Solid-state NMR studies of Ziegler-Natta and metallocene catalysts.

Solid State Nuclear Magnetic Resonance ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Contents lists available at ScienceDirect

Solid State Nuclear Magnetic Resonance journal homepage: www.elsevier.com/locate/ssnmr

Solid-state NMR studies of Ziegler–Natta and metallocene catalysts Koen C.H. Tijssen, E.S. (Merijn) Blaakmeer, Arno P.M. Kentgens n Radboud University, Institute for Molecules and Materials, Heyendaalseweg 135, Nijmegen, The Netherlands

art ic l e i nf o

a b s t r a c t

Article history: Received 23 December 2014 Received in revised form 30 March 2015

Ziegler–Natta catalysts are the workhorses of polyolefin production. However, although they have been used and intensively studied for half a century, there is still no comprehensive picture of their mechanistic operation. New techniques are needed to gain more insight in these catalysts. Solid-state NMR has reached a high level of sophistication over the last few decades and holds great promise for providing a deeper insight in Ziegler–Natta catalysis. This review outlines the possibilities for solid-state NMR to characterize the different components and interactions in Ziegler–Natta and metallocene catalysts. An overview is given of some of the expected mechanisms and the resulting polymer microstructure and other characteristics. In the second part of this review we present studies that have used solid-state NMR to investigate the composition of Ziegler–Natta and metallocene catalysts or the interactions between their components. & 2015 Elsevier Inc. All rights reserved.

Keywords: Ziegler–Natta Metallocene Solid-state NMR Polymerization Catalysis

1. Introduction In the last century polymers have become one of the main materials in our everyday life. The first synthetic polymer, bakelite, was made by Leo Baekeland [1] in 1907 and the synthesis of many other polymers soon followed. However, control over the molecular weight of the polymer and stereochemical control was missing. The

Abbreviation: AA, aluminum-reduced and activated; Cpn, 1,2,3,4,5-pentamethylcyclopentadienyl; CP, cross polarization; Cp, cyclopentadienyl; Cp0 , substituted cyclopentadienyl; CQ, quadrupolar coupling constant; CSA, chemical shift anisotropy; DFS, double frequency sweep; DIBP, diisobutyl phthalate; EB, ethyl benzoate; EFG, electric field gradient; eQ, electric quadrupole moment; Est-ME, MgCl2 5EtOH PhCOOEt; Et, ethyl; EXAFS, extended X-ray absorption fine structure; Fi, weight fraction of polymer i; Flu, fluorenyl; GPC, gel-permeation chromatography; H4Ind, tetrahydroindenyl; HETCOR, heteronuclear polarization; HR, high resolution; Ind, indenyl; iPr, isopropylidene; IR, infrared; m, meso; M, metal; MAO, methyl alumoxane; MAS, magic angle spinning; ME, MgCl2 6EtOH; Me, methyl; Mi, molecular mass of polymer i; Mn, number averaged molecular weight; MQMAS, multiple quantum magic angle spinning; MW, molecular weight; Mw, weight averaged molecular weight; MWD, molecular weight distribution; NMR, nuclear magnetic resonance; NQR, nuclear quadrupolar resonance; OAC, organoaluminum compound; P, polymer chain; Ph, phenyl; QCPMG, quadrupolar Carr–Purcell– Meiboom–Gill; r, racemic; SSNMR, solid-state nuclear magnetic resonance; T1, spin-lattice relaxation time; T 1ρ , spin-lattice relaxation time in the rotating frame; Tm, melting temperature; WG, wet ground; WURST, wideband uniform rate and smooth truncation; ZN, Ziegler–Natta; δaniso, anisotropic part of the chemical shift; δiso, isotropic part of the chemical shift; η, asymmetry parameter for the chemical shift; ηQ , quadrupole asymmetry parameter. n Corresponding author. E-mail addresses: [email protected] (K.C.H. Tijssen), [email protected] (E.S. Blaakmeer), [email protected] (A.P.M. Kentgens).

discovery of Ziegler–Natta (ZN) catalysts in 1953 brought about a breakthrough in polymer chemistry. These catalysts provided for the first time stereochemical control over the polymerization process. Furthermore, polymerization using a Ziegler–Natta catalyst is a socalled living polymerization and the polymers created this way can have extremely high molecular weights. A wide variety of monomers can be polymerized; however, the polymerization of ethylene and propylene are by far the best studied ZN catalyzed polymerizations. Other α-olefins, like styrene, and diolefins or cyclic olefins have also been polymerized with ZN catalysts as is reviewed by Coates [2]. ZN catalysts have been in use for over half a decade, yet their structure and mechanistic workings are even today not fully understood, despite the numerous research articles on this topic. Most research studies focus on the polymer, instead of the catalyst, to deduce the structure and especially the mechanistic workings of the catalyst. However, this yields circumstantial evidence at best. It might be more useful to focus on the ZN catalyst itself. Solid-state nuclear magnetic resonance (SSNMR) is an information rich spectroscopic technique that can provide information about local structure and dynamic processes, therefore SSNMR might be a good candidate to further unravel the structure and mechanism of Ziegler–Natta catalysts. The scope of this review article is thus the study of ZN catalysts using SSNMR as the main characterization technique.

2. Ziegler–Natta catalysts: a theoretical overview ZN catalysts are fascinatingly complex. In this section we will elaborate on the separate components of a ZN catalyst, the current

http://dx.doi.org/10.1016/j.ssnmr.2015.04.002 0926-2040/& 2015 Elsevier Inc. All rights reserved.

Please cite this article as: K.C.H. Tijssen, et al., Solid State Nucl. Magn. Reson. (2015), http://dx.doi.org/10.1016/j.ssnmr.2015.04.002i

K.C.H. Tijssen et al. / Solid State Nuclear Magnetic Resonance ∎ (∎∎∎∎) ∎∎∎–∎∎∎

2

view on their mechanistic operation, the different forms of chirality involved and the types of polymer they can produce. 2.1. Historic overview In 1953 Karl Ziegler was the first to discover that a mixture of metal alkyls and transition metal salts can easily be used to produce high density polyethylene [3]. One year later Giulio Natta demonstrated that with the same type of mixture one can synthesize isotactic polymers from α-olefins [4]. In 1963 Ziegler and Natta were awarded the Nobel prize in chemistry for their work on these catalysts. In honor of the two men these catalysts have ever since been called Ziegler–Natta catalysts. In 1957 Breslow [5] and Natta [6] independently discovered the first homogeneous Ziegler–Natta catalyst: bis(cylcopentadienyl)titanium dichloride. Their discovery forms the basis of an entire group of homogeneous catalysts: the metallocene catalysts. At the end of the 1960s the field of heterogeneous ZN catalysts was given a new boost with the invention of MgCl2-supported catalysts [7]. In 1988 Ewen et al. [8] produced the first bridged metallocene catalyst, the ansametallocene Me2C(Cp)(9-Flu)ZrCl2, that was able to produce highly syndiotactic polypropylene. A more extensive historic overview can be found in the excellent review of Mülhaupt [9], which was published in 2003 on the occasion of the 50th anniversary of Ziegler's discovery.

compounds. In general a Ziegler–Natta catalyst consists of three components: the catalyst, the cocatalyst and an electron donor. The catalyst is a transition metal compound of the group IV–VIII transition metals, usually a halide, alkoxide or alkyl derivative. Titanium (Ti) is by far the most used transition metal, other transition metals that are often used are vanadium (V), chromium (Cr) and zirconium (Zr) [10]. The cocatalyst is a metal alkyl or metal alkyl halide of the group I–III base metals, in practice this is almost always an aluminum compound. The third component, the electron donor, can vary from amines to ethers and esters, and is used to increase the catalysts activity and/or stereoselectivity. Two types of electron donors exist: internal and external donors. The internal donor is added to the solid catalyst, while an external donor is added to the total polymerization system. The two types of electron donors have slightly different roles (adapted from Ref. [11]). Roles of the internal donor: 1. Enhance the effective surface area by preventing coagulation of the MgCl2 support during synthesis of the catalyst. 2. Prevent formation of non-stereospecific sites. 3. Take part in the formation of highly isospecific sites. 4. To be later replaced by external donors, resulting in the formation of more isospecific sites. Roles of the external donor:

2.2. Polymer characteristics The molecular weight (MW) of a polymer is the result of competition between propagation and termination reactions. The molecular weight can be optimized by modifying the ligands sterically and electronically, or in case a lower MW is needed a terminating agent can be used. A method to increase the MW is to lower the polymerization temperature or to use heterogeneous instead of homogeneous catalysts. However, a drawback in most MW-increasing methods is that they decrease the catalyst's activity. The molecular weight distribution (MWD), or (poly)dispersity, is one of the most important characteristics of a polymer and is expressed as a dimensionless number calculated by dividing the weight averaged molecular weight (Mw) of the polymer by the number averaged molecular weight (Mn), or M w =M n . The number averaged molecular weight (Eq. (1)) is simply the total mass of all polymers divided by the number of polymer molecules. The weight averaged molecular weight (Eq. (2)) is determined by calculating the weight fraction (Fi, Eq. (3)) of each different polymer (i), multiplied by its molecular mass (Mi) and then sum all these products: P NM Mn ¼ P i i ð1Þ Ni Mw ¼

X

F i Mi

NM Fi ¼ P i i Ni Mi

ð2Þ ð3Þ

Typical values for the MWD are between 1 and 3. However, higher values, up to 30, are also possible. The MWD of a polymer is considered narrow if M w =M n o 2 and extremely narrow if its value is smaller than 1.2. Polymers with a large MWD flow more easily in their molten state, which is important for their processability. A standard method for the determination of the MWD is gel-permeation chromatography (GPC). 2.3. Heterogeneous Ziegler–Natta catalysts Heterogeneous Ziegler–Natta catalysts are complex mixtures of several metal compounds and optionally some small organic

1. 2. 3. 4.

Poison non-stereospecific sites selectively (see also Section 2.12). Convert non-stereospecific sites into isospecific sites. Convert isospecific sites into more highly isospecific sites. Increase the reactivity of the sites.

Terano et al. [12] demonstrated that the internal electron donor ethyl benzoate (EB) can be used to prevent the formation of aspecific sites, promote the formation of sites with a high isospecificity and even create new active sites. Often there is a fourth component present: the ‘support.’ Different catalyst species (single, dimeric, clusters, etc.) can be grown on different lattice planes of the support crystal (Fig. 1) and this results in a number of catalytic sites with different symmetry, activity and stereoregularity. A common support is crystalline MgCl2, because (titanium) catalysts supported on this material show very high activities. Kashiwa [13] proposed that this is caused by the similarities in crystal structure and ionic radius of MgCl2 and the titanium catalyst. Other commonly used supports are silicon oxide, SiO2, and a hybrid of both MgCl2 and silicon oxide [14,15]. The classical example of a supported ZN catalyst is TiCl4 supported on a MgCl2 crystalline phase. Extended X-ray absorption fine structure (EXAFS) measurements have shown that TiCl4 can be coordinated as monomer or dimer on the (100) face of MgCl2 support [17]. The Ti(IV) can reduce during the reaction to Ti(III) or Ti(II); however, Ti(II) is considered to be inactive. Density functional theory (DFT) calculations have shown that the TiCl4 and TiCl3 fragments preferentially adsorb to the (110) lateral cut of the crystal instead of the (100) face [17]. 2.4. Homogeneous Ziegler–Natta catalysts Homogeneous catalysts differ from the heterogeneous ones in the fact that they have the same phase as the reactants. Both homogeneous Ziegler–Natta catalysts and their reactants are thus dissolved in a solvent, or in some cases the solvent is the reactant (for instance liquid propene). However, these catalysts can become heterogeneous at low temperatures while the polymer is formed. The largest group of homogeneous Ziegler–Natta catalysts are the metallocenes and the focus has long been on this type of catalyst. Later, different homogeneous Ziegler–Natta catalysts were synthesized [18]. Most notable



C1-symmetric Ti C2-symmetric Ti (hindered)

C2-symmetric Ti (unhindered) (100)

3

summarized in what is now referred to as Ewen's symmetry rules [8,28–30]. 2.5. Heterogeneous vs. homogeneous (110)

dimeric Ti-species MgCl 2 Mg

} Cl Ti-species Ti

} Cl Vacancy

Ti-cluster

Fig. 1. Different Ti-centers of a Ti-catalyst grown on a MgCl2 support; reproduced from Ref. [16].

are those based on nitrogen and oxygen donor ligands, and octahedrally coordinated catalysts [19–21].

