Ultrafiltration and microfiltration operations are applied intensively in the
dairy and water cleaning industries. The main capacity limiting factors of
such operations are the flux and efficiency decline by irreversible adsorption
of foulants onto the membranes and the efficiency by which the reversible
fouling can be removed/cleaned. The aim of this thesis is to investigate the
residual fouling that is deposited on ultrafiltration and microfiltration
membranes after usage. The membrane surfaces are investigated using
infrared spectroscopy with an attenuated reflectance sampling unit and this is
thesis work highlights the strengths and weaknesses of using infrared
spectroscopy to investigate residual fouling on membranes and in particular
the challenges with the infrared penetration depth when layering in the
samples occurs.
Real size production membrane cartridges at different stages of use from
Danmark Protein, Arla amba were the target of the investigations. However,
in order to obtain samples sizes that fit in the sampling interface of the
infrared instrument the membranes were dissected into smaller pieces named
coupons. In total four ultrafiltration membrane cartridges and two
microfiltration membrane cartridges were investigated with Attenuated-
Total-Reflection Fourier-Transform-Infrared (ATR FT-IR) to map the
residual fouling on both types of cartridges. The height of the characteristic
amide peaks from proteins were used to determine the relative
concentrations.
The first investigation (Paper I) describes the concentration development
over the membrane leaves as a function of the distance from the feed inlet
and the distance from the center permeate tube. A non-homogenous
concentration distribution of residual fouling was observed with the highest
concentration of residual fouling present at the center tube decreasing in
concentration outwards in a flame-like shape. The relative concentration
calculations are based on the height of the amide II peak (1500-1550 cm-1)
which was chosen because it unlike the amide I band has no interference with
adsorbed water and other membrane constituents that can interfere with the
computation.
Based on the findings of the first investigation it was decided to develop a
new method to evaluate the concentration of the residual fouling on real size
production membranes as current best practice methods rely on univariate
height measurements that supply only information on the targeted residual
fouling peak(s). In a second study (Paper II), it was decided to investigate the
infrared data of the membrane by applying multivariate curve resolution
(MCR) in order to resolve the residual fouling from the membrane
components. Indeed the result showed that the MCR model needed three
factors to describe the system, one describing the membrane material
(polyethersulfone, PES), and two describing the residual fouling that is
present on the membrane. The MCR method improved the interpretation of
the models considerably compared to e.g. PCA or the univariate data
analysis. However, it also became evident that the penetration depth of the
infrared beam creates additional complexity when measuring semi-solid
layered samples.
In order to obtain an overview of the different analysis methods and data
analysis methods that have been employed by other researchers when
studying residual fouling on ultrafiltration and microfiltration membranes a
literature review was conducted (Paper III). ATR FT-IR turned out to be a
commonly used spectroscopic method to evaluate ultrafiltration membranes.
The data analysis is most commonly performed univariate by calculating
height of selected peaks along with identification of different chemical
entities especially when investigating grafting/grafted membranes. Paper III
gives an overview of these different approaches and data analysis methods
and their results.
In conclusion, the research in this thesis has shown how the application of
multivariate infrared spectroscopy combined with new data analysis methods
has augmented the knowledge about residual fouling on real size production
membranes. The information obtained can be used to investigate and monitor
residual membrane fouling and help in the design of new membranes and
membrane grafting that can be optimized for the purpose.