Swimming Pools Are Commonly Disinfected Using Chlorination Biology Essay

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

Swimming pools are commonly disinfected using chlorination, with Hypochlorous acid acting as a disinfectant. However, HOCl reacts with organic matter to form disinfection by-products, such as Trihalomethanes e.g. Chloroform, which can be dangerous for swimmers. Hence other methods are used to reduce the concentrations of THM, one of them being membrane processes, such as Ultrafiltration (UF) and Nanofiltration (NF) membranes that are studied in this research. The aim of the study was to find out how effective are Ultrafiltration and Nanofiltration membranes in reducing the concentrations of Trihalomethanes in swimming pool water. Pool water was collected from the children pool in Kopališče Pristan. Furthermore characterizations of the UF (Polyvinyldieneflouride) and NF (Thin-film composite polyamide) membranes were made using Xylose and Raffinose solutions (concentration 2.00 g dm-3) and the Memcell RO/NF/UF MCS 80 model device. By using Total Organic Carbon (TOC) quick-test retentions were calculated and a graph of retention versus molecular weight was constructed. The pool water sample was processed using the characterized UF and NF membranes. Samples were collected before and after a one hour membrane treatment. Using TOC and Adsorbable Organically Bound Halogens (AOX) quick-tests the Retention was calculated for each membrane. By measuring the volume change in 30 seconds the flux was calculated. It was found out that NF membrane retains 64.8% TOC and 45.2% AOX, whereas UF retains only 14.9% TOC and 31.4% AOX. NF is concluded to be more effective in terms of retention than UF. However, the flux of the NF membrane is, with 165 L m-2 h-1 before and 135 L m-2 h-1 after the water treatment, much less than the flux of the UF membrane which was measured to be 1119 L m-2 h-1 before and 912 L m-2 h-1 after the processing.

Abstract word count: 295

Contents Page

Table of Figures

Table of Tables

List of Abbreviations

Adsorbable Organically Bound Halogens


Disinfection By-Products


International Agency for Research on Cancer


Molecular Weight




Nominal Molecular Weight Cut-off [1] 




Thin Film Composite


Total Organic Carbon







Pool waters are regularly disinfected by various methods. The most commonly used method is disinfection by chlorination. However, besides disinfecting the pool, chlorine can also react with organic matter present in pool water, yielding disinfection by-products, such as Trihalomethanes (THM) (Bitenc et al. 2009). THM are potentially dangerous compounds that can cause, for instance, asthma (Bernard et al. 2003 quoted in Bitenc et al. 2009), and are, according to the IARC, listed as potentially carcinogenic (IARC 1987 quoted in Bitenc et al. 2009).

Hence much effort is made to find methods which could reduce the concentrations of THM, either by removing the precursors [2] before chlorination or by removing the THM formed by chlorine compounds. One of these methods is water treatment with synthetic membranes, such as Ulftrafiltration and Nanofiltration.

The objectives of the research were to study a removal of THM's using Ultrafiltration and Nanofiltration membranes. By analyzing the results it should be also possible to determine the fractional distribution of organic and halogenated molecules in pool water, similarly to the fractionalization done in Glauner et al. (2005).

Research question

How effective are Ultrafiltration and Nanofiltration membranes in reducing the Trihalomethanes concentration from swimming pool waters?


Based on the properties of the Ultrafiltration (UF) and Nanofiltration (NF) membranes [3] I believe that NF membrane will remove a higher percentage of THM than UF membrane. However, the flux through the NF membrane will be lower. We assume that the total organic carbon (TOC) and adsorbable organically bound halogens (AOX) are good measures of THM concentrations in the water.

Theoretical Background

Trihalomethanes production

In Slovenia it is prescribed by law that any swimming pool water needs to be at least disinfected by any method (Uradni List 2011). The most commonly used method is chlorination. This can be done in various ways, for example, by introducing gaseous chlorine (Glauner et al. 2005), or using Sodium Hypochlorite (NaOCl) (Judd and Jeffrey 1995). Both result in the formation of hypochlorite ions (-OCl) and hypochlorous acid (HOCl).

