Analysis of networks based on styrene and divinylbenzene containing iron anchored using variable pressure scanning electron microscopy

Authors


L. C. de Santa Maria. Fax: +55 21 2587 7227; e-mail: lcsm@uerj.br

Summary

There is great demand for the development of composite materials containing small metal or metal oxides particles, owing to their variable properties and wide application. However, microscopic evaluation of these materials using high-vacuum scanning electron microscopy is difficult because the samples must undergo a series of preparation steps to reach a high image quality and to avoid becoming shrunk inside the microscope vacuum chamber. Thus, in this study, we used variable pressure scanning electron microscopy to evaluate the morphology and iron distribution on the surface of magnetic microspheres based on poly(styrene-co-divinylbenzene). These materials were obtained by suspension copolymerization of styrene and divinylbenzene in the presence of fine iron particles. Energy-dispersive X-rays were also used to analyse distribution of the iron particles. The results indicate that, under the conditions used, magnetic microspheres with a relatively narrow size distribution were formed. Moreover, the micrographs show that agglomerated iron particles appeared only on the microsphere surface.

Introduction

Composite materials containing small metal particles have attracted a great deal of attention because of their interesting chemical and physical properties. These materials have potential technological applications due to their possible use in the development of new optical and electro-optical elements (filters, mode-locking devices, switches, etc.), cell labelling, cell separation, enzyme immunoassay, target drug, etc. (Robinson et al., 1973; Ugelstad et al., 1992; Ayers et al., 1996; Kroll et al., 1996; Chen et al., 1997; Ramesh et al., 1997; Ding et al., 1998; Gurin et al., 1998; Ely et al., 1999; Horak et al., 2000; Horak, 2001). Moreover, considering their magnetic properties, they can be separated relatively rapidly and easily, and require simple apparatus compared with centrifugal separation (Ding et al., 1998). Many different techniques have been developed for the preparation of magnetic particles (Robinson et al., 1973; Ugelstad et al., 1992; Gurin et al., 1998; Sun et al., 1999; Horak et al., 2000; Horak, 2001).

Network polymers based on styrene (STY) and divinylbenzene (DVB), and prepared using the well-developed suspension polymerization process, have many advantages for a microcomposite host matrix. For network polymers, this includes ease of preparation, controlled surface area and porous structure, good thermal resistance and a reasonable chemical inertness. However, contamination by remnants of the suspension agents, diluents, monomer, etc. used during the polymerization process may introduce experimental limitations for certain guest compositions and some composite properties.

The chemical and physical properties of network polymer composites containing metal particles depend on the size and distribution of the particles and on the presence (or otherwise) of interactions among their surface, matrix and contaminants (Ely et al., 1999). Therefore, this requires control of the size and distribution of the particles on the host material. These features are easily observed using microscopy, especially if variable pressure scanning electron microscopy (VP-SEM) can be used. Use of this kind of environment brings several advantages to the study of composites. There is no damage to organic materials because they are no longer exposed to high vacuums; moreover, preparation of the samples is simplified. Thus, uncoated and low conductive specimens can be imaged by the interactions between the electron beam and air molecules in the chamber, producing positive ions, and neutralizing excess charge at the sample surface. Also, the microscope can use accelerated voltages during imaging acquisition, which favours the performance of X-ray microanalysis (Mathieu, 1996).

In this respect, to the best of our knowledge, the VP-SEM has not been applied to evaluate the morphology of iron/network composite materials and the distribution of the iron particles. Therefore, this study was aimed at the morphological characterization of network poly(styrene-co-divinylbenzene) microspheres containing iron particles dispersed using VP-SEM. Energy dispersive X-ray analysis (EDX) was also performed in this mode of operation. The images were then digitally processed and the iron distribution on the spheres surface could be determined.

Materials and methods

Preparation and characterization of iron composite microspheres

The synthesis and characterization of iron composites have been described previously (Santa Maria et al., 2003a,b).