2.4.1. Metallocene catalysts Metallocene catalysts are organometallic complexes with a single transition metal, usually a group 4 metal. The metallocenes are named after their transition metal (titanocenes, zirconocenes and hafnocenes). Of these metals zirconium is the most active, followed by hafnium and titanium, respectively. Four ligands are coordinated to a metallocene. Two of them are per definition cyclopentadienyl (Cp) groups or Cp-substituted/Cp-based (Cp0 ) ligands. These Cp/Cp0 groups are coordinated with all five carbon atoms to the metal, this is called a hapticity of five, also indicated as η5-Cp. The other two are easily removable ligands, often chloride ions. In a metallocene/cocatalyst system the active species are 14-electron d0 cationic metallocene alkyls [10] and there is a tendency for the anion to coordinate to the vacant coordination site of the transition metal. To prevent this coordination fluorinated anions are used, especially [B(C6F5)4] . Substituting the Cp rings will affect the activity of the catalyst. The substituent may cause steric crowding and therefore decrease the catalysts activity. However, the substituent can also be electron donating, which would increase activity. The overall effect is the result of both steric and electronic factors [10]. To prevent the Cp or Cp0 groups from rotating and to gain better control over the properties of the metallocene catalyst the two Cp/Cp0 groups are often connected with a bridging group. This type of metallocene complexes are called ansa-metallocene complexes. In 1958 Lüttringhaus and Kullick [22] synthesized the first ansametallocene complexes. However, it was not until the pioneering work of Brintzinger et al. [23–25] in the late 1970s and 1980s and their synthesis of the first chiral ansa-metallocenes, that ansa-metallocenes really became an important field of research. The bridging group in ansa-metallocenes does not only prevent the Cp0 groups from rotating, but it constrains the Cp0 groups in fixed angles and distances from the metal center. The exact configuration of the Cp0 ligands, described in detail by Shapiro [26] and Conway et al. [27], and the type of bridge determine the catalyst's activity and stereoregularity. The symmetry of a metallocene is very important for the stereochemistry. The tacticity (see Section 2.8) of the polymer depends on the symmetry of the catalyst. Ewen, Kaminsky and both their co-workers did a lot of work in this field and their results are

In heterogeneous systems a variety of active centers with different structures and environments is present, while in homogeneous systems the ligand environment is uniform and well defined, and there is a single type of active center. As a result, heterogeneous and homogeneous catalysts differ in several respects: (1) In heterogeneous systems a fraction of the transition metals may be coordinated such that it is catalytically inactive, while in homogeneous systems all transition metal atoms form active centers. (2) Homogeneous catalysts are known to produce polymers with a narrower molecular weight distribution than those produced by the heterogeneous catalysts. This is explained by the multiple types of active centers in heterogeneous catalysts. (3) For exactly the same reason, heterogeneous catalysts are more difficult to study in terms of their kinetics and reaction mechanisms. However, metallocene catalysts are in general less active than the supported heterogeneous catalysts. On the other hand the properties of a homogeneous catalyst can be changed more easily. Simply modifying the ligands, allows one to change the activity or the stereoregularity of the catalyst. Catalyst handling is easier for heterogeneous catalysts. For industrial use homogeneous catalyst are therefore often made heterogeneous by supporting them on a carrier material, the support. MgCl2, SiO2, zeolites, clays and polymers are all common supports. Hence classification of a catalyst as homogeneous or heterogeneous becomes ambiguous. The catalytic activity and polymer properties are more extensively discussed in the review of Hlatky [31] for supported homogeneous ZN catalysts and in the review of Campos et al. [32] for metallocene catalysts supported on zeolites or mesoporous silicas. Both homogeneous and heterogeneous catalysts have their benefits. But homogeneous catalysts are easier to modify and study, therefore in the last three decades they have been subjected to more mechanistic studies than their heterogeneous cousins. Despite the differences chemists hope that by studying the homogeneous catalysts they will find out more about the heterogeneous catalysts as well. 2.6. Activation of the Ziegler–Natta catalyst The transition metal complex has to be activated before its participation in the polymerization reaction. The activated form of the complex must have both a vacant coordination site to accommodate the monomer, and an M–C or M–H bond to initiate the insertion. The creation of the latter is done by the cocatalyst. Trialkyl aluminum (AlR3, for example triethyl aluminum) and methyl alumoxane (MAO) are standard cocatalysts, we will briefly focus on the activation with MAO. MAO can be obtained by hydrolysis of AlMe3, which was discovered by accident in the group of Kaminsky [33]. Studies have shown that MAO is a mixture of several compounds, including some ring structures, of which many are in equilibrium with each other [34]. The aluminum and oxygen atoms are arranged alternately and the free coordination sites of the aluminum atoms are filled with methyl groups. However, the aluminum atoms are coordinatively unsaturated and the small MAO units, mainly [Al4O3Me6], form clusters and cage structures with molecular weights between 1200 and 1600 g/mol [33]. The role of MAO is threefold: (1) methylation of the catalyst; (2) activation of the methylated catalyst by removal of the chloride, creating a vacant site; and (3) stabilization of the activated cationic catalyst complex. Besides that MAO also acts as an impurity scavenger.



4

The drawbacks of MAO are its relatively high costs and the large amount needed [30]. Nevertheless MAO is used in industrial polymerization processes. For industrial purposes it is often heterogenized by absorption on silica, alumina and other supports to make gas phase polymerization possible [35].

2.7. Chirality and orientation Chirality plays a very important role in polymer chemistry and it is present in several parts of the catalyst–polymer system. In most heterogeneous ZN catalysts, the metal centers are octahedrally coordinated. This is the case in crystalline TiCl3, in which each chlorine atom is coordinated to two titanium atoms and each titanium atom is coordinated to six chlorine atoms, so there are vacant sites in the lattice (Fig. 2b). A catalytically active Ti-center is located on the surface and is coordinated to two other titanium atoms via four chlorine atoms. We can describe each –Cl–Ti–Cl– chain as a bidentate ligand, and the active titanium center is thus occupied by two of these bidentate ligands. The two other coordination sites are needed for coordination of the monomer and the growing polymer chain. The bidentate ligands make the Ti-center chiral and, depending on their relative orientation, the center is labeled Λ or Δ (Fig. 2a). Another important aspect is the orientation in which the monomer is coordinated. A monomer can be coordinated to the catalyst's metal center in several ways (Fig. 3): with the R-group (which in the case of propene is a methyl group) closest to the polymer chain, called primary coordination or 1,2-coordination, or with the R-group directed away from the growing polymer chain, called secondary coordination or 2,1-coordination. Also, the R-group will point to the left or right relative to the plane in which the metal center and the two carbon atoms of the double bond lie. These different orientations are named si and re. The orientation of the coordinated monomer will determine which way the monomer is inserted into the polymer, and therefore determine the microstructure of the polymer (see Section 2.8). A third form of chirality can be found in the conformation of the growing polymer chain. The dihedral angle X–M–C3–P (X is the center of the double bond of the alkene; M is the transition metal; C3 is the growing polymer chain-end carbon atom (after insertion of the next

M

M

Δ-complex

Λ-complex

Λ

Δ Λ

Δ Λ

Δ

Δ Λ

Λ

Δ Λ Δ

Fig. 2. (a) Enantiomeric metal complexes and (b) crystal structure of TiCl3, with Tiatoms (black spheres), Cl-atoms (white spheres) and vacant sites (white squares); adapted from Ref. [16].

Fig. 3. The four possible monomer coordinations to the catalyst's metal center; reproduced from Ref. [30].

monomer it is the third carbon atom in the polymer chain); and P is rest of the polymer) can be either positive (þ) or negative ( ). 2.8. Tacticity During polymerization the orientation of the monomer before insertion is relevant for the final polymer structure, also called tacticity. The classical example to demonstrate this is polypropylene. The propene monomer can be seen as an ethene molecule that is substituted with a methyl group. The carbon atom to which the methyl group is connected is a prochiral center. After insertion of the monomer, the prochiral center becomes a chiral center and two enantiomers are possible. R,S-nomenclature, based on the Cahn– Ingold–Prelog priority rules [36], can be used to assign the different enantiomers or, often preferred, the enantiomers can still be indicated with si and re. Polymers of the type –(CH2CHR)n– have wildly varying mechanical properties depending on the regularity in the configuration of the chiral centers. If the main carbon chain of the polymer is represented in a fully extended planar zigzag, then the R-groups can be on either side of the plane. The polymer is called atactic (Fig. 4c) if the configuration of the R-groups is random. If the configuration of each repeating unit is such that all the R-groups are on one side of the plane, then the polymer is isotactic (Fig. 4a). The configuration can also alternate for each subsequent unit, resulting in a syndiotactic polymer (Fig. 4b). A special case of tacticity is the hemiisotactic polymer [37] (Fig. 4d). In this polymer order and disorder coexist: each even R-group is on the same side of the plane, while the uneven R-groups have a random configuration. For other polymers, like dienes and 1,2-substituted ethylenes, the description of the isomers and regularity can be even more complicated. The coordination of the monomer to the complex determines how the monomer is inserted into the polymer and thus the tacticity. There are two mechanisms that control stereoregulation. The first type of stereoregulation is chain-end control, in which the steric interactions



R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

R

Fig. 4. A polymer chain in different tacticities: (a) isotactic, (b) syndiotactic, (c) atactic, and (d) hemiisotactic.

between the side groups of the last added and incoming monomer control if the monomer is inserted si or re. The second type of stereoregulation is catalytic-site control, also called enantiomorphic control or chiral-site control. In this type the asymmetry of the active site forces the monomer to coordinate either si or re. When the stereoregulation is catalytic-site controlled, it is possible that the catalytic site changes during the polymerization reaction; for instance because of the rotation of a Cp0 group. This may lead to stereoblock polymers consisting of isotactic and atactic blocks (Fig. 5). A wrong insertion of a monomer can reveal which stereoregulating mechanism is at work. Fig. 6 shows two isotactic polymer structures with stereo-errors. In structure B the errors are corrected and insertion with the old configuration is resumed. In this case the catalyst clearly affects the configuration in which the monomer is inserted and the stereoregulating mechanism is catalytic-site control. In structure A the stereo-errors are not corrected and after the errors are made the catalyst keeps inserting new monomers in the new configuration, resulting in blocks of si-inserted monomer and blocks of re-inserted monomer. This sort of stereoregulation is the result of chain-end control. Mechanistic studies [39,40] have shown that stereo-errors are often not caused by incorrect coordination and insertion of the monomer. Instead chain-end epimerization (Fig. 7) seems to cause these stereo-errors. In chain-end epimerization the stereochemistry of the last inserted monomer is changed in two consecutive series of βhydride elimination, rotation and insertion of the hydride. Stereoblock polymers can also be produced by heterogeneous ZN catalysts; however, in this case the exact mechanism is unknown [2]. Two reasonable mechanisms are (1) structural rearrangements of the catalytic sites and (2) chain transfer between sites that propagate the polymer with different tacticities. 2.9. The Cossee mechanism In 1960 Cossee proposed a mechanism for the propagation of the polymerization of ethene [41] and propene [42,43]. Roughly 50 years later the basics of Cossee's mechanism are still generally accepted. In this mechanism (Fig. 8) the monomer is first coordinated by π-bonding at the vacant coordination site of the activated transition metal complex. The active site of the Ziegler–Natta catalyst is the M–C σ-bond between the transition metal and the polymer chain, so the monomer will line up with this bond. Finally, the monomer is inserted into the M–C bond by migration

5

of the polymer chain from the transition metal to the double bond. In the resulting complex the polymer chain is now one unit longer and coordinated at the former vacant site. The site previously occupied by the polymer chain has now become the vacant site. After a repetition of the propagation steps and insertion of a second monomer unit, the coordination of the metal complex is equal to its original coordination. The polymer chain and the vacant coordination site can thus interchange. However, the other ligands of the transition metal complex (not drawn in Fig. 8 for clarity) do not change site and they are called spectator or ancillary ligands. These ligands have an important function though; they determine how each monomer is coordinated and inserted (si or re), and therefore the tacticity of the polymer. In general, stereoregulation of the catalyst is mainly determined by steric interactions of the Cp/Cp0 ligands (catalytic-site control) and by the last inserted monomer unit of the polymer chain (chain-end control). Another important factor controlling the stereoregulation is the polymerization temperature. Especially for homogeneous catalysts this is important, because at higher temperature the configuration of the ligands is more likely to change due to thermal disturbance. The polymer chain flips between two sites and the vacant coordination site consequently also alternates between the two sites, therefore the Ziegler–Natta catalyst has two active sites. It should be noted that these sites can be chemically different due to an asymmetry in the complex. Each active site can therefore insert the monomer with a different chirality and also the stereoregularity of both sites may be different. However, for heterogeneous ZN catalysts the mechanism might be a little bit different. The two sites are nonequivalent and therefore the polymer chain might have a strong preference for one of the two sites. In this case, after insertion of the monomer, the polymer chain can immediately flip back to its original site and so the monomer might always be coordinated to the same vacant site [44]. Many mechanistic studies have been performed, among which there are also some solid-state NMR studies. A noteworthy example is the nutation NMR study of Clarke et al. [45]. They studied the polymerization of acetylene using a titanium tetra-nbutoxide/triethylaluminum catalyst. A mixture of 96% normal acetylene and 4% doubly 13C-enriched acetylene was polymerized and a proton-decoupled 13C nutation NMR spectrum was acquired. From this spectrum the dipole–dipole coupling can be calculated, which is inversely proportional to the third power of (in this case) the distance between the neighboring 13C atoms. Clarke and Yannoni measured a bond length of 1.37 Å, which clearly showed that it was a double bond. So the double bond remains between carbon atoms of the acetylene unit, instead of being formed between the original units, thereby they disproved the so-called metallacycle mechanism. Unfortunately the Cossee mechanism cannot explain all observations and the mechanism itself predicts some problems. For example, the vacant coordination site is acidic and in principle it can be attacked by Lewis bases in the solution. In practice this is not observed. Also this mechanism predicts that the reaction rate is first order, however many groups have reported higher-order reaction rates [46]. Other mechanisms have been proposed, however the Cossee mechanism remains the most used mechanism to describe the propagation. Examples of other mechanisms are: ‘the trigger mechanism’ of Ystenes [47], in which a second monomer triggers the insertion of the first one; or the more popular mechanism proposed by Green, Rooney and Brookhart [48], where a hydrogen atom on the α-carbon of the growing polymer chain interacts with the catalyst's metal center. This so-called α-agostic interaction lowers the activation energy for olefin insertion, see Fig. 9. Piers and Bercaw showed by isotopic labeling that there is indeed an α-agostic interaction in some catalysts [49].



6

Fig. 5. A symmetry switching catalyst synthesized by Coates and Waymouth [38] that produces a stereoblock polymer; reproduced from Ref. [30].