The following reactions occur.(pKa value of HOCl according to Ripin and Evans 2005):

The abundance of the acid is higher in acidic environment (pH < 7.5) and lower in basic environment (pH > 7.5). This is important since hypochlorous acid is about 1000 times more effective as a disinfectant than the corresponding ion (Judd and Jeffrey1995).

However, as good as HOCl is as a disinfectant it is also highly chemically reactive. It readily reacts with organic matter in the pool, which is introduced by the swimmers or is naturally present in the water used in the pool, to produce disinfection by-products (DBP). A group of DBP are Trihalomethanes: volatile halogenated alkanes (Judd and Jeffrey1995).

The most common THM formed are Chloroform (CHCl3), bromodichloromethane (CHBrCl2), dibromochloromethane (CHBr2Cl) and Bromoform (CHBr3) (Rook 1977 quoted in Lourencetti et al. 2012). [4] 

Figure : Schematic representation of THMs (from left to right), Bromoform (CHBr3), dibromochloromethane (CHBr2Cl), bromodichloromethane (CHBrCl2) and Chloroform (CHCl3).C:\Users\Tomaž\Desktop\šola\MM\EE\dejanski ee\slike2\THMs.png

Although bromine is not directly inserted into the water by disinfection, it is usually naturally present in the water. As soon as there is any bromine in the water, some brominated THM are present as well (Cooper et al. 1985 quoted in Nokes et al. 1999). Bromine substitution occurs due to the rapid oxidation of bromine by the chlorine in hypochlourus acid (Wong and Davidson 1977 quoted in Nokes et al. 1999). A redox reaction occurs in which the bromine acts as the oxidizing agent and chlorine as a reducing agent (Brezonik and Arnold 2011):

HOBr is formed, which reacts with organic matter in pool water analogous to HOCl. (Nokes et al. 1999).

Aqueous HOX acids can react with many organic compounds, such as Methyl ketones, human acid substances or even algae, present in pool water to produce THM. (Chawla et al. 1983)

The formation of THM can be summarized by the following overall reaction, where X represents Bromine or Chlorine (Chawla et al. 1983):

Precursor (aq)+ HOX (aq) →CHX3 (aq)

However, the reaction mechanism involves halogenated organic molecules, which react further with the disinfectant to finally produce THM (Chawla et al. 1983).

Figure : Flow- Chart of the mechanism for THM production (Chawla et al. 1983)

The study of THM is important since it was found out that they are dangerous to humans. Swimmers are exposed to THM through inhalation, skin absorption or oral ingestion (Bitenc et al. 2009), where inhalation is the most dangerous (Lee et al. 2009). Since THM are volatile compounds equilibrium is established (take for instance chloroform).

Due to their density, THM then concentrate on the surface of the water, where they are easily inhaled by swimmers (Erdinger et al. 2004 quoted in Bitenc et al. 2009). The consequences of the exposure to THM are presented in Table 1.

Table : The Consequences to the exposure to various THM (Vaniek et al. 2002)

Four THM are all genotoxic, with Bromoform being the least dangerous, which correlates with the findings of Glauner et al. (2005), where it was found out that THM with Molecular mass/weight [5] below 200 g/mol are the most genotoxical.

Hence it is important to reduce the concentration of THM in pool waters to decrease the exposure of swimmers to these dangerous compounds. THM concentrations can be reduced by the usage of alternative disinfectant, removal of THM precursors or by removing THM after they have formed (Chawla et al. 1983).

In the following research Ultrafiltration in comparison with Nanofiltration will be used to remove THM after their formation.

Ultrafiltration (UF) and Nanofiltration (NF)

According to Mulder (1996), a membrane is a selective barrier between two phases, the term 'selective' being inherent to a membrane or a membrane process. The main physical and chemical properties of the membranes, used in this research are listed in Table 2 and schematically presented in Figure 3.


