Journal of Environmental Chemical Engineering 9 (2021) 105213

Contents lists available at ScienceDirect

Journal of Environmental Chemical Engineering

journal homepage: www.elsevier.com/locate/jece

Multipurpose treatment of landfill leachate using natural coagulants –

Pretreatment for nutrient recovery and removal of heavy metals
and micropollutants
Ilaria Righetto a, b, *, Raed A. Al-Juboori a, Juho Uzkurt Kaljunen a, Anna Mikola a
a
Water and Environmental Engineering Research Group, Department of Built Environment, Aalto University, P.O. Box 15200, Aalto, FI-00076 Espoo, Finland
b
Department of Environment, Land and Infrastructure Engineering, Politecnico di Torino, Corso Duca degli Abruzzi, 24, 10129 Torino, Italy

A R T I C L E I N F O A B S T R A C T

Editor: Dr. GL Dotto Landfill leachate is considered to be a potential source for nutrient recovery due to high ammonia concentration.
However, it contains concerning pollutants such as metals and micropollutants in high levels. The coagulation-
Keywords: flocculation process is commonly used as a pretreatment technique to remove such pollutants. Plant-based
Natural coagulant materials have recently been explored as an alternative for conventional coagulants as the latter leave re­
Landfill leachate
siduals that require further treatment. The use of natural coagulants for nutrient recovery purposes has not been
Nutrient recovery
investigated previously. This study evaluated the effectiveness of tannin-based product named HTH to improve
Micropollutants
Metals landfill leachate quality to suite the requirements for nitrogen recovery using membrane contactor technology.
Sludge characterization The optimal combination of dosage and pH was determined using Response Surface Methodology (RSM). The
parameters selected to evaluate the treatment were total nitrogen, total phosphorus, total organic carbon, and
ammonia. The obtained optimized combination of dosage and pH were 11.1 mL/L and 7.3, respectively. This
combination resulted in respective removal of TP, TOC and TN of 47%, 15%, and 4%, while NH3‒N increased by
20%. The potency of the treatment for removing micropollutants and metals was evaluated. Highest metals
removal was attained for Fe, Ti, Cr, Al, Ba and V, while other metals were removed in smaller percentages.
Decent removal of cationic surfactants, phenols and pharmaceuticals such as 5-methylbenzotriazole, ibuprofen,
carbamazepine, and metoprolol was also achieved with the treatment. The results of this study suggest that
tannin-based materials are effective coagulants for treating landfill leachate with the goal of preparing it for
nitrogen recovery.

1. Introduction Landfill leachate quality can vary, depending on the geographic

location, the disposed of waste, and the age of the landfill. However, the
The sanitary landfill is considered the most cost-effective and safest main classifications for compounds found in leachate are organic matter,
method for ultimate waste disposal. When combining the population humic substances, ammonia-nitrogen, heavy metals, organic and inor­
growth with the increasing waste generation, the need for well-designed ganic salts, halogenated compounds, phenols and pesticides and other
and functioning landfill sites is needed together with a good monitoring micropollutants [3,4]. The presence of these compounds is a potential
system [1]. Even though waste is decomposed in a controlled way in threat to water sources, aquatic life, soil and human health [3–5]. For
landfills, methane and leachate generation must be monitored and instance, endocrine disruptor chemicals (EDCs) that have been detected
efficiently collected to avoid environmental problems. Methane is pro­ in landfill leachate such as polycyclic aromatic hydrocarbon, phthalic
duced by waste fermentation, while leachate forms when water perco­ acid esters (PAEs) and others can interfere with the production and
lates through the landfill site. The production of methane and leachate is functionality of hormones in our bodies causing disorder in the organ
expected to continue even after a landfill site is closed. Therefore, good system that could result in serious diseases [6]. Similarly, the high
management of post-closure operations is always required [2]. alkalinity, anions, cations, and organic matter of the leachate can

* Corresponding author at: Water and Environmental Engineering Research Group, Department of Built Environment, Aalto University, P.O. Box 15200, Aalto, FI-
00076 Espoo, Finland.
E-mail address: [email protected] (I. Righetto).

https://doi.org/10.1016/j.jece.2021.105213
Received 6 December 2020; Received in revised form 2 February 2021; Accepted 9 February 2021
Available online 14 February 2021
2213-3437/© 2021 Elsevier Ltd. All rights reserved.
I. Righetto et al. Journal of Environmental Chemical Engineering 9 (2021) 105213

disturb the ecological systems in the neighboring areas to the landfill smaller environmental impacts. Sludge produced when using natural
sites [4]. The release of high dissolved solids to surface water bodies can coagulants in the coagulation-flocculation process can be used for soil
negatively impact light absorbing leading to stratification. In a similar amendment or as recycled coagulants [31–33].
fashion, the high concentration of organic matter can contribute to Natural coagulants are based on starch, tannin, seed extract, and
greenhouse gas emission, while elevated concentration of cations can guar gum can easily be modified to be more effective. Ahmed et al. [34]
affect soil cation exchange capacity. Biological systems are commonly compared seven different natural coagulants with alum, ferric chloride,
applied for landfill leachate treatment. However, the scarcity of organic and PAC, and measured COD, color; turbidity, TSS, heavy metals, and
compounds and the presence of toxic compounds (PAH, phenols, halo­ NH3‒N for all of them. It was reported that some of the natural co­
gens) hinder the microbial activity governing the biological process [6, agulants were more efficient than traditional coagulants. For example,
7]. In addition to biological processes, nutrient recovery techniques can the lowest COD removal was achieved with PAC, while tobacco leaf had
be applied to take off part of the contamination load of landfill leachate the highest COD removal. TSS removal was higher for all seven natural
in a sustainable way. The relatively high nutrient content in stabilized coagulants compared to alum and ferric chloride. Another study con­
landfill leachate makes it a valuable wastewater stream to which ni­ ducted by Cheng et al. [35] explored the use of guar gum as a coagulant
trogen and phosphorus recovery processes can be applied. Some of the to treat landfill leachate; they reported 22.6% COD removal when
nutrient recovery processes applied to landfill leachate are magnesium working in optimal conditions. Few other studies report the effective­
ammonium phosphate precipitation, microbial electrolysis cells com­ ness of natural products when used as coagulants or coagulant aids when
bined with forward osmosis, and a membrane contactor [8–11]. The treating landfill leachate, reporting mainly COD, color, SS, turbidity,
latter technique is of special interest to this study. and ammonia removal [36–38].
When considering ammonia extraction using gas-permeable mem­ When considering coagulation-flocculation as a pre-treatment prior
brane technology, the concentration of solids in the wastewater stream to a membrane contactor for nitrogen recovery, ammonia removal needs
can cause membrane fouling and hinder nitrogen recovery. However, to be minimized. It appears that there are no studies that explored the
due to the low SS concentration in landfill leachate, the membrane possibility to minimize ammonia removal using natural coagulants for
fouling possibility is reduced, leading to high nitrogen recovery effi­ such a purpose. Moreover, total phosphorus and organic carbon have
ciencies [11]. On the other hand, landfill leachate usually contains not been addressed in existing studies. For nutrient recovery processes,
substances such as organic matter, phenols, heavy metals, and surfac­ these are important parameters to be evaluated in order to accumulate
tants. These substances need to be removed using a pre-treatment phosphorus in the sludge for further extraction and prevent membrane
method to avoid interaction between the membrane and these com­ fouling by organic carbon. This study is designed to find a sustainable
pounds, which may lead to contamination of the final recovered prod­ alternative to landfill leachate coagulation-flocculation pretreatment
uct. Some of the possible pre-treatment methods are biological methods that also supports the needs of the subsequent nitrogen recovery process.
(aerobic or anaerobic biological degradation), physical/chemical The effective tannin-based coagulant named HTH was selected (based
methods (activated carbon adsorption, advanced oxidation processes, on screening tests) and tested for pretreating Ämmässuo landfill
the coagulation/flocculation process), or a combination of those [3, leachate for nitrogen recovery using a membrane contactor. To optimize
12–15]. The main drawback of biological processes, especially when two factors, dosage and pH, Response Surface Methodology (RSM) was
treating old landfill leachates, is the interaction between recalcitrant utilized. To evaluate the effectiveness of the treatment, the final con­
organic substances and the microbial activity [1,16]. The phys­ centration after treatment of four parameters was measured. These pa­
ical/chemical processes’ main disadvantages are the complexity of the rameters were TP, TOC, TN, and NH3‒N. The latter two parameters were
equipment and high operational cost [2,7]. In addition, ammonia is set to be maintained at the highest level to maximize nitrogen content in
reported to be removed with the aforementioned technologies [17], the subsequent recovery step, while the other two were set to be mini­
particularly when using a combination of those. Therefore, these are not mized to avoid membrane fouling and to accumulate phosphorus in the
suitable pre-treatment technologies for nitrogen recovery with a mem­ sludge for later recovery. Two scenarios were considered: scenario A,
brane reactor. both pH and dosage were set to be optimized within the range,
Coagulation-flocculation is considered a low-cost and effective so­ considering only the target set for each parameter (maximized/mini­
lution when targeting specific compounds in landfill leachate. Many mized), while in scenario B, pH was set to the maximum due to the need
studies have reported the effectiveness of traditional aluminum and iron for an alkaline environment in the nitrogen recovery unit, and the
salts to treat landfill leachate [18–20]. The application of metal co­ dosage was set to the minimum to minimize coagulant consumption.
agulants for landfill leachate treatment has recently witnessed an evo­ Samples collected from the optimum scenario were analyzed in order to
lution represented by the emergence of hybrid coagulants and the in-situ evaluate the effectiveness of the treatment on the removal of heavy
combination with electro-chemical and oxidation processes. For metals and micropollutants, which has only been investigated in a
instance, a recent study conducted by Mohd-Salleh et al. [21] found that limited number of studies using a synthetic sample matrix. The physical
a composite of polyaluminium chloride (PAC) and tapioca peel powder characteristics of the produced sludge with tannin were also determined.
was effective for treating landfill leachate. Peroxicoagulation is a good The objective of this study was to evaluate the effectiveness of the
example for the integrated process of coagulation, electrolysis, and suggested pretreatment in providing a multi-purpose process step for
oxidation. This process has successfully been applied for improving landfill leachate treatment – enhancing the nitrogen recovery while
landfill leachate quality especially at acidic pH [22,23]. Despite the removing micropollutants.
recent development in metal coagulants applications, they still have
significant shortcomings such as the concerns associated with the 2. Materials and methods
toxicity of Al to aquatic life especially in high concentrations [24,25]
and the high capital and operational cost of integrated processes [26]. 2.1. Ämmässuo landfill leachate
Therefore, in recent years, natural coagulants have been recognized as a
valuable alternative to traditional coagulants. Indeed, some of the ad­ Ämmässuo landfill is situated in the city of Espoo (Finland), and it is
vantages are their availability, biodegradability, and less voluminous one of the largest landfills in northern Europe. It is owned by the Hel­
and biodegradable sludge produced [27–29]. A recent study conducted sinki Region Environmental Authority (HSY). It is divided into the old
by Carlqvist et al. [30] compared the impact of tannin coagulant pro­ landfill area (which operated from 1987 and 2007 and covers about
duction with other traditional coagulants by performing a life cycle 52 ha) and the new landfill (in active use from 2007 to April 2014,
assessment (LCA). It was reported that the production of tannin co­ covering an area of 13 ha). There are approximately 10 million tons of
agulants in the Scandinavian countries could have comparable or even waste disposed of in the old landfill while it was in operation. This

