Advertisement

Clean Technologies and Environmental Policy

, Volume 20, Issue 9, pp 1951–1970 | Cite as

Treatment of palm oil mill effluent (POME) by coagulation flocculation process using peanut–okra and wheat germ–okra

  • Chee Yap Chung
  • Anurita Selvarajoo
  • Vasanthi Sethu
  • Apurav Krishna Koyande
  • Arvin Arputhan
  • Zhi Chien Lim
Original Paper
  • 128 Downloads

Abstract

Coagulation–flocculation has been proven as one of the effective processes in treating palm oil mill effluent (POME), which is a highly polluted wastewater generated from palm oil milling process. Two pairs of natural coagulant–flocculant were studied and evaluated: peanut–okra and wheat germ–okra. This research aims to optimize the operating parameters of the coagulation flocculation process in removing turbidity, total suspended solid and chemical oxygen demand (COD) from POME by using a central composite design in the Design Expert® software. Important parameters such as operating pH, coagulant and flocculant dosages were empirically determined using jar test experiment and optimized using response surface methodology module. Significant quadratic polynomial models were obtained via regression analyses (R2) for peanut–okra (0.9355, 0.9534 and 0.8586 for turbidity, total suspended solids and COD removal, respectively) and wheat germ–okra (0.9638, 0.9578 and 0.7691 for turbidity, total suspended solids and COD removal, respectively). The highest observed removal efficiencies of turbidity, total suspended solids and COD (92.5, 86.6 and 34.8%, respectively, for peanut–okra; 86.6, 87.5 and 43.6%, respectively, for wheat germ–okra) were obtained at optimum pH, coagulant and flocculant dosages (pH 11.6, 1000.1 mg/L and 135.5 mg/L, respectively, for peanut–okra; pH 12, 1170.5 mg/L and 100 mg/L, respectively, for wheat germ–okra). The coagulation flocculation performance of peanut–okra and wheat germ–okra were comparable to each other. Characterizations of the natural coagulant–flocculant, as well as the sludge produced, were performed using Fourier transform infrared, energy-dispersive X-ray spectroscopy and field emission scanning electron microscope. More than 98% of water was removed from POME sludge by using centrifuge and drying methods, indicating that a significant reduction in sludge volume was achieved.

Keywords

Coagulation–flocculation POME Water treatment Peanut Wheat germ Okra 

Introduction

Palm oil industry has gained attention over the years as the palm oil is a versatile product. Malaysia is currently one of the largest palm oil producers in the world. By 2016, the total oil palm plantation area was recorded as 5.74 million hectares, which covers more than 17.3% of the available land in Malaysia (MPOB 2016). The global consumption of palm oil in 2015 and 2016 was 58.46 and 60.96 million tonnes, respectively, and it is expected to increasingly grow in the next few years due to the increased world population (Statista 2016). According to Malaysian Palm Oil Board, the total production of crude palm oil (CPO) in Malaysia was 19.96 and 17.32 million tonnes in 2015 and 2016, respectively (MPOB 2016). The country has steadily increased the palm oil production in order to satisfy the global palm oil demand.

Palm oil mill effluent (POME), as one of the by-products in the palm oil industry, however, has gained much attention due to its potential harmful effect to human health, as well as the detrimental effect to environment. POME is a highly polluted wastewater produced from the wet palm oil milling process. High-temperature steam and water which are used for the palm oil extraction from fresh fruit brunch (FFB) eventually result in the formation of POME. Typically, palm oil mills produce 5–7.5 t of POME for the production of every ton of crude palm oil (Latif Ahmad et al. 2003). In Malaysia, it is estimated that at least 60 million tonnes of POME was produced annually since 2009 (MPOB 2009). POME is usually hot, thick brownish wastewater that contains oil, nutrients, high turbidity and total solids (Ahmad et al. 2006). This highly polluted wastewater can cause serious pollution of watercourses such as oxygen depletion (Latif Ahmad et al. 2003). POME can also pollute the land such as clogging and water logging of the soil and killing the vegetation upon contact (Rupani et al. 2010). Typical characteristics of raw POME wastewater and the regulatory discharge limit in Malaysia are presented in Table 1.
Table 1

