1. Introduction
The appearance of the MD process can be found in the Bodell patent from 1963 [
1,
2] and despite the intensive studies carried out for more than 50 years, durable MD membranes have not yet been developed. The hydrophobic porous membranes made from polytetrafluoroethylene (PTFE), polyethylene (PE), polyvinylidene fluoride (PVDF) and polypropylene (PP) are frequently used in the studies of the MD process [
2,
3,
4,
5,
6,
7]. In several works, the commercial membranes made from these polymers for microfiltration (MF) were tested as the MD membranes [
3,
8,
9,
10,
11,
12]. However, with regard to the MD process conditions (e.g., higher feed temperature and non-wetted pores) the membranes fabricated for MF, in the majority of cases, did not fulfil the MD process requirements. Generally, a membrane for the MD process has to simultaneously meet the following requirements [
2,
12,
13,
14,
15,
16,
17,
18,
19]:
high liquid entry pressure (LEP), which is the minimum transmembrane hydrostatic pressure that is applied on the membrane before a liquid solution penetrates into the pores;
good thermal stability—up to the boiling temperature of water;
high chemical resistance to separated solutions;
high permeability;
low thermal conductivity;
narrow pore size distribution.
A simultaneous fulfilment of these conditions is difficult, e.g., high permeate flux can be achieved for thin membranes, but a low thermal conductivity and higher energy efficiency can be obtained when thicker membranes are used, especially for brines desalination [
20,
21]. Moreover, a further progress of the MD process (industrial implementation) requires a substantial growth in studies on a semi-pilot or pilot scale [
18,
21,
22].
The pioneering activity within a scope of the MD process implementation is performed, among others, by SolarSpring GmbH (Freiburg, Germany), Memsys GmbH (Schwabmünchen, Germany), TNO (Den Haag, The Netherlands), Aquaver (Voorburg, The Netherlands), Scarab Development AB (Stockholm, Sweden), Xzero AB (Stockholm, Sweden), BlueGold Technologies (Largo, FL, USA) and Abengoa Water (Seville, Spain) [
18]. The pilot plants constructed by the above-mentioned companies utilize mainly the membranes made from PTFE, PE and PP recently produced for the MF process [
18,
23]. Although the applied membranes exhibit certain drawbacks, the realized pilot studies have a positive impact on the demand for the MD process and enhance the opportunity for implementation of novel membrane production in future.
An intensive development of the MD process investigation has been observed in recent years, and as a result, the number of publications has grown from 1000 to almost 5000 [
16,
18]. Several works have reported that the pores wettability can be restricted by application of composite membranes or membranes with modified surface which enhances the resistance to wetting [
8,
9,
16,
24,
25,
26]. Promising results were obtained for the membranes formed by electrospinning method and the membranes with addition of different fillers, e.g., carbon nanotubes, titanium and silicon dioxide or talc [
27,
28,
29,
30,
31,
32]. In addition to membranes made of polymers, ceramic membranes are also presented, which create a chance for applications requiring particularly durable membranes [
33,
34]. Many new types of membranes and methods for their preparation and the possibility of their use in the MD process are described extensively in review articles [
8,
16,
18,
34,
35].
Nevertheless, so far, no breakthrough results have been obtained, which indicates that durable membranes for the MD process could be difficult to fabricate. One of a reasons for the slow development of MD membranes is the fact that the majority of works present only ways to increase the LEP value or contact angle, but there are no studies demonstrating the durability of new types of membranes during a long-term module exploitation. Moreover, even the membranes with a very good resistance to wetting, can undergo the progressive degradation due to slow changes in the polymer after 100–200 h of the MD process [
36]. Moreover, the implementation of new types of membranes for the production is difficult, mainly due to a fact that the MD membranes market is just being created [
18]. Moreover, the manufacture of MD membranes on a larger scale is necessary because it enables the performing of pilot studies [
18,
22]. Demonstrating the possibility of realizing different applications of the MD process on a pilot scale gives the opportunity to attract industrial investors, what is necessary for process implementation.