Microscopy characterization

The microscopy characterization of the samples was carried out to observe variations in microstructure due to the different preparation procedures used (Santa Maria et al., 2003a). Samples were dispersed on a conductive tape and analysed directly using a VP-SEM LEO 1450 VP without being covered with any conductive coating. The beads were observed using both variable pressure secondary and backscattered electrons modes (VPSE and BSE, respectively) at several magnifications and a minimum chamber pressure of 100 Pa. VPSE images were produced with accelerate voltages of 15–20 kV, whereas the BSE images were obtained with 20–30 kV (Four Quadrant Backscattered Electron Detector Type 222). EDX maps were also acquired to detect the distribution of iron particles on the beads. Coated and uncoated samples were analysed. Both the microspheres and the dispersed iron particles sizes were evaluated by imaging processing procedures, using the software ks 400 Version 3.0 (Zeiss). The subroutine used for the analysis of the microspheres is described in Fig. 1, and the one employed to estimate the size of the iron particles is described in Fig. 2 (Gonzalez & Woods, 1993).

Figure 1.

Subroutine used to analyse the diameter of the microspheres: (a) original image; (b) contrast enhancement; (c) high pass filter to enhance the edges; (d) segmentation of the image; (e) scrapping of the binary image (< 25 pixels); (f) dilation (Gonzalez & Wood, 1993). Original image was taken using a VPSE detector.

Figure 2.

Subroutine used to analyse the diameter of the iron particles: (a) original image; (b) contrast enhancement; (c) high pass filter to enhance the edges; (d) segmentation of the image; (e) scrapping of the binary image (< 30 pixels); (Gonzalez & Wood, 1993). Original image was taken using a BSE detector.

Results and discussion

In a former study, we reported the synthesis of iron composites based on STY/DVB copolymer networks produced by suspension copolymerizations in the presence of different organic solvents as porogenic agents (Santa Maria et al., 2003a). Table 1 summarizes the chemical composition and characteristics of the composites (Santa Maria et al., 2003b). These previous results showed that the suspension copolymerization technique was suitable for producing networks with dispersed iron particles.

Table 1.  Characterization of the composite microspheres (Santa Maria et al., 2003b).
CompositeDb (g cm−3)CFe (%)Hc (Oe)St (%)Ss (m2 g−1)Dp (Å)
  1. Db, Bulk density; CFe, amount of iron determined using atomic absorption; Hc, coercive field (unit Oersted); St, toluene swelling; Ss, surface area determined using the BET method; Dp, pore average diameter determined using the BJH method.

FR010.625236.85 50 29208
FR020.422174.01 45 25180
FR030.584169.18 67 32152
FR040.737285.11180152 20
FR050.635328.69175189 16

The morphologies of the beads were very distinct and related to the synthesis process used. A general view of the spherical beads (Fig. 3) shows that they were also different sizes (Table 2). This difference in morphology may depend on the suspension agent used to produce the beads. These agents seemed to influence the surface tension of the dispersal medium (water), modifying the type of network formed and the number of iron particles anchored on the composite surface. Some difference can be noted between the mean diameter value of samples FR01 in Table 2 and their appearance in Fig. 3(a). This may be due to the last stage of the developed image processing method, because the dilatation stage may join the samples, producing no well-defined edge, so that two samples are measured as one. An improvement in the image processing is being developed to overcome this difficulty.

Figure 3.

Photomicrography of the composites described in Table 1 showing a general view of the beads: (a) 1; (b) 2; (c) 3; (d) 4; (e) 5.

Table 2.  Size diameters of the microspheres characterized in Table 1.
CompositeDiameter (10−6 m)No. of samples
FR01121.44 ± 74.24 404
FR02152.27 ± 39.10 283
FR03120.44 ± 46.38 538
FR04103.11 ± 52.06 538
FR05 84.06 ± 31.021483

