A new in situ method for the characterization of membranes in a wet state in the environmental scanning electron microscope

https://doi.org/10.1016/j.memsci.2012.01.031Get rights and content

Abstract

This work presents a new microscopic method for the characterization of microfiltration membranes. It allows the direct observation of the behavior of water at the membrane surface during wetting and drying. For this purpose, an environmental scanning electron microscope (ESEM) was equipped with a special cooling stage, enabling the investigation of wet samples inside the microscope chamber. The images recorded from the membrane surfaces provide information about the number and size distribution of dry and wet pores during the drying process. Additional information about the wetting and drying of the membrane interior was gained by simultaneously measuring the temperature of both membrane surfaces as a function of time. The basic mechanism is the strong influence of the evaporation of cooled water from the membrane interior on the membrane surface temperature. These temperature characteristics reflect the interior membrane structure. The correlation between the microscopic parameters, obtained from the observation of the wetting and drying of the surface pores and the temperature characteristic, a macroscopic parameter, enables a qualitative description of the membrane structure. This information also allows the liquid and gas transport inside the complex membrane structure to be studied. In the present work two different types of flat microfiltration membranes were investigated.

Highlights

► Visualization of the wetting and drying process of the membrane surfaces. ► Accurate measuring of the membrane surface temperature changes during the processes. ► The surface temperature characteristic reflects the interior membrane structure. ► We were gaining the size distribution of the dry pores during the drying process. ► Correlation of the membrane mean pore size and the drying behavior.

Introduction

Microfiltration is one of the most important processes in many industrial applications, for example in the beverage industry or waste water treatment as well as in pharmaceutical and medical applications. Membranes used in these applications should offer a well defined molecular weight cut off (MWCO), a high flow rate and low fouling tendency. As the membranes are often also used for filtration of aggressive substances, the used respective membrane material itself plays an important role. Especially membranes made from polyethersulfone (PES) are resistant to many aggressive substances even at high temperatures due to their glass transition temperature of 220 °C.

New designs of PES microfiltration membranes lead to an increase in filtration performance. Most of the recently developed microfiltration membranes offer a very asymmetric pore structure consisting of a thin separation layer and a broad backup or support layer. The separation layer inside the membrane is essentially responsible for the filtration process. The support layer is necessary for the mechanical stability and protects the separation layer from external damage. This asymmetric, so-called “hour glass” structure [1] can be tuned by varying the temperatures at the “roll side” and “air side” as well as the relative humidity at the “air side” during the production process. As a result, the position and thickness of the separation layer changes. Both the thin separation layer and the large-pored structures of the support layer, which strongly reduce fouling tendency and enable high permeability [1].

Scanning and transmission electron microscopy (SEM, TEM) play a major role in the characterization of this “hour glass” structure. The surfaces of the membranes [1] and their cross-section morphologies have already been studied in full detail in two dimensions (2D) [2]. But a complete description of the pore structure is only possible by reconstructing the membrane morphology in three dimensions (3D). 3D reconstruction enables a visualization of the pore network and the calculation of quantitative parameters like the mean pore diameter, the mean pore length, the average number of connected pores, the specific surface area and volume porosity [3]. Subsequently, both the reconstructed 3D model and the calculated parameters can be used as input for fluid simulations.

Such a static model does not, however, provide information about the nature of the membrane pore surfaces, e.g. whether they are hydrophilic or hydrophobic, or the interaction mechanisms between pore surfaces and fluid.

The focus of this work was thus on studying the dynamics of the system ‘liquid (water)–solid (membrane matrix)–gaseous (water vapor)’ during the wetting and drying process of the membrane. The membranes were mounted in an environmental scanning electron microscope (ESEM) and images of the membrane surfaces were recorded during the drying process. As a consequence, the change in the number of dry pores and their size distribution as a function of time can be determined. The liquid transport inside the pores is influenced not only by the membrane structure, but also by environmental parameters (pressure and temperature), as well as the properties of the pore surfaces (hydrophobic of hydrophilic). As the contact angle of the liquid is determined by the membrane material, its measurement can give relevant information about membrane hydrophilicity [4]. As PES is a hydrophobic material, manufacturers use different additives for enhancing the hydrophilic properties. The quantity and composition of these additives have an influence on the contact angle. The hydrophilic properties of the pore surfaces are essential for proper filtration performance.

Hydrophobic as well as hydrophilic membranes under wet conditions were studied by de la Parra [5], who used an ESEM for the recording of images of the membrane surfaces during the wetting process. Hydrophobic membranes show water droplet formation at the surface. The droplet shape is in correlation with the hydrophobic level of the membrane material. As the hydrophobicity of the membrane surface increases, the shape of the droplet becomes more and more spherical. The images also show droplet formation at hydrophobic spots at the surface of an otherwise hydrophilic membrane surface. The potential of using an ESEM for localization of less hydrophilic areas was clearly shown. However, the results are more qualitative and do not provide a direct correlation between the membrane structure and the flow of water.