Characteristics of raw POME wastewater (Ma 2000) and the regulatory discharge limits (Environmental Quality Act 1974, 2005) in Malaysia

Parameter

Unit

Raw POME

Discharge limit

pH

3.4–5.2

5.0–9.0

Biological oxygen demand (BOD)

mg/L

10,250–43,750

100

Chemical oxygen demand (COD)

mg/L

15,000–100,000

Oil and grease

mg/L

4000

50

Suspended solid

mg/L

5000–54,000

400

Total volatile solids

mg/L

9000–72,000

Temperature

°C

80–90

45

Research and development of POME wastewater treatment has been intensively carried out in order to achieve sustainable development of palm oil industry. Conventional POME wastewater treatments such as biological treatments using anaerobic or facultative digestion and physical treatment using sedimentation are commonly applied in the palm oil mills which acquire large area of lands (Quah et al. 1982). Furthermore, physico-chemical treatment such as coagulation and flocculation is also applied in the treatment due to its short retention time and low capital costs (Tajuddin et al. 2015). In order to achieve the increasingly stringent environmental regulations, most of the POME treatment utilises advanced technologies such as adsorption, ultrafiltration and membrane separation (Ahmad et al. 2009).

Pre-treatment processes of advanced treatment of POME such as membrane filtration, coagulation and flocculation play an important role in effectively removing most of pollutants, for example total suspended solid (TSS), turbidity (TUR) and chemical oxygen demand (COD) in POME. The formation of suspension is mainly due to the repulsive property of colloidal particles, which is caused by the presence of primary charges at the surface of particles. Coagulation is the unit process of POME wastewater treatment for destabilizing the mutual repulsion of the colloidal particles, thus causing the particles to bind together and to effectively settle down (Sincero and Sincero 2003). Therefore, surface charge reduction and formation of micro-flocs will take place simultaneously. This process usually requires addition of chemicals known as coagulants to remove the organic matter and suspended particles which cause turbidity and colour of the wastewater. The conventional coagulants used in POME treatment are alum, copperas (ferrous sulphate), ferric sulphate and ferric chloride (Sincero and Sincero 2003). Coagulation is normally followed by flocculation to agglomerate the small particles formed from the coagulation. Without the flocculation process, the micro-flocs formed have low strength and will easily break down when subjected to physical forces, thereby reducing the efficiency of the process. (Lee et al. 2014).

Flocculation provides agitation in coagulant-treated POME in order to induce agglomeration of the smaller particles into larger flocs, thus improving removal efficiency of particles as sediment in sedimentation process (Sincero and Sincero 2003). The addition of flocculants not only can increase the density of the floc formed, but also reduce the coagulant consumption, thus improving the overall efficiency of the treatment. The application of coagulation–flocculation with chemical coagulant–flocculant in wastewater treatment has been proved as an effective treatment to reduce the environmental concerned parameters of the wastewater such as TUR, TSS and COD (Lee et al. 2014). Furthermore, as compared to other water treatment processes, coagulation–flocculation process is relatively low cost, simpler, more reliable and lower energy consuming (Oladoja 2015).

The use of metal- and polymer-based coagulant–flocculant in conventional water treatment has exhibited some shortcomings, for example production of voluminous sludge after the treatment, which is harmful and difficult to treat (Oladoja 2015). The residual aluminium in alum-treated water has also raised other concerns related to human health and environmental issues (Oladoja 2015). There was strong evidence relating the application of aluminium-based coagulants in water treatment with the development of Alzheimer’s disease (AD) (Flaten 2001). The cost of importing these chemical coagulant–flocculants for developing countries, as well as the relatively high treatment cost for its sludge, also becomes one of the drawbacks for their applications in POME treatment. It is pertinent to replace these chemical coagulant–flocculants with natural plant-based coagulant–flocculant in order to counteract the aforementioned problems (Sanghi et al. 2006). Natural coagulants are largely low cost, highly biodegradable, non-toxic, eco-friendly and generally results in less sludge volume (Ndabigengesere et al. 1995).

