1. Introduction
Biofiltration is one of the most promising and economically profitable methods currently used in drinking water treatment technologies. Due to the biological processes carried out by microorganisms forming the biological membrane, biofilters can operate effectively for many years without the need to replace/regenerate the bed. The process of biodegradation of adsorbed pollutants enables the continuous renewal of the adsorption capacity of the bed, and thus prolongs the effective operation of biofilters. Biological transformations are slow, but thanks to the sorption properties of the biofilter filling, pollutants can be retained on the biosorption bed for a longer time. As the sorption capacity of the bed gradually depletes, the microbial biomass takes over its functions, sorbing and biodegrading substances present in the water [
1,
2].
The presence of difficult-to-remove contaminants in water abstracted for municipal needs means that technological systems must be expanded to include coagulation, chemical oxidation, and sorption processes on activated carbons, which significantly increases the costs of water production. Due to the level of water pollution and the insufficient effectiveness of these processes, biofiltration has become an indispensable complement to water purification systems [
3]. Including the biofiltration process in the water purification system improves the quality and biological stability of the treated water. The biofiltration process reduces both health risks and risks related to threats to technical infrastructure. The risk of exceeding the permissible water quality parameters is significantly reduced, which confirms the validity of introducing the biofiltration process into the water purification system [
4].
One of the most troublesome compounds found in water is ammonium nitrogen. Its presence hinders, among other things, the water chlorination process (ammonium ion reacts with chlorine to form chloramines, which may have carcinogenic properties [
5,
6]), causes problems with the removal of manganese and iron [
7], accelerates the corrosion of water pipes and affecting the biological stability of water [
8,
9]. The presence of nitrogen compounds, e.g., ammonia, causes the eutrophication of surface waters and increases toxicity towards fish [
10]. The increased content of ammonium ions in treated water consumes oxygen in the water supply network, which may lead to the development of harmful anaerobic bacteria and secondary water pollution. The maximum permitted concentration of ammonium ions in drinking water is 0.5 mg/dm
3 [
11,
12]. Therefore, it is important to investigate various techniques that can effectively remove nitrogen compounds, including ammonium ions, from drinking water. The removal of ammonium nitrogen can be achieved by physicochemical treatment (including chlorination, ion exchange, and membrane filtration) or biological treatment (nitrification process) [
8,
13,
14,
15]. Due to the cost, biological methods are increasingly being proposed to remove nitrogen pollutants. Many technological systems use the nitrification process where there is a need to remove ammonium nitrogen without the need to completely remove nitrogen. The nitrification process carried out using the biofiltration method is highly effective, provided that the carrier and process parameters are properly selected. The limitation in its use is the oxygen content necessary to carry out the full nitrification process. Nitrification is a process of biochemical oxidation of ammonium ions to nitrite or nitrate nitrogen. The process of oxidation of NH
4+ ions to NO
3− takes place in two phases and is carried out by two different but cooperative groups of autotrophic bacteria: AOB and nitrite-oxidizing bacteria (NOB) [
16,
17]. In the first stage, AOB bacteria use NH
4+ ions as electron donors, oxidizing them to NO
2− ions. This reaction can be represented by the equation [
18,
19]:
Microbial community studies have shown that the dominant AOB population includes, among others:
Nitrosomonas communis and
Nitrosomonas oligotropha [
17,
20,
21]. AOB must have at least two key enzymes to efficiently oxidize ammonium nitrogen to nitrite nitrogen. The first enzyme is ammonia monooxygenase (AMO), which allows the oxidation of the ammonium ion to the intermediate hydroxylamine. The second enzyme is hydroxylamine oxidoreductase (HAO), which then converts the intermediate hydroxylamine into nitrite nitrogen. AOB bacteria comprise several genera within the phylum Proteobacteria, including
Nitrosomonas,
Nitrosospira, and
Nitrosococcus [
16,
22].
In the second stage, NOB bacteria (e.g.,
Nitrobacter) oxidize nitrite nitrogen to nitrate nitrogen using NO
2− ions as electron donors [
23]. The reaction is presented in equation [
16].