Structure A

P M

mmmmrmmmmmmrmmmm P

Structure B

M

P*

M

mmmmrrmmmmrrmmmm Fig. 6. Two isotactic fragments with insertion errors caused by different stereoregulation mechanisms: chain-end control (structure A) and catalytic-site control (structure B).

P M

P *

M P M

Fig. 8. The Cossee mechanism for propylene polymerization: M¼ transition metal, □¼ vacant coordination site, P ¼polymer chain (uneven number of propylene units), and Pn ¼polymer chain (even number of propylene units).

Fig. 7. The mechanism of chain-end epimerization; reproduced from Ref. [2].

2.10. Polymer microstructure Besides the molecular weight and (poly)dispersity of the polymer (Section 2.2) the microstructure is also of great importance. Polymers with different tacticities have different properties. The standard technique for characterization of the microstructure of a polymer is 13 C NMR. Especially the characterization of the microstructure of polypropylene has been well documented [28,50,51]. The chemical shifts of the methyl groups of polypropylene are dependent on the relative configurations of the neighboring propylene units, also called

steric arrangement. For each methyl group the exact chemical shift is determined by the steric arrangement of the unit itself and, on both sides, its two closest neighboring units. The five units in total are called a pentad and the pentad's steric arrangement can be described by a four-letter code. Each letter, r or m, describes the steric arrangement of two neighboring units, called a diad. The two units of the diad can either have the same configuration, called meso (m), in which case the methyl groups are directed in the same direction. Or they can have a different configuration, called racemic (r), and the methyl groups are directed in opposite directions (Fig. 10). There are ten unique pentads (see Fig. 11) that correspond to nine separate signals in the 13C NMR spectrum (the pentads mmrm and rmrr give rise to one single signal). Often the microstructure is expressed as the percentage of a certain pentad, for instance mmmm for isotactic polymers and rrrr for syndiotactic polymers. The NMR data can also be used to measure insertion errors, such as stereo-errors (Section 2.8) and 2,1-insertion. Besides NMR also mass spectrometry [52–55], IR and UV spectroscopy [56,57], chromatography [58,59], atomic force microscopy and scanning tunneling microscopy [60,59] are used to determine polymer microstructure. Other common physical properties to characterize a polymer are the melting temperature (Tm) and the solubility. Again polypropylene is a commonly used example. Polypropylene with a Tm higher than 165 1C or insoluble in refluxing n-heptane is considered isotactic propylene [10]. Regiodefects can have a strong effect on the



H

H

H

H

H

‡

H

rrrr

P M

P

M

P

1.462

M

rrrm 1.659

mmrr 0.771

mmrm

7

mmmm 0.000

0.971

mmmr

rmrr

0.248

mrrm

0.971

1.843

rmmr

mrmr

0.469

1.170 H H M

P

P M H

H

Fig. 9. The Green–Rooney–Brookhart mechanism.

r-diad

R m -diad

R

r-diad

R

R

R

m -diad

Fig. 10. All four possible configurations of two neighboring monomer units, together the four diads form an mrmr-pentad.

crystallinity and melting temperature of a polymer, as is the case for isotactic polypropylene [30]. 2.11. Termination and end-functionalization The molecular weight of a polymer is an important property, because it determines the mechanical properties of the polymer. The molecular weight is determined by chain termination, which is the effect of several processes. These processes [61] are: (1) thermal cleavage by β-hydride elimination; (2) thermal cleavage by β-CH3 elimination; (3) chain transfer to the monomer; (4) chain transfer to a metal alkyl, for instance MAO; and (5) chain transfer to a transfer agent, often H2. To prevent chain transfer to the metal alkyl/MAO the use of the borane B(C6F5)3 as an activator instead of the metal alkyl/ MAO has been reported [62,63]. A special case of terminating the growing chain is by endfunctionalization. A polymer is end-functionalized to use it as a precursor for advanced polymeric materials or when the polymer is used as a hydrogen bonding unit for a supramolecular assembly [64]. 2.12. Catalyst poisoning As discussed before, certain compounds, when added to a heterogeneous catalyst, have an electron donating effect and increase the catalyst's activity and stereoselectivity. The activity of a catalyst can be expressed as the amount of polymer produced per mole of catalyst per time unit, often in kg/(mol h). The opposite is also possible: a molecule coordinating on the catalyst's surface in the vicinity of the active species can lower the catalyst's activity and/or stereoselectivity due to steric or electronic effects. In this case we speak of catalyst poisoning. It is even possible to selectively poison the aspecific or isospecific sites. Tangjituabun et al. showed that small organic molecules only affect the activity and not the stereoselectivity of a MgCl2-supported TiCl4 [65]. Homogeneous catalysts can also be poisoned. Kallio et al. [66] showed that both O2 and CO can reduce the activity of a

Fig. 11. Top: the four-letter codes of the ten pentads, their schematic structure and their relative chemical shift and bottom: 13C NMR spectrum of the methyl region of a polypropylene polymer, showing the nine pentad signals; adapted from Ref. [28].

metallocene; however, polymers with a higher molecular weight were obtained when the catalyst was poisoned with either gas.

3. Solid-state NMR studies of ZN catalysts Olefin polymerization with Ziegler–Natta catalysts is a well-studied topic. However, nearly all the research is based on the analysis of the produced polymer. In far fewer cases the catalyst itself is directly studied. In this section an overview is given of the investigations of ZN catalysts by SSNMR. 3.1. SSNMR of heterogeneous ZN Catalysts In the field of heterogeneous ZN catalysis most SSNMR studies have been performed for MgCl2-supported catalysts. Research is especially focused on the aluminum coordination states of the cocatalyst and coordination and interaction of the electron donors, although also some research is done on the MgCl2 support [67]. 3.1.1. Composition determination A simple NMR analysis of ZN catalysts is the investigation of the catalyst's composition. Thushara et al. [68] investigated the TiCl4 catalyst supported on an MgCl2 5EtOHEB (Est-ME) molecular adduct, which contains the internal electron donor ethyl benzoate (EB). This internal donor was formed by a reaction of benzoyl chloride and ethanol. However, the synthesis of the catalyst complex consists of several reaction steps and the group was interested if the catalyst complex formed contains EB, or if the EB is lost in one of the reaction steps. They compared this catalyst system with the catalyst system TiCl4 =MgCl2 6EtOH (with the


8


molecular adduct MgCl2 6EtOH ¼ ME), which does not contain an internal electron donor. Ramped 13C CP-MAS experiments were performed and from the carbonyl peak at 169 ppm (Fig. 12, iv) it was concluded that EB is indeed present in the catalyst complex. From the 13C singlepulse excitation MAS spectrum which can be interpreted quantitatively they determined that one out of six ethanol molecules was converted to EB. Ohashi et al. [69] performed 47,49Ti solid-state NMR on heterogeneous ZN catalysts. The two NMR active isotopes of titanium (both in low natural abundance) have very similar Larmor frequencies; 54.420 (47Ti) and 54.434 (49Ti) MHz at 21.8 T. This allows for the simultaneous detection of both isotopes. Ohashi et al. started with the measurement of a model system consisting of co-milled MgCl2 and TiCl4. They measured fast MAS spectra at 21.8 T and discovered two narrow lines corresponding to the two titanium isotopes. A pronounced broadening of the lines and upfield shift of the peaks of both isotopes is observed with increased co-milling time. However, in all cases only one resonance per isotope is observed, indicating one type of titanium on the MgCl2 surface. The ratio of the linewidths, 47Ti/49Ti, is lower than the theoretical value for solid titanium species. Some kind of mobility is thus present, which the authors explain via chemical exchange between two different phases of the liquid and solid state, that is, free TiCl4 and TiCl4 on MgCl2. Real ethylene slurry polymerization is performed with these catalysts. Higher activity is observed for the catalysts with longer co-grinding times. How this is related to the observed spectral changes is still under investigation. Garoff et al. [70] studied the effect of calcination of silica support surfaces on the activity of ZN catalysts and looked at two calcination temperatures (300 and 590 1C). After calcination the silica is reacted (in consecutive steps) with MgR2, HCl and TiCl4. This process is performed to create catalyst particles in which the good support morphology of silica is combined with the high activity of MgCl2 based catalysts. Garoff et al. studied these supports after each reaction step with 1H, 13C and 29Si MAS NMR. The 1H MAS NMR spectra were

deconvoluted to determine the amount of isolated and hydrogen bonded hydroxy groups. Results show that there is a ring opening of the siloxane bridges in the silica calcined at 300 1C and this causes the formation of low-activity Si–O–Ti(Cl)2–O–Si groups, and thus the low activity of these catalysts. Static 35Cl NMR spectra of silica supported TiCl4 have been reported by Johnston et al. [71]. They claim to have a monografted [RSiO–TiCl3 ] supported species. Fig. 13 shows the resulting spectrum which is about 400 kHz wide. It has been obtained with the WURST-QCPMG (wideband uniform-rate smooth truncation quadrupolar Carr–Purcell–Meiboom–Gill) technique with the aid of strong 1H decoupling at a field of 18.8 T. Johnston et al. give a single site fit (shown above the experimental data) with quadrupolar parameters of C Q ¼ 14:31 MHz and ηQ ¼ 0:15. The lack of sharp features in the experimental data is attributed to a small distribution in quadrupole and chemical shift parameters, even though there is only one titanium species. This results from the lack of long-range order for these surface sites. This initial result is promising as it shows the possibility to observe a dilute species in a reasonable time (20 h). Differences can be expected with changes in the titanium coordination. Hence it would be interesting to see the changes upon addition of electron donors and/or the addition of aluminum alkyls. 3.1.2. Coordinative states of the cocatalyst As noted before the MWD of a polymer is an important factor. In some cases it is advantageous to have a polymer with a broad MWD. Several approaches exist to create polymers with broad or bimodal/multimodal MWD. One approach is to use a catalytic system with multiple transition metal species. Another approach is to use a multi-stage process. In this process each stage produces polymers with a different MWD. Liu et al. [72] noticed that they could also control the MWD of the polymers produced with their SiO2-supported ZN catalyst by changing the cocatalyst. They investigated the effect of different alkyl-Al cocatalysts on the polymerization and 27Al NMR was used to determine the coordinative states of the alkyl-Al cocatalysts on the activated SiO2-supported titanium catalyst. The 27Al MAS SSNMR experiments were done in a 7 mm silicon nitride rotor, which contains Cl Cl Ti Ti Cl

SiO2

O

Cl Cl

Cl Cl

Si O

O O

Fig. 12. 13C MAS NMR spectra of (a) the molecular adducts (i) ME and (ii) Est-ME and (b) the titanated catalysts (iii) Ti/ME and (iv) Ti/Est-ME; reproduced from Ref. [68].

300

200

100

0

−100

−200

−300

kHz

35

Fig. 13. Static Cl NMR spectrum of silica supported TiCl4 and an analytical simulation (smooth line on top), reproduced from Ref. [71].



some aluminum impurities whose signals (peaks 1 and 2 in Fig. 14) were used as internal reference. 27Al is a quadrupolar nucleus (spin I¼ 5/2, 100% natural abundance) with moderate quadrupolar coupling constants [73]. The aluminum chemical shift is sensitive to the coordination. Experiments were done on the SiO2 support with and without the titanium catalyst and with diethylaluminum chloride (AlEt2Cl) or triethylaluminum (AlEt3) as cocatalyst, Fig. 14. Peaks 5, 4 and 3 were assigned to 6, 5 and 4 coordinated aluminum species, respectively. For AlEt2Cl peak number 3 is significantly higher and peaks 4 and 5 are lower for the supported catalyst (Fig. 14c) compared to the support's spectrum (Fig. 14a). However, large overlap of the peaks complicates the assignment. More unambiguous spectra can be obtained with a Multiple Quantum Magic Angle Spinning (MQMAS) experiment [73]. So AlEt2Cl is dominantly four-coordinated. For AlEt3 peaks 5 and 4 are higher and so AlEt3 is dominantly five- and six-coordinated. Peak 3 is also a little higher, indicating a small amount of four-coordinated aluminum. The authors link these results to their MWD measurements, which show a single narrow band for the polymers produced with AlEt2Cl and a broad trimodal band for the polymers produced with AlEt3 as a cocatalyst. Changing the chloride ligand of the AlEt2Cl cocatalyst to other halogens strongly affects the MWD of the polymer. Diethylaluminum bromide (AlEt2Br) broadens the MWD of the low MW polymers and causes polymers with a high MW to be produced. Diethylaluminum iodine (AlEt2I) strongly broadens the MWD and makes it trimodal, just like AlEt3. The 27Al NMR experiments seem to agree with these results. The peaks assigned to five- and six-coordinated aluminum are higher in the catalyst/AlEt2Br complex than in that of the catalyst/AlEt2Cl complex. In the AlEt2I spectrum these peaks are even higher, which is in agreement with the broader MWD. Potapov et al. [74] also investigated the structures of adsorbed organoaluminum compounds (OACs) using 27Al NMR. They first looked at structures formed on MgCl2. Different coordination environments were observed depending on the type and concentration of the OAC used. With a low AlCl3 concentration the main adsorbate was found to be four coordinated, corresponding to monomers of AlCl3 adsorbed on MgCl2. Increasing the AlCl3 concentration leads to dominantly six coordinated species. Six coordinated species are also found for AlEtCl2 originating from (AlEtCl2)m chains. For AlEt2Cl three types of aluminum are observed attributed to one five coordinated species and two six coordinated species. Potapov et al. [74] found an aluminum loading corresponding to the number of Lewis acid centers on MgCl2, which is similar to the titanium loading. Therefore they suggest that OACs adsorb on the same Lewis acid sites of MgCl2 as TiCl4. However, the aluminum loading on TiCl4/MgCl2 catalyst systems exceeds the number of free surface centers. It is therefore proposed

Fig. 14. 27Al MAS SSNMR spectra of (a) the SiO2 support with AlEt2Cl, (b) the SiO2support with AlEt3, (c) the SiO2-supported Ti catalyst activated with AlEt2Cl, and (d) the SiO2-supported Ti catalyst activated with AlEt3, peaks 1 and 2 are rotor background; adapted from Ref. [72].