Moringa oleifera (MO) is one of the most studied natural coagulants. A high removal efficiency of TSS (> 98%) in POME treatment was achieved at optimum pH 5 and settling time of 114 min with the optimum MO dosage of 4000 mg/L and flocculant (NALCO 7751) dosage of 7000 mg/L (Garde et al. 2017). Another study reported that MO was able to remove 95% of suspended solids (SS) and 52.2% of COD from POME treatment, and it also increased the removal efficiency of SS and COD to 99.3% and 52.5%, respectively, when MO combined with flocculant (NALCO 7751) (Bhatia et al. 2007). Wheat germ is a by-product of flour milling process and also a good source of vegetable protein. Its potential as natural coagulant for POME treatment has been explored and studied (Daud et al. 2014). The highest removal efficiencies of TUR, TSS and COD were achieved at 99.1, 95.6 and 61.7%, respectively, at optimum dosage of 12,000 mg/L wheat germ coagulant and optimum pH 2. It also reported that the best extraction method of active component in wheat germ can be performed using 1 M sodium chloride (NaCl) solution (Daud et al. 2014). Moreover, peanut cake, which is the residual after oil extraction of peanut seed, has the potential to be applied as natural coagulant for POME treatment. Birima et al. (2013) studied the extraction of active agent in peanut as natural coagulant in the treatment of high turbid water and concluded that peanut seeds extracted with NaCl achieved 92% TUR removal. Furthermore, it performed better coagulation performance than peanut seeds extracted with distilled water which removed only 31.5% TUR (Birima et al. 2013). The coagulation activity of peanut seed was evaluated, and the study concluded that the higher the NaCl concentration used for the extraction of coagulation active component, the lower the optimum dosage and the higher percentage of removal efficiency of TUR, TSS and COD in POME treatment (Birima et al. 2014).

Chitosan as a natural based flocculant was able to achieve efficiency of 80% TUR removal in najayote wastewater treatment at pH 5.5 and at low dosage less than 3 g/L (Meraz et al. 2016). Furthermore, the effectiveness of okra mucilages (Hibiscus esculentus) on the removal of TUR from synthetic and biologically treated effluent was studied and the result showed that okra mucilage was able to remove 61–74% TUR for both synthetic wastewater and effluent (Anastasakis et al. 2009). The performance of okra mucilage with ferric chloride coagulant in the ratio of 27.5:1 at pH 6 was studied for textile wastewater treatment. This combination of coagulant–flocculant achieved the high removal efficiencies of colour (93.6%), TUR (97.2%) and COD (85.7%). It also produced less sludge and increased floc size by using low dosage of okra mucilage (Freitas et al. 2015). Table 2 shows the previous studies of coagulation–flocculation with respect to the maximum observed removal efficiencies of parameters such as TUR, TSS and COD.
Table 2

Previous studies of coagulation–flocculation process

Types of effluent

Parameters studied

Maximum removal efficiency (%)

Coagulant–flocculant and optimum conditions

Reference

Palm oil mill effluent (POME)

TUR

99.1

Wheat germ (12,000 mg/L); pH 2

Daud et al. (2014)

TSS

95.6

COD

61.7

TSS

93.6

Peanut extracted with 1 M NaCl solution (6000 mg/L)

Birima et al. (2014)

COD

67.0

TSS

82.2

Alum (730 mg/L); pH 5.3

Teh et al. (2014b)

COD

49.1

 

TSS

92.5

Rice starch (740 mg/L); pH 2.6

COD

30.9

TSS

86.7

Alum (380 mg/L); rice starch (280 mg/L); pH 4.5

COD

49.2

TUR

97.7

Chitosan (370 mg/L); pH 6

Saifuddin and Dinara (2011)