2
I. Righetto et al. Journal of Environmental Chemical Engineering 9 (2021) 105213

landfill collected waste and disposed of it in the Helsinki metropolitan inorganic complexes, but they are totally biodegradable. The extract is a
area, which has about one million residents. The leachate that is polyphenolic compound modified by cationization, which is a reaction
generated is sent to the Suomenoja WWTP for treatment. The efficiency with formaldehyde and ammonium chloride, also known as the Mannich
of the leachate collection is estimated to be around 90%. It is anticipated reaction [40]. The positively charged modified tannin-based products
that leachate will no longer be generated after 2026 [39]. At the are obtained by adding a tertiary ammonium group to the ionized
Ämmässuo landfill site, there are different collection points for leachate. phenols. This combination is effective for destabilizing negatively
The collection point chosen for this study was T2. Based on information charged particles found in wastewater [41]. The polymeric structure is
provided by HSY, this collection point has the highest NH3‒N content. involved in the bridging mechanisms in the coagulation process [42].
This point is located in the old landfill area. Table 1 shows the The possible products coming from this modification process are sug­
Ämmässuo leachate characteristics measured in the lab from the sam­ gested by Lugo et al. [40]. The reaction is illustrated in Fig. 1. Consid­
ples collected from batch tests during February and March 2020. Sam­ ering the carbonaceous structure of tannin-based coagulant, its
ples were also sent to external labs (MetropoliLab in Helsinki and biodegradable nature [43] and the fact of being used for treating
Eurofins Lab in Lahti) and analyzed for a wide list of pollutants and wastewater with high organic load, its residual in this case does not
metals. However, only the compounds that were detected are reported impose serious environmental or health risks and hence it has not been
in Table 1. The standard methods for the measurements are reported in measured in this study [44,45].
Section SM.1.
2.3. Design of experiments and statistical analysis
2.2. Coagulant characteristics
Design Expert software was used to perform RSM, which is an effi­
The tannin-based product selected for this study was provided by cient technique for modeling response behavior as a function of exper­
Haarla Oy. According to the provider’s information, HTH is produced by imental factors. In this study, pH and dosage were selected as factors to
combining extract from the bark of the Acacia mearnsii tree with be investigated, as they have been selected in many optimization studies
for coagulation-flocculation experiments [16,42,46,47]. The use of RSM
for two-factors study such as ours is deemed suitable and commonly
Table 1
Landfill leachate parameters measured in the lab*; pollutants analyzed by applied in the literature. In fact, screening is normally applied when
MetropoliLab**; metals analyzed by Eurofins Lab***. there is higher number of factors prior to RSM optimization in order to
reduce the number of experimental runs [48,49]. TN, TP, TOC, and
Constituents Value* Unit
NH3‒N are the six responses measured to evaluate the treatment per­
TN 1019 mg/L formance. These responses were selected because of their relevance to
TP 7.74 mg/L
nutrient recovery using membrane contactor technology. For instance,
TOC 1189 mg/L
SS 20 mg/L TN and NH3‒N are measured as the main constituents to be maintained
VSS 25 mg/L at the lowest level of removal. TP needs to be removed for further
NH3‒N 536 mg/L extraction and recovery. TOC is indicative of the compounds causing
pH 7.95 –
membrane fouling that decreases the overall ammonia transfer rate
Constituents Value** Unit
Ca 110 mg/L
across the membrane [50].
Mg 83 mg/L Models generated for a single response are based on experimental
K 550 mg/L data collected in the lab and thus fit a quadratic model (Eq. (1) [51]).
Na 1100 mg/L The complexity of the model depends on the significance of the terms for
Al 310 μg/L
each response. For example, when the quadratic (x2) and interaction
Ba 310 μg/L
B 7200 μg/L (xixj) terms are insignificant, the model can be contracted to the linear
Cd <5 μg/L form [51].
Cr 260 μg/L
Cu < 20 μg/L ∑k ∑ ∑ ∑
k k k

Pb < 60 μg/L y = β0 + β xi
i=1 i
+ βii x2i + βij xi xj + ε (1)
Mn 450 μg/L i=1 i=1 j=i+1

Ni 140 μg/L
Si 24,000 μg/L where y is the predicted response, x represents the factors, k is the
Fe 10,000 μg/L number of factors, β0 is the constant coefficient, and βi, βii, and βij are the
S 30,000 μg/L regression coefficients of linear, quadratic, and interaction terms,
Zn 18 μg/L
Sr 720 μg/L
respectively [52]. To select the amount of experimental data to be
Ti 360 μg/L collected, Central Composite Design (CCD) was applied [53]. CCD is the
U < 1000 μg/L most popular class of designs used for fitting second-order models. It is
V 36 μg/L well suited for fitting quadratic surfaces and thus it is a very efficient
Constituents Value*** Unit
design tool used for process optimization [54]. This design defines 2k
Surfactants
Anionic surfactants 1.7 mg/L corner points, 2k axial points, where k is the number of independent
Non-ionic surfactants 2.5 mg/L variables (or factors) selected, and a center point as it is shown in Fig. 2
Cationic surfactants 6.2 mg/L [55]. In this study, two factors (dosage and pH) were selected, which
Pharmaceuticals resulted in 4 axial points, 4 corners points and 1 central point. Each axial
5-Methylbenzotriazole 3.2 μg/L
Phenazone 2.3 μg/L
and corner point was duplicated, and the center point was repeated five
Ibuprofen 65 μg/L times. This gives a total of 21 experiments, which are a combination of
Carbamazepine 1.5 μg/L different dosage and pH levels defined by the coded values calculated by
Ketoprofen 6.3 μg/L applying the formulas in Table SM2. The center point was repeated 5
Metoprolol 0.6 μg/L
times to obtain good predictions throughout the region of interest and
4-Nonylphenol 3.7 μg/L
Tributyltin 0.0099 μg/L identify the variance range of the data. It is indeed recommended to
PAH EPA 16 compounds μg/L repeat the center point 3–5 times [54]. The coded value associated with
Anthracene 0.087 μg/L α is representative of the rotatability of the model and it is defined as α =
Naphthalene 0.42 μg/L √ ̅̅̅̅̅
2k , which in this case is equal to 1.414, which represents the distance
4

3
I. Righetto et al. Journal of Environmental Chemical Engineering 9 (2021) 105213

Fig. 1. Possible products of the Mannich reaction with Acacia tannin [40].

coagulation/flocculation parameters were chosen based on a thorough

literature review survey as reported in Table 2 [13,42,59–61]. The rapid
mixing was selected to be 250 rpm for 60 s, followed by slow mixing at
50 rpm for 15 min, and a final settling time of 40 min 2 M H2SO4 and
4 M NaOH solutions were used to adjust the pH of the samples. Sludge
analysis were conducted in order to characterize the produced sludge.
Two parameters were used as indicators of the physical properties of the
sludge: sludge volume index (SVI) and sludge filterability by calculating
cake resistance. SVI was calculated applying Eq. (2) [62]:
1000∙V30
SVI = [ml/g] (2)
MLSS

where, V30 is the volume of the sludge settled after 30 min in mL/L and
MLSS is the mixed liquor suspended solids in mg/L.
Sludge filterability was determined following the method proposed
in [63]. The cake resistance was calculated using Eq. (3).
Fig. 2. Schematic representation of CCD adapted from Montgomery [54].
2∙A2 ∙P∙b
r= [m/kg] (3)
c∙μ
of the point α from the center. It is possible to calculate coded values for
each data point in the same way since the region of interest is defined
where A is the area of the filter paper (m2), P is the vacuum pressure
(Xmin and Xmax) [51,56,57]. Analyses of variance (ANOVA) was per­
applied (Pa), c is the mass of solids deposited onto the filter paper per
formed to evaluate the significance of the responses’ change with the
unit volume of filtrate passed through the paper (kg/m3), μ is water
main, quadratic, and interaction effects of the factors.
viscosity (Pa s), and b is the slope of t/V vs V plot in (s/m6) [63].

