New technique of sample preparation for the morphological investigation of Ziegler-Natta supports and catalysts


Marco Casinelli1

1. LyondellBasell Research Center G.Natta, P.le Donegani 12 44122 Ferrara, Italy

Marco Casinelli is an experienced specialist in the field of optical and electron microscopy. He studied morphology in the R&D department at the Research Center G. Natta of Lyondellbasell in Ferrara. He has twenty years of experience which has allowed him to perfect several techniques of observation and studies related to the polymeric materials science, as well as the customization of detectors and the research on improved analytical settings. His skills include the use of SEM (scanning electron microscope), EDS (energy dispersive X-ray spectroscopy), TEM (transmission electron microscope), OM (optical microscopy), AFM (atomic force microscopy), OCA (optical contact angle), XRM (X-ray microscopy) and various preparation techniques in the field of polymers.
Ziegler-Natta catalysts, discovered in 1954 by Karl Ziegler and Giulio Natta (who received the Nobel prize in chemistry in 1963), are the basis for the industrial production of polyolefins. Nowadays, the complete control of morphology from support to catalyst and finally within the growing polymer particles, is a key consideration in leading industrial polymerization processes. This paper presents a morphological study based on an interdisciplinary approach involving different microscopy techniques, resulting from the development of a new sample preparation procedure; which in combination with complementary information has enabled a detailed understanding of the morphology observed. 
I would like to thank Fiorella Pradella, Massimo Gnani, Testoni Fabio for proofreading and granting me permission to publish this paper, Roberta Selleri for supporting me in this endeavor, Fabrizio Piemontesi for collaborating in the revision of some results and Holger Blank (Carl Zeiss Microscopy GMBH) for performing the XRM trials.
Corresponding author
Marco Casinelli, Technical Expert AA/R&D, LyondellBasell Research Center G.Natta, P.le Donegani 12 44122 Ferrara (IT)
Good control of the native Polymer particles in Ziegler-Natta polymerization industrial processes can be obtained by a detailed morphological comprehension of the replica phenomena.1,2
An investigation using ESEM (operating in variable pressure, low vacuum mode) and AFM microscopy techniques, applied to the sample chain comprising support -> catalyst -> polymer,3 allows a connection between the final polymer porosity obtained and the original support, highlighting differences present due to the typical micro globule network formed during the reaction process.4,5 A further interpolation of the sample images using image analysis software generates semi-quantitative models of the morphological behaviour of different support and catalyst types. 
For a greater understanding of polymer growth there is a need to develop a technique that enables an investigation of specific surface. The concept of specific surface takes on a different meaning depending on whether it refers to a porous or non-porous solid.
For non-porous solids it represents the external surface, whereas for a porous solid it consists of an external surface comparable to that of a non-porous solid plus an internal surface that is significantly greater than the former, which consists of the surface of the pores. The latter is usually of interest in catalysis since it has been found that the porosity of the mass (bulk) and therefore, the overall surface area, play a vital part in determining the catalytic activity of a given material. 
It is important to note that the term ‘internal surface’ is limited to cavities which have an opening that is external to the grain (and includes the walls of all cracks, pores and cavities which are deeper than they are wide). These systems display a heterogeneity of distribution of the dimensions of the pores, which can be classified as: micropores with a width that is less than 20 Angstrom, meso-pores with a width of between 20 and 500 Angstrom and macro-pores that are greater than 500 Angstrom. 
The growth of porous particles (polyolefins) supported by catalysis may be divided into 2 main stages.4 From a theoretical view point, there is a first stage in which the porous magnesium chloride particle, the support for the catalytic sites, is gradually destroyed in small fragments due to the production of polymer in the restricted space of the pores. At the same time, the overall shape of the particle is maintained by the nascent polymer acting as a glue (a cement).
In the second stage, the growth of porous particles of polymer is due to the formation of polymer around the fragmented and localized (immobilized) active catalyzing sites on the surface. In accordance with the “multigrain model” concept, the catalyzer grain (macroparticle) consists of many smaller fragments (microparticles) disseminated in the bulk of the catalyzer. These microparticles are accessible by the monomer and the polymerization takes place on their surface with the consequent formation of a shell (micro-globule) of polymer.
