Removal behaviour of residual pollutants from biologically treated palm oil mill effluent by Pennisetum purpureum in constructed wetland

Removal of COD in constructed wetland

The CW had a steady state of outflow values after 75 days of operation, and good plant growth was observed. The system started to release effluent on the third day after the first influent was added. COD in the effluent was in adaptation process in the beginning because microorganisms and plants had not adapted and matured in the conditions provided by the FD²⁶. Generally, the COD removal efficiencies in wetlands were 88.4 ± 0.24% at day 75. COD concentration gradually decreased with time from day 27 to 75 as shown in Fig. 2. COD concentration showed a continuous reduction as compared to the input. Although a higher COD concentration of POME FD was supplied at day 24, the CW managed to reduce it by 15.84 ± 0.22% at day 28.

The concentration of influent had an effect on the elimination of COD. The removal of COD can be explained by the function of the P. purpureum in the CW as the plant roots exuded additional organic matters²⁷, which led to the changes of COD concentration. After 27 days, the living roots had flourished and became sturdier, which had stimulatory effect on soil organic matter decomposition, thus indicating development of roots-induced microbiota^28,29. On top of that, the microbial communities within the CW ecosystem were seeded by the presence of bacteria in the POME FD that were involved in anaerobic digestion. One of the major bacterial constituents that can be found in the POME FD was Pseudomonas sp. This bacterium is capable of dye degradation or colour removal, biodegradation of hydrocarbon, and insecticide degradation such as Malathion^30,31. It has been reported that Pseudomonas aeruginosa is capable to reduce the COD level from POME in a microbial fuel cell³², as well as the COD level of urban wastewater³³.

Removal of total suspended solid in constructed wetland

Figure 3 shows the removal of TSS from POME FD. Different concentrations of POME FD were supplied throughout this study to simulate a real palm oil mill. Even though the wastewater input into the wetland did not have the same initial TSS concentration, the system still managed to reduce the TSS up to 88.1 ± 13.3%. The POME FD concentrations supplied were 129 ± 3.2 mg/L (day 0–24), 295 ± 4.3 mg/L (day 27–39) and 553.7 ± 11.2 mg/L (day 46–72). The concentrations of TSS from the CW outlet for day 25, day 46, and day 75 were 13.6 ± 3.6 mg/L, 30.2 ± 8.5 mg/L, and 19.4 ± 3.3 mg/L, respectively.

Dhulap and Patil³⁴ have shown the potential of P. purpureum to minimise TSS in sewage by 55.17%. The removal strategies for TSS in CW are via sedimentation and filtration. Particulates were removed whenever the POME FD seeped through the developed root masses of P. purpureum, gravel, and sand. This finding matches current finding as the sedimentation of the solids happened at the top of the CW, and the roots of P. purpureum also showed massive growth in the entire CW, as shown in Fig. 4. Typically, the CW has a lengthy hydraulic retention time, often several days, to extract settle-able and floatable solids while bacterial growth and interactions with adsorbed species remove non-settling solids.

Removal of tannin and lignin in constructed wetland

Tannins, also known as polyphenols, are by far the most abundant macromolecules contained in POME that greatly affect its colour, and create musty odour³⁵. The tannins leaching onto the run-off wastewater from the palm oil processing give the brown colouring effect to the POME. From the results obtained, tannin and lignin concentrations were decreasing in three phases as the inlet of the CW was taken in different batches for three consecutive months at Felda Pasoh, Negeri Sembilan.

Overall, the tannin and lignin contents decreased from the initial concentration as shown in Fig. 5. The concentrations of tannin and lignin in the inlet of CW were 8.7 ± 0.6 mg/L, 14.5 ± 0.2 mg/L, and 21.8 ± 0.4 mg/L. The removal of tannin and lignin was showing as early as day 3, which then reduced to 6.4 ± 0.9 mg/L and 4.5 ± 0.6 mg/L at day 15. The concentration drastically increased on day 27 due to the high concentration of the inlet of CW. Overall, the removal of tannin and lignin was 57.5 ± 22.3% using this CW, and believed to be related to the loading rates and hydraulic retention time (HRT). According to Grismer and Sherpherd³⁶, a smaller loading rate and greater HRT can influence the removal efficiency to reach nearly 100% provided that the recirculation system of effluent is applied. It was reported that CW have been proven to have the ability to reduce the concentration of tannin and lignin in wood waste leachate³⁷. The study reported that there was a total reduction of 42% of tannin and lignin using the CW planted with a cattail (Typha sp.). This is in agreement with Grismer and Sherpherd³⁶ who reported successful reduction of up to 77.9% of tannin and lignin from the winery process wastewater by subsurface flow wetland. Comparing both studies with this study, better removal of tannin and lignin was achieved on day 52 to day 75 of 82.9 ± 0.47%, which is higher than reported literature^36,37.