TSS

91.7

COD

42.7

TUR

98.8

Chitosan–magnetite (250 mg/L); pH 6

TSS

97.6

COD

62.5

TSS

95.0

Moringa oleifera seeds (6000 mg/L); pH 5

Bhatia et al. (2007)

COD

52.2

TSS

99.3

Moringa oleifera seeds (4000 mg/L); flocculant (NALCO 7751) (7000 mg/L); pH 5

COD

52.5

Textile wastewater

TUR

93.8

FeCl3 6H2O (320 mg/L); pH 6

Freitas et al. (2015)

COD

50.0

TUR

97.2

Okra (3.2 mg/L); FeCl3 6H2O (88.0 mg/L); pH 6

COD

85.7

Landfill leachate

COD

65.7

Lateritic soil (14,000 mg/L); pH 2

Syafalni et al. (2012)

COD

85.4

Alum (10,000 mg/L); pH 4.8

Nejayote wastewater

TUR

80.0

Chitosan (3000 mg/L); pH 5.5

Meraz et al. (2016)

Aggregation of particles in a solution can occur via four basic coagulation mechanisms such as (1) double-layer compression; (2) intraparticle bridging; (3) adsorption; and (4) charge neutralization (Sincero and Sincero 2003). The addition of NaCl solution for extracting active agent in coagulant–flocculant can cause the double-layer compression, resulting in destabilization of suspended particles which cause high turbidity in POME. The repulsive electrostatic interactions between the particles are overcome by the attractive van der Waals force, resulting in the agglomeration between the particles and sedimentation (Miller et al. 2008).

In this research, the coagulation–flocculation performance of peanut–okra (PN–OK) and wheat germ–okra (WG–OK) was studied and compared. The objective of this research was to investigate the effectiveness (removal efficiencies of TUR, TSS and COD) of these two pairs of natural coagulant–flocculant (PN–OK and WG–OK) on the POME wastewater treatment under different operating conditions (pH, coagulant and flocculant dosages). Optimization work was carried out in order to determine the optimum operating conditions for highest removal efficiencies of TUR, TSS and COD in POME treatment.

Methodology

Preparation of materials

POME sample

Raw POME sample was collected from Seri Ulu Langat Palm Oil Mill, Dengkil. The sample was collected from facultative pond, which is the effluent to the final pond before discharging to water bodies. The POME sample was kept in an airtight plastic container and stored in a dry, dark area in the laboratory in order to prevent biodegradation due to the microbial activity.

Three hundred millilitres of raw POME sample was required for each jar test. The sample was homogenized by inverting the plastic container for several times before pouring it into the beakers for jar test experiment. The quality of fresh POME may vary according to time, which affected the accuracy on the calculation of removal efficiency. Therefore, the initial parameter values of fresh POME in different sets of experiment were tested and recorded in order to obtain the average parameter values, as shown in Table 3.
Table 3

Characteristic of raw POME sample collected

Parameter

Unit

Range

Average value

Turbidity (TUR)

NTU

197–394

299

Total suspended solid (TSS)

mg/L

370–762

525

Chemical oxygen demand (COD)

mg/L

1083–1359

1226

Peanut (PN) and wheat germ (WG) coagulant

Fresh peanut (PN) and wheat germ (WG) were bought from a local supermarket in Semenyih, Selangor. Both PN and WG were dried in an oven at 60 °C overnight as high drying temperature (above 80 °C) may cause the loss in efficacy of coagulant agent (Miller et al. 2008). The weight of dried sample was constantly measured and recorded until a constant weight of sample was obtained in order to ensure complete drying. The dried PN and WG samples were ground into solid powder separately using a domestic blender. Fine powder of WG was obtained, while crushed, coarse peanut powder was obtained. The solid powder of PN and WG was stored in airtight sealed plastic bag separately and kept in cool, dry place in the laboratory.