With regard to the above-mentioned issues, at the initial stage of MD development it is essential to apply the methods already industrially employed for the production of MD membranes. A thermally induced phase separation (TIPS) process is such a method, commonly used to fabricate the membranes from hydrophobic polymers [
1,
37,
38]. In the TIPS method, the polymer with/without fillers is introduced into the mixer, where is mixed with proper amounts of different kinds of oils [
37,
38,
39,
40,
41,
42,
43]. The polymer granulates are completely melted in the diluents system, and obtain a hot (e.g., 453–493 K) homogenous dope solution that flows through the spinning nozzle into a coagulation bath. Subsequently, the oils are extracted from the membrane matrix.
It has been demonstrated in numerous papers, that the membranes manufactured by the TIPS method have the appropriate properties for the MD process. The membranes made from PP are most often used for MD [
10,
18,
23,
37,
38,
41,
43]. Moreover, the application of other polymers, such as PVDF, PE and polyethylene chlorinetrifluoroethylene (ECTFE) for MD membrane preparation via the TIPS method is also possible [
39,
44,
45]. In several works it has been reported that improvement of the properties of MD membranes could be obtained by modification a composition of dope solution and by using different conditions for the membrane formation. However, in the majority of cases, the studied membranes were formed under laboratory conditions. Therefore, certain phenomena taking place during the membranes production on the industrial scale, such as the row arrangement of polymer chains resulting in the formation of linear structure of the membrane matrix, can affect the polymer mechanical properties [
46]. The possibilities of preventing the polymer degradation and methods for improvement of the mechanical properties of PP membranes by introducing fillers into the dope solution, are presented in this paper.
The polypropylene capillary membranes represent the membranes manufactured via the TIPS method on an industrial scale for the MF process [
10,
23]. The flow of melted polypropylene through a spinning nozzle causes a more linear arrangement of the polymer chains. A cooling-down of formed capillary in the coagulation bath proceeds rapidly [
38], which favours the freezing of polymer linear structure. A result is a deterioration of tensile strength manifested by a longitudinal creaking of polymer. The addition of crystals nucleus (e.g., inorganic fillers) into a dope solution is one of the methods of disturbance of the linear arrangement of polymer chains [
46,
47]. It is well-documented in the literature that the blending of polymers with mineral fillers is considered a useful way to improve the mechanical properties and talc is a popular mineral filler used for this purpose [
46]. Talc is a cheap filler and more importantly, it has the hydrophobic properties that enable its very good dispersion in polyolefin. For this reason, talc is often used to improve the properties of materials produced from polyethylene or polypropylene [
30,
47].
In current MD pilot studies, hydrophobic membranes produced for microfiltration are often used. However, the feed temperature in the MD process, as a rule, is higher than that which is applied in the MF process. As a result, the membrane matrix is subjected to a larger thermal expansion during the MD process, what can lead to damages in the membrane structure. In order to enhance the thermal resistance of produced polymeric materials from PP, the polymer is usually blended with different inorganic fillers, such as talc, which exhibits many actions improving the PP properties [
29,
47]. It was demonstrated that the application of talc enhanced the thermal resistance and tensile strength of PP membranes [
29,
39,
48,
49]. However, as the thickness of the film produced from PP decreases, then the beneficial effects of the fillers used, such as talc, are reduced [
46]. Moreover, the structure of the membranes (pore walls) has a much smaller thickness compared to the films produced. For this reason, the objective of this work is to determine the effect of talc addition on the properties of PP membranes used in the MD process. Moreover, PP membranes have been shown to undergo a significant transformation, such as surface wettability, during initial 50–100 h of the MD process [
10,
36,
50]. Therefore, the possibility of application of used membranes in this work over significantly longer periods (350 h) was studied.