2.4. Process optimization

3. Results and discussion

Single responses were modeled, and 3D surfaces were generated

3.1. Experimental results
together with contour plots. This study aims to produce a combination of
dosage and pH that could satisfy the requirements set for each response.
Based on screening tests, HTH was selected among five natural co­
Therefore, the desirability function was used to predict the optimal
agulants (2 tannin-based and 3 starch-based) to be used to treat landfill
combination of dosage and pH that results in the best treatment
leachate. Existing studies report the effectiveness of tannin-based co­
outcome. The desirability function requires setting a target for each
agulants in a slightly acidic environment [42,64]. Moreover, when pH
response (minimized or maximized) along with the importance ranking
[57]. Setting the importance ranking depends on their relevance to the
overall goal of the treatment, the extent of their change, and the good­ Table 2
ness of the fit of their generated models. The function results in a nu­ Coagulation flocculation parameters from literature.
merical value varying between 0 and 1. The closer the value to 1, the
Study Rapid Rapid Slow Slow Settling
better it is [58]. A combination of dosage and pH is then provided mixing mixing mixing mixing time (min)
together with the predicted final concentration for each response. These (rpm) (min) (rpm) (min)
predicted values are then verified with validation experiments. Verifi­
Guo et al. 250 2 80 15 60
cation experiments represent the ultimate evaluation for the validity of [13]
the generated statistical models. Banch et al. 250 15 60 30 30
[42]
binti 120 3 50 20 60
2.5. Laboratory analysis and equipment
Saharudin
et al. [59]
All analyses were conducted at room temperature (25 ◦ C). Hameed et al. 200 1 60 5 10
Coagulation-flocculation experiments were carried out by using a mix­ [60]
ing apparatus composed of a set of six impellers in series and six 1 L Razali et al. 200 2 35 10 15
[61]
beakers (Kemira Flocculator 2000) filled with 900 mL of leachate. The

4
I. Righetto et al. Journal of Environmental Chemical Engineering 9 (2021) 105213

was adjusted lower than 6, excessive foam formation was observed. Supplementary Materials in Section SM.4 (Figs. SM3–SM5). The graph
Foam formation observed in this study is likely to be associated with the shown in Fig. 3(a) is used to check the normal distribution of residuals. It
decomposition of HCO−3 , CO2− 3 and organic foaming agents (e.g. sur­ can be seen that all data points fall on a straight line, indicating a normal
factants) under acidic environment [65–67]. This is reasonable given the distribution of residuals [71]. Fig. 3(b) shows that all points are
high alkalinity (5500 mg/L CaCO3 as provided by site operators) and randomly scattered around the 0 y-axis, so the residuals’ variance is
surfactants concentration (as shown in Table 1) of the leachate sample considered constant (homoscedasticity). In Fig. 3(c), residuals vs. run
used in this study. Thus, pH lower than 6 could not be investigated in points do not follow a specific pattern, which suggests that responses are
this study. Considering that the purpose of the pre-treatment is to not dependent on the order of runs [56,57]. Lastly, Fig. 3(d) illustrates
recover nitrogen in the form of ammonia, an alkaline pH is favorable. that predicted values vs. actual values lay close on the straight line,
However, alkaline pH did not result in satisfactory results (see Section which indicates good predictability of the model [72]. With all the
SM.3). This is likely due to the effect of pH on tannin charge density, presented results in Figs. 3 and SM3–SM5, it can be said that the re­
which decreases as pH increases [29]. When charge density decreases, it quirements were fulfilled for all selected responses.
negatively affects charge neutralization and bridging mechanisms. Table 4 shows p-values from the ANOVA results. With a 95% confi­
Therefore, the effective pH range was selected to be 6–8. Figs. SM1 and dence level, the significance of the models can be checked by analyzing
SM2 demonstrates with photos the effect of pH on treatments within and the p-values. When its value is less than 0.05, it means that either one or
outside the selected range. In addition to the ineffectiveness of coagu­ both factors (pH and dosage) have a significant effect on the response.
lation process outside the selected pH range, studying such a narrow The p-values related to linear (A), quadratic (A2, B2), and interaction
range can be justified by the significant effect of pH on (AB) terms are analyzed to understand the effect of factors on the model
ammonia-ammonium equilibrium in the treated leachate which is for each response [46]. When examining the ANOVA table, it can be seen
important for the subsequent nitrogen recovery. Eq. (4) shows that a that the p-values for TOC, TP, and NH3‒N models are < 0.05, which
change of one pH unit can result in ten times higher conversion of means that these factors have a significant effect on the generated
ammonium to ammonia [68,69]. Moreover, the fitting of nonlinear model. However, the p-value for TN is > 0.05. Therefore, the model is
functions of RSM works better with narrow ranges [70]. Similar narrow considered not significant. It can be seen that some of the responses’
range of pH has also been applied in previous coagulation optimization models include only the linear terms, whereas others include both the
studies using RSM [46]. Based on preliminary testing, the selected linear and quadratic terms. The inclusion and exclusion of the models’
dosage range was found to be 5–20 mL/L. The treatment was less terms are determined through a built-in model-fitting function in Design
effective when the coagulant was dosed outside this range, which could Expert that suggests the best model for each response. The goodness of
be attributed to overdosing that causes particle restabilization, as fit for the responses’ models will be scrutinized further in the following
observed in other studies [36,42,61]. After defining the region of in­ section by applying several criteria.
terest for each factor, uncoded values were calculated for all 21 exper­
iments, as explained in Section 2.3. These values, together with the
results obtained from batch tests for all responses, are reported in 3.3. Empirical models for individual responses
Table 3.
The equations of the models in coded values (A = dosage, B = pH) for
[NH 4 − N]
(4) both trials are presented in Table 5. Several determining statistical pa­
[NH 3 − N] =
10pK NH4 + 1 rameters for the goodness of the fit of the models are presented in
Table 6. The coefficient of variance (CV) is defined as the ratio of the
3.2. Statistical analysis standard error of estimate and the mean value of the observed response.
A model with a CV less than 10% is deemed to be reproducible [46]. In
Before performing ANOVA and describing the generated 3D models, this case, all the responses’ models can be reproduced. R2 is the coeffi­
it is important to demonstrate through diagnostic checks that residual cient of determination; it represents the percentage of variation of re­
errors are normally distributed and independent of time and order of sponses explained by the adopted model. The closer it is to 1, the better
runs. Fig. 3 shows the diagnostic plots for the TN response as an the model is. In this study, R2 was high enough to indicate a small
example. The diagnostic checks for other responses are provided in variation in the predicted responses from the actual ones except for TN.

Table 3
Experimental design and results from batch experiments.
Runs Dosage (coded) pH (coded) Dosage mg/L (uncoded) pH (uncoded) TN (mg/L) TP (mg/L) TOC (mg/L) NH3‒N (mg/L)

1 − 1 1 7.2 7.7 810 4.65 1060 654.1

2 − 1.414 0 5 7 1059 4.59 888.4 529.8
3 0 0 12.5 7 1076 3.52 860.9 514.1
4 − 1 1 7.2 7.7 1062 4.69 1074 615.8
5 0 0 12.5 7 1080 3.65 879.6 511.9
6 − 1 − 1 7.2 6.3 1069 4.71 574.4 451.9
7 0 0 12.5 7 1072 3.84 917.4 527.5
8 − 1 − 1 7.2 6.3 1065 4.77 575.2 395.5
9 0 0 12.5 7 1070 3.77 888 490.3
10 1 1 17.8 7.7 1069 3.92 1086 553.1
11 1 − 1 17.8 6.3 1083 3.15 582.5 442.3
12 1 1 17.8 7.7 1067 3.75 1089 539
13 0 − 1.414 12.5 6 1050 4.34 489.1 482
14 − 1.414 0 5 7 1028 4.88 928.9 490.3
15 0 − 1.414 12.5 6 1061 4.04 488.3 438.5
16 1 − 1 17.8 6.3 1068 3.67 554.9 387.1
17 1.414 0 20 7 1068 3.75 903.9 473.8
18 1.414 0 20 7 1064 3.79 940.5 514.1
19 0 1.414 12.5 8 1045 4.23 1115 550.7
20 0 0 12.5 7 1055 3.86 862.9 442.3
21 0 1.414 12.5 8 1015 4.21 1083 527.5

5
I. Righetto et al. Journal of Environmental Chemical Engineering 9 (2021) 105213

Fig. 3. Diagnostic plots for TN response: (a) normal % probability vs. residuals; (b) residuals vs. predicted; (c) residuals vs. run order; (d) predicted vs. actual.

