU-1 [41 54 62 ], silicalite (MFI [41 52 62 ], ITQ-4 (IFR [42 64 ]) SSZ-23 [43 54 ], and so on.

This is illustrated in Figure 7.16 for F− in the [41 54 62 ] cage of EU-1 [51]. In this



191



192



7 Structural Chemistry of Zeolites



Figure 7.16 Part of the as-prepared EU-1 structure prepared in fluoride medium. The fluoride ion is located within

a small cage and is connected to one of the Si atoms, raising its coordination fivefold. (Reproduced with permission

from [50]).



way F− acts as an inorganic SDA. At the site where it coordinates to the silicon,

the tetrahedrally arranged SiO4 group is distorted so that the three OTO angles

that are closest to the F− ion are increased to minimize the O–F repulsion [8].

Once incorporated into the structure, the F− ion balances the positive charge of

the alkylammonium ion template. Calcination results in simultaneous removal

of both the organic cation and the F− , leaving a SiO2 framework with very few

framework defects. Such materials are hydrophobic and as a direct result have

potential applications in adsorption.

Titanium can substitute for Si at low concentrations in pure silica zeolites,

and adopts tetrahedral coordination once the as-prepared solid is calcined and

dehydrated. Upon exposure to, for example, aqueous hydrogen peroxide [52], it

expands its coordination, acting as a Lewis acid, and the local geometry is distorted.

In this way, titanosilicate zeolites can act as important oxidation catalysts, especially

when they activate hydrogen peroxide [49].



7.3 The Expanding Library of Zeolite Structures: Novel Structures, Novel Features



7.3.4

The Current Limits of Structural Complexity in Zeolites



The ongoing synthetic efforts described above continue to give rise to structures

with increasing complexity, both in terms of their crystallographic description

and the diversity of their framework architecture. In this section, we discuss

in detail the current limits of structural complexity, as defined by the number

of crystallographically distinct tetrahedral cation environments (T-sites) present

in the repeat unit of the structure. In terms of crystallographic complexity, the

structures of both zeolite A and zeolite Y are very simple, for although there are

many tetrahedra within the unit cells of each framework, they are all related by the

many symmetry elements present in the two structures, so there is only one unique

position in each case. By contrast, the structure of ZSM-5 is far more complex and its

framework structure is built from layers with 12 crystallographically distinct T-sites.

As described in Section 7.3.1, each layer is linked to symmetrically equivalent layers

by centers of symmetry. (In fact, high silica materials with this structure distort to

lower symmetry at low temperatures so that they have 24 different sites, but all the

sites retain their original coordination sequences and the distortion is a subtle one)

[53, 54].

Recently two zeolite structures have been solved which have 24 crystallographically distinct sites, each with a different coordination sequence, TNU-9 [38] and IM-5

[40]. A third reported structure, SSZ-74 [39], is similarly complex, with 23 different

T-sites and an ordered vacancy. None forms as single crystals and their complexity

proved a major challenge to structure solution. Fortunately, in a tour-de-force of

powder crystallography, by combining experimental X-ray diffraction (XRD) and

phases determined from high-resolution electron microscopy with crystallographic

computing algorithms, the groups of McCusker, Baerlocher, and Terasaki (for

TNU-9) and McCusker, Baerlocher, and Zou (for IM-5 and SSZ-74) have successfully elucidated the structures of these materials. For TNU-9, high-resolution

powder XRD and HRTEM have been successfully combined with the FOCUS

program, designed to search trial electron density maps for tetrahedra. Sufficient

correct phase information was obtained from the electron micrographs to make

this possible. For IM-5 and SSZ-74, similar types of experimental diffraction and

imaging data have been combined with so-called enhanced charge flipping crystallographic algorithms adapted from other application programs. These charge

flipping programs can be applied without applying structural constraints to obtain

solutions (and so are generally applicable, regardless of framework connectivity).

The structure of TNU-9 has a very similar projection (down [010]) to that observed

for ZSM-5, but has two different sets of straight 10-ring channels, labeled A and

B in Figure 7.17, rather than one. Perpendicular to [010], channels of type B are

linked to each other via short 10-ring channels and to channels of type A via 10-ring

windows. It is instructive to investigate how a structure as complex as TNU-9 is

built from repeating units, and as a consequence to suggest how it may assemble

from solution, in the presence of organic SDAs [55]. The TNU-9 framework can

be built up from a single kind of chain, similar to what is seen in ZSM-5. These



193



194



7 Structural Chemistry of Zeolites



B



A



y

x



(a)



z



4.9 Å



5.4 Å



y

z

x

y



x



(b)



(c)



z



Figure 7.17 The complex topology of TNU-9 (a) contains

two kinds of channel (b), as depicted in pink and blue,

which show complex 3D connectivity. The framework itself

is built from one kind of chain that links to form layers (c),

which join to give the final structure. (Reproduced with permission from [38] and [52]).



chains are connected via mirror planes to give sheets which are asymmetric (i.e.,

one side is different from the other). These sheets can only be linked to each other

via their similar sides, so that there are two different intersheet regions. Modeling

indicates that the organic SDA is able to interact favorably in different sites in both

intersheet regions, so indicating a mechanism by which stacking of layers can be

favored during synthesis [55].

The high silica zeolite IM-5 is, like ZSM-5 and TNU-9, highly thermally stable

and shows interesting performance for hydrocarbon processing and selective

NO reduction [40]. The structure has a 2D-connected 10-ring channel system.

