The exploration of the potential of grafting for fouling reduction of TiO2 NF membranes

Datum: 20 juni 2016

Locatie: UAntwerp - Campus Drie Eiken - Building Q - Promotiezaal - Universiteitsplein 1 - 2610 Wilrijk

Tijdstip: 15 uur

Promovendus: Ghulam Mustafa

Promotor: Prof Vera Meynen

Co-promotor: Dr Anita Buekenhoudt (VITO)

Korte beschrijving: PhD defence Ghulam Mustafa - Faculty of Science


Membranes have attracted a lot of attention in terms of research and development, because they are considered as a very promising and sustainable separation technology. Membrane technology offers several advantages over conventional separation techniques such as low energy consumption, no additive requirements, no phase change and better permeate quality etc.. Recently, application of membrane technology to molecular separation processes has stimulated interest and showed great potential in a number of different fields such as solvent filtration, water treatment, gas separation and catalyst recovery. Specifically for water treatment, membrane filtration is of high value as it is a physical method of separation, reducing the need for chemicals to provide clean water while operating at room temperature with low energy requirements and being able to handle large volumes in short times. One type of membrane process is nanofiltration (NF), indicating the nanopore size of the membrane. One of the main applications in last two decades has been in water treatment. NF membranes can be used to treat all kinds of water including surface, ground and waste water as well as a pre-treatment for desalination. Interest for applying NF membranes in the water sector is intensifying day by day owing to their high flux (compared to reverse osmosis membranes) combined with low molecular weight cut off, capable of removing bacteria, viruses, organics, fractions of natural organic matter (NOM), micro pollutants (e.g. pesticides), and importantly, without complete removal of inorganic salts. Above all, these NF membranes make it possible to do all this in one step. It is worth to mention that next to several polymeric membranes, also commercially available ceramic NF membranes (i.e. Inopor TiO2 NF membranes) perfectly fit to this profile.

Despite their large potential, one of the major and most critical issues in the development of effective membrane processes is the decline in system performance due to membrane fouling, which limits the economic efficiency of the operation, requires harsh chemical cleanings and slows down large scale industrial applications of the NF membranes. Fouling of NF membranes depends upon many parameters like operational conditions (e.g. pressure, temperature, flux or flow of water, filtration type), feed water quality, membrane material, membrane surface roughness, chemistry and porosity.

Fouling of membranes and especially tighter membranes such as NF membranes, is attributed to the physicochemical interactions between the fluid (foulants) and the membrane. As soon as the membrane surface comes into contact with the feed stream, different fouling mechanisms occur. Normally, a part of the fouling or fouling layer can be removed from the membrane if an appropriate physical washing protocol is employed (known as reversible fouling). However, also irreversible fouling is a persistent problem, caused by the strong adsorption of dissolved matter onto the surface and/or into the membrane pores and can generally only be removed by strong chemical cleaning or even cannot be removed.

The complex process of membrane fouling and mitigation options have been the subject of many research efforts and industrial developments since the early 1960’s when industrial membrane separation processes emerged. Classical solutions to remediate fouling are the optimisation of feed pre-treatment (e.g. ultrafiltration (UF), microfiltration (MF), ozonation or UV oxidation, flocculation, coagulation), regular physical cleaning (e.g. back pulsing, air scrubbing) and/or chemical cleaning of the membranes at regular times. Additionally, a careful selection of the appropriate membrane, module design, operating and process conditions is important to decrease the fouling tendency of membranes.

Nevertheless, the most sustainable approach to mitigate membrane fouling is the prevention of the undesired adsorption or adhesion processes by surface modification. To date, a number of papers have been published relevant to membrane surface modification avoiding membrane fouling. The general aim is to modify the membrane surface hydrophilicity, reduce surface roughness, reduce surface charge density etc. Moreover, researchers also tried to introduce structured polymer chains and incorporate nanoparticles in NF toplayer.

The majority of the membranes in water filtration have been polymeric membranes, and therefore the antifouling strategies as well. However, currently, also ceramic membranes are finding their way in water purification, as they achieve better performance than polymeric in terms of treated water quality and flux stability. The main benefits of ceramic membranes are their high chemical and thermal stability enabling chemical and/or thermal regeneration and sterilization by aggressive chemicals and/or hot steam. Moreover, their high mechanical stability enables high pressure back-flushing. As a consequence, despite of their higher cost price, ceramic membranes become economically feasible in water treatment especially in the more difficult waters (e.g. humic surface waters).

Fouling studies of ceramic (TiO2) NF membranes and their surface modification for antifouling action is scarce but becomes more and more important. Indeed, it is already reported that the ceramic membrane market is growing increasingly and is projected to register a market size of $5.1 Billion by 2020, signifying a firm annualized growth of 11.7% between 2015 and 2020.

Therefore, main objective of this PhD is to fill in that gap and investigate the fouling tendency of ceramic NF membranes in conditions similar to drinking water production and difficult industrial waste waters treatment. Innovative solutions of surface grafting have been developed and validated to diminish the fouling sensitivity of the ceramic NF membranes. Our grafting approach of ceramic NF membranes in this perspective of low-fouling is unique in the state of the art. In addition, the aim has been to acquire important new insights into the fouling mechanism of the NF membranes, making the phenomenon or the processes of fouling and anti-fouling more clear in order to be able to design the membrane grafting for each specific fouling situation.