9

that two classes of OACs exist: bonded to MgCl2 surface sites or bonded to Ti-centers, Ref. [75]. Indeed AlEt2Cl adsorbs on α-TiCl3. The resulting NMR spectrum is similar to that of adsorbed AlEt2Cl on MgCl2. The interaction of AlEt3 with a MgCl2-supported TiCl4 catalyst (TiCl4/MgCl2) also leads to a similar NMR spectrum. This is explained by the formation of AlEt2Cl resulting from the alkylation and reduction of TiCl4. The spectrum of the sample AlEtCl2/TiCl4/MgCl2 has signals from both adsorbed AlEtCl2 and AlEt2Cl. The effect of internal and external electron donors on the coordinative states of alkyl-Al cocatalysts adsorbed on TiCl4/MgCl2 was also investigated [76]. 27Al MAS SSNMR spectra of the catalyst system with AlEt2Cl as a cocatalyst and with or without internal and/or external donor were recorded. From these spectra Potapov et al. concluded that both the internal and the external electron donor do not change the coordinative states of the cocatalyst. The donors only compete for surface sites. Hasan et al. [77] investigated several wet ground (WG) processed AA-TiCl3 (aluminum-reduced and activated TiCl3) catalysts with 27Al MAS NMR. Pure AlCl3 (Fig. 15a) is octahedrally coordinated. Their results show that the aluminum in both WG AA-TiCl3 with and without AlEt2Cl as a cocatalyst are also octahedrally coordinated. In the spectrum of WG AA-TiCl3 activated with AlEt3 the octahedral peak (0–10 ppm) is decreased and several new peaks are visible at lower field. These peaks correspond to four- and five-coordinated aluminum species. Hasan et al. noticed that the catalysts with a high-coordinated cocatalyst (mostly octahedral) are more isospecific than the catalysts that have mostly four- and five-coordinated Al cocatalysts. So possibly isospecificity can be linked to the coordination state of the cocatalyst. Besides 27Al NMR, 35Cl NMR offers a second possibility to observe the AlCl3 cocatalyst. Sandland et al. [78] acquired a solid state 35Cl spectrum of anhydrous AlCl3, which shows a site with an extremely high isotropic chemical shift (δiso) of 2880 ppm. A large quadrupolar coupling constant (CQ) of 9.4 MHz was found. 35Cl spectra of heterogeneous catalyst systems are likely to be dominated by signals from the MgCl2 support, but due to its high shift AlCl3 might be distinguishable. Similar research was also performed on other types of catalyst. For instance, Xia et al. [79] performed 1H and 27Al MAS NMR experiments on CrOx/SiO2 Phillips catalysts with AlEt3 as a cocatalyst and they determined the coordinative state of the Al centers at several Al/Cr molar ratios.

3.1.3. Structure of MgCl2/EtOH and other MgCl2/ROH adducts MgCl2-supported heterogeneous ZN catalysts can be more efficient by addition of a Lewis base, for instance ethanol, to the support. The new MgCl2 nEtOH support (where n is between 0 and 6) is called an (molecular) adduct and the MgCl2/EtOH ratio (or the value of n) turns out to be of great importance for the activity and stereoselectivity of the ZN catalyst. Sozzani et al. [80] synthesized and investigated several MgCl2 nEtOH adducts with different MgCl2/EtOH ratios. The exact composition of the adducts was determined by solution NMR. The molar ratios in the five MgCl2 nEtOH molecular adducts, labeled with the letters h–l, are (h) n ¼2.73; (i) n ¼2.48; (j) n ¼2.01; (k) n ¼1.83; and (l) n ¼1.45. These adducts are mixtures of pure compounds. The magnesium sites in these pure compounds are octahedrally coordinated and each Mgsite can be labeled Ln, with n equal to the number of EtOH ligands. The chemical shift of ethanol will depend on the number of EtOH molecules coordinated to the magnesium site. The 13C CP-MAS NMR spectra (Fig. 16) of the adducts clearly show the different magnesium sites (Ln's) and their change in intensity between adducts with a different molar ratio n. Sozzani et al. were able to assign the methylene signals to the magnesium sites Ln. Furthermore they noticed simple ratios between the peak integrals in the spectra of adducts h and l. By further research they



10

Al (VI)

(a)

(b)

(c) (d) 300

200

100

0

-100

-200

-300

ppm Fig. 15. The 27Al MAS NMR spectra of (a) AlCl3, (b) WG AA-TiCl3, (c) WG AA-TiCl3/ AlEt2Cl, and (d) WG AA-TiCl3/AlEt3; reproduced from Ref. [77].

Fig. 17. 13C HETCOR NMR spectrum of adduct k, methylene and methyl regions are shown separately, the traces in the carbon domain, collected at 1H chemical shifts of 3.93 and 4.37 ppm, are shown on the right side of the figure; reproduced from Ref. [80].

Fig. 16. 13C CP-MAS NMR spectra of the methylene (CH2) and methyl (CH3) regions of the MgCl2 nEtOH adducts h–l, the peaks are deconvoluted with Lorentzian lineshapes; reproduced from Ref. [80].

could determine that adducts h and l are the pure compounds 5MgCl2 14EtOH and 2MgCl2 3EtOH, respectively. Adducts i, j and k are mixtures of these two compounds. 1 H–13C HETCOR experiments were done and according to Sozzani et al. these can in principle be used to determine the pure compounds in a mixed adduct. Fig. 17 shows the 1H–13C HETCOR NMR spectrum of adduct k. Traces in the carbon domain, at a 1H chemical shift of 3.93 and 4.37 ppm show the profiles of the pure compounds 5MgCl2 14EtOH (h) and 2MgCl2 3EtOH (l), respectively. The mobility of the ethanol ligands was investigated with variable temperature and exchange measurements. These show that the ethanol ligands exchange between different conformations and sites. For instance the 2D exchange spectrum (5 s mixing time) of 5MgCl2 14EtOH shows an exchange between the L3 and L4 sites. The group of Gopinath has investigated several other porous MgCl2/ROH adducts. In the paper of Thushara et al. [81] they give

the results of 1H MAS and 13C CP-MAS NMR experiments on a MgCl2–2-butanol adduct and its corresponding TiCl4/MgCl2–2butanol catalyst. The measurements of the catalyst show that the alcohol molecules are removed during the catalyst preparation and that chlorobenzene and hexane, both solvents, enter the pores. The 13C CP-MAS experiments on the adduct show that the coordinated 2-butanol molecules are magnetically inequivalent. In the spectrum the peak of the methyl groups furthest away from the hydroxy group is split in two. Also the CH2 peak shows a shoulder. The same group also published [82] their results of solid-state NMR experiments of a MgCl2 6CH3 OH adduct and its corresponding titanium catalyst. In the adduct magnesium is octahedrally coordinated with methanol, which gives rise to a single carbon signal. By heating the adduct the stoichiometry changes. Methanol is removed from the adduct and the methanol groups become magnetically inequivalent, as can be seen from the 13C CP-MAS spectra (Fig. 18) taken of adducts treated at different temperatures. At higher temperatures multiple peaks are visible and the chemical shift changes. From this data the authors concluded that mixed phases are present. When they added TiCl4 to the adduct they observed the formation of oxygenated species as well as free methanol. The interaction between TiCl4 and the MgCl2 6CH3 OH adduct can yield HCl which in turn can cause a cascade of side reactions in which the oxygenated species are formed.

3.1.4. Internal donor interaction Several studies have looked into the interaction between the electron donors and the metal centers of the support and/or the catalyst. For this they used the 13C chemical shift of the electron donor. An interaction between carbon and the metal changes the electronic environment which leads to a change in the chemical shift tensor. However, this does not necessarily lead to a change in the isotropic



Fig. 18. (a) 1H MAS and (b)

13

11

C CP-MAS NMR spectra of MgMeOH adducts treated at different temperatures; reproduced from Ref. [82].

chemical shift (the diagonal of the chemical shift tensor) which is observed at high MAS rates. The determination of the full chemical shift tensor is necessary, which is generally done at low MAS rates. Heikkinen et al. [83] investigated the coordination of internal electron donors to a MgCl2-supported TiCl4 catalyst. They recorded 13 C CP-MAS NMR spectra of several catalyst/electron-donor combinations and of several so-called model compounds (a mixture of unsupported TiCl4 and an electron donor, or MgCl2 and an electron donor). At MAS rates of 4–5 kHz no differences in the isotropic chemical shift were observed, except for one model compound. In that case the carbonyl signal was split up in two peaks corresponding to coordination of the donor to two different lateral cuts, (100) and (110), of the crystal. Then they repeated the experiments at a lower MAS rate of 1.4 kHz. From these experiments they determined the principal values of the chemical shift tensor and calculated the CSA (δaniso) and asymmetry parameter. From the CSA values they could determine to which metal, Mg or Ti, the electron donor coordinates. A positive value for δaniso indicates coordination to Mg and a negative value indicates coordination to Ti, as was confirmed by the model compounds, which only contain Mg or Ti. All studied catalysts had a positive δaniso and the electron donor was thus coordinated to Mg. The relevance of the magnitude of δaniso is unknown. The magnitude of the asymmetry parameter is a measure for the symmetry of the electron distribution around the carbonyl moieties of the electron donors and gives information about the symmetry of the coordination between donor and metal. A lower value for the asymmetry parameter indicates a more symmetric coordination. The coordination of the Lewis bases ethyl benzoate (EB) and diisobutyl phthalate (DIBP) to the MgCl2 support, TiCl4 and TiCl4/ MgCl2 adducts have been investigated by Sormunen et al. [84]. They used 13C CP-MAS for this purpose and looked specifically at chemical shift changes and line widths. From the chemical shift changes they could not support the view of complexation via the carbonyl. Unlike Heikkinen et al. [83] they did not determine the full 13C tensor, so potential useful chemical shift information is lacking. It is nevertheless still possible to estimate relative interaction strengths from the shifts. A stronger coordination to the Lewis acids for DIBP compared to EB was concluded from the larger

chemical shift change. When complexes of MgCl2–EB and TiCl4–EB are compared, it can be seen that the stronger Lewis acid, TiCl4, leads to a stronger effect on the chemical shift. The stronger electron withdrawing effect leads to more deshielding. For the bidentate Lewis base DIBP a splitting of lines was sometimes observed. This line splitting can be explained by coordination to two different crystal sites. The shift from free donors is also observed. The treatment of MgCl2 has significant influence on the line widths of the 13C resonances of the Lewis bases. Broadening of the lines is observed going from unactivated to mechanically activated and chemically activated MgCl2. The broadening is attributed to an increased crystal disorder and/or the formation of amorphous phases. When TiCl4 is added a narrowing of the lines is observed. The environment of the ester thus becomes more ordered. Busico et al. [85] also investigated the binding of electron donors to a MgCl2 support. They made use of 1H High-Resolution (HR) MAS to acquire a spectrum of a so-called slurry, a suspension of ball-milled MgCl2 in cyclohexane-d12. However, this is not really a SSNMR technique, because only the liquid or liquid-like (very mobile) components are detected. Electron donors adsorbed to the MgCl2 surface will be less mobile and their lines are therefore broadened. Different electron donors can be added to the slurry and in principle the electron donor will bind with different affinities to the Mg sites of the crystal support, depending on the plane in which the site is located. The 1H HR-MAS NMR spectra of a slurry with octadecylmethyldimethoxysilane (RMeSi(OMe)2, Fig. 19a and b) and of a slurry with octadecyldimethylmethoxysilane (RMe2Si(OMe), Fig. 19c and d) show the adsorption of both electron donors to the support. At low molar ratios (RSi=Mg o 0:7%) both electron donors are absorbed to the (110) plane of the MgCl2 support, as can be seen from the broad lines in Fig. 19a and c. The signals of the methoxy protons (3.96 ppm), the methylene protons (0.59 ppm) and the protons of the methyl group bound to the silicon atom (0.10 ppm) are extremely broadened. The signals of the protons on the methyl group of the octadecyl group (0.94 ppm) are a little less broadened, due to higher mobility. It was determined by DFT calculations that electron donors adsorb to the (110) plane. At higher molar ratios there is free electron donor in the slurry with RMeSi(OMe)2, as can be seen from the sharp lines in Fig. 19b


12


active and stereoselective catalyst has the shortest relaxation values. While the opposite was true for the least active catalyst, which had the longest relaxation time. These results hint that there may be a link between the rigidity of the complex and the performance of the catalyst, however, more extensive studies are needed to confirm this.

Fig. 19. 1H HR-MAS NMR spectra of an MgCl2/cyclohexane-d12 slurry with the electron donor RMeSi(OMe)2 in a molar ratio RSi=Mg of (a) 0.4% and (b) 1.7%; and of an MgCl2/cyclohexane-d12 slurry with the electron donor RMe2Si(OMe) in a molar ratio RSi=Mg of (c) 0.2% and (d) 2.0%; adapted from Ref. [85].