Okra (OK) flocculant

Fresh okra was brought from a local supermarket in Semenyih, Selangor. The okra was cleaned with distilled water to remove impurities. The okra pods were halved to remove the fibres and seeds. The okra was cut into short, small pieces and was oven-dried at temperature of 60 °C overnight. The weight of dried samples was constantly measured and recorded until a constant weight of sample was obtained. The dried okra was ground into fine powder using domestic blender. The solid okra powder was stored in airtight sealed plastic bag kept in cool, dry place in the laboratory.

Coagulant and flocculant stock solution

50,000 mg/L (PN and WG) coagulant and 5000 mg/L (OK) flocculant stock solutions were prepared for the jar test experiment. The stock solution was prepared by dissolving the solid sample of natural coagulant–flocculant using the 1 M sodium chloride (NaCl) solution, which was obtained by dissolving 29.2 g of NaCl into 500 mL distilled water. For PN and WG coagulant stock solutions, 25 g of solid powder was added separately into 500 mL of 1 M NaCl solution in order to obtain a concentration of 50,000 mg/L. For OK flocculant stock solution, 4 g of okra powder was added into 200 mL of 1 M NaCl solution in order to obtain a concentration of 5000 mg/L. The stock solutions were stirred separately using magnetic stirrer for 10 min and filtered using Filtres Fioroni 125-mm filter paper to remove any solid residual in the solution. The fresh stock solutions were prepared using solid powder 1 day before every jar test experiment and were stored in refrigerator at temperature of 4 °C to prevent biodegradation.

Jar test experiment

Jar test experiment was performed using the flocculator (Phipps & Bird) with 6 units of 500 mL beakers. Six individual jar tests can be carried out simultaneously in one set of experiment. For each individual jar test, 300 mL fresh POME was required and poured into 500-mL breaker. Each jar test experiment was carried out using one pair of coagulant–flocculant, which was either PN–OR or WG–OR. The three manipulating variables, namely pH, coagulant dosage and flocculant dosage, were adjusted accordingly for each jar test. Different combinations of manipulating variables were generated by Design Expert® software and are shown in Table 4. The required volume of coagulant and flocculant stock solution was added simultaneously into POME sample to obtain the responding dosage (Birima et al. 2014).
Table 4

Manipulating variables for jar test experiment

No. of jar test

pH

Coagulant dosage (mg/L)

Flocculant dosage (mg/L)

1

10

1000

500

2

10

3000

300

3

11

3000

300

4

12

3000

300

5

12

5000

500

6

11

3000

500

7

11

3000

300

8

11

3000

300

9

12

1000

100

10

11

3000

300

11

10

1000

100

12

11

3000

300

13

11

3000

100

14

11

5000

300

15

10

5000

500

16

11

3000

300

17

10

5000

100

18

12

5000

100

19

12

1000

500

20

11

1000

300

The required pH value of the solution was adjusted using Ohaus ST20 pH pen meter after the required dosages of coagulant–flocculant were added and the solution was stirred at 150 rpm for 2 min. The range of pH required was 10–12; thus, pH value of POME sample was adjusted accordingly by adding 1 M NaOH solution. After the desired values of manipulating variables were achieved, each set of experiment had undergone three major stages: rapid mixing, slow mixing and sedimentation. The mixing speed and time required for each stage were set as constant variables of the experiment, as shown in Table 5. The collection of treated POME samples was performed after completing the stage of sedimentation. Treated POME samples were collected at a depth of 3 cm below the surface of POME sample in the beaker using pipette to avoid taking the floc on the surface. The samples were sent for the analyses of TUR, TSS and COD.
Table 5

Operating variables for jar test experiment

Stages

Speed (rpm)

Time (min)

Rapid mixing

150

5

Slow mixing

30

30

Sedimentation

60

Analytical method

Turbidity (TUR), total suspended solid (TSS) and chemical oxygen demand (COD)

TUR readings of POME samples were measured and recorded using Martini Mi415 turbidity meter. TSS readings of POME samples were measured and recorded using the Hach DR 3900 spectrophotometer. In order to determine COD readings, 2 mL of treated POME samples was mixed with high-range Hach COD vials separately. The mixtures were mixed well and heated up using Hach DRB 200 Reactor at 150 °C for 2 h. The COD readings were determined using Hach DR 3900 spectrophotometer after the temperature of the mixtures reached the room temperature. Three readings were taken for each parameter in order to obtain average value for the final reading. The TUR, TSS and COD removal efficiencies using each pair of coagulant–flocculant were calculated using Eq. (1).
$${\text{Removal efficiency }}\left( \% \right) = \frac{{{\text{Initial reading}}\,{-}\,{\text{Final reading}}}}{\text{Initial reading}} \times 100\%$$
(1)
where initial reading refers to average value of the parameters of raw POME, as shown in Table 2.