It is, like TNU-9, similar in projection to ZSM-5 but rather than complete 3D

˚

connectivity there are slabs (thickness 25 A) that contain channels that show

connectivity in three dimensions, complete with complex channel intersections.

Each slab is isolated from adjacent sheets by impermeable silicate sheets. The

catalytic performance could only be properly explained once this connectivity was

understood. The work also described how, because the charge flipping structure

solution methodology did not need to make use of symmetry, it determined the

location of 288 silicon atoms and 576 oxygen atoms in the unit cell. This is very

encouraging, as ever more complex structures are prepared, and suggests that an

inability to grow single crystals should not limit our discovery of their intricate



7.3 The Expanding Library of Zeolite Structures: Novel Structures, Novel Features



architectures. Both TNU-9 and IM-5 crystallize in the presence of the same organic

SDA, bis-N-methylpyrrolidiniumbutane, which has also been shown to favor the

crystallization of several other zeolites depending on solution pH, Si/Al ratio,

additional cations, and so on [29]. This is a clear indication that the detailed

composition of the hydrothermal reaction mixture plays a crucial role in supplying

species for the growth of these complex solids.

Finally, a third highly complex zeolite, SSZ-74, prepared using a template very

similar to that used for TNU-9 and IM-5 (bis-N-methylpyrrolidiniumheptane) has

recently been solved by the same combination of XRD, TEM, and application of

the charge flipping algorithm used successfully for IM-5. The structure possesses

23 different T-sites and an ordered tetrahedral vacancy (effectively a 24th T-site

that is vacant). The structure contains an undulating 10-ring channel connected by

straight 10-ring channels and leading to 3D connectivity limited by 10-rings similar

to that for ZSM-5. The implications of the observed vacant site are discussed in

more detail in Section 7.3.6.

7.3.5

Chirality and Mesoporosity



As described above, considerable progress has been made toward the synthesis of

zeolite structures with different connectivities and larger pore sizes, at least up to

˚

the 12.2 A circular pore openings observed in ITQ-33. One of the key remaining

challenges is to produce batches of chiral zeolites which consist entirely of crystals of

one enantiomer, so that these might find application in enantioselective separation

and catalysis. Very few silicate zeolites with potentially chiral porous structures are

known, and the most important examples are described below.

Zeolite beta, described in Section 7.3.3, typically exhibits stacking faults and does

not crystallize as an ordered polymorph. Theoretically, though, there is a regular

stacking sequence that would give a polymorph (Beta-A) that would be chiral, and

efforts have been made to obtain this by the use of chiral templates. No fully ordered

chiral polymorph A has yet been prepared, and one of the difficulties is to achieve

chiral recognition over the long helical pitch of the channel that this polymorph

would have, using either single molecules or molecules that could order.

More recently, two zeolitic silicates have been prepared as mixtures of chiral

crystals. The silicogermanate SU-32 is one of a family of fully tetrahedrally

connected frameworks templated by the achiral ammonium ion (CH3 )2 CHNH3 +

that are built up from chiral layers with 12-ring openings that consist of ‘‘4−1’’

repeating units of tetrahedra, which point up and down alternately (Figure 7.18)

[56]. Relating adjacent layers of this type across inversion centers results in the

achiral 12-ring structure SU-15, whereas stacking layers with a ±60◦ rotation

between adjacent sheets (which enables fortuitous coincidence) results in chiral

polymorphs of SU-32 (space groups P61 22 or P65 22). Remarkably, crystals form

in a mixture of pure enantiomorphs, which is different from what is observed for

zeolite Beta. The resulting structure comprises only D4Rs and 46 58 82 102 cavities

that share 10-ring openings and make up helical channels which are either right



195



196



7 Structural Chemistry of Zeolites



˚

or left handed, with pore openings 5 × 5.5 A along the channels and intersected

˚

by eight-ring channels that run parallel to the channel axis (4.7 × 3 A). The

challenge now is to prepare this topology in the form of a more stable silicate or

aluminosilicate, and even to prepare only one of the two enantiomeric forms.

The second type of zeolitic solid that crystallizes in chiral form is the silicogermanate ITQ-37 that crystallizes in the chiral space groups P41 32 and P43 32, with

a structure related to the gyroidal (G) periodic minimal surface. This G-surface

is exhibited by micelle-templated mesoporous amorphous silicas. In these mesoporous solids two pore systems of interconnected channels of opposite hand are

separated by a silica wall located at the G-surface, whereas ITQ-37 can be thought

of as having one of these pore systems empty and the other filled by a chiral zeolitic

framework. The remarkable structure is beautifully illustrated by Sun et al. [37], and

the cavity of this structure and the entrance window are illustrated in Figure 7.19.

Besides possessing chiral channels, the structure has the lowest framework density

˚

observed for a zeolite (10.2 T/1000 A3 ) and cavity dimensions in the mesoporous

˚ As in other germanosilicate structures, D4Rs are important SBUs.

regime (>20 A).

In ITQ-37 these have one or two terminal hydroxy groups, and this interrupted

nature of the framework is crucial in enabling the large cages to form.

The generation of a material that is both intrinsically chiral and with mesoporous

cavities is a significant step forward, but before chiral zeolites can find application,



(a)



a

c b



(b)

Figure 7.18 The chiral germanosilicate SU-32 is built from

sheets of tetrahedra made up from 4−1 units alternatively

pointing in opposite directions (a). These stack to give a

chiral structure that possesses helices of cages (b). Adjacent

cages are depicted in different colors.