In this PhD, we thus pioneered the exploration of the potential of grafting for fouling reduction of TiO2 NF membranes. The grafting is based on the replacement of the interactive -OH groups on the membrane surface by less-interactive groups without too much altering the hydrophilic character (water flux) of the membranes. Particularly, the grafting was aimed at reduction of the fouling sensitivity of strongly hydrophilic TiO2 NF membranes and thus maintaining sufficient water flux to be economically viable.

We grafted commercially available TiO2 NF membranes (i.e. Inopor TiO2 NF membranes with pore size 0.9 nm) using different grafting groups (R) and two methodologies. The applied methodologies are able to form a covalent bonding on titania, leading to membranes stable in water applications. The first method is a new method based on Grignard chemistry, recently developed at VITO and UAntwerpen, and allowing the formation of a direct M-R bond, stable on transition metal oxides. The second method used is phosphonic acid grafting, known from the literature (but not on membranes for antifouling purpose) to result in stable modified metal (M) oxide surfaces, except for silica. The commonly used silanation cannot be used here as the produced M-O-Si-R bond has been proven to be highly sensitive to re-hydroxylation on transition metal oxides.  

In order to maintain sufficient hydrophilicity of the grafted membranes, only small grafting groups have been chosen: methyl groups and phenyl groups. Characterization of the grafted membranes showed a decrease in water flux corresponding to the increase of hydrophobicity, while the MWCO remained almost the same.

To evaluate the performance of the grafting, especially its antifouling tendency, native membranes and all grafted NF membranes with sufficient water flux were subjected to a variety of fouling measurements. Focus was on experiments with different model foulants (organics and inorganics) mimicking the foulants in real surface and ground water for drinking water production. After studying the model foulant solutions, real surface water (water from “Waterproductiecentrum De Blankaart Belgium”) was used to validate the antifouling tendency of the graftings in real streams. Moreover, we also measured fouling using model solutions mimicking effluents from pulp and paper industry and real olive oil waste water with the purpose of recovering the valuable solutes as well.

We did not only graft TiO2 NF membranes for antifouling action for water treatment, but also more open TiO2 membranes (e.g. micro- and ultra-filtration membranes) were  successfully grafted. These grafted TiO2 ultrafiltration membranes (Inopor TiO2 UF membranes with pore size 30 nm) were tested for their antifouling tendency compared to the native membranes using two different model oil/water emulsions mimicking oil/gas produced waste water. Throughout the study, next to the antifouling/fouling measurements, also the quality of permeates of both grafted and native membranes were analyzed.

The obtained results suggested that grafting of ceramic membranes by the mentioned techniques definitely enhanced the antifouling tendency, though at the expense of some water flux decrease (typically less than or equal to 50%). The best results are obtained using methyl grafting of a membrane applying the GR (Grignard) technique (MGR membrane). This grafting makes the membrane very inert to irreversible fouling, not only in cases of drinking water production, but also for waste water processing and oil produced water treatment. While for other modifications, some fouling remains but in a much lesser extent than the native membrane. Taking into account chemical structure and the physico-chemical properties of the foulants and the different membrane surfaces, led to a qualitative understanding of the experimentally obtained results allowing basic predictability for the choice of grafting.

Moreover, a detailed study of dissolved organic matter (DOM) fouling with and without inorganic ions such as calcium ions, on grafted and native TiO2 NF membranes, allowed us to gain valuable new insights in the details of the membrane fouling process. The results, in contrast to the current knowledge in the state-of-the-art, can again be explained by the details of the physico-chemical properties of both the foulant and the membrane surface determining their type and strength of interactions

Next to anti-fouling capacity of course the obtained permeate water quality of the grafted membranes is important. These were found to be quite comparable with the native membranes. However, in the specific case of olive oil waste water, the permeate quality of the MGR membrane was also far better than the native membrane.

After the performance tests of the graftings, a lot of effort was dedicated to determine the stability of the graftings using several different acidic, basic and oxidative solutions with different concentrations at different temperatures. The obtained results made it obvious that the stability of grafting is sufficiently high to allow proper cleaning of the grafted membranes in real surface and ground water, and in different waste waters treatments.

This PhD project was running in parallel to a European FP7 research project named 'CeraWater' aiming to produce fouling resistant ceramic honeycomb nanofilters (with extra high surface to volume ratio) for efficient water treatment. In this project, by other researchers, also upscaling of the MGR grafting was successfully realized. In this FP7 CeraWater project, several grafted commercial scale membranes were tested in different real water streams (both difficult real waste water streams such as pulp and paper waste effluents and olive oil waste water and real surface water stream i. e. Blankaart river water). The obtained results confirmed the strong antifouling effect, and revealed the potential to lead to higher overall process fluxes of the grafted membrane, despite the lower water flux (~50% decreased by the grafting).

To support the developed technology perspectives, a proper techno economical evaluation both for drinking water production and difficult waste waters processing is in progress.

The acquired knowledge from this study is valuable to steer further research for understanding the fouling behavior of membranes by others types of foulants and to develop adequate antifouling strategies, enhancing the industrial applicability of (ceramic) NF membranes.