Ultrafiltration membrane

Nanofiltration membrane

Separation of particles

Separation of Macromolecules (bacteria, yeasts)

Separation of low molecular weight solutes (salts, glucose)

Range of Applied pressure [kPa]




Asymmetric porous

Composite membrane

Thickness of membrane

150 μm

Top layer: 1 μm

Sub layer: 150 μm

Pore size range [nm]



Membrane material

Polymers (e.g. polyvinyldienefluoride) or ceramic (Zirconium oxide)

aromatic polyamide

Table : Basic physical and chemical properties of UF and NF membranes (adapted from Mulder 1996)

From Table 1 we can see that the Nanofiltration membrane separates particles with lower molecular weight (MW), and should be hence more effective in reducing the concentration of

Figure : Schematic presentation of Microfiltration, Ultrafiltration, Nanofiltration and reverse osmosis, where only UF and NF will be used (Mulder 1996)

THM than Ultrafiltration. However we can see that the size of the pores in UF membranes is greater, hence the flux of permeate will be greater, meaning that a higher mass of water can be cleaned per unit time. Also the applied pressure in UF is less, meaning that the UF method consumes less energy (Mulder 1996). C:\Users\Tomaž\Desktop\šola\MM\EE\dejanski ee\slike2\285a.jpg

Figure : Schematic representation of dead-end and cross-flow filtration (Mulder 1996)However, the filtration is going to be different from conventional, dead-end filtration (as used in Mulder 1996). In chemistry we are used that the filtration is exerted through the force of gravity which acts perpendicular to the membrane. We will use the cross-flow method in which the pressure is applied parallel to the membrane and the difference in pressure and diffusion make the separation process possible. By applying the pressure parallel to the membrane we minimize the effect of polarization - the increased concentration of the solute near the membrane which increases the resistance of the flow and thus reduces the flux through the membrane. (Mulder 1996). The dead-end and cross-flow methods are presented in Figure 4. C:\Users\Tomaž\Desktop\šola\MM\EE\dejanski ee\slike2\291ka.jpg

The use of cross-flow method is rapidly increasing over the last few years because:

No additional chemicals needed - environmental friendly.

Low energy consumption.

Require a small amount of space.

Easily installable.

Membranes can be designed for different purposes and desired effectiveness.

There are also some disadvantages which can affect the membrane performance on a long term:

Leftovers have to be removed and stored.

Fouling of membranes. [6] 

Low lifetime of membranes.

Low selectivity of particles which pass through the membrane or low membrane flow of Permeate. [7] 

(Adapted from Mulder 1996, quoted in Lobnik et al. 2008)

Membranes have already been used in studies to determine their effectiveness and the fractions of THM in pool water. For example, Glauner, T., et al (2005) in Swimming pool water - fractionation and genotoxicologial characterization of organic constitutients, reported , by the use of UF and NF membranes, that most of the molecules in pool waters in Germany had MW from 200 g mol-1 to 1000 g mol-1. In the same study they discovered that THM with MW below 200 g mol-1 are the most genotoxic and hence most dangerous to swimmers.

For an accurate determination of THM concentration in pool water a Gas Chromatograph with a connected mass spectrometer could be used (Stack et al. 2000). However, as no access to such equipment was possible, quick tests for Adsorbable organically bound halogens (AOX) were used.

Practical Work

Sampling and On-site measurements

The samples were collected at Kopališče Pristan, a local swimming pool, where the water has been chlorinated. The sample was taken on June 20th 2012 at 8.20 AM. The sample was collected in eight 1 dm3 flasks, which were filled to the top and sealed tightly. Otherwise the volatile THM would evaporate and the concentration of halogens in the samples would be too low. On-site pH value, conductivity and temperature were measured.

Characterization of membranes

Two membranes, with area of (80.0 ± 0.1) cm2, were used, a UF membrane and a NF membrane.