NOB have a specific enzyme, nitrite oxidoreductase (NXR), that allows them to oxidize nitrite nitrogen to nitrate nitrogen [
24]. NOB classified within the phylum Proteobacteria include the genera
Nitrobacter,
Nitrotoga, and
Nitrococcus.
Nitrospira spp. are included in the phylum Nitrospirae [
16,
25]. The nitrification process takes place at temperatures from 4 to 45 °C, with an optimal temperature of 35 °C for Nitrosomonas and an optimal temperature of 35–42 °C for Nitrobacter. An increase in temperature by approximately 10 °C leads to a doubling of the growth rate of microorganisms [
8]. Nitrification is an aerobic process, and oxygen consumption indicates that biological processes are taking place in the filters. The total stoichiometric oxygen consumption in nitrification reactions is 4.57 mg O
2/mg N-NH
4+ [
9]. Oxygen is not only a key factor shaping the conditions for the development of microorganisms but also determines the possibility of their growth. Oxygen is also the final electron acceptor in catabolic metabolism, providing microorganisms with the most efficient energy source management. In single-stage filtration with the water table above the filtration layer, under conditions of full water saturation with oxygen (10–16 °C), it is possible to oxidize no more than approx. 2 g N-NH
4+/m
3. This limitation applies to biofiltration processes carried out in filters with filling that does not or only to a small extent participate in the process of removing contaminants. Taking into account the fact that water with an ammonium nitrogen content exceeding 2 g N-NH
4+/m
3 is increasingly used for economic purposes, nitrification carried out by biofiltration using biosorption beds may create the possibility of removing larger amounts of ammonium nitrogen than indicated by the oxygen content in treated water. Moreover, it significantly shortens the time of biofilm development by creating permanent connections between the carrier and microorganisms. During the nitrification process, the content of all forms of nitrogen should be monitored because they are limited in drinking water, i.e., ammonium nitrogen < 0.5 mg/dm
3, nitrite nitrogen < 0.5 mg/dm
3 and nitrate nitrogen < 50 mg/dm
3.
The use of the biofiltration process in drinking water purification technology, and in particular the use of sorption materials with oxidation reduction, ion exchange, sorption, and buffering properties, can solve two problems:
(1) significantly shorten the biofilm formation time by creating permanent connections between the carrier and microorganisms,
(2) create the possibility of removing larger amounts of ammonium nitrogen than indicated by the oxygen content in the treated water.
The research aimed to evaluate the method of conducting biofiltration by determining the impact of the direction of water flow (gravity and counter-gravity) on oxygen consumption, on the time of operation of nitrification beds (nitrification filter startup), and the efficiency of ammonia nitrogen removal.
3. Results and Discussion
The nitrification process, carried out using the biofiltration method on biosorption beds, is complex. The way it is implemented turns out to have a significant impact on its course. The analysis of changes in the controlled parameters showed comparable effectiveness of ammonium nitrogen removal, regardless of the direction of water flow used during biofiltration (
Figure 2) In both cases, the concentration of ammonium nitrogen in the treated water did not exceed the permissible value of 0.39 mg N-NH
4+/dm
3 from the 73rd day of the deposits’ operation. The maximum efficiency achieved was 97% for flow opposite to gravity and 99% for gravity flow (
Table 3).
In both analyzed variants, the process of ammonium nitrogen removal was accompanied by a simultaneous increase in nitrite nitrogen. In analyzing the changes in the concentration of this compound during gravity flow, it should be noted that until the 23rd day, the differences between their concentration in water before and after the biofiltration process were small. After this period, there was a significant increase in the content of nitrite nitrogen in the filtrate, which indicated the beginning of the first phase of nitrification (
Figure 3). In the water after biofiltration with the flow opposite to gravity, the first phase of nitrification was observed much earlier, from the 5th day of the experiment (
Table 3). A similar relationship was found for the second phase of the nitrification process.