(for instance the methoxy line). The spectrum of the slurry with RMe2Si(OMe) as an electron donor (Fig. 19d) becomes more resolved and the electron donor is now adsorbed on the (104) plane. At an even higher molecular ratio (RSi=Mg o 3:3%) free RMe2Si(OMe) is observed. The weaker electron donor RMeSi (OMe)2 does not adsorb to the (104) plane. These results were confirmed by diffusion filtered NMR experiments. Pakkanen et al. [86] studied the adsorption and interaction of silyl ethers with MgCl2-supported Ziegler–Natta catalysts with 13C CP-MAS NMR and elemental analysis. The silyl ethers, RSi(OMe)3 (R¼Et, Ph or OMe), served as internal and external electron donors. From the study of only isotropic chemical shifts (no full tensors) they claimed that the majority of the silyl ethers is mobile and weakly bound to the support. A second species is aluminumbound. Treatment with TiCl4 removes the mobile species, but leaves the aluminum-bound silyl ether surface complexes. It was observed that titanium does not bind to the silyl ether species and therefore the titanium can only bind to sites formerly occupied by the mobile silyl ethers. In this way the silyl ethers direct the coordination of TiCl4 on the support.

3.1.5. Relaxation times In the study of Ziegler–Natta catalysts hardly any use has been made of relaxation measurements. However, relaxation measurements are an excellent tool for studying local mobility [87,88]. A good example is the report of Vizzini et al. [89]. They studied three TiCl4 catalysts supported on a MgCl2 ð2ethylhexanolÞ adduct that was further modified with phthalic anhydride and diisobutylphthalate. By doing proton T1 (spin-lattice relaxation) and T 1ρ (spin-lattice relaxation in the rotating frame) measurements they found that the most

3.1.6. In situ monitoring The potentially most informative way to investigate the workings of a ZN catalyst is not by studying the produced polymers, or by studying the ZN catalyst itself, but to study the catalyst in situ, while the propagation is ongoing. The in situ study of ZN catalysts is difficult, especially the study of its surface. This is because most techniques require a high-vacuum environment. NMR is an exception to the rule and only requires a few hardware modifications. Mori et al. [90] modified an NMR probe for in-line addition of propene gas into an MAS rotor. Also the sensitivity of the catalyst to moisture and air had to be taken into account. They used their setup to study the polymerization of propene on a supported Ti catalyst and on an AA-TiCl3 catalyst [91]. First Mori et al. recorded a 13C CP-MAS NMR spectrum of the supported catalyst every 200 s, for 70 min. After 10 min propene was introduced to the rotor to start the polymerization and from then on the rotor was flushed with propene till the end of the experiment. The signals at 44, 26 and 22 ppm (Fig. 20a–c) are of the CH2, CH and CH3 groups of the growing polypropylene chain. The signal at 10 ppm is of the alkyl groups of the AlEt3 cocatalyst. During the first 10/15 min a rapid growth is observed. During the next 30 min the intensity increases more steadily. Next Mori et al. used 1H MAS NMR to determine the mobility of the growing polymer chain. By acquiring a spectrum of the system before and after polymerization and subtracting the first from the latter, a spectrum of the so-called nascent polymer (Fig. 20f) is obtained. The line contains a broad and a narrow component, suggesting there are rigid and mobile segments. By integration it was determined that 35% of the nascent polymer is mobile. In situ 1H MAS NMR experiments were performed to monitor the polymerization using non-supported AA-TiCl3. It was observed that the increase in signal intensity over time was roughly the same as for the supported catalyst. It should be noted that the

Fig. 20. 13C CP-MAS NMR spectra of the in situ polymerization with a supported catalyst at: (a) t¼ 0, (b) t ¼10 min and (c) t ¼60 min; 1H MAS NMR spectra of the in situ polymerization with a supported catalyst at: (d) t ¼0, (e) t ¼60 min and (f) spectrum of the nascent polymer obtained by subtraction of spectrum d from spectrum e; adapted from Ref. [91].



lines in the 1H MAS NMR spectra were broadened, presumably due to the paramagnetic Ti3 þ ions. Finally Mori et al. did in situ 27Al MAS NMR experiments of the supported catalyst. There were no significant changes in the spectrum during the polymerization and it was concluded that the aluminum states of the cocatalyst do not change. They repeated the experiment, flushing the rotor with air instead of propene to measure the deactivation of the catalyst by reactions with oxygen and moisture. This changed the aluminum coordination states; however, the peaks could not be assigned with certainty, because the chemical shift depends both on the coordination state of the aluminum and on the kind of atom coordinated to the aluminum (Cl or O).

Zr-Cp

Zr-CH3

d isobutyl

>C=CH2

3.2. SSNMR of metallocene catalysts Metallocene catalysts are soluble and therefore most NMR experiments have been done using solution NMR. Most SSNMR examples are of supported metallocene catalysts, in particular silica supported metallocenes, but there are also some examples of solid state model compounds. 3.2.1. SiO2-supported metallocenes Moroz et al. [92] supported the zirconocene catalysts Cp2Zr (CH3)2 and Cp2ZrCl2 on silica (SiO2). The activity of a catalyst supported on unmodified SiO2 is very low. Therefore the SiO2 was modified with trimethylchlorosilane ((CH3)3SiCl) or triisobutylaluminum (Al(i-C4H9)3). They used 1H MAS NMR to study the SiO2 and all modified supports. The spectrum of SiO2 (Fig. 21a) consists of two peaks: silanol (RSi–OH, 2.0 ppm) and QSi–OH (4.5 ppm). The spectrum of (CH3)3SiCl treated silica (Fig. 21b) contains a single peak at 0.0 ppm, belonging to RSi–CH3 . From the intensities it could be concluded that the treatment with (CH3)3SiCl fully substituted the hydroxyl groups with trimethylsilyl groups. The spectrum of Al(i-C4H9)3 treated silica (Fig. 21c) shows, besides the peaks of the butyl groups (0.9 and 1.5 ppm), a signal at 4.6 ppm belonging to adsorbed isobutene. Finally, the spectrum of SiO2-supported Cp2Zr(CH3)2 (Fig. 21d) shows a Zr–Cp signal at 5.7 ppm and a Zr–CH3 signal at 0.17 ppm. The two peaks at 0.7 and 0.9 ppm are most likely physically adsorbed n-hexane solvent peaks. The catalyst/support structure was expected to be RSi–O–ZrðCH3 ÞCp2 ; however, the NMR spectrum shows there are fewer Zr–CH3 groups than is expected based on the previous structure. The SiO2-supported zirconocenes show sufficiently high activities, even without adding any cocatalyst. Kröger-Laukkanen et al. [93] have reported solid state 13C and 29 Si CP-MAS experiments of zirconocenes supported on silica surfaces. They observed several resonances for the Cp carbons and have attribute those to different binding modes of the zirconium metal to the silica, for example, X3Zr(OSi) and X2Zr (OSi)2 with X ¼Cp, Cl. Jezequel et al. [94] tested partially dehydroxylated SiO2 as a n support material for the zirconocenes CpnZrMe3 (Cp ¼pentamethylcyclopentadienyl) and Cp2ZrMe2. A solid state spectrum of the supported catalyst was compared with a solution NMR spectrum of the unsupported CpnZrMe3. In both spectra three lines are visible at roughly the same chemical shift. The signals at 120 and 9 ppm are of the Cpn-ring and the methyl groups on the Cpn-ring, respectively. The signal at 37 ppm is of the Zr–Me groups, as was confirmed by spectra of the selectively 13C-labeled catalyst (only the Zr–Me groups were labeled). The chemical shift difference of the Zr–Me group between the supported and non-supported catalyst may indicate that the zirconium atom is more electron rich in the supported catalyst due to back-bonding of the surface oxygen atom. Based on this data the structure of the supported catalyst is RSi–O–ZrCpnðCH3 Þ2 .

13

c ≡ Si-CH3

*

*

b

≡ Si-OH

=Si(OH)2

15

10

a

*

* 5

0 (ppm)

-5

-10

-15

Fig. 21. 1H MAS NMR spectra of: (a) original silica, (b) silica after silylation with (CH3)3SiCl, (c) silica after treatment with Al(i-C4H9)3, and (d) Cp2Zr(CH3)2 adsorbed silica, asterisks denote spinning side bands; reproduced from Ref. [92].

Uusitalo et al. [95] investigated a broader range of SiO2supported metallocene catalysts. They premodified the silica with the (EtO)3Si(CH2)3Cp coupling agent to create a cyclopentadienyl covered surface. The surface was activated with BuLi and then the metallocene catalyst could be immobilized on the modified silica surface. During this process several different catalyst structures are formed (see Ref. [95] for the actual structures). Catalyst species attached to the surface via RSi–O–M bonds are expected to be inactive in polymerization, unless the cocatalyst is able to break the M-O bond. Uusitalo et al. used 13C CP-MAS NMR to study the modified support and the immobilization of metallocene catalysts on this support. In Fig. 22a the spectrum of the modified silica support is given. The signals of the –(CH2)3Cp group and the ethoxy (EtO–) groups are clearly visible. The spectrum of the BuLi-activated support (Fig. 22b) is a little more complex, due to the overlap of the butyl signals (12, 18, 26 and 34 ppm) with the Cp group and ethoxy signals. Furthermore the lines are broadened due to the presence of BuLi. After immobilizing CpHfCl3 to the support the spectrum remains similar to spectrum of the BuLi-activated support (Fig. 22b). One important difference is the peak at 113 ppm which belongs to the Cp group coordinated to Hf. Unreacted ethoxy groups diminish the catalyst's activity; therefore the 13C CP-MAS NMR spectra can be useful to determine the amount of unreacted ethoxy groups.


14


Fig. 22. 13C CP-MAS NMR spectra of (a) (EtO)3Si(CH2)3Cp-modified silica support (S1), (b) S1 after activation with BuLi; adapted from Ref. [95].

Also 29Si CP-MAS NMR was used to study the modified silica support and the immobilization of the hafnocene catalyst on this support. The silicon coupling agent, (EtO)3Si(CH2)3Cp, was detected in both the monodentate and bidentate form (Fig. 23a). Two thirds of the coupling agent is in its monodentate form. Several spectral changes are visible in the BuLi activated spectrum. New signals ( 17, 3 and 14 ppm) arise due to silicons with a butyl group. A signal of the (–O)3SiBu group was also expected; however, either this group is not as abundant as expected or its signal is overlapping with the other signals. The (–O)3SiBu signal could not be assigned with certainty. The immobilization of CpHfCl3 hardly changes the spectrum (Fig. 23c) compared to the one of the BuLi activated support.

3.2.2. Other supports Jezequel et al. [94] tested a SiO2–Al2O3 mix, Al2O3 and Nb2O5 (niobia) as support materials for the zirconocenes CpnZrMe3 and Cp2ZrMe2, and compared them to silica. However, reliable NMR spectra of the Nb2O5-supported catalysts could not be obtained due to a low catalyst loading. A [SiO2–Al2O3]-supported (molar ratio: Si/Al¼ 3/1) CpnZrMe3 catalyst was studied. This support has a slightly higher Lewis acidity compared to silica and is expected to produce more electrophilic supported zirconium complexes. The NMR spectrum of the [SiO2–Al2O3]-supported catalyst was roughly the same as the spectrum of the SiO2-supported catalyst. The spectrum of the catalyst with 13C-labeled methyl groups showed that there is in fact a difference. At 4 ppm there is a broad signal, which corresponds to the methyl groups in interaction with a Lewis functionality of the surface. These methyl groups bridge between zirconium and aluminum, or in some cases the methyl group transfers to the aluminum. In both cases a positive charge on the zirconium atom is induced. Next, the Al2O3-supported catalyst was investigated. In the spectrum of the unlabeled catalyst (Fig. 24a) only the signals of the Cpn-ligand (122 and 10 ppm) are visible. The 13C-labeled supported catalyst spectrum shows three additional peaks, so

Fig. 23. 29Si CP-MAS NMR spectra of (a) (EtO)3Si(CH2)3Cp-modified silica support (S1), (b) S1 after activation with BuLi, and (c) CpHfCl3 immobilized on S1; reproduced from Ref. [95].

several different catalyst species exist on the surface of the support. The signal at 35 ppm is again assigned to methyl groups of the neutral complex (Fig. 24c, structure A). The signal at 42 ppm is of the methyl groups on the (partially) cationic structure (assignment by comparison with literature) and the peak at 11 ppm belongs to the bridging methyl group, Zr–(μ-CH3)–Al (experimentally proved). These two lines are very broad and Jezequel et al. were unable to determine exactly to which of the species B–D (Fig. 24c) they belong. Structure E was believed not to be present on the supported catalyst; however, it cannot be excluded on basis of the NMR spectra. Beck et al. [96] designed several [α-zirconium phosphonate]supports for zirconocene catalysts. They investigated the zirconocene catalysts Cp2ZrCl2 and Cp2ZrMe2 on a Zr(O3PMe)2 support with 13C CP-MAS NMR. The cyclopentadienyl ligands of both supported catalysts show a large chemical shift anisotropy. The spectra do not provide information about the mode of the metallocene attachment to the Zr(O3PMe)2 support; however, the presence of spinning sidebands indicates that the catalysts are tightly bound and there is no isotropic tumbling.



15

Fig. 25. The P–CH3 region of the 13C CP-MAS NMR spectra of (a) α-Zr(O3PMe)2, (b) ZrðO3 PMeÞ2 ðCp2 ZrMe2 Þ0:01 , and (c) ZrðO3 PMeÞ2 ðCp2 ZrCl2 Þ0:24 ; reproduced from Ref. [96].

support and the coordination of metallocenes. 1H NMR could identify the changes after modifying a silica support and showed the adsorption of zirconocene. 29Si NMR was used to study another modified silica support. The adsorption of hafnocene did not significantly change the support. The same conclusion could be drawn from 13C NMR which also indicated that the hafnocene really coordinated to the support. In some cases 13C NMR could even be used for the identification of the multiple catalyst species on a support.

Fig. 24. 13C CP-MAS NMR spectra of (a) Al2O3-supported CpnZr(CH3)3, (b) Al2O3supported CpnZr(13CH3)3, and (c) zirconocene species on the surface of the Al2O3 support; adapted from Ref. [94].