Wang et al. give additionally quantitative data of the swelling behavior of fibers inside a cellulose based Sartobind® cation exchange membrane during water uptake and release in the gas phase. The membrane was not studied under complete wet conditions; however it was shown that small fibers inside the macropores show significant volume change [6].

Jenkins and Donald [7] use a special sample holder for the investigation of cellulose fibers. The water swelling behavior of the fibers was documented by series of images. To get accurate and repeatable data of the swelling behavior, nearly each sample shape (fibers, flat or particles) needs a special “adapted” holder. But contrary to cellulose membranes at polymer membranes a swelling behavior is not to be expected. For the polyethersulfone membranes investigated notable swelling could not be observed.

As porous solid materials show characteristic drying rate curves, the drying behavior of membranes should also give information about the interior porous structure [8]. Staude and Passlack use drying rate curves for characterizing different types of membranes [9]. The change in membrane weight during the drying process was measured by a precise balance to obtain characteristic drying rate curves, which are influenced by both the membrane structure and the membrane material. The direct observation of the membrane surface during the drying process and the simultaneous measurement of membrane weight were not implemented in our actual microscope setup. Instead of the change in membrane weight, we recorded the change in temperature at both membrane surfaces during the wetting and drying of the membrane with cooled water.

It is vital for the investigations that the dry membrane has a much higher temperature than the cooled water. These temperature–time characteristics provide information about the successive wetting and drying of membrane layers with different pore size distributions and thus about the membrane structure, especially for asymmetric membranes. This is mainly due to the dependence of the capillary force on the pore diameter. Thus small pores will dry more slowly than large ones. As soon as the surface of the membrane is dry, it starts warming up, even if inner membrane layers still contain cooled water. This and other effects entail characteristic features in the temperature–time characteristic, which reflect the membrane structure. The advantage of this method over 3D reconstructions is the much larger volume investigated (membrane volume around 15 mm × 5 mm × 0.15 mm). It provides the additional opportunity to study the interaction between pore surface and water. Changes in the properties of the pore surfaces or membrane fouling may result in changes in the temperature characteristic.

If the membrane properties change substantially during operation these characteristics could thus help to find the cause for that change. The correlation of microscopic parameters, obtained by direct observation of the membrane surface, and the temperature, a macroscopic parameter, can provide maximum information about both the membrane structure and the interaction between membrane and water.

Section snippets

Materials and methods

The ESEM analysis provides detailed information about the membrane structure in the dry [3] as well in the wet states. Contrary to the conventional high vacuum (HV) mode, this instrument can also be operated in the ESEM mode [10], where the pressure in the specimen chamber can be varied between 0.1 and 20 Torr. This enables the investigation of wet specimens or even liquids (Fig. 1). In this study water vapor is used as the chamber gas. The wetting and drying process can be controlled by

Instrumentation and experimental setup

The experiments were performed in an ESEM Quanta 600 FEG (FEI Company, Eindhoven, Netherlands). A special cooling stage was developed for investigating the flat membranes. It consists mainly of two brass clamps, mounted on Peltier elements (Global Component Sourcing) which are used for cooling the clamps. The clamps have a dimension of 6 mm × 6 mm × 4 mm and feature a slit of 0.3 mm width for holding the membrane (area: 15 mm × 5 mm) by use of two brass fixation plates. Both clamps are cooled to 4 °C. The

Results and discussion

The drying behavior of porous materials is a very complex process. A quantitative theoretical description of the evaporation of water in capillary structures can be found in [20]. Here only a qualitative description of the process is given. The drying of the membrane starts at the surface, with large pores drying first. Water beneath the membrane surface is transported primarily via smaller pores due to their larger capillary forces described by:h=2γcosθρgRwith ρ, liquid density; θ, contact

Conclusions and outlook

This work presents a new method for membrane characterization under wet conditions. As the pressure range in the microscope chamber is limited, the investigations were performed under low vacuum conditions. With water vapor as chamber gas, the pressure determines the relative humidity in the specimen chamber. However, the ESEM sample chamber is a micro laboratory, where well defined conditions (pressure, humidity, temperature) can be adjusted, in order to obtain reproducible and comparable data.

Acknowledgments

The authors thank the company Membrana (Wuppertal, Germany) for providing the samples and Mrs. Tatsch and Mr. Schuster for many helpful discussions.

References (22)

  • R. de la Parra

    A method to detect variations in the wetting properties of microporous polymer membranes

    Microsc. Res. Tech.

    (1993)
  • Cited by (0)

    View full text