Table 4 Table 5
Summary of responses’ significance values extracted from the ANOVA table for Model equations.
the measured responses. Response Model equation
Source TN p-value TP p-value TOC p-value NH3‒N p-value
TN 1061.3 + 5.8⋅A − 6.1⋅B
Model 0.1102 < 0.0001 < 0.0001 < 0.0001 TP 3.7 − 0.4⋅A + 0.05⋅B + 0.1⋅AB + 0.25⋅A2 + 0.2 * B2
A-Dosage 0.1402 < 0.0001 0.5869 0.1524 TOC 881.8 + 4.2⋅A + 234.3⋅B + 6.7⋅AB + 9.3⋅A2 − 51.5⋅B2
B-pH 0.1231 0.3052 < 0.0001 < 0.0001 NH3‒N 501.5 − 15.1⋅A + 56.8⋅B
AB 0.0831 0.5430
A2 0.0007 0.3386
B2 0.0017 < 0.0001 Adequate precision (AP) is the signal-to-noise ratio; it compares the
range of predicted values to the prediction error. When AP is greater
than 4, the model is characterized as adequate [73]. Another important
parameter is the lack of fit (LOF), which describes the distribution of
data in the model and how well the data fit the model. If the value is

6
I. Righetto et al. Journal of Environmental Chemical Engineering 9 (2021) 105213

Table 6 3.4. Single-response optimization

Statistical parameters related to all selected responses.
Response Std. Dev. Mean C.V.% R2 AP LOF Fig. 4 shows 3D plots generated with Design-Expert software. When
considering the initial concentration of TN (1019 mg/L), it can be seen
TN 15.01 1061.33 1.41 0.2173 4.181 0.0319
TP 0.19 4.08 4.58 0.8874 12.656 0.1159 that the measured value after the treatment was always higher. In
TOC 30.22 849.61 3.56 0.9853 41.020 0.0035 general, an increase in TN concentration with a pH drop was observed.
NH3‒N 40.40 501.49 8.06 0.6527 10.516 0.0398 The higher TN concentration could be attributed to the dosage applied,
as nitrogen is present in the tannin’s structure [40]. The observed in­
crease in nitrogen concentration was small, and it is not considered a
greater than 0.05, then the LOF is insignificant, which means that the
negative outcome of the treatment in the case of this study. TP con­
data are well-represented in the model. In this case, only the LOF for TP
centration followed a non-linear trend with both dosage and pH. At high
is insignificant, while other parameters’ LOF is significant.
dosage, TP concentration decreased with decreasing pH. However, the
When reviewing all the assessment criteria for the goodness of fit, it
trend is different at low dosage, as the minimum P concentration was
seems that the models fulfill most of the criteria and only failed some.
observed in the middle of the pH range. This could be due to the effect of
However, to improve the reliability and reproducibility of the work, the
pH on P solubility. These results are in agreement with the observations
statistical models were not used for predicting data points outside the
reported in Hameed et al. [60] and Turunen et al. [74]. It can be noticed
tested region of the factors. Furthermore, the desirability function was
that TOC removal does not seem to be significantly affected by dosage,
used to estimate the optimum factor levels for achieving the best multi-
but it is highly affected by the pH [74]. The increase in TOC concen­
response. The goodness of fit of each model is taken into account when
tration with the increasing pH might be caused by the carbon residues
setting the importance ranking for each response in the desirability
associated with the tannin structure, which contains mainly carbon
function (Section 3.5).

Fig. 4. 3D surface response plots for (a) total nitrogen, (b) total phosphorus, (c) total organic carbon, and (d) ammonia nitrogen.

7
I. Righetto et al. Journal of Environmental Chemical Engineering 9 (2021) 105213

[75]. As explained earlier, the higher pH negatively impacts the charge lowest possible coagulant consumption while fulfilling the target and
density of tannin and consequently hinders its interaction with water importance ranking requirements.
constituents. As the pH decreased, TOC decreased significantly, espe­ Fig. 5 shows the results of the desirability function for the two sce­
cially as it approached the enhanced coagulation region (pH ca. 6.0) narios. The translation from coded values to uncoded values for dosage
[76]. The 3D plot of ammonia shows that the NH3‒N concentration and pH reported in Fig. 5 are as follows: Scenario A produced a dosage of
dropped noticeably as the pH decreased from 8 to 6. Moreover, the in­ 16.8 mL/L, a pH level of 6.6, and the resulting desirability was 0.64. In
crease in the dosage increased NH3‒N removal. These results are in Scenario B, the dosage was 11.1 mL/L, the pH was 7.3, and the desir­
agreement with the findings reported by Banch et al. [42]. The reason ability was 0.57. Table 7 shows the predicted values of each response for
why the ammonia concentration decreased with the decrease in pH is both scenarios. It can be seen that the large variation between the two
likely due to the effect of pH on ammonia/ammonium equilibrium, scenarios was observed in TOC and NH3‒N, while TP and TN removal
which shifts toward ammonium ions as pH decreases [77]. However, it is were almost the same. It appears that regardless of which scenario is
important to keep in mind that in this study, landfill leachate was pre­ considered, pH adjustment is needed for the subsequent membrane
pared for nutrient recovery. Therefore, ammonia should not be removed contactor process. However, the dosage requirement in Scenario B is
with the treatment. 34% less than that of Scenario A.
In this process, the behavior of coagulants and their effectiveness can
be explained by different destabilization mechanisms. When using nat­ 3.6. Optimization verification
ural polymers, these mechanisms include charge neutralization,
adsorption, and particle bridging [78]. These mechanisms could occur The optimization of the process was validated by running jar tests in
simultaneously, or one can be predominant over others, depending on duplicates and measuring the responses. Table 8 shows the predicted
the experimental conditions and type of wastewater. Charge neutrali­ values for each response along with the average measured values and
zation is expected to occur due to the presence of the tertiary amine the calculated removal percentage. The variation between the measured
group added onto the tannin structure by the Mannich reaction [29]. As and the predicted removals was also calculated.
suggested by Lopes et al. [38], a tannin-based coagulant modified by Even though the TN model was not significant from the statistical
cationization is more effective in acidic pH, and its charge decreases due analysis, the variation between the measured concentration and the
to deprotonation of the amine and hydroxyl groups when pH increases predicted one is low (− 3% in Scenario A and − 2% in Scenario B), and
(pH > 5), which suggests that bridging is the main destabilization the TN reduction in both scenarios is negligible. The TP variation is also
mechanism. Graham et al. [29] reported that at pH 7, the main mech­ low (+7% and +4% for scenarios A and B, respectively), and the
anisms could be a combination of electrostatic interaction and bridging. measured removal is not that different in both scenarios (around 49% on
However, these studies focused on the effect of the coagulant on syn­ average). The TOC variation is low. However, the measured removal in
thetic solutions. Therefore, it is important to consider that wastewater Scenario A is more than double the measured removal in Scenario B
characteristics could also affect the optimal range of pH and dosage. For (34% in A and 15% in B). Lastly, the NH3‒N variation is greater than
example, the tannin-based coagulant was found to be most effective in 20% in Scenario A and 18% in Scenario B, which is acceptable. In both
the pH range of 4–7.5 [79], but applying an acidic range is not possible scenarios, ammonia concentration increased after the treatment. From
with leachate due to the foaming issue. Hence, the effectiveness of samples taken from the verification experiments, the change in color
tannin-based products needs to be tested and optimized based on the was also measured. The color reduction measured for Scenario B was
operation and characteristics of the water to be treated. 78%, which is in accordance with the values reported by Grenda et al.
[80], who treated wastewater using only tannin coagulants. In their
3.5. Multi-response optimization work, they used a product similar to the one used in this study.
When comparing the performance of tannin to traditional coagulants
The aim of this step is to optimize the process by combining the such as FeCl3, existing studies reported TP removal of ~ 80–90% and
selected responses and generating a desirability function that takes into TOC removal of ~ 50–70% [15,81,82], which are higher than the ones
account the importance ranking and the target set for each response. obtained in this study. However, their dosages were high (3–6 g/L of
Considering the aim of the process is to treat the landfill leachate for coagulant) compared to our dosages. Moreover, the initial concentration
subsequent nitrogen recovery with a membrane contactor, TN and NH3‒ of TOC and TP in these studies were different than the ones measured in
N removal needs to be minimized. The removal of other parameters is set this study, which may affect the treatment efficiency. Therefore,
to be maximized to mitigate membrane fouling. In particular, TP considering the lower dosage and other advantages related to natural
removal needs to be maximized to utilize the produced sludge for coagulants such as biodegradability and no toxicity of residuals after
phosphorous recovery. When setting the importance ranking for each treatment, their usage is highly recommended as valuable alternative for
response, the statistical analysis performed for each response was conventional coagulants.
considered as well as the importance of the response for the treatment Considering the verification results, Scenario B is recommended for
goal. The TP response is considered the most important response the treatment as it resulted in similar TP and TN changes and higher
because its model is statistically significant, and all the statistical pa­ ammonia production. Despite the lower TOC removal, the lower coag­
rameters fulfill the requirements. In addition, TP removal is important ulant and pH adjustment needed in this scenario makes up for this un­
for the subsequent process, as mentioned earlier. The TP response is desirable outcome. Hence, Scenario B settings will be tested for micro-
followed by NH3‒N and TOC since their models are statistically signif­ pollutants and metals removal from landfill leachate.
icant, but LOF is not in the desired range. Lastly, the TN response is set at
the minimum because the change in its concentration was small and its 3.7. Micropollutants and metals removal
statistical model was not significant. The target and importance ranking
for all responses are presented in Table SM4. After analyzing the results obtained from the optimization verifica­
After defining the settings for the optimization of the process, two tion for common constituents, Scenario B showed great potential and
different scenarios were constructed. The first case scenario (Scenario A) hence was tested further for metal, metalloids and micropollutant
suggests the recommended pH and dosage, based on the criteria of removal. Fig. 6 presents the selected elements and their relative removal
meeting only the targets and the importance ranking. In the second case percentages. The elements that are not reported in Fig. 6 were not
scenario (Scenario B), pH was set to the maximum, considering the affected by the treatment, except for K and Co, where a respective in­
preferred alkaline environment for subsequent membrane contactor crease of 2% and 5% was observed. However, this could be associated
treatment [11], and the dosage was set to the minimum to achieve the with uncertainties in the measurements since the chemical structure of