Regardless of the flow used, in the first days of the experiment, the concentrations of nitrate nitrogen in raw and treated water were comparable. This indicated a slow start of the second phase of the nitrification process. The full second phase of the nitrification process probably did not occur because the increase in nitrate nitrogen was not accompanied by a simultaneous decrease in the nitrite nitrogen content. In biofiltration with unconventional flow, the increase in nitrate nitrogen content was achieved already on the 40th day, while with gravity flow only achieved an increase on the 62nd day (
Figure 4).
The flow in biofilters, both gravity and counter-gravity flow, can have a significant impact on the efficiency of the process of removing contaminants from water. The discussion on the differences between these two types of flow in biofilters is important from the point of view of the design, operation, and optimization of these systems. In gravity biofilters, water flows freely through the filter bed under the influence of gravity, while in anti-gravity biofilters, water is pumped through the filter bed. The efficiency of biofilters may vary depending on the type of flow. Gravity biofilters may be more susceptible to clogging, which can lead to an uneven distribution of the water stream and reduced contact between contaminants and microorganisms. In contrast, anti-gravity biofilters can have better control over the flow of water through the filter bed, leading to a more even-flow distribution and better efficiency of the biological process. The use of unconventional flow makes it possible to adjust the thickness of the biological membrane, to which it is possible to maintain a relatively constant activity of microorganisms. Research by Lu et al. confirms the better efficiency of pollutant removal using anti-gravity flow (the NH
3−N removal capacity has been doubled). Moreover, in this variant, there were better oxygen conditions and there was no leakage of microorganisms, as was the case in the biological filter with gravity flow [
26].
However, many factors affect the performance of biofilters. Factors such as the type of filter materials used, the composition of contaminants, temperature, pH, retention time, and hydraulic and organic load may also have a significant impact on the efficiency of water purification in biofilters [
26,
27]. In practice, the appropriate selection of design parameters and regular control and monitoring of the process are crucial for the effective operation of biofilters, regardless of the type of flow.
The number and activity of microorganisms present in the deposit have a significant impact on the effectiveness of water treatment. The chemical composition of the biofilter filling has a decisive influence on the growth of biofilm. Diatomite, due to its buffering properties and the content of substances that stimulate the development of microorganisms, allows for the rapid formation of a biofilm and the start-up of a biofilter. The biofilm formation time also depends on the chemical composition of the water directed to the biofilters. The presence of iron(III) in water treated in the nitrification process increases the effectiveness of ammonium nitrogen removal and shortens the time of biofilm formation. The presence of iron has a positive effect on the start time of the second phase of the nitrification process but increases the oxygen consumption during the processes (higher biofilm activity). However, the presence of iron inhibits the development of heterotrophs in the biological membrane, which improves the bacteriological quality of treated water [
28]. The final effect of the treatment process is determined not only by the surface of the active biofilm but also by the filtration speed. The conducted research shows that the efficiency of ammonium nitrogen removal increases with a decrease in the flow rate. Hydrodynamic conditions, such as flow rate and velocity, harm biofilm stability. Higher flow rates may destroy microbial colonies (AOB) responsible for the direct reduction in ammonium nitrogen [
6,
29,
30]. To maintain the maximum activity of microorganisms, the thickness of the biofilm covering the bed should be appropriate. The biofilm includes both active bacteria and inactive substances. The former are responsible for removing contaminants from the interfacial zone, while the latter influence the thickness of the biofilm. The accumulation of inactive substances in biofilm contributes to the reduction in average microbiological activity. The recommended thickness for an aerobic biofilm is 15–25 µm (the ammonium nitrogen oxidation reaction occurs already at a thickness of 7 μm), the remainder are the so-called inactive supporting substances [
31].
Regardless of the flow direction used, the purified water did not meet the standard requirements, especially since exceedances could be observed in the case of nitrite nitrogen (values exceeded 0.5 mg NO
2−/dm
3) (
Figure 3). High concentrations of nitrite nitrogen in the filtrate are an unfavorable phenomenon because they have a significant impact on inhibiting the growth of nitrifying bacteria present in the biofilter filling. Additionally, nitric acid(III) formed from NO
2− ions may be toxic to both Nitrosomonas bacteria and Nitrobacter [
4]. The accumulation of a significant amount of NO
2− ions may also result from differences in the growth rates of phase I and phase II nitrifying bacteria. The causes of nitrite accumulation can also be found in the structure of the biofilm, in which an anoxic and an oxygen layer are distinguished. An oxygen deficit causes reactions involving heterotrophic bacteria to occur on the grain surfaces and in the pores, which reduces the biochemical activity of nitrifying bacteria. The oxidation capacity of nitrite nitrogen also depends on other factors: temperature (below 7 °C, oxidation processes are inhibited) [
17], the amount of oxygen [
32], the content of nutrients (phosphates) [
33], and the design and operation of filters [
34].