The membrane matrix durability studies were often omitted in the previous works describing the MD process. One of the main applications of the MD process is water desalination in order to obtain drinking water. Works in which research is undertaken to determining whether the fillers introduced into the membrane matrix or compounds used to modify the membranes surface will not be released into the produced water are scarce [
51]. This information is important since many of the components, such as nanomaterials, used to improve the properties of MD membranes, are suspected of being carcinogenic [
52,
53]. There are well-known works suggesting that a long-term contact with talc may also have a carcinogenic effect [
54]. Until now, such an action has been attributed to talcum forming asbestos-moulded structures [
54,
55]. However, recent studies indicate that the so-far considered safe platelet forms of talc may also have a negative effect on human health [
56]. For this reason, it is important to obtain a kind of membrane matrix that would prevent the leaching of fillers. The present study investigated whether there was no leaching of talc from the surface of the membranes during a long-term MD process.
Taking into account a significant impact of production installation design and the conditions of its exploitation on the properties of fabricated membranes, the studies were performed using two types of capillary MF membranes (with/without talc) manufactured in industrial installation under similar conditions of the TIPS process.
2. Materials and Methods
Two types of capillary PP membranes purchased from PolyMem (Warszawa, Poland) were used in the studies. These membranes have the internal diameter of 1.8 mm and the outer diameter of 2.6 mm. Although the composition of dope solution and the process parameters of membrane production were similar, about 10% of talc (as a nucleation agents) was added to one dope solution. The membranes utilized in this work were designated as PP-N (net polypropylene) and PP-T (polypropylene with talc). The parameters of used membranes were summarized in
Table 1. The manufacturer stated that the average pore diameters equal to 0.2 μm were represented in both cases. The membrane porosity was determined using the gravimetric method [
57]. In this case, the membrane samples with a length of 25 cm were soaked in isopropanol for 1 h to achieve a pore wetting.
The MD module configuration, such as submerged modules, was applied for direct contact MD studies with preferred option. Each module was equipped with two capillaries, and the working length of capillaries was 40 cm (membrane area 45 cm
2). The experimental set-up is shown in
Figure 1. The submerged MD module was assembled inside the feed tank, and distillate flowed inside the capillaries (linear velocity of 0.55 m/s). The MD installation was operated, at the feed temperature equal to 353 K, for 350 h in a continuous mode (day and night). Since mainly the thermal effects were studied, a dilute NaCl solution (1 g/L) was used as a feed. When certain pores are wetted in the membrane, the feed may leak to the distillate. Therefore, the presence of NaCl increases the electrical conductivity of the distillate, which allowed for assessing the degree of wetting of tested membranes. The distillate was cooled by tap water and its temperature was maintained at 288–295 K. The permeate flux was calculated on a basis of changes in the distillate volume over a studied period of time (20–24 h). The volume of obtained distillate was in the range of 500–1000 mL over this period. Assuming that the volume was measured with an accuracy of 5 mL and the error of the membrane area calculation was 2%, the error of the MD measurements (permeate flux) did not exceed 2–3%.
The thermal properties of used PP membranes were determined by differential scanning calorimetry (DSC). These tests were performed by means of a device DSC Q100 (TA Instruments, New Castle, DE, USA), at heating and cooling rate of 10 K/min within the temperature range 200–523 K. The samples were examined in heating-cooling-heating cycles after previous drying.
The mechanical strength and elongation at break of the PP capillary membrane were measured with a tensile tester (3366 Universal Material Testing Machine, Instron, Norwood, MA, USA) according to PN-EN ISO 527-1:1998 method. The specimens of membrane with an initial length of 10 cm were clamped at both ends obtaining the measurement length 3 cm, and pulled in tension at the constant elongation rate of 10 mm/min. In each case, the measurements were performed for ten samples of PP membrane.
The changes of membrane hydrophobicity were determined using a Sigma 701 microbalance (KSV Instrument, Ltd., Espoo, Finland). Based on the Wilhelmy plate method, the dynamic contact angles were measured.
A 6P Ultrameter (Myron L Company, Carlsbad, CA, USA) was applied to measure the values of the electrical conductivity.
The membrane morphology and composition of inorganic fillers were investigated using a Hitachi SU8000 Scanning Electron Microscope (SEM) equipped with Energy-dispersive X-ray Spectrometer (EDS). The specimens for cross-sectional examinations were prepared by fracture of the capillary membranes in liquid nitrogen. Before the SEM examination the membrane samples were sputter coated with chromium (about 3–5 nm) using Q150T ES coater (Quorum Technologies Ltd., Lewes, UK).