8
I. Righetto et al. Journal of Environmental Chemical Engineering 9 (2021) 105213

Fig. 5. Desirability functions for (a) Scenario A, where the uncoded values for dosage and pH are 16.8 mL/L and 6.6, respectively, and (b) Scenario B where the
uncoded values for dosage and pH are 11.1 mL/L and 7.3, respectively.

treatment, while compounds such as 4-tert-octyl phenol, ketoprofen,

Table 7
anthracene, and naphthalene showed only a slight increase in concen­
Predicted response values for suggested scenarios.
tration, which could also be related to uncertainties in measurements.
Scenario Scenario A Scenario B Therefore, assuming tannin does not contain any of the above-
Response Raw Predicted Predicted Predicted C Predicted mentioned compounds, it can be stated that they are not affected by
Conc. Conc. [mg/ removal [mg/L] removal the treatment. The initial and final concentrations for all these com­
[mg/L] L] [%] [%]
pounds are reported in Supplementary Materials in Table SM5.
TN 1074 1069 0 1057 − 2 As shown in Fig. 6, Fe, Ti, Cr, and Al were removed at a high per­
TP 7.74 3.52 − 55 3.90 − 50 centage (92%, 91%, 68%, and 52%, respectively). Lugo et al. [40] re­
TOC 1189 753 − 37 964 19
ported a very low Cr removal percentage (< 20%) using tannin to treat
−
NH3‒N 536 460 − 14 528 − 2
landfill leachate at pH 7. However, the removal increased to 87% when
leachate was treated at pH 10. Similar findings were reported by Banch
tannin does not contain any of these elements. Surfactants and micro­ et al. [42], where Fe removal of 89.76% was reported. Other than Fe, Ti,
pollutants were also measured. For micropollutants, the following Cr, and Al, Fig. 6 reports decent removals of Ba, V, and Mn, while other
groups were selected for the analysis: alkylphenols and ethoxylates, metals were only removed by small percentages. Compared to the report
phthalates, pharmaceuticals, organo-tin, and PAH. Fig. 7 reports only by Beltrán-Heredia and Sánchez-Martín [83], when treating surface
the removed micropollutants, while all the others were either under the water using tannin coagulants, more than double the Ni removal was
detection limit or not removed. Phenazone was not removed after achieved in this study.

9
I. Righetto et al. Journal of Environmental Chemical Engineering 9 (2021) 105213

Table 8
Predicted concentration, measured concentration, and variation of the measured value from the predicted value for each response.
Scenario Scenario A Scenario B

Response Predicted Conc. Measured Conc. Measured removal Variation Predicted Conc. Measured Conc. Measured removal Variation
[mg/L] [mg/L] [%] [%] [mg/L] [mg/L] [%] [%]

TN 1069 1034 ± 14 − 4 − 3 1057 1033 ± 4 − 4 − 2

TP 3.52 3.76 ± 0.04 − 51 +7 3.90 4.08 ± 0.06 − 47 +4
TOC 753 784 ± 5 − 34 +4 964 1016 ± 5 − 15 +5
NH3‒N 460 652 ± 13 +22 +29 528 642 ± 5 +20 +18

Fig. 6. Metals and metalloids removal when treating landfill leachate dosing 11.1 mL/L HTH at pH 7.3.

Fig. 7. Micropollutant removal when treating landfill leachate dosing 11.1 mL/L HTH at pH 7.3.

Metals removal is believed to be mainly due to their complexation Fe and Al were found to form mono-di-, and trisdentate complexes with
with tannin [42]. The ortho-dihydroxyl group is thought to be the tannin. Ti was also found to form colored complexes with tannin. This
responsible moiety of the tannin structure for metals complexation [84]. may explain the observed high removal of these metals in this study. The

10
I. Righetto et al. Journal of Environmental Chemical Engineering 9 (2021) 105213

large removal of barium could be associated with high affinity of this the value obtained was 4.8 × 107 m/kg. There are no studies where cake
element with tannin as the latter was able to remove Ba even with very resistance of landfill leachate treated with coagulation-flocculation
low concentration of 7.2 μg/L [85]. Metals could bind to the organic process was determined. However, this value is 3–4 order of magni­
carbon present in landfill leachate. It was shown in a study conducted by tude smaller than the values reported for treating petrochemical
Li et al. that leachate organic matter favors binding to Mg than Ca [86, wastewater with FeCl3 (2.49 × 1011 m/kg) and FeCl3 + Cationic poly­
87]. This suggests that the small removal of Mg was facilitated mainly by acrylamide (1.14 × 1010 m/kg) [92].
leachate organic matter and not tannin. The zero removal of Ca indicates
the unfavored binding of tannin and leachate organic matter to this 4. Conclusion
element in the applied experimental conditions. The other possible
removal pathway is metal trapping via cation exchange capacity [4]. Treating landfill leachate with natural coagulants for nutrient re­
The removed metals and metalloids end up in the P-rich sludge. Valu­ covery purposes has not been addressed by previous studies and was
able metals along with P can be recovered from the sludge using thoroughly investigated in this work. Among the tested natural co­
chemical leaching or other metallurgical techniques. However, this agulants, a tannin-based product named HTH showed the best perfor­
subject is outside the scope of this study and planned to be investigated mance and was chosen for carrying out the rest of the investigation. The
in our future research work. process was then optimized for pH and dosage ranges of 6–8 and
The removal of cationic surfactants is a valuable outcome, consid­ 5–20 mL/L, respectively. Two optimization scenarios were examined —
ering the ability of those compounds to interact with the hydrophobic Scenario A, where pH and dosage were optimized within the range; and
membrane owing to their hydrophobic head. The removal of surfactants Scenario B, where pH was maximized for subsequent nutrient recovery
is in accordance with the findings of Beltrán-Heredia et al. [64], where purposes and dosage was minimized to keep HTH consumption as low as
surfactants were removed by around 56% when working at pH 7 using a possible. The best results were obtained from Scenario B, where an op­
cationic tannin coagulant. In that case, mainly anionic surfactants were timum combination of dosage and pH were 11.1 mL/L and 7.3,
removed. Assuming that the cationic tannin interacts with the anionic respectively. The removal percentages measured with Scenario B were
head of surfactants and considering that a prepared synthetic solution close to the predicted percentages, as follows: TOC − 15%, TP − 47%,
contains only anionic surfactants, the results reported by Beltrán-Her­ and TN − 4%. Ammonia-nitrogen increased by +20%, which is a good
edia et al. [64] are reasonable. On the other hand, Fig. 7 shows that result when considering that the purpose of the treatment was to prepare
cationic surfactants were mainly removed while anionic surfactants the landfill leachate for nitrogen recovery. After verifying the optimized
were not removed. The wastewater stream used was landfill leachate dosage and pH, metals and micropollutants removal were evaluated
and thus contains several compounds other than anionic and cationic under these conditions. The following elements were removed at high
surfactants. Therefore, a different mechanism could have occur­ percentages: Fe − 92%, Ti − 91%, Cr − 68%, Al − 52%, Ba 48%, and V
red—negatively charged particles found in landfill leachate could have 44%. Other elements were removed in smaller percentages: Mn 33%, B
been attracted to the cationic tannin first, and then positively charged 15%, Ni 14%, Sr 8%, and Si 4%. For micropollutants, the following
heads would attach to those particles. removal percentages were achieved: cationic surfactants − 56%, 4-Non­
Nonylphenols were removed by 27% using tannin. These compounds ylphenol − 27%, TBT − 52%, and four different pharmaceuticals such as
are used as lubricants, detergents, manufacturing of antioxidants, and 5-methylbenzotriazole, ibuprofen, carbamazepine, and metoprolol
are toxic to aquatic organisms. Tributyltin (TBT) removal is also rela­ (− 16%, − 14%, − 13%, and − 3%, respectively). Other measured
tively high, but the concentration limit given by the Finnish industrial micropollutants and metals were not removed. Considering that in this
wastewater guide is 0.0015 μg/L [88], which was not reached. For case the process was not optimized for the removal of metals, metalloids
pharmaceuticals and cationic surfactants, no limits are reported in the and micropollutants, the removal of these compounds with tannin co­
guide, and the removal percentage was also relatively low. agulants could be further investigated and optimized. Furthermore, the
It is important to keep in mind that these analyses were conducted accumulation of all these compounds in the sludge makes its charac­
after the process was optimized based on the evaluation of other pa­ terization important if direct use of the sludge is considered. Physical
rameters (TN, TP, TOC, NH3‒N). Therefore, the measured removal characterization of the sludge was performed by conducting SVI and
percentages of metals, surfactants, and micropollutants are considered a filterability tests of the produced sludge. It was found that tannin pro­
bonus and highlight the potential of tannin as an effective coagulant, not duced sludge with quality on par with or better than that produced with
only to prepare landfill leachate for nutrient recovery but also to conventional coagulants. Our findings suggest that natural coagulants
improve the quality of produced wastewater from the hazardous mate­ have the potential to offer a sustainable solution for micropollutant
rials perspective. When using landfill leachate as a feed stream to a removal and improvement of nutrient recovery using membrane con­
membrane contactor, the removal of these substances can decrease the tactor technology. Tannin coagulation did not remove NH3 rather
chances of membrane fouling and improve the quality of the leachate for increased it slightly while removing other constituents. With cleaner
subsequent treatment processes. leachate and higher concentration of NH3, mass transfer of the latter
across the membrane would significantly increase leading to better re­
3.8. Sludge characteristics covery. Tannin was also capable of accumulating P in the sludge to
prepare it for recovery with further processing. The recovery of P and
To evaluate the sludge settleability, the sludge was left to settle for other elements from the sludge is an interesting topic we endeavor to
30 min while the volume occupied by the sludge was progressively unravel in our future studies.
measured. Fig. SM6 shows sludge settleability represented by the sludge
volume change with time. The settleability of the sludge produced with CRediT authorship contribution statement
tannin is comparable to the observed results in the literature for con­
ventional coagulants. A study conducted by Li et al. (2010) [89] used Ilaria Righetto: Conceptualization, Methodology, Investigation,
FeCl3 for treating landfill leachate, and reported a sludge volume per­ Formal analysis, Validation, Writing - Original Draft and Writing - Re­
centage of around 30% after 1 h of settling, while in our case the volume view & Editing. Raed A. Al-Juboori: Conceptualization, Investigation,
percentage was 20% after 30 min of settling. The SVI measured in this Formal analysis, Writing - Original Draft, Writing - Review & Editing,
study of 185 ± 4 mL/g was in the same range or smaller as the one re­ Project administration and supervision. Juho Uzkurt Kaljunen: Project
ported by other studies where traditional coagulants, such as FeCl3 and administration, Investigation and supervision. Anna Mikola: Concep­
PACl were used (~ 100–700 mg/L) [90,91]. Cake resistance was also tualization, Supervision, Resources, Project administration, Funding
calculated to determine the filterability of the sludge using Eq. (3) and acquisition and Writing - Review & Editing.