Figure : Monomer of polyvinyldienefluoride (Informatzione 2013)The UF membrane used was PVDF 400-PET and was produced by Rotreat GmbH. The membrane is polymeric - polyvinyldienefluoride (Figure 5).C:\Users\Tomaž\Desktop\šola\MM\EE\dejanski ee\slike2\200px-Polyvinylidenfluorid.svg.png

Figure : Chemical composition of the Nanofiltration membrane (Bauman 2010)The NF membrane used was TFC - thin film composite and was manufactured by SEPRO membranes Inc. (Figure 6). C:\Users\Tomaž\Desktop\šola\MM\EE\dejanski ee\slike2\nova.jpg

Figure : Schematic representation of the functional device: Memcell RO/NF/UF MCS 80 (ÄŒeplak 2013)The experiment was modeled on the functional device: Memcell RO/NF/UF MCS 80, from the company OSMO Membrane Systems. The device is specially designed for flat membrane separations and allows us to make a model of the real life when using membranes. F:\EE 001.jpg

Figure : Rotreat's Functional device that was used in the research (Čeplak 2013) C:\Users\Tomaž\Desktop\šola\MM\EE\dejanski ee\slike\u-IMAG0143.jpg

For the determination of the nominal molecular weight cut-off (nMWCO), which is the molecular mass at which 90% of the molecules are retained by a particular membrane (as defined by Lee et al. 2002 quoted in Glauner et al. 2005), two sugars were used:

D-(+)-Xylose with molecular mass 150.13 g/mol [8] and D-(+)-Raffinose pentahydrate with the molecular mass of 594.42 g/mol [9] , both supplied by Sigma-Aldrich Company.

Sugar solutions with the concentration of 2 gmol-1 were prepared and processed by the functional device. We collected samples of the sugar solutions before and after the processing. From these samples we measured the pH value, the Conductivity and temperature of the solutions. We also measured the Total Organic Carbon (TOC), by using LCK 386 quick test for TOC 30-300 mg dm-3 and LCK 387 quick test for TOC 300-3000 mg dm-3, both manufactured by Hach-Lange, which photometrically determine the concentration of carbon in the sample. By measuring the change in TOC concentration we can determine the retention [10] of each membrane and construct the Retention versus MW graph, from which we can determine the nMWCO.

Pool water analysis

For each membrane 3 dm3 of pool water is taken and processed by either UF or NF membrane. We take samples for the analysis before and after each membrane treatment. Again the pH, conductivity and Temperature are measured. We also measure the average volume, which is processed by a membrane in 30 seconds. From this we can calculate the flux [11] of a membrane. In addition to the TOC test we also test for the concentration of Adsorbable organically bound halogens (AOX), which is performed by the LCK 390 quick-test for AOX, manufactured by Hach-Lange. By measuring the change in concentration of both TOC and AOX we determine the retention of the membranes when treating swimming pool water. From this we infer the extent to which the membranes are able to remove the dangerous THM from pool waters.

The data processing and presentation was done on Microsoft excel and Logger pro 3.1. by Vernier company.


On-site measurements of the sample

Concentration of dissolved O2

([mg/ dm-3] ± 0.01 mg/ dm-3)

Conductivity ([μS/cm] ± 1 μS/cm)

pH (± 0.1)


([°C] ± 0.1°C)





Table : The on-site measurements

We see that the pH is somewhere in the region where hypochlorite ion and hypochlorous acid are in equilibrium and that the temperature is slightly elevated.

Membrane characterization

Before the experiment, each membrane was drowned into water for 24 hours. The properties of a membrane get adapted to the water environment and should not change significantly during the experiment.

We prepared 1 dm-3 solutions of Xylose and Raffinose.


Molecular mass [g mol-1]

Concentration [g dm-3] ± 0.01 g dm-3




D-(+)-Raffinose pentahydrate



Table : The molecular masses and mass concentrations of the sugar solutions used

The solutions were processed for one hour with the functional device. We collected samples before the treatment and collected the permeate after one hour.

