Microscopy Observation

At present, microscopy observation and image processing of micrographs can be considered a reliable source of information for many membrane morphological characteristics, such as pore shape and size, their distribution and density, surface porosity, cross-sectional structure. Nevertheless, these tools are not suitable for obtaining information about pore length and tortuosity.

Microscopic techniques, like scanning electron microscope (SEM) or transmission electron microscope (TEM), present significant difficulties, mainly concerning the preparation of membrane sample. In order to observe, via SEM, the cross sections of the membrane without causing a collapse of the structure, the sample must be dried and then fractured at liquid nitrogen temperature.

In TEM observation, instead, the dried sample is first embedded, if necessary, and then cut by microtome. Care has to be given in the choice of embedding material, to avoid interaction with the membrane, and then to cut very thin sections (in the order of tens of nm).

Although the first electron micrographs of a membrane documented in literature can be attributed to Riley et al. [4], only in the 1980 Merin and Cheryan [5] succeeded to visualize membrane surface pores by using a replica technique and TEM.

Both SEM and TEM expose the membrane to an high electron beam energy, which causes damages, thus hindering the correct inspection of the surface. A significant improvement has been brought in the 1980's by the introduction of field emission scanning electron microscopy (FESEM), which can achieve very high resolution (< 1 nm) even with low beam energy/accelerating voltage.

Kohtake et al. [6] have first used FESEM imaging technique to observe pores on UF membrane surfaces. Kim et al. [7] exploited FESEM micrographs to calculate the pore size, density and surface porosity of various types of UF membranes, using an higher resolution than Kohtake et al. and analyzing the differences between membrane surfaces before and after protein filtration [8].

Although it is possible to measure and calculate the mean pore size and pore size distribution from SEM micrographs, it also results an extremely boring and time consuming

X1 X2 X3 X4 X5 X6 X7 X8X9 X10V X1 X2 X X* X X X7X8X9X1qX11

X1 X2 X3 X4 X5 X6 X7 X8X9 X10V X1 X2 X X* X X X7X8X9X1qX11

FIGURE 7.1. Schematic representation of electron micrographs of the plane parallel to the membrane surface. The vacant regions surrounded by dotted area indicate pores: (a) a membrane which has ellipsoidal pores; (b) an actual membrane which has complicated pores; xi, X2, X3, ..., xii and xmax indicate cut-off length (Reprinted from [3] with permission from Elsevier).

FIGURE 7.1. Schematic representation of electron micrographs of the plane parallel to the membrane surface. The vacant regions surrounded by dotted area indicate pores: (a) a membrane which has ellipsoidal pores; (b) an actual membrane which has complicated pores; xi, X2, X3, ..., xii and xmax indicate cut-off length (Reprinted from [3] with permission from Elsevier).

work. Manabe et al. [9] have developed a method for characterizing the pores of a polymeric membrane having a mean pore diameter larger than 10 nm via electron micrographs. They devised ellipsoidal, spherical and straight-through cylindrical pore models and formulated theoretical equations relating the pore radius distribution function N(r) to the distribution function F(x), where x is the length of test segments cut off by pores in an electron micrograph as illustrated in Figure 7.1.

The theoretical relations are illustrated in Table 7.1; results obtained with a regenerated cellulose (RC) membrane are reported in [9] along with a comparison with other methods, such as mercury intrusion (MI) and bubble pressure (BP).

Image analysis techniques allow fast evaluation of the pore size and distribution from micrographs. Vivier et al. [10] have exploited quantitative image analysis for characterizing a microporous PTFE hollow fiber membrane. More recently, several software tools have been developed for this purpose, e.g. the NIH Image (developed by the National Institute of

TABLE 7.1. Pore radius distribution function for various pore typologies. $e is a shape factor.

Ellipsoidal pores Spherical pores

Straight-through cylindrical pores

Health, USA, Division of Computer Research and Technology). This program allows the determination of the porosity (AK ), the pore density (N), the mean pore radius (rp) and the pore size distribution of a membrane. This tool has been exploited by Masselin et al. [11] to measure the porosity, the pore density, the mean pore radius and the pore size distribution of five asymmetric ultrafiltration membranes by means of FESEM images. The analysis is based on the digitalization of the original grey-scale image, followed by measurement of the area of each pore. In [11], calculations were performed with a 0.5 nm interval. This value appeared to be a good compromise between the calculation precision and the number of intervals. The following parameters have been evaluated:

• The maximum frequency gives an information on the gathering of data in the same radius interval. This means that it gives an idea on the confidence that can be attributed to the modal pore interval and radius.

• The modal interval is the pore radius interval for which the pore radii are the most frequent. A 10% variation around the maximum frequency value is accepted in order to take into account wide distribution. The center of this interval is called the modal pore radius (its frequency is assumed to be maximum).

• The pore radius distribution extent is the maximum difference between the superior boundary of the pore radius interval for which the frequency is minimum (except zero), and the inferior boundary of the modal interval.

Finally, the membrane thickness has been also measured on images. Results allowed the comparison between the AK/Ax values obtained from image analysis and the AK/Ax values obtained by diffusion experiments.

In the last years, atomic force microscopy (AFM) has become a popular method for investigating the surface microstructure of polymer and inorganic membranes [12]. AFM characterization of membranes has focused on measurement of nanometer sized surface pores [13-16] and comparison with pore size analysis from scanning electron microscopy (SEM) [17-18], correlation of surface structure with membrane properties (such as molecular weigh cut-off, MWCO) [19], the effects of surface roughness on fouling [20-21], and adhesion force between model particles and the membrane surface [22].

It is worth mentioning the studies made by Smorgonskaya et al. [23] on the structural organization of bulk nanoporous carbon (NPC) materials produced from carbides. The studies were carried out by the small angle X-ray scattering (SAXS), X-ray diffraction (XRD), and high resolution transmission electron microscopy (HRTEM) techniques, which will not be described further in this survey.

0 0

Post a comment