11
I. Righetto et al. Journal of Environmental Chemical Engineering 9 (2021) 105213

Declaration of Competing Interest [20] L.M. Rui, Z. Daud, A.A.A. Latif, Treatment of leachate by coagulation-flocculation
using different coagulants and polymer: a review, Int. J. Adv. Sci. Eng. Inf. Technol.
2 (2012) 114–117.
The authors declare that they have no known competing financial [21] S.N.A. Mohd-Salleh, M.Z.N. Shaylinda, N. Othman, G. Yashni, A.B. Norshila,
interests or personal relationships that could have appeared to influence Coagulation performance and mechanism of a new coagulant (polyaluminium
the work reported in this paper. chloride-tapioca peel powder) for landfill leachate treatment, J. Eng. Sci. Technol.
15 (2020) 3709–3722.
[22] D. Venu, R. Gandhimathi, P.V. Nidheesh, S.T. Ramesh, Treatment of stabilized
Acknowledgment landfill leachate using peroxicoagulation process, Sep. Purif. Technol. 129 (2014)
64–70.
[23] X. Zhou, Z. Hou, L. Lv, J. Song, Z. Yin, Electro-Fenton with peroxi-coagulation as a
The authors are grateful to Haarla Oy and Chemigate Oy for feasible pre-treatment for high-strength refractory coke plant wastewater:
providing coagulants and supporting information for their products. The parameters optimization, removal behavior and kinetics analysis, Chemosphere
238 (2020), 124649.
research was partially financed by Maa- ja Vesitekniikan Tuki Ry
[24] Y. Zhang, B. Shi, Y. Zhao, M. Yan, D.A. Lytle, D. Wang, Deposition behavior of
(MVTT). residual aluminum in drinking water distribution system: effect of aluminum
speciation, J. Environ. Sci. 42 (2016) 142–151.
[25] S.D.W. Comber, M.J. Gardner, J. Churchley, Aluminium speciation: implications of
Appendix A. Supplementary material wastewater effluent dosing on river water quality, Chem. Speciat. Bioavailab. 17
(2005) 117–128, https://doi.org/10.3184/095422905782774874.
Supplementary data associated with this article can be found in the [26] A.K. Darban, A. Shahedi, F. Taghipour, A. Jamshidi-Zanjani, A review on industrial
wastewater treatment via electrocoagulation processes, Curr. Opin. Electrochem.
online version at doi:10.1016/j.jece.2021.105213.
(2020).
[27] S.Y. Choy, K.M.N. Prasad, T.Y. Wu, M.E. Raghunandan, R.N. Ramanan, Utilization
References of plant-based natural coagulants as future alternatives towards sustainable water
clarification, J. Environ. Sci. 26 (2014) 2178–2189, https://doi.org/10.1016/j.
jes.2014.09.024 (China).
[1] S. Renou, J.G. Givaudan, S. Poulain, F. Dirassouyan, P. Moulin, Landfill leachate
[28] C.Y. Teh, P.M. Budiman, K.P.Y. Shak, T.Y. Wu, Recent advancement of
treatment: review and opportunity, J. Hazard. Mater. 150 (2008) 468–493,
coagulation-flocculation and its application in wastewater treatment, Ind. Eng.
https://doi.org/10.1016/j.jhazmat.2007.09.077.
Chem. Res. 55 (2016) 4363–4389, https://doi.org/10.1021/acs.iecr.5b04703.
[2] M.A. Kamaruddin, M.S. Yusoff, H.A. Aziz, Y.-T. Hung, Sustainable treatment of
[29] N. Graham, F. Gang, G. Fowler, M. Watts, Characterisation and coagulation
landfill leachate, Appl. Water Sci. 5 (2015) 113–126, https://doi.org/10.1007/
performance of a tannin-based cationic polymer: a preliminary assessment,
s13201-014-0177-7.
Colloids Surf. A Physicochem. Eng. Asp. 327 (2008) 9–16, https://doi.org/
[3] A.A. Abbas, G. Jingsong, L.Z. Ping, P.Y. Ya, W.S. Al-Rekabi, Review on landfll
10.1016/j.colsurfa.2008.05.045.
leachate treatments, J. Appl. Sci. Res. 5 (2009) 534–545.
[30] K. Carlqvist, M. Arshadi, T. Mossing, U. Östman, H. Brännström, E. Halmemies,
[4] B.P. Naveen, D.M. Mahapatra, T.G. Sitharam, P.V. Sivapullaiah, T.
J. Nurmi, G. Lidén, P. Börjesson, Life-cycle assessment of the production of
V. Ramachandra, Physico-chemical and biological characterization of urban
cationized tannins from Norway spruce bark as flocculants in wastewater
municipal landfill leachate, Environ. Pollut. 220 (2017) 1–12.
treatment, Biofuels Bioprod. Biorefin. 14 (2020) 1270–1285.
[5] J. Krook, N. Svensson, M. Eklund, Landfill mining: a critical review of two decades
[31] J.J. Feria-Díaz, L. Polo-Corrales, E.J. Hernandez-Ramos, Evaluation of coagulation
of research, Waste Manag. 32 (2012) 513–520, https://doi.org/10.1016/j.
sludge from raw water treated with Moringa oleifera for agricultural use, Ing.
wasman.2011.10.015.
Investig. 36 (2016) 14–20.
[6] D. Seibert, H. Quesada, R. Bergamasco, F.H. Borba, L. Pellenz, Presence of
[32] S.S. Moghaddam, M.R.A. Moghaddam, M. Arami, Coagulation/flocculation process
endocrine disrupting chemicals in sanitary landfill leachate, its treatment and
for dye removal using sludge from water treatment plant: optimization through
degradation by Fenton based processes: a review, Process Saf. Environ. Prot. 131
response surface methodology, J. Hazard. Mater. 175 (2010) 651–657.
(2019) 255–267.
[33] L. Mazari, D. Abdessemed, A. Szymczyk, Evaluating reuse of alum sludge as
[7] J. Wiszniowski, D. Robert, J. Surmacz-Gorska, K. Miksch, J.V. Weber, Landfill
coagulant for tertiary wastewater treatment, J. Environ. Eng. 144 (2018),
leachate treatment methods: a review, Environ. Chem. Lett. 4 (2006) 51–61,
04018119.
https://doi.org/10.1007/s10311-005-0016-z.
[34] Z. Ahmed, M.S. Yusoff, M.K. NH, Application of natural coagulants for sustainable