We repeat the procedure with the second sugar solution with the same membrane. After the characterization of one membrane is completed, we repeat the whole procedure for the second membrane. [12] It is important to note that the characterization of the NF membrane was already made beforehand by my external mentor, Dr. Maja Bauman, since I was using the same membrane, no additional measurement was neither necessary.




± 0.1 μS/cm)

pH (±0.1)


[°C] ± 0.1°C


([mg dm-3] ± 1 mg dm-3)























Table Measurements for Ultrafiltration process


Conductivity ([μS/cm]

± 0.1 μS/cm)



Temperature [°C] ± 0.1°C


([mg dm-3] ± 0.01 mg dm-3)























Table Measurements for Nanofiltration [13] 

By comparing the TOC concentrations we can see that a fraction of molecules is rejected by each membrane, where the rejection of the NF membrane is much greater. We can also see that the conductivity decreases, which means that the membranes also reject some ions in the water. Summing up from this table we construct of retentions.

Retention is, according to Mulder (1996) calculated by:



Retention of Xylose

([%] ± 0.1%)

Retention of Raffinose

([%] ±0.1%)




Nanofiltration [14] 



Table Calculated Retentions (%) for UF and NF membranes for Xylose and Raffinose

We see that the retention of the NF membrane is much higher than the corresponding retention of the UF membrane. NF rejects 67.4% of Xylose molecules, whereas UF rejects only 5.8% of Xylose molecules. The difference is even bigger with Raffinose, where NF rejects 95.8% of all molecules and UF rejects only 5.0%. It is a bit unusual that the retention of Raffinose is lower than the retention of Xylose with the UF membrane. Almost constant retention of Xylose and Raffinose by the UF membrane indicates that the UF membrane's pores are much bigger than the size of molecules with the mass of 600 g/mol. We can assume that the UF membrane will significantly reject only macromolecules, [15] for example fat molecules. Sterlitech (2013) suggests that the nMWCO of PVDF UF membranes range from 30000 to 120000 g mol-1.

On the basis of these results a MWCO graph can be constructed, which is a graphical representation of retention as a function of Molecular weight (the fraction of molecules rejected by the membranes versus their molecular weight).

Figure : Graph of Retention (%) versus Molecular weight (g mol-1)

From the graph we can see that the retention of the Nanofiltration membrane is higher at lower molecular masses than Ultrafiltration, which retention is almost unchanged when comparing the Xylose and Raffinose. Thus UF will only reject molecules with much higher molecular mass. Thus we can assume that UF membrane only retain molecules that have a greater size than Raffinose. From the graph we can determine the nMWCO value for NF membrane, which is at the intersection between the 90% line and the Nanofiltration Retention and lies at approximately 500 g mol-1. Although one could argue that only 2 sugar solutions are too few to accurately determine the Retentions of a membrane, we see that for UF the line over the interval from 150 to 600 g mol-1 is constant and that for NF the slope of the interval from 150 to 600 g mol-1 is less compared to the slope in the first interval. Hence even if there is an uncertainty in the determination of nMWCO for NF, it is negligible.

Pool water analysis

The first sample was taken before the treatment and the second sample is the permeate after a one hour treatment with the membranes. We analyze the collected samples, by measuring the pH, conductivity, temperature and the flow, from which we can determine the flux, and the TOC and AOX concentrations.


Sample was taken

Applied Pressure [kPa]




([μS cm-1]

± 1 μS cm-1)


[°C] ± 0.1°C

Volume of permeate per 30 s [cm3] ± 0.1 cm3


([mg dm-3] ± 0.01 mg dm-3)


([mg dm-3] ± 0.01 mg dm-3)

































Table : Measured data from the collected pool water treatment

From these data we can calculate the retention according to the AOX and TOC concentrations listed in table 7, by using equation (1). When analyzing the measured values we can see that both the mass concentration of halogens and the mass concentration of carbon atoms decrease upon treatment with the membranes. It was expected that the retention of the NF membrane will be higher than that of the UF membrane, which is confirmed by the results.