The supports and catalyst must necessarily be processed in inert atmospheres (dry box with N2 gas), since they are susceptible to the external atmosphere and deteriorate quickly upon contact with oxygen and humidity (liquescence – highly hygroscopic). 
This paper describes an alternative technique to possible instruments such as FIB, array tomography in situ SEM, that might allow an investigation into the aspects related to the internal morphologies of porous materials, processed in a controlled atmosphere. To this end, an ultramicrotome and a transfer system (sample transfer) for analysis by scanning electronic microscope (ESEM) was customized. During the study, we used a low-viscosity epoxy resin which allowed us to fill all the cavities with an average radius dimension of the 200nm pore (meso-macro). The use of resin with a low viscosity together with low interfacial tension allowed faster wettability of the system to be inserted at ambient pressure (101 kPa). The speed of the absorption of the sample depends on the viscosity of the resin, translated as greater wettability of the specific surface before it reaches the “gel point” (consequent change of viscosity and loss of fluidity). 
Table 1. Plot 1 Log-linear plot of skirt radius rs as function of E0 (primary beam energy).
The materials used in this study are supports produced in our industrial plants that differ in terms of ethanol content (MgCl2·nEtOH), in association with the catalyst (cat = MgCl2/TiCl4) derived from one of the selected supports.
The study of morphologies was performed using Environmental Scanning Electron Microscopy (ESEM), Thermo Fisher Scientific FEI 200 Quanta FEG. All images included in this paper have been acquired via BSE (backscattered electron). The section linked to analytical complementarity is assessed via AFM (multimode V-Bruker).
An additional cross-check test on the preparation was performed through X-ray microscopy analysis using Zeiss XRadia 520 Versa. The resin used in the study was a low viscosity bicomponent “Epofix” in a cold state, which is particularly suitable for impregnating porous samples which require low shrinkage and good rigidity in the hardened state.
In order to be able to process the samples, two customized tools were built that allowed a conservative investigation of the materials under examination to be performed: sample transfer from the preparatory environment to the ESEM chamber and modification of the ultramicrotome (Leica UC6) in order to section the surface to be analyzed. The transfer system (tPod – figure 1) was designed based on the requirement to obtain a tool that would preserve the characteristics of miniaturization, handling and ease of use.
Figure 1. tPod. 1) sample holder, 2) cap lock.
The sample transfer is a type of pod which can contain up to 3 sample holders, which will be isolated from the external environment through the use of a cap lock. Once the sample is placed on the sample holder, the cap lock is inserted in the appropriate location and a vacuum of about 90 kPa is created inside the chamber, which is sufficient to keep the pod isolated/closed during transportation through the various stages of preparation. 
The cap lock may be opened when needed by opening a tightening valve. Once the preparation is finished, the sample transfer is positioned inside the microscope chamber. The lid is released once the microscope reaches the vacuum in the chamber (10-5 Pa). 
Lastly, the cap lock can be extracted using a magnetic slider arm flanged to the microscope chamber (figure 2). The cutting system (ultramicrotome) was customized by building a dry-box (figure 3). The isolation of the cutting equipment allowed the safe handling of the samples, thereby avoiding the morphological deterioration of the section surfaces of the support/cat. The blade used for the preparation is a diamond knife with 45° cutting angle. Below are the various steps of the preparation procedure:
•Embedding in resin of approx. 30 mg of a sample in a silicone mold (dimensions 14 mm l x 5 mm w x 4 mm). Complete polymerization after 12 hours (ambient temperature).
•Portioning of a part of the composite (resin + support/cat) using a manual saw, and gluing to the sample holder.
•Sample transportation (using sample holder), inside the dry-box containing the ultramicrotome
•Preparation of the cutting surface (shape dimensions 2 x 1 mm). To avoid chattering, the overall height of the block will not exceed the longest side of the base relative to the surface of the block.
•The cutting parameters must be set based on the composite to be prepared. For support/cat samples, a cutting speed of 2 mm/s may be used, together with an advancement step (section thickness) equal to 2-3 um.
•Re-load the sample in the sample transfer and recreate the vacuum to tighten the cap lock.