[8] X.Z. Li, Q.L. Zhao, Recovery of ammonium-nitrogen from landfill leachate as a
treatment of semi-aerobic landfill leachate, in: AIP Conference Proceedings, AIP
multi-nutrient fertilizer, Ecol. Eng. 20 (2003) 171–181, https://doi.org/10.1016/
Publishing LLC, 2020, p. 20024.
S0925-8574(03)00012-0.
[35] S.Y. Cheng, P.-L. Show, J.C. Juan, T.C. Ling, B.F. Lau, S.H. Lai, E.P. Ng, Sustainable
[9] M. Qin, H. Molitor, B. Brazil, J.T. Novak, Z. He, Recovery of nitrogen and water
landfill leachate treatment: optimize use of guar gum as natural coagulant and floc
from landfill leachate by a microbial electrolysis cell-forward osmosis system,
characterization, Environ. Res. 188 (2020), 109737.
Bioresour. Technol. 200 (2016) 485–492, https://doi.org/10.1016/j.
[36] A. Ibrahim, A.Z. Yaser, Colour removal from biologically treated landfill leachate
biortech.2015.10.066.
with tannin-based coagulant, J. Environ. Chem. Eng. 7 (2019), 103483, https://
[10] C. Di Iaconi, M. Pagano, R. Ramadori, A. Lopez, Nitrogen recovery from a
doi.org/10.1016/j.jece.2019.103483.
stabilized municipal landfill leachate, Bioresour. Technol. 101 (2010) 1732–1736,
[37] M.S. Yusoff, H.A. Aziz, M.F.M.A. Zamri, A.Z. Abdullah, N.E.A. Basri, Floc behavior
https://doi.org/10.1016/j.biortech.2009.10.013.
and removal mechanisms of cross-linked Durio zibethinus seed starch as a natural
[11] M.C. Amaral, N.C. Magalhães, W.G. Moravia, C.D. Ferreira, Ammonia recovery
flocculant for landfill leachate coagulation-flocculation treatment, Waste Manag.
from landfill leachate using hydrophobic membrane contactors, Water Sci.
74 (2018) 362–372.
Technol. 74 (2016) 2177–2184, https://doi.org/10.2166/wst.2016.375.
[38] E.C. Lopes, S.C.R. Santos, A.M.A. Pintor, R.A.R. Boaventura, C.M.S. Botelho,
[12] K.Y. Foo, B.H. Hameed, An overview of landfill leachate treatment via activated
Evaluation of a tannin-based coagulant on the decolorization of synthetic effluents,
carbon adsorption process, J. Hazard. Mater. 171 (2009) 54–60, https://doi.org/
J. Environ. Chem. Eng. 7 (2019), 103125.
10.1016/j.jhazmat.2009.06.038.
[39] A. Niskanen, S. Manfredi, T.H. Christensen, R. Anderson, Environmental
[13] J.-S. Guo, A.A. Abbas, Y.-P. Chen, Z.-P. Liu, F. Fang, P. Chen, Treatment of landfill
assessment of Ämmässuo landfill (Finland) by means of LCA-modelling
leachate using a combined stripping, Fenton, SBR, and coagulation process,
(EASEWASTE), Waste Manag. Res. 27 (2009) 542–550, https://doi.org/10.1177/
J. Hazard. Mater. 178 (2010) 699–705.
0734242×08096976.
[14] C. Amor, E. De Torres-Socías, J.A. Peres, M.I. Maldonado, I. Oller, S. Malato, M.
[40] L. Lugo, A. Martín, J. Diaz, A. Pérez-Flórez, C. Celis, Implementation of modified
S. Lucas, Mature landfill leachate treatment by coagulation/flocculation combined
acacia tannin by mannich reaction for removal of heavy metals (Cu, Cr and Hg),
with Fenton and solar photo-Fenton processes, J. Hazard. Mater. 286 (2015)
Water 12 (2020), https://doi.org/10.3390/w12020352 (Switzerland).
261–268.
[41] J. Beltrán-Heredia, J. Sánchez-Martín, M.C. Gómez-Muñoz, New coagulant agents
[15] A.R. Ishak, F.S. Hamid, S. Mohamad, K.S. Tay, Stabilized landfill leachate
from tannin extracts: preliminary optimisation studies, Chem. Eng. J. 162 (2010)
treatment by coagulation-flocculation coupled with UV-based sulfate radical
1019–1025.
oxidation process, Waste Manag. 76 (2018) 575–581.
[42] T.J. Banch, M.M. Hanafiah, A.F. Alkarkhi, S.S. Abu, Factorial design and
[16] X. Liu, X.-M. Li, Q. Yang, X. Yue, T.-T. Shen, W. Zheng, K. Luo, Y.-H. Sun, G.-
optimization of landfill leachate treatment using tannin-based natural coagulant,
M. Zeng, Landfill leachate pretreatment by coagulation–flocculation process using
Polymers 11 (2019) 1349.
iron-based coagulants: optimization by response surface methodology, Chem. Eng.
[43] M. Özacar, İ.A. Şengil, Evaluation of tannin biopolymer as a coagulant aid for
J. 200 (2012) 39–51.
coagulation of colloidal particles, Colloids Surf. A Physicochem. Eng. Asp. 229
[17] A.G. Capodaglio, P. Hlavínek, M. Raboni, Physico-chemical technologies for
(2003) 85–96.
nitrogen removal from wastewaters: a review, Rev. Ambient. Agua 10 (2015)
[44] S.N. Ugwu, A.F. Umuokoro, E.A. Echiegu, B.O. Ugwuishiwu, C.C. Enweremadu,
481–498, https://doi.org/10.4136/ambi-agua.1618.
Comparative study of the use of natural and artificial coagulants for the treatment
[18] H. Bakraouy, S. Souabi, K. Digua, O. Dkhissi, M. Sabar, M. Fadil, Optimization of
of sullage (domestic wastewater), Cogent Eng. 4 (2017), 1365676.
the treatment of an anaerobic pretreated landfill leachate by a
[45] A. Benalia, K. Derbal, A. Panico, F. Pirozzi, Use of Acorn Leaves as a natural
coagulation–flocculation process using experimental design methodology, Process
coagulant in a drinking water treatment plant, Water 11 (2019) 57.
Saf. Environ. Prot. 109 (2017) 621–630.
[46] S. Ghafari, H.A. Aziz, M.H. Isa, A.A. Zinatizadeh, Application of response surface
[19] A.A. Tatsi, A.I. Zouboulis, K.A. Matis, P. Samaras, Coagulation–flocculation
methodology (RSM) to optimize coagulation-flocculation treatment of leachate
pretreatment of sanitary landfill leachates, Chemosphere 53 (2003) 737–744.