Sludge dewatering

Sludge dewatering tests were performed on the experiment sets which were obtained from the highest observed removal efficiencies of TUR, TSS and COD using PN–OK and WG–OK, respectively. The sludge after sedimentation was collected by removing most of the supernatant. Hundred millilitres of sludge was collected for each POME treated sample and put into two 50-mL plastic centrifuge tubes. The two plastic tubes were then placed into the fixed-angle Eppendorf 5430 centrifuge, with the setting of 4000 rpm and 4 min as rotation speed and time, respectively, for further recovery of supernatant. The concentrated sludge was separated from the supernatant and oven-dried at 70 °C overnight. Finally, the solid sludge powder was formed after centrifuge and oven-drying stages. The weight of sludge at different stages was measured and recorded. The dried solid sludge powder was collected and sent to further analyses such as characterization by FESEM, EDX and FTIR.

Characterization tests

Fourier transform infrared (FTIR) spectra were determined and plotted using Perkin-Elmer Frontier FTIR spectrometer for the peanut, wheat germ and okra, as well as the sludge resulted from coagulation–flocculation. The spectra from range of 4000–400 cm−1 were obtained to analyse the nature of the functional groups on the surface of the samples. Morphological studies of samples were conducted by FEI Quanta 400F field emission scanning electron microscope (FESEM). In addition, the elemental compositions in the sample were determined using Oxford-Instrument INCA 400 with X-Max Detector, which performed energy-dispersive X-ray spectroscopy (EDX) analysis.

Design of experiment (DoE)

Response surface methodology (RSM) by Design Expert®

Design Expert® version 6.0.5 software (Stat-Ease Inc., Minneapolis, USA) was used for design, mathematical modelling and optimization. The effects of three process variables: operating pH (A), coagulant and flocculant dosages (B and C, respectively) on the removal efficiencies of TUR (X), TSS and COD (Y and Z, respectively) in POME treatment, were evaluated using RSM with minimum number of experiment. The range for the selected variables was obtained via preliminary experiment and literature. pH range of 10–12, coagulant dosage of 1000–5000 mg/L and flocculant dosage if 100–500 mg/L was selected are shown in Table 6.
Table 6

Factors and their levels for central composite design

Variables

Symbol

Low level (− 1)

Zero (0)

High level (+ 1)

pH

A

10

11

12

Coagulant dosage (mg/L)

B

1000

3000

5000

Flocculant dosage (mg/L)

C

100

300

500

Based on the principle of CCD, the design consists of ‘2k’ fractional factorial points, ‘2k’ axial points and ‘1’ central point, where ‘k’ is the number of variables. For this research, the effects of the three variables were studied (k = 3). Therefore, 20 experiments were required for each set of coagulant–flocculant, which consisted of eight fractional factorial points, six axial points and one central point, as well as five additional replicated central points for a final estimation of experimental error. The experimental design matrix is tabulated in Table 7.
Table 7

Arrangement of process variables

Run order

No. of test jar

pH (A)

Coagulant dosage (B)

Flocculant dosage (C)