Retention of TOC

([%] ± 0.1%)

Retention of AOX

([%] ± 0.1%)







Table : Retention of TOC and AOX

Figure : Retention of TOC and AOX

We see that NF has a higher percentage of retention in both the TOC and AOX. From my results NF rejects 64.8 % of carbon and 45.2% of all halogens in the pool water. UF rejects merely 14.9 % of all carbon and 31.4 % of all halogens in the pool water. Based on these results, NF is more efficient than UF, when decreasing solute concentrations in pool water. We can also deduce the size of the molecules present in the pool water. UF shows that around 15 % of the molecules are large scaled with molecular weight over 10000 g/mol, by assuming that the nMWCO value of the UF membrane lies in that region. The AOX value shows that halogens also form products with heavy molecules since 31 % of halogens in the sample are retained by the UF membrane. We can also deduce that 33% of all carbon molecules in pool water have a lower molecular mass than 500 g/mol, since they didn't get rejected by the NF.

We can also determine the flux through a given membrane. [16] The area of a membrane is (80.0 ± 0.1) cm2 and the time interval was 30 seconds. Change in volume can be seen from Table 7.



Sample was taken


([L m-2 h-1] ± 3 L m-2 h-1)





912 (81.5% of the initial)





135 (81.8% of the initial)

Table : The Calculated Flux of UF and NF membranes

NF has a major drawback - the low flux. From Table 9 we see that the flux of the UF membrane is in about 10 fold greater than the flux of NF membrane. Hence much more time is needed for NF to process the same volume of pool water than UF for the same area of membrane. This effectively means that more pool water can flow through the membrane and be filtered. We can also see that due to the treatment, the flux decreases to about 82% of the initial value. We can assume that the decrease continues with longer time intervals.

From Table 8 the determination of the ratio of masses of the carbon-containing molecules and molecules containing halogens in the pool water can be made. This is shown on Figure 11 and Figure 12. [17] 

Figure : TOC distribution of molecules according to their molecular weight

Figure : AOX distribution of molecules according to their molecular weight

A similar analysis was done by Glauner et al. (2005), however, a comparison cannot be made, since their nMWCO for the UF and NF membranes were different from mine, so the mass distribution was also different. However both studies show that the major part the halogenated compounds in the solutions have a low molecular mass. In Glauner et al (2005) it was also stated that molecules with low molecular mass are the most genotoxical and hence the most dangerous to swimmers.

By comparing Figure 13 and 14 we can see that the most molecules containing carbon have a molecular weight over 500 g mol-1 and are not retained by the UF membrane. On the other hand, most on the molecules, which contain halogens, have a molecular weight less than 500 g mol-1 and are retained by neither NF nor UF membrane. Thus we can assume that halogens react mostly with organic molecules with low molecular mass.


Judging by retention, the NF membrane showed to be more efficient by rejecting a higher percentage of both TOC and AOX. Assuming that TOC and AOX concentrations represent the THM concentration in pool waters, we deduce that NF reduces the concentration of hazardous THM to a greater extent than UF. However, by comparing the flux through the membranes, we see that the NF membrane's flux is about one tenth of the flux of UF membrane. Hence even though NF may produce a permeate with lower THM concentration, it will process about ten times less water in a given time interval then UF in the same membrane area.

Figure 14 gives us another important piece of information; approximately 55% of halogen-containing molecules are retained by neither of the membranes. Combining this with the result of Glauner et al. study (2005), which reports that the low molecular mass THM are the most dangerous, we can deduce that the most dangerous molecules are still present in the water, despite the membrane treatment. For example, neither of the molecules in Figures 1 and 2 are retained by any of the used membranes. [18] 

According to the study of Lee et al. (2009), the most danger is presented through inhaled THM, while membranes retain only dissolved molecules and hence do not significantly contribute to the reduction of danger. However, in every pool the following equilibrium occurs (e.g. chloroform).