Removal of effluent colour in constructed wetland

The colour of wastewater is one of the parameters that must be treated. A high concentration of colour in wastewater has an impact on the environment by reducing the penetration of sunlight into the water body, and this is often caused by organic matters which include suspended and dissolved particles. Dissolved organic matters, such as humus, peat, and plant decay contribute to the yellow or brown colour of the wastewater, whereas as the dissolved organic acids including tannin and lignin give the wastewater a tea colour. According to Mohammed and Chong³⁵, there is a correlation between TSS, and tannin and lignin contents of POME, and the resulting effluent colour.

By reducing the concentration of TSS, and tannin and lignin in the POME FD, the colour concentration may also decrease. In this study, the colour started to reduce from day 3 to day 75. The colour concentration in this study is represented by platinum-cobalt (Pt–Co). Initially, the colour of POME FD recorded was 493 ± 3 Pt–Co (day 0–24), 1330 ± 31 Pt–Co (day 27–39), and 2501 ± 47 Pt–Co (day 46–75). Based on Fig. 6, the colour of treated POME FD reduced to 310 ± 16 Pt–Co (day 24), to 491 ± 16 Pt–Co (day 45), and finally to 238 ± 4 Pt–Co (day 75).

Overall, the CW managed to yield 57.5 ± 22.3% of colour removal, thus indicating the effectiveness of using P. purpureum in removing the colour³⁴. It is expected that the dissolved solids that add to the colour will be washed out by the CW. The presence of bacteria in the CW often also leads to a decrease in colour quality³².

Removal of ammonium nitrogen in constructed wetland

As the concentration of ammonium nitrogen (NH₄⁺-N) was high inside the POME FD (14.26 ± 0.04 mg/L), the CW was used to reduce its concentration. Figure 7 exhibits the ammonia removal after 75 days in the CW. The influents of the CW were different at 3.66 ± 0.01 mg/L, 14.26 ± 0.04 mg/L, and 8.92 ± 0.04 mg/L according to the date of sampling for every batch which was in July, August, and September, respectively. An extended HRT would lengthen the contact interval between microorganisms and organic matters in wastewater, thus increasing the efficiency of NH₄⁺-N, contrary to a shortened HRT which would inadequately remove the ammonia as a result of reduced time for nutrient uptake by the nitrifying bacteria, thereby slowing down the proliferation of autotrophic bacteria³⁸.

Even though the initial concentration had higher values, at the end of the treatment, the CW managed to reduce the concentration of NH₄⁺-N to 1.4 ± 0.4 mg/L, 3.8 ± 0.4 mg/L, and 1.2 ± 0.1 mg/L, of about 64.8 ± 9.6%, 56.6 ± 4.6%, and 86.2 ± 1.7%, respectively. The NH₄⁺-N concentrations were only fractionally declined within the first few days by 7.3 ± 0.4% due to a large amount of dissolved oxygen that has been consumed by the organic matters for degradation. NH₄⁺-N concentrations stabilised after 57 days, at 1.7 ± 0.2 mg/L, and on day 60 at 1.74 ± 0.04 mg/L. On days 63, 66, and 75, there were stable NH₄⁺-N concentrations at 1.3 ± 0.3 mg/L, 1.2 ± 0.1 mg/L, and 1.2 ± 0.1 mg/L, respectively. This suggests that there has been a combination of NH₄⁺-N sorption onto CW substrates, microbial assimilation, and nitrification at the air–water interface³⁹. NH₄⁺-N concentrations steadily declined as HRT increased, with removal performance recorded at 74.4 ± 0.4% and 84.2 ± 1.8%.

A study has reported the low reduction trend of hydrocarbon in the first 60 days, and greatly decreased towards the end of the experiment once the maturity phase has been achieved⁴⁰. This explains that the removal of NH₄⁺-N did not only rely on the adsorption by the CW substrates, but also on photosynthesis acclamation and root development of the plants, as well as sufficiency of oxygen supply to rhizosphere aerobic microorganisms¹³. In context of biogeochemical nitrogen cycle, there are three oxidative steps involved in converting ammonia to nitrate (NH₄⁺  → NH₂OH → NO₂⁻  → NO₃⁻)⁴¹. P. purpureum will absorb the nitrates and enhance its growth. In this study, the CW managed to reduce the concentration of NH₄⁺-N in POME FD by 62.3 ± 24.8%.