The nitrification process, which is classified as a biological water treatment process, is accompanied by the consumption of oxygen. The total amount of oxygen necessary to carry out the full nitrification process is 4.57 mgO
2/mgN. For the first phase of the process, 3.43 mgO
2/mgN is needed, while for the second phase, where the nitrite ion is oxidized to nitrate, the amount is 1.14 mgO
2/mgN. It has been repeatedly confirmed that oxygen consumption is lower and is not stoichiometric to the amount of nitrogen removed [
35]. The reason for this phenomenon may be, among others, a biosynthetic process during which oxygen is released. Additionally, it was shown that in the first phase of the oxidation of ammonium nitrogen to hydroxylamine (NH
2OH), oxygen can come from both molecular oxygen and water molecules. Attention should also be paid to the fact that part of the nitrogen is incorporated into the cells of microorganisms involved in the process, thanks to which the oxygen demand will be lower than it would result from the stoichiometric equations.
The Initial oxygen concentration in the water flowing to the biofilters ranged from 6.18 to 7.8 mg O
2/dm
3. After biofiltration with gravity flow, the oxygen content decreased, reaching values from 2.14 to 4.84 mg O
2/dm
3 (the highest consumption of 5.37 mg O
2/dm
3 was observed on the 74th day) (
Figure 5). The actual oxygen consumption in this case ranged from 1.48 to 5.37 mg O
2/dm
3 and was higher than the theoretical oxygen consumption resulting from the stoichiometry of the reaction (1.21 to 3.52 mg O
2/dm
3) (
Table 4). In the case of biofiltration with counter-gravity flow, the actual oxygen consumption ranged from 0.39 to 2.5 mg O
2/dm
3 and was lower than the theoretical one, which should be in the range of 1.03–5.97 mg O
2/dm
3 (
Table 4). The oxygen content in water after biofiltration with counter-gravity flow ranged from 5.13 to 6.38 O
2/dm
3 (
Figure 5).
Based on the prepared balance, it can be concluded that oxygen consumption differed significantly in both analyzed variants. During biofiltration with gravity flow, the average actual oxygen consumption was 2.37 for each mg of nitrogen removed, while with counter-gravity flow, this value was twice as low and amounted to 1.02 mg O
2/dm
3 (
Figure 6). This is because in the first case, dead biomass was retained in the lower part of the biofilter, which could have resulted in higher oxygen consumption. However, the actual oxygen consumption in biofiltration with counter-gravity flow was two times lower than the theoretical one and was usually lower than 2 mg O
2 for each mg N removed. For comparison, in the biofiltration process in a sand bed with gravity flow, the oxygen consumption for each 1 mg of ammonium nitrogen removed ranged from 50% to 150% of the theoretical oxygen demand [
36]. Such high oxygen consumption was explained by the biodegradation of dead organic matter accumulated in the bed. In the discussed case of biofiltration, the use of anti-gravity flow promoted the natural regulation of the biofilm thickness. It resulted in the gradual removal of dead organic matter from the biofilter and, consequently, reduced oxygen consumption.