12
I. Righetto et al. Journal of Environmental Chemical Engineering 9 (2021) 105213

using poly-aluminum chloride (PAC) and alum, J. Hazard. Mater. 163 (2009) [70] C.M. Todaro, H.C. Vogel, Fermentation and Biochemical Engineering Handbook,
650–656, https://doi.org/10.1016/j.jhazmat.2008.07.090. William Andrew, 2014.
[47] T.K. Trinh, L.S. Kang, Response surface methodological approach to optimize the [71] H. Pashaei, A. Ghaemi, M. Nasiri, B. Karami, Experimental modeling and
coagulation–flocculation process in drinking water treatment, Chem. Eng. Res. Des. optimization of CO2 absorption into piperazine solutions using RSM-CCD
89 (2011) 1126–1135. methodology, ACS Omega 5 (2020) 8432–8448.
[48] P.K. Bhagwat, S.B. Jhample, P.B. Dandge, Statistical medium optimization for the [72] N. Manmai, Y. Unpaprom, R. Ramaraj, Bioethanol production from sunflower
production of collagenolytic protease by Pseudomonas sp. SUK using response stalk: application of chemical and biological pretreatments by response surface
surface methodology, Microbiology 84 (2015) 520–530. methodology (RSM), Biomass Convers. Biorefin. (2020) 1–15.
[49] S.B. Jhample, P.K. Bhagwat, P.B. Dandge, Statistical media optimization for [73] M.Y. Noordin, V.C. Venkatesh, S. Sharif, S. Elting, A. Abdullah, Application of
enhanced production of fibrinolytic enzyme from newly isolated Proteus penneri response surface methodology in describing the performance of coated carbide
SP-20, Biocatal. Agric. Biotechnol. 4 (2015) 370–379. tools when turning AISI 1045 steel, J. Mater. Process. Technol. 145 (2004) 46–58,
[50] M. Gryta, Fouling in direct contact membrane distillation process, J. Membr. Sci. https://doi.org/10.1016/S0924-0136(03)00861-6.
325 (2008) 383–394, https://doi.org/10.1016/j.memsci.2008.08.001. [74] J. Turunen, A. Karppinen, R. Ihme, Effectiveness of biopolymer coagulants in
[51] R.A. Al-Juboori, V. Aravinthan, T. Yusaf, L. Bowtell, Assessing the application and agricultural wastewater treatment at two contrasting levels of pollution, SN Appl.
downstream effects of pulsed mode ultrasound as a pre-treatment for alum Sci. 1 (2019) 1–9, https://doi.org/10.1007/s42452-019-0225-x.
coagulation, Ultrason. Sonochem. 31 (2016) 7–19, https://doi.org/10.1016/j. [75] M. Dela Justina, B. Rodrigues Bagnolin Muniz, M. Mattge Bröring, V.J. Costa,
ultsonch.2015.11.028. E. Skoronski, Using vegetable tannin and polyaluminium chloride as coagulants for
[52] M.N. Adlan, P. Palaniandy, H.A. Aziz, Optimization of coagulation and dissolved dairy wastewater treatment: a comparative study, J. Water Process Eng. 25 (2018)
air flotation (DAF) treatment of semi-aerobic landfill leachate using response 173–181, https://doi.org/10.1016/j.jwpe.2018.08.001.
surface methodology (RSM), Desalination 277 (2011) 74–82, https://doi.org/ [76] J. Beltrán-Heredia, J. Sánchez-Martín, A. Muñoz-Serrano, J.A. Peres, Towards
10.1016/j.desal.2011.04.006. overcoming TOC increase in wastewater treated with Moringa oleifera seed
[53] B. Wu, Reliability Analysis of Dynamic Systems: Efficient Probabilistic Methods extract, Chem. Eng. J. 188 (2012) 40–46.
and Aerospace Applications, Academic Press, 2013. [77] S. Syafalni, H.K. Lim, N. Ismail, I. Abustan, M.F. Murshed, A. Ahmad, Treatment of
[54] D.C. Montgomery. Design and Analysis of Experiments, Fifth Edition, 2001, landfill leachate by using lateritic soil as a natural coagulant, J. Environ. Manag.
https://doi.org/10.1002/9780470191750. 112 (2012) 353–359, https://doi.org/10.1016/j.jenvman.2012.08.001.
[55] M. Ghanbarzadeh Lak, M.R. Sabour, E. Ghafari, A. Amiri, Energy consumption and [78] C.-Y. Yin, Emerging usage of plant-based coagulants for water and wastewater
relative efficiency improvement of photo-Fenton – optimization by RSM for landfill treatment, Process Biochem. 45 (2010) 1437–1444.
leachate treatment, a case study, Waste Manag. 79 (2018) 58–70, https://doi.org/ [79] J. Sánchez-Martín, M. González-Velasco, J. Beltrán-Heredia, Surface water
10.1016/j.wasman.2018.07.029. treatment with tannin-based coagulants from Quebracho (Schinopsis balansae),
[56] R.L. Mason, R.F. Gunst, J.L. Hess. Statistical Design and Analysis of Experiments – Chem. Eng. J. 165 (2010) 851–858.
With Applications to Engineering and Science, 2nd ed., John Wiley & Sons, 2003. [80] K. Grenda, J. Arnold, J.A.F. Gamelas, M.G. Rasteiro, Up-scaling of tannin-based
[57] R.H. Myers, D.C. Montgomery, C.M. Anderson-Cook. Response Surface coagulants for wastewater treatment: performance in a water treatment plant,
Methodology: Process and Product Optimization Using Designed Experiments, 4th Environ. Sci. Pollut. Res. 27 (2020) 1202–1213, https://doi.org/10.1007/s11356-
ed., John Wiley & Sons, 2016. 018-2570-5.
[58] R. Ramanujam, R. Raju, N. Muthukrishnan, Taguchi Multi-Machining [81] S.S. Kumar, N.R. Bishnoi, Coagulation of landfill leachate by FeCl3: process
Characteristics Optimization in Turning of Al-15% SiC p Composites using optimization using Box–Behnken design (RSM), Appl. Water Sci. 7 (2017)
Desirability Function Analysis, 1, 2010, pp. 120–125. 1943–1953, https://doi.org/10.1007/s13201-015-0372-1.
[59] N.F.A. binti Saharudin, R. Nithyanandam. Wastewater Treatment by Using Natural [82] J.S. Yadav, A.K. Dikshit, Effect of pretreatment by coagulation on stabilized landfill
Coagulant, Chemical Engineering Department, 2014, pp. 2–3. leachate during anaerobic treatment, Cogent Environ. Sci. 2 (2016), 1209993.
[60] Y.T. Hameed, A. Idris, S.A. Hussain, N. Abdullah, A tannin-based agent for [83] J. Beltrán-Heredia, J. Sánchez-Martín, Removing heavy metals from polluted
coagulation and flocculation of municipal wastewater: chemical composition, surface water with a tannin-based flocculant agent, J. Hazard. Mater. 165 (2009)
performance assessment compared to polyaluminum chloride, and application in a 1215–1218.
pilot plant, J. Environ. Manag. 184 (2016) 494–503, https://doi.org/10.1016/j. [84] N. Slabbert, Complexation of condensed tannins with metal ions. Plant
jenvman.2016.10.033. Polyphenols, Springer, 1992, pp. 421–436.
[61] M.A.A. Razali, A. Ariffin, Polymeric flocculant based on cassava starch grafted [85] A. Bello, V. Virtanen, J.-P. Salminen, T. Leiviskä, Aminomethylation of spruce
polydiallyldimethylammonium chloride: flocculation behavior and mechanism, tannins and their application as coagulants for water clarification, Sep. Purif.
Appl. Surf. Sci. 351 (2015) 89–94, https://doi.org/10.1016/j.apsusc.2015.05.080. Technol. 242 (2020), 116765.
[62] F.R. Spellman, Handbook of Water and Wastewater Treatment Plant Operations, [86] Z. Li, Q. Xue, L. Liu, J. Li, Precipitates in landfill leachate mediated by dissolved
CRC Press, 2013. organic matters, J. Hazard. Mater. 287 (2015) 278–286.
[63] J. Bratby, Coagulation and Flocculation in Water and Wastewater Treatment, IWA [87] Z. Li, Modeling precipitate-dominant clogging for landfill leachate with NICA-
Publishing, 2016. Donnan theory, J. Hazard. Mater. 274 (2014) 413–419.
[64] J. Beltrán-Heredia, J. Sánchez-Martín, C. Solera-Hernández, Anionic surfactants [88] Ympäristöministeriö (SYKE). Vesiympäristölle vaarallisista ja haitallisista aineista
removal by natural coagulant/flocculant products, Ind. Eng. Chem. Res. 48 (2009) an-nettujen säädösten soveltaminen. Kuvaus hyvistä menettelytavoista, 2015.
5085–5092, https://doi.org/10.1021/ie801913y. [89] W. Li, T. Hua, Q. Zhou, S. Zhang, F. Li, Treatment of stabilized landfill leachate by
[65] I.-S. Choi, U. Wiesmann, Aerobic degradation of a non-ionic surfactant in a the combined process of coagulation/flocculation and powder activated carbon
membrane bioreactor (MBR), Environ. Eng. Res. 12 (2007) 81–91. adsorption, Desalination 264 (2010) 56–62.
[66] Y. Deng, J.D. Englehardt, Treatment of landfill leachate by the Fenton process, [90] V. Oloibiri, I. Ufomba, M. Chys, W.T.M. Audenaert, K. Demeestere, S.W.H. Van
Water Res. 40 (2006) 3683–3694. Hulle, A comparative study on the efficiency of ozonation and
[67] X. Liu, J.T. Novak, Z. He, Synergistically coupling membrane electrochemical coagulation–flocculation as pretreatment to activated carbon adsorption of
reactor with Fenton process to enhance landfill leachate treatment, Chemosphere biologically stabilized landfill leachate, Waste Manag. 43 (2015) 335–342.
247 (2020), 125954. [91] M. Assou, A. Madinzi, A. Anouzla, M.A. Aboulhassan, S. Souabi, M. Hafidi,
[68] B. Lauterböck, M. Ortner, R. Haider, W. Fuchs, Counteracting ammonia inhibition Reducing pollution of stabilized landfill leachate by mixing of coagulants and
in anaerobic digestion by removal with a hollow fiber membrane contactor, Water flocculants: a comparative study, Int. J. Eng. Innov. Technol. 4 (2014) 20–25.
Res. 46 (2012) 4861–4869. [92] S. Verma, B. Prasad, I.M. Mishra, Pretreatment of petrochemical wastewater by
[69] F. Wäeger-Baumann, W. Fuchs, The application of membrane contactors for the coagulation and flocculation and the sludge characteristics, J. Hazard. Mater. 178
removal of ammonium from anaerobic digester effluent, Sep. Sci. Technol. 47 (2010) 1055–1064, https://doi.org/10.1016/j.jhazmat.2010.02.047.
(2012) 1436–1442.

13