A closer look at the P-CH3 region of the 13C CP-MAS NMR spectrum of the Zr(O3PMe)2 support (Fig. 25a) reveals a doublet separated by 150 Hz, consistent with the 31P–13C J-coupling constant for phosponic acids. This doublet is covered with a fine structure that is lost when the zirconocene catalysts are attached to the surface of the support (Fig. 25b and c). This and some other experiments suggest that the fine structure is caused by methyl groups located on the edges of the microcrystalline support particles. De Camargo Forte et al. [97] have produced hybrid ZN/metallocene catalysts by immobilizing CpTiCl3, pretreated with triisobutylaluminum (Al(i-C4H9)3), on TiCl4/MgCl2. They have studied their hybrid catalysts with 27Al MAS NMR, however its spectrum is nearly identical to the one of the TiCl4/MgCl2 catalyst, so they had to conclude that the chemical surrounding of the alkyl aluminum is the same in all their catalysts. The exact conformation with which the catalyst is attached to the support remains a fundamental question. Different support materials might lead to different coordination of the metallocene and also multiple sites might exist on a single support. Modification of the support is used to yield more active catalysts. Different supports have been measured by 1H, 13C, 29Si NMR to investigate the modifications of the

3.2.3. NMR on low-γ quadrupolar nuclei The group of Schurko performed a number of studies on different low-γ quadrupolar nuclei involved in metallocenes, among which: 93 Nb, 139La, 35Cl, 47,49Ti and 91Zr (Refs. [98–103]). These publications mostly concern model systems (measured as solid powders) instead of real heterogeneous catalysts. However, the aim is to explore the possibilities of quadrupolar NMR and quantum chemical calculations. This is a crucial first step in the characterization of real heterogeneous catalyst systems. Rossini et al. [101] performed 47,49Ti solid-state NMR and quantum chemical calculations on five titanocene or related chlorides (CpTiCl3, CpnTiCl3, Cp2TiCl2, Cpn2TiCl2 and Cp2TiMe2). The latter is the most interesting and important species from a catalytic point-of-view. The CQ value of this compound is calculated to be very large (22–25 MHz). Also a large CSA (δ11 δ33 4 1800 ppm) is anticipated for Cp2TiMe2. No NMR experiments were done for this compound. The acquisition of an NMR spectrum will be rather challenging, especially in real heterogeneous catalysts with a low titanium content, and will certainly require high magnetic fields. The chloride containing titanocenes, which serve as model systems, have smaller quadrupolar couplings (1.6–5.5 MHz) and hence smaller linewidths. Both MAS and static spectra are relatively easy acquired at 21.1 T with standard echo sequences. MAS spectra were acquired in 1.5 h and CQ, ηQ , and δiso values were extracted via simulations of the 49Ti resonance, see Fig. 26. The 47Ti resonance assists in the simulations, as the parameters found for 49Ti must also fit to the 47Ti lineshape (after scaling of CQ). At lower fields (9.4 T) signal enhancement techniques are necessary (for instance the double frequency sweep quadrupolar Carr–Purcell–Meiboom–Gill (DFSQCPMG) sequence) and even so the experimental times increase to days, but quadrupolar parameters are more accurately determined for CpTiCl3 and CpnTiCl3 (CQ ¼1.6 resp. 3.0 MHz). It is observed that the mono-Cp0 species possess a higher chemical shift compared to the bisCp0 species irrespective of the substitution pattern. An increase in isotropic chemical shift is observed for the Cpn complexes compared to the analogous Cp complexes. Static experiments show a substantial influence of CSA on the titanium spectra (Fig. 27). Rossini et al. [102,103] performed 91Zr solid-state NMR on zirconocenes. 91Zr is a low-γ quadrupolar nucleus (ν0(91Zr)¼ 83.7 MHz at 21.1 T) with a natural abundance of approximately 11%. They investigated a broad range of complexes ranging from ‘simple’ Cp2ZrCl2 to the complex [Cp2ZrMe][MeB(C6F5)3] mimicking an active olefin


16


Fig. 26. Experimental MAS 47,49Ti spectra of Cp2TiCl2 (1), Cpn2TiCl2 (2), CpTiCl3 (3) and CpnTiCl3 (4) (black) at 9.4 and 21.1 T. Simulations are shown in red. The insets in spectrum 4 provide the isotropic peaks that are obtained by adding the spinning sidebands (marked with n) to the central transition; reproduced from Ref. [101]. (For interpretation of the references to color in this figure caption, the reader is referred to the web version of this paper.)

polymerization catalyst. Both quadrupolar and chemical shift interaction parameters have been extracted from the spectra using simulations and quantum mechanical calculations (Fig. 28). For a series of zirconocenes with varying halide ligands (Cl and Br ) and Cp substitution pattern relatively small quadrupolar coupling constants (2.5–6.5 MHz) were found, with bromine yielding smaller CQ values than chlorine. Consequently relatively narrow powder patterns were obtained which allowed for the relative easy acquisition of both MAS and static spectra, even at fields of 9.4 T. Both the isotropic chemical shift and the chemical shift anisotropy are strongly influenced by the coordinating heteroligands, with bromine leading to more positive δiso and with a larger anisotropy. The Zr shift also seems to be sensitive to the substitution pattern of the Cp ring. Reaction of MAO with a precatalyst (Cp2ZrCl2) leads to methylation of the zirconium atom. A dramatic increase in quadrupolar coupling constant (20–40 MHz) was observed going from the halides to methyl groups. This was also observed in the calculations for the titanocenes. An important reason for this seems to be the decreased Zr–X distance when going from halides to methyl groups. A large quadrupolar coupling constant is also found for Ind2ZrCl2 which has chloride substituents, but in which there is an asymmetry in the Zr–C bond distances of the indenyl group. These ultrawideline spectra possess additional challenges for the acquisition of 91Zr spectra with a good signal-to-noise. Standard echo sequences are no longer sufficient. Rossini et al. therefore employed the (Wideband Uniform Rate and Smooth Truncation) WURST-QCPMG sequence, however experimental times increased from minutes to hours. The zirconium nucleus has thus proven itself to be sensitive to the coordination environment concerning both the Cp0 ligands as well as the X-ligand in Cp2'ZrX2 metallocenes. However dilution of the 91Zr nucleus in real heterogeneous catalysts implies considerable challenges, especially when the quadrupolar coupling constant is in the

range of tens of MHz's. It is also of importance to choose an appropriate rotor. ZrO2 is frequently used as rotor material, giving strong and broad background signals hindering the observation of the zirconium signals of interest. Si3N4 rotors are a good alternative. One other problem that is encountered in the acquisition of 91Zr ultrawideline NMR spectra of zirconocenes is the detection of satellite transitions from 35/37Cl, which makes the interpretation of the zirconium powder pattern more difficult. Chlorine has two NMR active isotopes, 35Cl and 37Cl. Both are low-γ quadrupolar nuclei. Due to its higher natural abundance (E75% vs. 25%) 35Cl is usually the preferred isotope, despite its higher quadrupole moment. Solid-state 35Cl NMR of group IV transition metallocenes (Ti, Zr, and Hf) have been acquired by Rossini et al. [100] at 9.4 and 21.1 T. From Nuclear Quadrupole Resonance (NQR) experiments it is known that these complexes exhibit extremely high chlorine quadrupolar constants in the order of 20 MHz due to low local symmetry around the chlorine nucleus. Only static experiments were performed. Spectra were acquired in a frequency-stepped fashion using the QCPMG sequence. Long chlorine T2 allowed for the acquisition of many echoes so that experimental times of the piecewise collected spectra are reasonable. The static powder patterns have widths from 1 MHz to 2.5 MHz at 9.4 T (of which the latter corresponds to CQ ¼ 22.1 MHz), see Fig. 29. Lower quadrupole interactions are found for the Zr and Hf analogs which possess significant longer M–Cl distances compared to the titanocenes. Due to the widths it is not possible to distinguish between chemically similar, but crystallographic distinct chlorine sites. However, chlorines in chemically different sites are resolved as was seen for bridging and terminal chlorines. 35Cl NMR thus proved to be sensitive to the differences in the first coordination sphere. This also follows from the differences between Cp2TiCl2 vs. CpTiCl3 with the latter having a much smaller CQ (15 MHz). For Cp2ZrMeCl, in which the 91Zr CQ is larger compared to Cp2ZrCl2, a notably smaller 35Cl CQ is observed than for the dichloride species which might be related to an increased Zr–Cl bond. At 21.1 T a significant reduction in the linewidths is observed. Experimental time decreases to minutes, indicating that heterogeneous catalysts might be observable with 35Cl solid-state NMR in reasonable experimental time. However, the effect of chlorine CSA increases and simulations without accounting for this are poor. Again attention should be paid to the support material, which cannot be MgCl2 since this will dominate the spectrum. Hung et al. [104] also measured wideline 35Cl QCPMG spectra of metallocenes, Cp2MCl2 (M¼Zr, Ti, Mo, W). They acquired broad chlorine spectra at 19.6 and 25 T, with the latter being a resistive magnet. Their experiment is mainly methodologically driven as it intends to show that resistive magnets can be used for the acquisition of wideline spectra. The field fluctuations and spatial homogeneity of this powered magnet are much smaller (20 ppm) than the quadrupole induced broadening (several ten thousands of ppm) and are therefore negligible. Quadrupole coupling and chemical shift anisotropy parameters can readily be extracted from these spectra. Simple titanocenes and zirconocenes generally yield spectra with relatively narrow resonances. It appears that catalytically active species have considerably larger quadrupolar interactions. This complicates experiments in real heterogeneous systems in which the concentration of the active metal is much lower. However, the nuclei 35 Cl, 47,49Ti and 91Zr have proven to be sensitive to changes in the first coordination sphere and might therefore be considered as probes for the investigation of catalytic active materials.

4. Conclusions and outlook Ziegler–Natta catalysts have been around since the late 1950s. Today, more than 50 years later, a lot of work has been done to



17

Fig. 27. Experimental static 47,49Ti spectra of Cp2TiCl2 (1), Cpn2TiCl2 (2), CpTiCl3 (3) and CpnTiCl3 (4) (black) at 9.4 and 21.1 T. Spectra at 21.1 T were acquired with a standard echo sequence. The spectra of compounds 1 and 2 were acquired at 9.4 T with the DFS-QCPMG pulse sequence. Simulations are shown in red with deconvolutions of the 49Ti and 47Ti in green and blue; reproduced from Ref. [101]. (For interpretation of the references to color in this figure caption, the reader is referred to the web version of this paper.)

determine their exact structure and mechanistic operation; however, chemists have published contradicting results and have not yet reached a full, generally accepted understanding of these catalysts. Several discoveries, for instance those of the MgCl2 support, the metallocenes and the electron donor, have already drastically changed the field of ZN catalysis. Maybe now it is time for a revolution in the techniques used to study ZN catalysts. SSNMR has steadily progressed over the last decades, however the application of this technique in the field of ZN catalysis has been minimal. Thus far NMR spectroscopists have mainly studied the spin-1/2 nuclei 1H, 13C and 29Si. This gave them information about the electron donor or the (silica) support. However, this rarely gave information about the catalyst itself. It is clear that the possibilities of SSNMR in the field of ZN catalysts have not been fully explored yet. There are far more NMR active isotopes present in the catalyst systems, including 17O, 25Mg, 27Al, 35,37Cl, 47,49Ti, 91Zr and 177,179Hf. Studying these nuclei directly could give information about the active constructs. 13 C CP-MAS NMR has been shown to be useful for studying the composition, the structure of molecular adducts and the

interaction of electron donors in ZN catalysts. However, most studies focus on the isotropic carbon shift, while the full chemical shift tensor provides much more information about the coordination of electron donors to metals. In the field of ZN catalysts, 27Al is the most studied quadrupolar nucleus. Several studies tried to study the coordination state of the cocatalyst in heterogeneous ZN catalysts. However, the results of these experiments are difficult to interpret and the assignments made are not unambiguous. The experiments do not live up to today's standards, but they show the potential of the 27Al nucleus for the investigation of the coordinative state of the cocatalyst. State-of-the-art experiments (high magnetic fields available nowadays, combined with faster spinning and more advanced pulse sequences such as MQMAS) should be able to separate the different sites. Preliminary measurements on the quadrupolar nuclei in metallocene model compounds (35Cl, 47,49Ti and 91Zr) have been performed and show both the potential as well as the difficulties of studying those nuclei. The combination of low natural abundance, diluted species and broad lines hamper the straightforward observation of those nuclei in real catalyst systems. On the other hand it has been



18

8

21.1 T

shown that the spectra are sensitive to the first coordination sphere so that it might yield valuable information about the catalytic centers. A preliminary study on 47,49Ti has shown that the resonance of both isotopes can also be detected in ZN catalysts.

Acknowledgment 2000

0

200

100

ppm

-2000

0

-100

-200

We would like to thank Victor Litvinov (senior scientist DSM Resolve in Geleen, The Netherlands) for his useful comments and suggestions on the paper.

kHz

8

9.4 T

References

10000 600

0

400

200

0

ppm

-10000 -200

-600

-400

kHz

9

21.1T

4000 600

400

0 200

0

-4000 -200

ppm

-400

-600

kHz

91

Fig. 28. Static Zr SSNMR spectra of Cp2ZrMe2 (8) and [Cp2ZrMe][MeB(C6F5)3] (9). Analytical simulations (red traces) overlaid on the experimental WURST-QCPMG spectra of (8) acquired at fields of (a) 21.1 T and (b) 9.4 T. (c) Analytical simulation (red trace) overlaid on the experimental WURST-QCPMG spectrum of (9) acquired at a field of 21.1 T; reproduced from Ref. [103]. (For interpretation of the references to color in this figure caption, the reader is referred to the web version of this paper.)