Figure 2. Magnetic transfer arm.
Figure 3. Ultramicrotome (Leica UC6) customized in a dry box in preparation for cutting; this configuration avoids sample degradation.
The objective is to process samples without the aid of a surface coating (e.g. Au/Pd, C), with a view to differentiate the structure using compositional contrast (Cz - compositional contrast); BSE image. 
The morphological study of internal macro-porosity, in both samples, was assessed using low vacuum mode (30 Pa) and Helium gas (He) as a precursor inside the electronic microscope chamber (ESEM). The gas was chosen by assessing the radius of the skirt produced by the interaction between the electrons of the primary beam and the atoms/molecules of the gas in the chamber. In accordance with Danilatos,7 the radius of the skirt Rs can be described analytically by the following expression:
Where Rs is the radius of the skirt, Z is the atomic number of the gas, E is the incident energy of the beam (kV), P is the pressure (Pa), T is the temperature (K) and L is the gas path length (m). The equation described above was plotted on a log-linear graphic, on which it is possible to evaluate the radius of the skirt variation based on primary electron beam E0, assuming that: L = 2 mm, P = 100 Pa and T = 293 K.
The scattering radius is directly proportional to the atomic number of the gas used. In addition, increasing the energy of the primary beam (E0) significantly reduces the distance over which the electrons are scattered.6
The Helium (He) produced the correct quantity of positive ions that counterbalance the negative charge placed on the surface of the sample using the primary electron beam. The radius Rs, with an acceleration tension of 10 kV at a distance of L = 10 mm and pressure of P = 30 Pa, shall equal 23um with a percentage of electrons in focused probe fp/% = 94 (fp/% = e-(L/l) x 100). These values are considered optimal to obtain BSE images with an adequate contrast.
The compositional differences between the epoxy resin and the nature of the samples analyzed (MgCl2/ TiCl4) are fundamental to be able to quickly distinguish between the different phases or the generation of element distribution maps. Based on the equation: Cz = (η2-η1)/η2, Cz expresses the contrast based on atomic number (Z), η1 coefficient of background backscattering, η2 coefficient of backscattering of the sample area to be analyzed. The relationship between Z and η was experimentally determined for pure elements (Z>~10)7: ηlnZ/6)-(1/4). 
This expression was subsequently reconsidered for non-pure composites (Heinrich, 1966) as a background coefficient of a homogenous mixture on the following atomic scale: η=∑iCiηi. (i any constituent, Ci the concentration in terms of weight of the individual constituents and ηi the backscattered coefficient of the element).8 The result obtained is clearly visible (figure 4), where the brightness can be correlated to the morphology of the support (figure 4a) in contrast with the areas of porosity filled by resin (dark region). 
Figure 4. Comparison of support (A) vs catalyst (B) particles. Well defined the differences in term of pores distribution. The sample’s morphology is resolved from brightness contrast.
A comparison of the images of the support and catalyst (figures 4 a-b) clearly shows that there is a decrease in the average radius of the pore in the catalytic system. To confirm the effectiveness of the preparation regarding the preservation of the porous morphology, the support sample using X-ray microscopy (XRM) has been assessed. The sample, subsumed in resin beforehand, is rotated relative to an X-ray source (cone beam) and multiple detector objectives (two-stage magnification process using both geometric magnification and optical magnification).
The acquired images are 2D-projections and are collected at discrete angles based on the set rotation steps. The set of 2D-projections is reconstructed to a tomographic data set, using “filtered back projection”. The images clearly show the embedded support particles (figure 5 a). By restricting the area of observation to a single particle (figure 5 b), it is possible to see the areas of pores not filled by resin (dark region). 
Figure 5. Particles of support embedded in epoxy resin. The sample can be imaged with significant contrast between material and resin. Voxel size adopted 0.5 um. A) 3D X-ray Microscopy (XRM) image, volume rendered, B) Detail view of virtual slice, black color  represents empty voids.