By reducing the concentration of aqueous THM and applying Le Chatellier's principle, more gaseous molecules will dissolve. Hence they are less dangerous and can be processed with membranes to further reduce the concentration of THM in the pool environment.

My suggestion to improve the efficiency of the membranes would be to use bromine disinfection instead of chlorine. This can be done by introducing Sodium Bromide (NaBr) into the solution a chlorinated pool or use Sodium hypobromite (NaOBr). (Judd and Jeffrey 1995). Brominated THM are less volatile than chlorinated (Judd and Jeffrey 1995) and have a greater molecular weight, thus they would be retained by a higher percentage than chlorine analogs. A drawback of bromination is the fact that a higher overall concentration of THM is formed (Ram et al. 1986; Heller-Grossman et al. 1993 quoted in Judd and Jeffrey 1995) which is not desirable.

I would recommend a system of membranes as presented in figure 15.

Figure A schematic representation of my suggestion for pool cleaning by the use of membranesC:\Users\Tomaž\Desktop\šola\MM\EE\dejanski ee\slike2\nova spet.jpg

The system would have a series of membranes with progressively higher retention and lower flux. In between them, there would be a switch, which determines how the water is processed. The whole volume would be processed by the UF membrane, due to the higher flux, but only a fraction would be treated also with NF, which with higher retention but lower flux. One could also add another membrane with nMWCO less than 500 g mol-1.

However, we might argue that the results are biased. The time of sampling was at 8.20 a.m. We can expect that there weren't many swimmers by that time in the pool, which could have introduced any body fluids or other organic molecules, which would have altered the concentration and distribution of TOC and AOX molecules in pool water.

The procedure itself proved to be environmental friendly. No additional chemicals were needed in order for the membrane to function, almost no waste material except for the precipitated on the membrane.


The aim was to determine how effective are Ultrafiltration and Nanofiltration membranes in reducing the concentrations of Trihalomethanes in pool water. By assuming that TOC and AOX concentrations represent the THM concentration in pool waters, it can be concluded that the retention of the Nanofiltration membrane observed is higher than the retention of the Ultrafiltration membrane. NF had 64.8% retention of TOC and 45.2% retention of AOX, while UF had 14.9% retention of TOC and 31.4% retention of AOX. Therefore, from the retention point of view, NF is the more effective method.

Also, the flux of the UF membrane is much higher than the flux of the NF membrane. The flux of UF membrane was 1119 L m-2 h-1 before and 912 L m-2 h-1 the pool water treatment, while the flux of NF membrane was only 165 L m-2 h-1 before and 135 L m-2 h-1 after the water processing. That means that UF membrane can process a greater volume of pool water per area per time than NF.

However, it is observed that after the treatment with UF and NF membranes, 68.6% and 54.8% (respectively) halogen containing molecules with low molecular mass still remained in the pool water.

According to the results Ultrafiltration and Nanofiltration methods are very effective when it comes to reducing Trihalomethanes concentrations in pool waters. Because not only that they can reduce the concentration of halogen containing atoms (AOX) by 31.4% and 45.2% respectively, the procedure is environmental friendly as well. I do believe that membrane technologies can significantly decrease the concentration of Trihalomethanes in pool waters and hence reduce the healthy dangers connected with them.

Furthermore I believe that with the right choice of membranes, for example a UF membrane with high nMWCO and high flux and a NF membrane with low flux and low nMWCO connected as in figure 15, the concentrations of THM can be reduced to an even greater extent. Hence I believe that membrane technologies are the future of purifying swimming pool waters.


I would like to give special thanks to my external mentors dr. Maja Bauman and dr. Mojca Pobrežnik for their time and patience, that was lost while I was doing my research and for sharing their knowledge, experience and their gathered data with me. I would also like to thank the IOS institute for letting me perform my experimental part in their laboratories. I would also like to thank my EE supervisor prof. Zdenka Keuc, for her time and the freedom she gave me while I was writing my Extended essay and for not giving up, when a due date was not held.