Simulation Experiment

Cp 2TiCl 2

Cp 2ZrCl 2

Cp 2HfCl 2

Cp*2 ZrCl 2

CpTiCl 3

30000 1000

20000

10000 500

0

-10000

0

-500

-20000

-30000

-1000

-40000 -1500

ppm kHz

Fig. 29. 35Cl QCPMG SSNMR spectra and analytical simulations of the spectra (solid traces) for selected metallocenes. Satellite transitions are visible in the spectra of Cp2ZrCl2, Cpn2ZrCl2, and Cp2HfCl2. The asterisk in the spectrum of Cpn2ZrCl2 denotes a discontinuity of a satellite transition; reproduced from Ref. [100].

[1] D. Crespy, M. Bozonnet, M. Meier, Angew. Chem. Int. Ed. 47 (18) (2008) 3322–3328. http://dx.doi.org/10.1002/anie.200704281. [2] G.W. Coates, Chem. Rev. 100 (4) (2000) 1223–1252. http://dx.doi.org/10.1021/ cr990286u, arXiv:http://pubs.acs.org/doi/pdf/10.1021/cr990286u, http://pubs. acs.org/doi/pdf/10.1021/cr990286u. [3] K. Ziegler, E. Holzkamp, H. Breil, H. Martin, Angew. Chem. 67 (19–20) (1955) 541–547. http://dx.doi.org/10.1002/ange.19550671902. [4] G. Natta, P. Pino, P. Corradini, F. Danusso, E. Mantica, G. Mazzanti, G. Moraglio, J. Am. Chem. Soc. 77 (6) (1955) 1708–1710. http://dx.doi.org/10.1021/ ja01611a109. arXiv:http://pubs.acs.org/doi/pdf/10.1021/ja01611a109. [5] D.S. Breslow, N.R. Newburg, J. Am. Chem. Soc. 79 (18) (1957) 5072–5073. http://dx.doi.org/10.1021/ja01575a066. arXiv:http://pubs.acs.org/doi/pdf/10. 1021/ja01575a066. [6] G. Natta, P. Pino, G. Mazzanti, R. Lanzo, La Chim. L'Ind. 39 (1957) 1032–1036. [7] Montedision, British Patent 1286867, 1968. [8] J.A. Ewen, R.L. Jones, A. Razavi, J.D. Ferrara, J. Am. Chem. Soc. 110 (18) (1988) 6255–6256. http://dx.doi.org/10.1021/ja00226a056. arXiv:http://pubs.acs. org/doi/pdf/10.1021/ja00226a056. [9] R. Mülhaupt, Macromol. Chem. Phys. 204 (2) (2003) 289–327. http://dx.doi. org/10.1002/macp.200290085. [10] J. Huang, G. Rempel, Prog. Polym. Sci. 20 (3) (1995) 459–526 http://dx.doi. org/10.1016/0079-6700(94)00039-5. [11] K. Soga, T. Shiono, Prog. Polym. Sci. 22 (7) (1997) 1503–1546. doi:http://dx. doi.org/10.1016/S0079-6700(97)00003-8. [12] B. Liu, T. Nitta, H. Nakatani, M. Terano, Macromol. Chem. Phys. 204 (3) (2003) 395–402. http://dx.doi.org/10.1002/macp.200390005. [13] N. Kashiwa, Polym. J. 12 (9) (1980) 603–608. http://dx.doi.org/10.1295/ polymj.12.603. [14] K. Fukuda, B. Liu, H. Nakatani, I. Nishiyama, M. Yamahiro, M. Terano, Catal. Commun. 4 (12) (2003) 657–662. http://dx.doi.org/10.1016/j.catcom.2003.10.001. [15] J. Wang, L. Wang, H. Gao, W. Wang, W. Wang, Z. Zhao, T. Sun, L. Feng, Polym. Int. 55 (3) (2006) 299–304. http://dx.doi.org/10.1002/pi.1952. [16] P. Corradini, G. Guerra, L. Cavallo, Acc. Chem. Res. 37 (4) (2004) 231–241. http://dx.doi.org/10.1021/ar030165n. arXiv:http://pubs.acs.org/doi/pdf/10. 1021/ar030165n. [17] G. Monaco, M. Toto, G. Guerra, P. Corradini, L. Cavallo, Macromolecules 33 (24) (2000) 8953–8962. http://dx.doi.org/10.1021/ma000988h. arXiv:http:// pubs.acs.org/doi/pdf/10.1021/ma000988h. [18] G.J.P. Britovsek, V.C. Gibson, D.F. Wass, Angew. Chem. Int. Ed. 38 (4) (1999) 428–447 http://dx.doi.org/10.1002/(SICI)1521-3773(19990215)38:4o428::AID-A NIE42843.0.CO;2-3. [19] E.Y. Tshuva, I. Goldberg, M. Kol, J. Am. Chem. Soc. 122 (43) (2000) 10706–10707. http://dx.doi.org/10.1021/ja001219g. arXiv:http://pubs.acs.org/doi/pdf/10.1021/ ja001219g. [20] J. Tian, G.W. Coates, Angew. Chem. Int. Ed. 39 (20) (2000) 3626–3629 http:// dx.doi.org/10.1002/1521-3773(20001016)39:20 o3626::AID-ANIE3626 43. 0.CO;2-U. [21] J. Saito, M. Mitani, M. Onda, J.-I. Mohri, S.-I. Ishii, Y. Yoshida, T. Nakano, H. Tanaka, T. Matsugi, S.-I. Kojoh, N. Kashiwa, T. Fujita, Macromol. Rapid Commun. 22 (13) (2001) 1072–1075 http://dx.doi.org/10.1002/1521-3927 (20010901)22:13 o1072::AID-MARC1072 43.0.CO;2-7. [22] A. Lüttringhaus, W. Kullick, Angew. Chem. 70 (14) (1958) 438. http://dx.doi. org/10.1002/ange.19580701407. [23] J.A. Smith, J. von Seyerl, G. Huttner, H.H. Brintzinger, J. Organomet. Chem. 173 (2) (1979) 175–185 http://dx.doi.org/10.1016/S0022-328X(00)81287-5. [24] F.R. Wild, L. Zsolnai, G. Huttner, H.H. Brintzinger, J. Organomet. Chem. 232 (3) (1982) 233–247 http://dx.doi.org/10.1016/S0022-328X(00)89067-1. [25] F.R.W.P. Wild, M. Wasiucionek, G. Huttner, H.H. Brintzinger, J. Organomet. Chem. 288 (1) (1985) 63–67 http://dx.doi.org/10.1016/0022-328X(85)80104-2. [26] P.J. Shapiro, Coord. Chem. Rev. 231 (1–2) (2002) 67–81 http://dx.doi.org/10. 1016/S0010-8545(02)00114-5. [27] S.L.J. Conway, M.L.H. Green, S.I. Pascu, H.O. Peake, Polyhedron 25 (2) (2006) 406–420 http://dx.doi.org/10.1016/j.poly.2005.08.025. [28] J.A. Ewen, J. Am. Chem. Soc. 106 (21) (1984) 6355–6364. http://dx.doi.org/ 10.1021/ja00333a041. arXiv:http://pubs.acs.org/doi/pdf/10.1021/ja00333a041. [29] W. Kaminsky, K. Külper, H.H. Brintzinger, F.R.W.P. Wild, Angew. Chem. Int. Ed. Engl. 24 (6) (1985) 507–508. http://dx.doi.org/10.1002/anie.198505071.


K.C.H. Tijssen et al. / Solid State Nuclear Magnetic Resonance ∎ (∎∎∎∎) ∎∎∎–∎∎∎ [30] L. Resconi, L. Cavallo, A. Fait, F. Piemontesi, Chem. Rev. 100 (4) (2000) 1253–1346. http://dx.doi.org/10.1021/cr9804691. [31] G.G. Hlatky, Chem. Rev. 100 (4) (2000) 1347–1376. http://dx.doi.org/10.1021/ cr9902401. [32] J.M. Campos, J.P. Lourenço, H. Cramail, M.R. Ribeiro, Prog. Polym. Sci. 37 (12) (2012) 1764–1804 http://dx.doi.org/10.1016/j.progpolymsci.2012.02.006. [33] W. Kaminsky, Macromolecules 45 (8) (2012) 3289–3297. http://dx.doi.org/ 10.1021/ma202453u. arXiv:http://pubs.acs.org/doi/pdf/10.1021/ma202453u. [34] T. Sugano, K. Matsubara, T. Fujita, T. Takahashi, J. Mol. Catal. 82 (1) (1993) 93–101 http://dx.doi.org/10.1016/0304-5102(93)80074-5. [35] R. Hoff, R.T. Mathers, Handbook of Transition Metal Polymerization Catalysts, Wiley and Sons, New York, 2010. [36] V. Prelog, G. Helmchen, Angew. Chem. Int. Ed. Engl. 21 (8) (1982) 567–583. http://dx.doi.org/10.1002/anie.198205671. [37] M. Farina, G.D. Silvestro, P. Sozzani, Prog. Polym. Sci. 16 (2–3) (1991) 219–238 http://dx.doi.org/10.1016/0079-6700(91)90017-F. [38] G.W. Coates, R.M. Waymouth, Science 267 (5195) (1995) 217–219. http://dx. doi.org/10.1126/science.267.5195.217. [39] V. Busico, R. Cipullo, L. Caporaso, G. Angelini, A.L. Segre, J. Mol. Catal. A: Chem. 128 (1–3) (1998) 53–64 http://dx.doi.org/10.1016/S1381-1169(97)00162-3. [40] M.K. Leclerc, H.H. Brintzinger, J. Am. Chem. Soc. 118 (38) (1996) 9024–9032. http://dx.doi.org/10.1021/ja961157n. arXiv:http://pubs.acs.org/doi/pdf/10. 1021/ja961157n. [41] P. Cossee, Tetrahedron Lett. 1 (38) (1960) 12–16 http://dx.doi.org/10.1016/ S0040-4039(01)99340-2. [42] P. Cossee, Tetrahedron Lett. 1 (38) (1960) 17–21 http://dx.doi.org/10.1016/ S0040-4039(01)99341-4. [43] P. Cossee, J. Catal. 3 (1) (1964) 80–88 http://dx.doi.org/10.1016/0021-9517 (64)90095-8. [44] E.J. Arlman, P. Cossee, J. Catal. 3 (1) (1964) 99–104. http://dx.doi.org/10.1016/ 0021-9517(64)90097-1. [45] T.C. Clarke, C.S. Yannoni, T.J. Katz, J. Am. Chem. Soc. 105 (26) (1983) 7787–7789. http://dx.doi.org/10.1021/ja00364a076 arXiv:http://pubs.acs. org/doi/pdf/10.1021/ja00364a076. [46] J.C.W. Chien, Z. Yu, M.M. Marques, J.C. Flores, M.D. Rausch, J. Polym. Sci. Part A: Polym. Chem. 36 (2) (1998) 319–328 http://dx.doi.org/10.1002/(SICI) 1099-0518(19980130)36:2 o 319::AID-POLA154 3.0.CO;2-R. [47] M. Ystenes, J. Catal. 129 (2) (1991) 383–401 http://dx.doi.org/10.1016/ 0021-9517(91)90043-4. [48] R.H. Grubbs, G.W. Coates, Acc. Chem. Res. 29 (2) (1996) 85–93. http://dx.doi. org/10.1021/ar9501683. [49] W.E. Piers, J.E. Bercaw, J. Am. Chem. Soc. 112 (25) (1990) 9406–9407. http: //dx.doi.org/10.1021/ja00181a060 arXiv:http://pubs.acs.org/doi/pdf/10.1021/ ja00181a060. [50] V. Busico, R. Cipullo, Prog. Polym. Sci. 26 (3) (2001) 443–533 http://dx.doi. org/10.1016/S0079-6700(00)00046-0. [51] A.E. Tonelli, NMR Spectroscopy and Polymer Microstructure: The Conformational Connection, Wiley-VCH, New York, 1989. [52] G. Hart-Smith, Anal. Chim. Acta 808 (0) (2014) 44–55 http://dx.doi.org/10. 1016/j.aca.2013.09.033. [53] M.R.L. Paine, P.J. Barker, S.J. Blanksby, Anal. Chim. Acta 808 (0) (2014) 70–82 http://dx.doi.org/10.1016/j.aca.2013.10.001. [54] P. Rizzarelli, S. Carroccio, Anal. Chim. Acta 808 (0) (2014) 18–43 http://dx.doi. org/10.1016/j.aca.2013.11.001. [55] G. Montaudo, F. Samperi, M.S. Montaudo, Prog. Polym. Sci. 31 (3) (2006) 277–357 http://dx.doi.org/10.1016/j.progpolymsci.2005.12.001. [56] W. Bai, X. Xiao, Q. Chen, Y. Xu, S. Zheng, J. Lin, Prog. Org. Coat. 75 (3) (2012) 184–189 http://dx.doi.org/10.1016/j.porgcoat.2012.04.016. [57] R. Satapathy, H. Padhy, Y.-H. Wu, H.-C. Lin, Chem.—Eur. J. 18 (50) (2012) 16061–16072. http://dx.doi.org/10.1002/chem.201201437. [58] A. Baumgaertel, E. Altuntaş, U.S. Schubert, J. Chromatogr. A 1240 (0) (2012) 1–20 http://dx.doi.org/10.1016/j.chroma.2012.03.038. [59] Y. Liu, Z. Wang, X. Zhang, Chem. Soc. Rev. 41 (2012) 5922–5932. http://dx.doi. org/10.1039/C2CS35084J. [60] A. Salleo, R.J. Kline, D.M. DeLongchamp, M.L. Chabinyc, Adv. Mater. 22 (34) (2010) 3812–3838. http://dx.doi.org/10.1002/adma.200903712. [61] L. Resconi, I. Camurati, O. Sudmeijer, Top. Catal. 7 (1–4) (1999) 145–163. http: //dx.doi.org/10.1023/A:1019115801193. [62] H. Hagihara, T. Shiono, T. Ikeda, Macromolecules 31 (10) (1998) 3184–3188. http://dx.doi.org/10.1021/ma971697k. arXiv:http://pubs.acs.org/doi/pdf/10. 1021/ma971697k. [63] J.D. Scollard, D.H. McConville, J. Am. Chem. Soc. 118 (41) (1996) 10008–10009. http://dx.doi.org/10.1021/ja9618964. [64] A. Bertrand, F. Lortie, J. Bernard, Macromol. Rapid Commun. 33 (24) (2012) 2062–2091. http://dx.doi.org/10.1002/marc.201200508. [65] K. Tangjituabun, S.Y. Kim, Y. Hiraoka, T. Taniike, M. Terano, B. Jongsomjit, P. Praserthdam, Sci. Technol. Adv. Mater. 9 (2) (2008) 024402. [66] K. Kallio, A. Wartmann, K.-H. Reichert, Macromol. Rapid Commun. 23 (3) (2002) 187–190 http://dx.doi.org/10.1002/1521-3927(20020201)23:3 o 187:: AID-MARC18743.0.CO;2-Y. [67] M. Vittadello, P.E. Stallworth, F.M. Alamgir, S. Suarez, S. Abbrent, C.M. Drain, V.D. Noto, S.G. Greenbaum, Inorg. Chim. Acta 359 (8) (2006) 2513–2518 http://dx.doi.org/10.1016/j.ica.2006.01.044. [68] K.S. Thushara, E.S. Gnanakumar, R. Mathew, R.K. Jha, T.G. Ajithkumar, P. R. Rajamohanan, K. Sarma, S. Padmanabhan, S. Bhaduri, C.S. Gopinath, J. Phys.