Effluent characteristics in the constructed wetland

Overall, the effluent characteristics after being treated in the CW showed a higher reduction in FD concentration as compared to its initial concentration. Table 2 shows the different initial concentrations of FD in July, August, and September due to the different seasons. In July, the concentration of FD was quite low as compared to the concentrations in August and September since it was a rainy season. Thus, the concentration of the FD was diluted by the rain. On the contrary, August and September were dry seasons, hence the concentrations of FD were quite high. However, even with different initial concentrations, 52.4 ± 8.2%, 66.9 ± 2.2%, and 85.2 ± 0.5% removal of COD was recorded in each three different months. Concentrations of TSS were reduced to 88.9 ± 2.9%, 90.5 ± 2.7%, and 96.4 ± 0.6% in those months, respectively. For the concentration of colour, September showed the highest removal (90.4 ± 0.18%), followed by 63.1 ± 0.36% in August. and 50.4 ± 4.57% in July. The ammonia concentration also showed an increment in removal in July, August, and September by 31.1 ± 3.8%, 72.9 ± 3.4%, and 86.2 ± 1.7%, respectively.

Overall, the CW managed to reduce the concentration of COD, TSS, tannin and lignin, colour, and ammonia in the FD. Approximately 62.2 ± 14.3%, 88.1 ± 13.3%, 57.5 ± 22.3%, 66.6 ± 13.2%, and 62.3 ± 24.8% of COD, TSS, tannin and lignin, colour, and ammonia were reduced from the effluent sample, as summarised in Table 2. All parameters met the river water quality except for COD. For future consideration, this study suggests an extended HRT to achieve a better efficacy in reducing the COD.

The designed CW treatment using P. purpureum showed a comparable removal with other studies, while some of the parameters showed better removal as shown in Table 3. A comparable COD removal efficiency of 66.2% was achieved, which was comparable to 66.7%⁴³ and 65.9%⁴⁴. For NH₄⁺-N removal, this study recorded 62.3% removal from POME FD, which is within the range of other studies (65.9%⁴⁸ and 67.2%⁴²). Interestingly, the TSS showed a greater removal (85%) as compared to a previous study (68%)⁴³.

Removal of trace elements in constructed wetland

The POME FD could be used as a source for nitrogen, phosphorus, potassium, magnesium, and calcium for plant growth⁴⁸. The elements that were present in the POME FD are listed in Table 4. In phytoremediation studies, grass is the most common plant that has been used as it has high tolerance towards heavy metals. For this study, P. purpureum was used in the CW owing to its extensive fibrous roots (Fig. 4) that could reduce leaching and erosion⁴⁹. All these characteristics give P. purpureum phytoremediation advantage where it can help to reduce the trace elements inside the POME FD. The phytoremediation concept managed to reduce the concentrations of potassium, silica, caesium, rubidium, sodium, copper, magnesium, calcium, strontium, and manganese in POME FD by 82.19%, 80.51%, 71.17%, 66.07%, 58.46%, 59.27%, 56.81%, 35.80%, 35.57%, and 20.81%, respectively, as shown in Table 4.

The concentration of silica in POME FD was quite high (25.51 ± 1.05 ppm) before treatment. However, after treatment with CW, the concentration of silica was reduced by 80.51%. This result shows the capability of P. purpureum to remediate silica by absorbing and immobilising it. Silica will be uptaken by the plant, and processed to form phytoliths or silica bodies which infill the cell walls and lumina of certain plant cells⁵⁰. The uptake of silica by the plant improves its quality by showing increased chlorophyll content and plant growth⁵¹. On the other hand, the plant in the CW also shows a rapid growth in “Plant physiology” section. Based on this, it can be concluded that the silica uptake by the plant managed to increase the chlorophyll content and growth of P. purpureum while at the same time reduced the concentration of silica inside the POME FD. It has been proven that plants are capable to reduce and remediate caesium by 59.3% through accumulation at the dry shoots, which is similarly shown in this study with caesium removal of 71.17% by P. purpureum in the CW⁵².

Plant physiology

The growth performance of P. purpureum is measured in term of height, of as shown in Fig. 8. The initial heights of P. purpureum were 92 cm and 86 cm for treatment with POME FD and control with rainwater, respectively. On the last day of experiment, the height of P. purpureum treated with POME FD increased by 135 ± 12%, but only 88 ± 9% in control. This is due to the uptake of nutrient from the water sources⁵³. A study⁵⁴ reported that plant height increment of 61 ± 2% can be observed when it was treated with POME FD. As mentioned in “Removal of trace elements in constructed wetland” section, silica increases the growth of the plant, while potassium functions as the major osmoticum of the plant cells which controls cellular expansion⁵³.