Small particles are dispersed on the bead surfaces, as seen in Fig. 4(a), in which a backscattered electron detector was used. Because the contrast in this type of image formation is due to differences in the atomic number of the elements in the sample, it was an interesting technique to observe the iron distribution on the spheres. EDX microanalysis was then used to map the dispersed iron on the surface of the spheres, as shown in Fig. 4(b). However, the use of uncoated samples produced variable and unreliable iron responses. Another consequence was local heating of the sample, even at smaller accelerated voltages and higher spot sizes. It is known that both the image quality and the accuracy of X-ray analysis in VP-SEM are often limited by the spread of the primary electron beam due to the introduced gas (Stowe & Robinson, 1998). Moreover, the intensity of a characteristic X-ray line is highly dependent on the primary electron beam landing energy (Reimer, 1985). Toth et al. (2002) have shown that changes in this landing energy can be caused by the electrical field generated by ionized gas molecules located above the sample surface. This effect introduces errors into X-ray microanalysis, which can be alleviated by the provision of efficient ion neutralization routes, as the use of new sample–electrode geometries (Toth et al., 2002). In our case, it was achieved by covering the samples with a conductive gold film, permitting more reliable results be observed (Fig. 4b).

Figure 4.

(a) Backscattered photomicrography of composite 1, showing the iron particles on its surface (magnification: 4500×). (b) Iron EDX map of the region shown in (a).

Conversely, no iron particles were seen in the interior of the beads. We have already shown that the different porogen agents used to increase the size and depth of the pores did not succeed in trapping iron in the interior of the microspheres, which may be very useful to increase the magnetic properties of these composites (Santa Maria et al., 2003a,b). Thus, the iron particles were observed only on the surface of the beads.

Electron micrographs of the composites also showed that the small iron particles on their surfaces can present two different features, as shown in Fig. 5. As seen in Fig. 5(a), the small Fe particles are randomly dispersed on the surface of the bead. By contrast, Fig. 5(b) shows clusters of Fe particles on preferred regions of the bead. The suspension agent was considered to be responsible for modifying the amount of iron anchored to the composite surface and the type of iron distribution found on the surface of the beads was also dependent on the preparation conditions (Santa Maria et al., 2003a). It seems that the filamentous microstructure favours dispersion of the particles, whereas smooth beads, where no filamentous microstructure could be detected, stabilize the agglomerated form of these particles.

Figure 5.

Micrographies showing two different features of the iron particle dispersed on the surface of the beads: (a) randomly dispersion of the particles (FR02), shown by the arrows; (b) formation of agglomerates (FR04).

Using the BSE images of samples showing the morphological differences presented in Fig. 5, the iron distribution in both types of composites could be quantified using image-processing procedures. The histograms showing the particle size distribution can be observed in Fig. 6. For samples whose morphologies were similar to those in Fig. 5(a), a narrow distribution is noted in Fig. 6(a), showing that the large number of iron particles in samples presenting the filamentous microstructure was < 2.00 × 10−6 m. By contrast, a broader normal curve can be seen in Fig. 6(b) for samples morphologically comparable with those in Fig. 5(b), due to different sizes of the agglomerates formed. However, it must be emphasized that even samples produced under the same conditions were not totally homogeneous, which may explain the large standard deviation values observed.

Figure 6.

Histograms showing the diameters distribution of the particles: (a) randomly dispersed particles (FR02); (b) agglomerated particles (FR04).

Conclusions

It is important to point out that the size of the iron particles and their distribution on the surface of copolymer network-based composites can be attained using VP-SEM techniques. Good quality images were obtained at high acceleration voltages. In addition, the use of BSE images permitted better observation of the iron particles dispersed on the surface of the microspheres. However, EDX maps could produce reliable results only after covering the samples with a conductive gold film. The iron particle distribution depended on the type of suspension agent used to modify the sorption capacity of the copolymer host.

Acknowledgements

The authors thank CNPq, FAPERJ and CETREINA/UERJ for financial support. They also wish to express their thanks to Mr Jonas Britto (LABMEL/IBRAG – UERJ) for the SEM characterizations, to Instituto de Fisica da UFRJ for the magnetic measurements to Petroflex, Nitriflex and Metracril for monomers and initiator donations and to CENPES/PETROBRAS and Instituto Militar de Engenharia for BET/BJH analyses.

Ancillary