The contrast, in gray-scale, derives from a different X-ray absorption and attenuation compared to the bulk of the sample. On a preliminary basis, the empty regions (dark area) can be associated with macro pores present in the particle. The dynamic range on the sample is homogeneous due to a uniform X-ray absorption, in part due to the filling of the resin inserted into the cavities (modelled on a network of open pores). One of the aspects being studied involves the correct evaluation of the voxel size, which can be correlated to the maximum obtainable resolution of porosity (radius of the pore).
The BSE images obtained can be used as models to perform quantitative measurements, through the opportune segmentation of the more important regions (figures 6 a-b). Among the quantitative morphological evaluations of interest, it is possible to determine the volumetric fraction (voids area/total area), the interfacial area (perimeter/area) and the fractal dimension (Hausdorff irregular grade of object d = lgN/lgK).
Figure 6. Image analysis of the particles with voids segmentation A) support, B) catalyst.  C) Area distribution of voids in the support, D) Area distribution of voids in the catalyst.
Using the Anasys software for image analysis (by Olympus), the voids segmentation was evaluated for purely explanatory purposes. The values found have been placed on chart and shown as a histogram of frequencies for a certain interval of area, with related cumulative frequency (figures 6 c-d).
The evaluation highlights an increase in the frequency of the area of low dimensional voids in the catalyst sample compared to the support sample, which can be correlated to a decrease in the average pore radius.  
As a result of the conservative aspects of the preparation, it was possible to perform an observation of the same areas on the same samples investigated by SEM using the AFM technique. The instrument used in tapping mode (collected as amplitude and phase data), permitted the identification of a structure organized into variable dimension nano-particles (figures 7 a-b) for the first time.  To perform the analysis, the instrument was opportunely isolated inside a dry-box with a constant flow of N2 gas. 
Figure 7. AFM in tapping mode (phase data). In both samples a globular morphology probably derived from the unit element of the crystalline structure A is visible) The nano-particles in support are estimated 30nm, B) The nano-particles in catalyst are estimated 60nm.
This preparatory technique could help in achieving a more in-depth knowledge of internal morphologies relating to the porous micro particle (PSD 50-100um) which must be processed in an inert atmosphere. 
The use of the preparation and the various techniques of microscopy, applied to Ziegler-Natta supports and catalysts, have resulted in a greater understanding at the nano/micrometric level. In fact, interesting structures have been revealed by applying ESEM and AFM analyses on real spherical samples. The investigation of catalyst support primitive morphology was possible thanks to a conservative approach used in the sample preparation procedure involving a low viscosity epoxy resin in association with ESEM analysis in low vacuum mode (with helium gas as a precursor).
By using atomic contrast (Cz) which is obtainable with an ESEM backscattered detector (BSD), the sectioned surfaces of catalyst support highlighted the internal porosity of the samples whilst offering clearly contrasted images for image analysis interpolation. The porosity network information revealed by ESEM in association with the EDS (data not show in this paper) resulted in a powerful tool for a more detailed understanding of the growth mechanism in the industrial process from support to catalyst.
Owing to a conservative sample preparation, the same real catalyst support particles observed by ESEM were also investigated by AFM in tapping mode, revealing a peculiar morphology with typical grain structure at nanometric dimensions; subsequently the granular structure morphology was also confirmed by TEM analyses.3
1. P. Blais, R. St. J. Manley, Journal of Polymer Science. Part A-1(1968), 6 p.291-334.
2. A. Munoz-Escalona, C. Villamizar and P. Frias, Polymer Science and Technology (1982), p. 95-113.
3. R. Selleri and M. Casinelli, MC2013 Regensburg, MS.1.P010.
4. G. Cecchin, E. Marchetti and G. Baruzzi, Macromol.Chem.Phys.(2001),202 p.1987-1994.
5. J. Kosek, B. Horackova Basell Symposium, (2006), p.65-81.
6. D.J. Stokes (2008), Principles and practice of Variable Pressure/Environmental SEM (VP-SEM), p.126-130.
7. Heinrich, 1964; Bishop, 1966 and 1974
8. J. Goldstein, D.E. Newbury, D. C. Joy, Charles E. Lyman, P. Echlin, E. Lifshin, L. Sawyer, J.R. Michael, Scanning Electron Microscopy and X-ray Microanalysis: Third Edition. 
Website developed by S8080 Digital Media