Bacterial community analysis

The other component of wetlands such as microorganisms have been proven to help on the reduction of the pollutants or contaminants in a CW, such as ammonium, nitrate, organic matters, industrial organics, and emerging organic contaminants/pollutants^15,16,17,18. The submerged plant litter did affect the soil properties and microbial diversities in a CW⁵⁵. As shown in Fig. 9, Proteobacteria yielded the highest percentage in the phylum, ranging from 37 to 70.7%. Proteobacteria is one of the three most abundant soil bacteria phyla, and the most widespread and ubiquitous. These phyla are found dominant in wastewater treatment^23,56. It was reported that Proteobacteria had a high resistance towards heavy metals⁵⁷. The Proteobacteria and Firmicutes phyla can live together in harsh environments, and have been shown to have the potential to remove a broad-spectrum of heavy metals⁵⁸. Proteobacteria are copiotrophic and root-associated bacteria that influence NO₃⁻-N and NH₄⁺-N in the soil⁴⁵. Proteobacteria do not show a consistent change in depth^59,60. Proteobacteria including Nitrosomonas, Nitrosococcus, and Nitrosospira have been reported as ammonia-oxidising bacteria⁶¹.

Chloroflexi is a phylum of primarily gliding, filamentous bacteria possessing a wide diversity of metabolisms and ecological roles, but are best known as photoheterotrophs⁶². Chloroflexi percentage at the beginning of the system was low (0.96%) but increased thereafter to 18.11%. Wetland soils are characterised by high soil moisture, and this habitat may be suitable for some bacterial phyla such as Chloroflexi to grow⁶³. Since it is photoheterotrophic, it cannot use CO₂ as a carbon source, but use the organic compounds instead, thus helping in the removal of pollutants. Chloroflexi are aerobic and anaerobic thermophiles, filamentous anoxygenic phototrophs, and anaerobic organohalides⁶⁴. Since this phylum is anaerobic, it can survive in the CW even though the system does not have aeration.

Acidobacteria in this study also increased from 0.14 to 5.2% which shows the capability of these phyla to survive and adapt in the CW condition. These phyla have a unique ability including the ability to use nitrite as a nitrogen source, which gives positive impacts to soil macro- and micronutrients, and these phyla can also express multiple active transporters.

Cyanobacteria are phyla that contain members that possess two photosystems that are coupled in series to perform oxygenic photosynthesis. Cyanobacteria occupy a privileged position among microorganisms because of their role in carbon and nitrogen cycles⁶⁵. These phyla also showed an increment in abundancy in the CW after treatment for 102 days. The phyla increased from 0.39 to 5.29%, which gives a clear indication that Cyanobacteria did involve in the pollutant reduction.

Based on Fig. 10, Anaerolineaceae, Cyanobacteria, Acidobacteria, and Nitrosomonadaceae were detected in the samples of POME FD and treated POME FD. These four genera showed an increment in the CW after the treatment of POME FD for 75 days. Given the fermentative metabolism-capability to consume carbohydrates and proteinaceous carbon sources under anaerobic conditions, Anaerolineaceae growth had increased from 0.67 to 13.21%, signaling its participation in COD and ammonia removals⁶⁶.

Cyanobacteria also showed an increment in the CW up to 93.9%. In the beginning, it was only 0.3%, but at the end of the system, it increased to 4.92%. This also shows the effectiveness of this genus to grow in this condition, and aid in the removal of pollutants in the POME FD. Cyanobacteria may affect the abundance of several genera across several phyla, in which most are associated with nitrogen metabolism. For example, the increase in the abundance of Cyanobacteria under nitrogen addition may be due to its great demand for nitrogen. Cyanobacteria cells are composed of nitrogen, nitrate and ammonium, which are common sources of nitrogen for Cyanobacteria. Cyanobacteriaare facultative bacteria.

At the beginning of the experiment, Acidobacteria only showed 0.06% abundance, but after 102 days, the value increased to 2.44%. This also shows the ability of this genus to grow in the CW, which suggest its association in the removal of pollutants in the POME FD. Meanwhile, Nitrosomonadaceae, which belongs to β-proteobacteria, is also an ammonium-oxidising bacteria involved in the nitrification process⁶⁷. This genus’ increment from 0.07 to 1.1% provides strong proof that nitrification had occurred in the CW which led to the decrement of ammonia nitrogen and total nitrogen inside the POME FD.