[69]

[70] [71]

[72]

[73] [74] [75] [76] [77] [78] [79] [80]

[81] [82]

[83]

[84] [85]

[86] [87]

[88]

[89] [90]

[91] [92]

[93] [94]

[95] [96] [97]

[98]

[99]

[100]

19

Chem. C 115 (5) (2011) 1952–1960. http://dx.doi.org/10.1021/jp1078289 arXiv: http://pubs.acs.org/doi/pdf/10.1021/jp1078289. R. Ohashi, M. Saito, T. Fujita, T. Nakai, H. Utsumi, K. Deguchi, M. Tansho, T. Shimizu, Chem. Lett. 41 (12) (2012) 1563–1565. http://dx.doi.org/10.1246/ cl.2012.1563. T. Garoff, M. Väänänen, A. Root, M. Stålhammar, K. Pesonen, Eur. Polym. J. 43 (12) (2007) 5055–5064 http://dx.doi.org/10.1016/j.eurpolymj.2007.09.011. K.E. Johnston, C.A. O'Keefe, R.M. Gauvin, J. Trébosc, L. Delevoye, J.-P. Amoureux, N. Popoff, M. Taoufik, K. Oudatchin, R.W. Schurko, Chem.—Eur. J. 19 (37) (2013) 12396–12414. http://dx.doi.org/10.1002/chem.201301268. B. Liu, K. Fukuda, H. Nakatani, I. Nishiyama, M. Yamahiro, M. Terano, J. Mol. Catal. A: Chem. 219 (2) (2004) 363–370 http://dx.doi.org/10.1016/j.molcata. 2004.05.028. A.P.M. Kentgens, Geoderma 80 (3–4) (1997) 271–306 http://dx.doi.org/10. 1016/S0016-7061(97)00056-6. A.G. Potapov, V.V. Terskikh, G.D. Bukatov, V.A. Zakharov, J. Mol. Catal. A: Chem. 122 (1) (1997) 61–65 http://dx.doi.org/10.1016/S1381-1169(97)00017-4. A.G. Potapov, V.V. Terskikh, V.A. Zakharov, G.D. Bukatov, J. Mol. Catal. A: Chem. 145 (1–2) (1999) 147–152 http://dx.doi.org/10.1016/S1381-1169(99)00024-2. A.G. Potapov, V.V. Terskikh, G.D. Bukatov, V.A. Zakharov, J. Mol. Catal. A: Chem. 158 (1) (2000) 457–460 http://dx.doi.org/10.1016/S1381-1169(00)00124-2. A.T.M.K. Hasan, Y. Fang, B. Liu, M. Terano, Polymer 51 (16) (2010) 3627–3635 http://dx.doi.org/10.1016/j.polymer.2010.05.053. T.O. Sandland, L.-S. Du, J.F. Stebbins, J.D. Webster, Geochim. Cosmochim. Acta 68 (24) (2004) 5059–5069 http://dx.doi.org/10.1016/j.gca.2004.07.017. W. Xia, B. Liu, Y. Fang, T. Fujitani, T. Taniike, M. Terano, Appl. Catal.: Gen. 389 (1–2) (2010) 186–194 http://dx.doi.org/10.1016/j.apcata.2010.09.023. P. Sozzani, S. Bracco, A. Comotti, R. Simonutti, I. Camurati, J. Am. Chem. Soc. 125 (42) (2003) 12881–12893. http://dx.doi.org/10.1021/ja034630n arXiv: http://pubs.acs.org/doi/pdf/10.1021/ja034630n. K.S. Thushara, T.G. Ajithkumar, P.R. Rajamohanan, C.S. Gopinath, Appl. Catal. A: Gen. 469 (0) (2014) 267–274 http://dx.doi.org/10.1016/j.apcata.2013.10.002. E.S. Gnanakumar, R.R. Gowda, S. Kunjir, T.G. Ajithkumar, P.R. Rajamohanan, D. Chakraborty, C.S. Gopinath, ACS Catal. 3 (3) (2013) 303–311. http://dx.doi. org/10.1021/cs300730j arXiv:http://pubs.acs.org/doi/pdf/10.1021/cs300730j. H. Heikkinen, T. Liitiä, V. Virkkunen, T. Leinonen, T. Helaja, P. Denifl, Solid State Nucl. Magn. Reson. 43–44 (0) (2012) 36–41 http://dx.doi.org/10.1016/j. ssnmr.2012.02.006. P. Sormunen, T. Hjertberg, E. Iiskola, Die Makromol. Chem. 191 (11) (1990) 2663–2673. http://dx.doi.org/10.1002/macp.1990.021911115. V. Busico, M. Causà, R. Cipullo, R. Credendino, F. Cutillo, N. Friederichs, R. Lamanna, A. Segre, V. Van Axel Castelli, J. Phys. Chem. C 112 (4) (2008) 1081–1089. http://dx.doi.org/10.1021/jp076679b arXiv:http://pubs.acs.org/doi/ pdf/10.1021/jp076679b. T.T. Pakkanen, E. Vahasarja, T.A. Pakkanen, E. Iiskola, P. Sormunen, J. Catal. 121 (2) (1990) 248–261 http://dx.doi.org/10.1016/0021-9517(90)90235-C. R. Powers, G.M. Clore, S.J. Stahl, P.T. Wingfield, A. Gronenborn, Biochemistry 31 (38) (1992) 9150–9157. http://dx.doi.org/10.1021/bi00153a006 arXiv: http://pubs.acs.org/doi/pdf/10.1021/bi00153a006. A.D. Meltzer, D.A. Tirrell, A.A. Jones, P.T. Inglefield, D.M. Hedstrand, D.A. Tomalia, Macromolecules 25 (18) (1992) 4541–4548. http://dx.doi.org/10.1021/ ma00044a013. arXiv:http://pubs.acs.org/doi/pdf/10.1021/ma00044a013. J.C. Vizzini, J.-F. Shi, J.C.W. Chien, J. Appl. Polym. Sci. 48 (12) (1993) 2173–2179. http://dx.doi.org/10.1002/app.1993.070481211. H. Mori, H. Kono, M. Terano, A. Nosov, V.A. Zakharov, Macromol. Rapid Commun. 20 (10) (1999) 536–540 http://dx.doi.org/10.1002/(SICI)1521-3927 (19991001)20:10o 536::AID-MARC536 4 3.0.CO;2-Z. H. Mori, H. Kono, M. Terano, A. Nosov, V.A. Zakharov, J. Mol. Catal. A: Chem. 164 (1–2) (2000) 235–243 http://dx.doi.org/10.1016/S1381-1169(00)00282-X. B.L. Moroz, N.V. Semikolenova, A.V. Nosov, V.A. Zakharov, S. Nagy, N. J. O'Reilly, J. Mol. Catal. A: Chem. 130 (1–2) (1998) 121–129 http://dx.doi. org/10.1016/S1381-1169(97)00206-9. M. Kröger-Laukkanen, M. Peussa, M. Leskelä, L. Niinistö, Appl. Surf. Sci. 183 (3–4) (2001) 290–300 http://dx.doi.org/10.1016/S0169-4332(01)00573-6. M. Jezequel, V. Dufaud, M.J. Ruiz-Garcia, F. Carrillo-Hermosilla, U. Neugebauer, G.P. Niccolai, F. Lefebvre, F. Bayard, J. Corker, S. Fiddy, J. Evans, J.-P. Broyer, J. Malinge, J.-M. Basset, J. Am. Chem. Soc. 123 (15) (2001) 3520–3540. http://dx.doi.org/10.1021/ja000682q arXiv:http://pubs. acs.org/doi/pdf/10.1021/ja000682q. A.-M. Uusitalo, T.T. Pakkanen, E.I. Iskola, J. Mol. Catal. A: Chem. 177 (2) (2002) 179–194 http://dx.doi.org/10.1016/S1381-1169(01)00274-6. S. Beck, A.R. Brough, M. Bochmann, J. Mol. Catal. A: Chem. 220 (2) (2004) 275–284 http://dx.doi.org/10.1016/j.molcata.2004.06.011. M.M. de Camargo Forte, F.V. da Cunha, J.H.Z. dos Santos, J. Mol. Catal. A: Chem. 175 (1–2) (2001) 91–103 http://dx.doi.org/10.1016/S1381-1169(01) 00172-8. H. Hamaed, A.Y.H. Lo, D.S. Lee, W.J. Evans, R.W. Schurko, J. Am. Chem. Soc. 128 (39) (2006) 12638–12639. http://dx.doi.org/10.1021/ja0645180. arXiv: http://pubs.acs.org/doi/pdf/10.1021/ja0645180. A.Y.H. Lo, T.E. Bitterwolf, C.L.B. Macdonald, R.W. Schurko, J. Phys. Chem. A 109 (32) (2005) 7073–7087. http://dx.doi.org/10.1021/jp0521499 arXiv:http:// pubs.acs.org/doi/pdf/10.1021/jp0521499. A.J. Rossini, R.W. Mills, G.A. Briscoe, E.L. Norton, S.J. Geier, I. Hung, S. Zheng, J. Autschbach, R.W. Schurko, J. Am. Chem. Soc. 131 (9) (2009) 3317–3330. http://dx.doi.org/10.1021/ja808390a. arXiv:http://pubs.acs.org/doi/pdf/10. 1021/ja808390a.


20


[101] A.J. Rossini, I. Hung, R.W. Schurko, J. Phys. Chem. Lett. 1 (20) (2010) 2989–2998. http://dx.doi.org/10.1021/jz1012017 arXiv:http://pubs.acs.org/doi/pdf/10.1021/ jz1012017. [102] I. Hung, R.W. Schurko, J. Phys. Chem. B 108 (26) (2004) 9060–9069. http://dx. doi.org/10.1021/jp040270u.

[103] A.J. Rossini, I. Hung, S.A. Johnson, C. Slebodnick, M. Mensch, P.A. Deck, R. W. Schurko, J. Am. Chem. Soc. 132 (51) (2010) 18301–18317. http://dx.doi. org/10.1021/ja107749b. arXiv:http://pubs.acs.org/doi/pdf/10.1021/ja107749b. [104] I. Hung, K. Shetty, P.D. Ellis, W.W. Brey, Z. Gan, Solid State Nucl. Magn. Reson. 36 (4) (2009) 159–163 http://dx.doi.org/10.1016/j.ssnmr.2009.10.001.


Heterogenization of metallocene catalysts for alkene polymerization.

The Influence of Ziegler-Natta and Metallocene Catalysts on Polyolefin Structure, Properties, and Processing Ability.

NMR studies of metabolism.

Metallocene Based Polyolefin Nanocomposites.

NMR studies of ligand binding.

NMR studies of prebiotic polypertides.

NMR studies of membrane proteins.

NMR studies of retinal proteins.

Metabolic transient studies by NMR.

NMR studies of nucleic acid dynamics.

Multinuclear NMR studies of the trp-repressor.

Three-dimensional heteronuclear NMR studies of RNA.

13C NMR studies of bacterial fermentations.

NMR studies of fluorinated visual pigment analogs.

NMR studies of dynamic biomolecular conformational ensembles.

1H-NMR assignments and conformational studies of melanin concentrating hormone in water using two-dimensional NMR.

1H-, 13C- and 31P-NMR studies of dioctanoylphosphatidylcholine and dioctanoylthiophosphatidylcholine.

NMR studies of structure and dynamics of isotope enriched proteins.

19F NMR studies of enflurane elimination and metabolism.

Protein-Inhibitor Interaction Studies Using NMR.

H-NMR studies of native and fragmented subtilisin Carlsberg.

NMR and CD studies of triple-helical peptides.

1H NMR studies on lanthanides substituted transferrins.

19F NMR studies on 8-fluoroflavins and 8-fluoro flavoproteins.