The cause of the oxygen paradox, consisting of lower oxygen consumption than results from stoichiometric calculations, may be the removal of ammonium nitrogen by a route other than nitrification. An example may be the process of nitrogen assimilation by the cells of microorganisms forming a biofilm, which incorporates part of the ammonium nitrogen into their biomass using the carbon dioxide present in the water. Yu et al. claim that phosphorus can be used to show the role of assimilation in the biofiltration process. If there is a correlation between nitrogen losses and phosphorus consumption, the assimilation process is the basic method of ammonium nitrogen removal [
37]. In the case under consideration, due to the lack of control of the phosphorus concentration, it cannot be clearly stated what the contribution of the assimilation process was to the biofiltration process, but this process could have contributed to nitrogen losses. Another type of process may be denitrification. The lack of oxygen causes reactions involving heterotrophic bacteria to occur in the pores, and on the grain surface, the biochemical activity of nitrifying bacteria is inhibited and the denitrification process may occur. Denitrification is anaerobic respiration where nitrates are an electron acceptor and are reduced to N
2 or N
2O. Typically, biofilters for drinking water are aerobic reactors, but the diffusion of oxygen into the inner layer of the biofilm may be difficult, so anaerobic niches favorable to the denitrification process may theoretically be created. Since most denitrifying microorganisms are chemoheterotrophs, the parameter informing about the role of denitrification in nitrogen losses may be the removal of COD
Mn [
38]. The last process that should be mentioned is sorption. Sorption processes are related to the properties of the biofilter filling. Diatomite is a mineral that has a certain sorption capacity for ammonium ions [
38]. However, during the exploitation of the deposits, the sorption capacity is exhausted and may affect nitrogen losses only in the initial phase of the deposit operation. However, it cannot be completely ruled out that the previously retained ammonium ions do not participate in the cycle of transformations of nitrogen compounds. There are known methods of biological regeneration of deposits used in ammonium nitrogen removal processes. Theoretically, ammonium nitrogen retained through ion exchange may undergo further biochemical transformations. This is evidenced by the appearance of more oxidized forms of nitrogen compounds in the outflow from biofilters than results from the amount of ammonium nitrogen removed at a given time.
To sum up, it can be said that with counter-gravity flow it is possible to remove larger amounts of ammonium nitrogen than the oxygen content in the water. The use of an unconventional flow direction during water biofiltration resulted in lower oxygen consumption, created better conditions for the full nitrification process, and resulted in a faster biofilm formation. Automatic regulation of the biofilm thickness during water flow prevented excessive growth of dead organic matter and limited the development of heterotrophic bacteria. The controlled water quality parameters also showed that the biofiltration process contributed to a reduction in the color (by 10% on average) and turbidity of the water (by 55% and 17% on average for gravity and reverse-gravity flow). Despite many advantages, the counter-gravity flow reduced the zone of the second phase of nitrification, causing the accumulation of nitrate(III) ions.
The use of the second stage of biofiltration allowed us to obtain water that meets the requirements for water intended for human consumption. The effectiveness of two-stage biofiltration is shown in
Figure 6,
Figure 7 and
Figure 8. The determinations were made in raw and treated water from 44 to 87 days of filter operation. Analyzing the test results, it can be concluded that after the first stage of biofiltration, the ammonium nitrogen content ranged from 0.07 to 1.42 mg N/dm
3 (removal efficiency ranged from 31 to 97%) (
Figure 7). The introduction of the second stage of filtration allowed us to obtain ammonium nitrogen values in the range of 0 to 0.34 mg N/dm
3 and the water quality corresponded to the normative value for water intended for human consumption (<0.39 mg N/dm
3).
In the filtrate after the first stage of biofiltration, the concentration of nitrite nitrogen was in the range of 0.6–1.188 mg N-NO
2−/dm
3 (higher values were obtained than in the water directed to the bed, which indicated the occurrence of the first stage of the nitrification process). However, after the second stage of filtration, a decrease in the concentration of this form of nitrogen in the filtrate was observed. From the 75th day of the experiment, the concentration of nitrite nitrogen gradually decreased (from the 79th day the water reached the normative value < 0.15 mg N-NO
2−/dm
3) (
Figure 8). In the case of nitrate nitrogen, it was observed that the increase in the concentration of this form of nitrogen occurred much faster after the second stage of filtration. The concentration of nitrate nitrogen ranged from 2.3 to 5 mg N-NO
3−/dm
3 (
Figure 9). In this case, the second phase of the nitrification process started very slowly—this could be because the bacteria of the second phase of nitrification were less effectively immobilized on the bed.