In addition to SEM-EDS examinations, the membranes were tested using FTIR, XPS and elemental analysis. This analysis was performed using Elemental Analyzer FLASH 2000 CHNS/O (Thermo Scientific, Waltham, MA, USA), which operates according to the dynamic flash combustion of the sample. For C and H determination the obtained gases were carried by a helium flow to a layer filled with copper, then swept through a GC column, and finally, a Thermal Conductivity Detector (TCD) detected them. In the case of oxygen, the samples were introduced into the pyrolysis chamber via the MAS Plus Autosampler. The reactor contains nickel-coated carbon maintained at 1060 °C. The oxygen in the sample forms carbon monoxide, which is then separated from other products using gas chromatograph and detected by the TCD Detector.
The functional groups presented on the membrane surface were identified by a Fourier transform infrared (FTIR) spectroscopy. The method of Attenuated Total Reflection (ATR) was applied. In the performed studies, a Nicolet 380 FTIR spectrofotometer connected with Smart Orbit diamond ATR instrument (Thermo Electron Corp., Waltham, MA, USA) was used.
Chemical composition of the membrane surfaces was evaluated by X-ray photoelectron spectroscopy (XPS) (Prevac, Rogów, Poland). Prevac electron spectrometer, equipped with an SES 2002 (VG Scienta, Uppsala, Sweden) electron energy analyzer working in a Constant Energy Aperture mode was applied in these studies. The concentration of detected elements (expressed in atomic percents) was calculated in CasaXPS software.
X-ray diffraction (XRD) studies were performed in order to determine a crystal structure of the membranes. In these studies was applied an EMPYREAN diffractometer (PANanalytical, Almelo, The Nederland) using a monochromatized CuKα radiation (35 kV, 30 mA). The obtained peaks that presented the kth Bragg reflection were described by Pseudo-Voigt profile function, which can be assigned a fixed shape of any type between their limiting Gaussian and Lorentzian forms. Based on the Pseudo-Voigt profile, the High Score Plus 3.0 software was applied for estimation of a full-width at half maximum (FWUH) parameter. XRD allowed for indicating a material with different electron density. Assuming an existence of a difference between PP and talc, a trial was undertaken to apply software Easy SAXS 2.0 for determination of talc particle size.
4. Conclusions
The obtained MD results demonstrated that both types of PP membranes used in the work possess the appropriate properties for the MD process, such as a resistance to wetting and almost 100% salt retention. However, the additional tests of the membrane matrix stability revealed that only the PP membranes that were reinforced by talc addition had the properties suitable for application to pilot plant MD studies. Moreover, the MD tests carried out for 350 h did not allow for revealing the occurrence of membrane defects and for this purpose, an even longer MD study should be performed, when the industrial application of membranes is considered.
During the membrane formation via the TIPS method, the flow of melted polypropylene through a spinning nozzle caused a linear arrangement of the polymer chains. A cooling-down of the formed capillary in the coagulation bath proceeded rapidly, favoring the freezing of polymer linear structure. A result is a deterioration of the tensile strength manifested by a longitudinal cracking of polymer. The addition of talc (crystals nucleus) into a dope solution caused the disturbance in the linear arrangement of the polymer chains, and PP membranes with a surface morphology more akin to a sponge-like structure was formed. In this case, the MD process efficiency was higher by more than 10% in comparison to the utilization of membranes without talc addition.
In the MD process, the membranes are in contact with the feed at a high temperature, e.g., 353 K, which causes the formation of thermal stress resulting in the numerous cracks which were observed on the surface of membranes having a linear structure. Such problems were not found in the case of membranes where the linear arrangement of polymer chains was disturbed by the addition of talc into a dope solution.
It has been confirmed that the uniform dispersion of talc was achieved within the membrane formed by the TIPS method. Talc was permanently incorporated into the polypropylene matrix and did not leach from the PP membranes during the MD process.