1

5

− 1

− 1

− 1

2

9

1

− 1

− 1

3

17

− 1

1

− 1

4

10

1

1

− 1

5

8

− 1

− 1

1

6

14

1

− 1

1

7

16

− 1

1

1

8

19

1

1

1

9

2

− 1

0

0

10

20

1

0

0

11

1

0

− 1

0

12

15

0

1

0

13

13

0

0

− 1

14

12

0

0

1

15

7

0

0

0

16

18

0

0

0

17

3

0

0

0

18

4

0

0

0

19

6

0

0

0

20

11

0

0

0

ANOVA was the default analysis method used to interpret the relationship between three independent and three dependent responses variables. The results from PN–OK and WG–OK experiment were analysed using statistical analyses such as Fisher’ test (F test), probability (P value) at 95% confidence level and alpha of 5% (0.05) in order to evaluate and to compare their performance in coagulation–flocculation process of POME treatment. The removal percentage of TUR, TSS and COD can be predicted by using the empirical second-order quadratic polynomial model, which is expressed in Eq (2).
$$Y = \beta_{0} + \mathop \sum \limits_{i = 1} \beta_{i} x_{i} + \mathop \sum \limits_{i = 1} \beta_{ii} x_{i}^{2} + \mathop \sum \limits_{i = 1} \mathop \sum \limits_{i \ne j = 1} \beta_{ij} x_{i} x_{ij} + \varepsilon$$
(2)
where Y is the response (removal efficiency); \(x_{i}\) is the independent variables (pH, coagulant and flocculant dosages); \(\beta_{o}\) is constant coefficient; \(\beta_{i}\), \(\beta_{ii}\) and \(\beta_{ij}\) represent coefficients for the linear, quadratic and interaction effect; and \(\varepsilon\) is the random error (Mohajeri et al. 2010).

Result and discussion

Process model

The removal efficiencies of TUR, TSS and COD from POME were recorded and fitted into model using Design Expert® as shown in Table 8. The highest observed removal efficiencies of TUR, TSS and COD in PN–OK experiment were 92.3, 90.8 and 38.8%, respectively. The highest observed removal efficiencies of TUR, TSS and COD in WG–OK experiment were 89.2, 88.5 and 46.8%, respectively. The observed (OB) and the predicted (PR) removal efficiencies were compared and are tabulated in Table 8. The predicted removal efficiencies can be obtained by using Eq. (2). Each model for percentage TUR, TSS and COD removal is shown as follows:

For PN–OK experiment,
$$\begin{aligned} {\text{TUR removal}}\;\left( \% \right) & = 78.47 + 19.67A\,{-}\,5.58B\,{-}\,2.93C\,{-}\,13.43A^{2} \\ & \quad + \,0.18B^{2} + 1.02C^{2} + 7.1AB\,{-}\,1.81AC + 1.56BC \\ {\text{TSS removal}}\;\left( \% \right) & = 73.85 + 22.12A\,{-}\,4.49B\,{-}\,2.32C\,{-}\,10.03A^{2} \\ & \quad {-}\,0.25B^{2} + 1.76C^{2} + 6.57AB + 0.67AC + 0.38BC \\ {\text{COD removal}}\;\left( \% \right) & = 31.59 + 6.76A\,{-}\,6.11B\,{-}\,3.9C\,{-}\,1.06A^{2} \\ & \quad {-}\,0.98B^{2} \,{-}\,2.48C^{2} + 3.84AB + 4.48AC\,{-}\,1.34BC \\ \end{aligned}$$
For WG–OK experiment,
$$\begin{aligned} {\text{TUR removal }}\left( \% \right) & = 76.14 + 14.19A\,{-}\,0.1B + 0.2C\,{-}\,4.92A^{2} \\ & \quad {-}\,0.045B^{2} + 0.12C^{2} \,{-}\,1.56AB\,{-}\,3.32AC\,{-}\,1.56BC \\ {\text{TSS removal }}\left( \% \right) & = 69.38 + 16.87A\,{-}\,1.32B + 0.77C\,{-}\,3.56A^{2} \\ & \quad {-}\,0.59B^{2} + 3.34C^{2} + 0.14AB\,{-}\,2.3AC\,{-}\,1.97BC \\ {\text{COD removal }}\left( \% \right) & = 28.82 + 7.35A\,{-}\,4.06B\,{-}\,3.46C\,{-}\,0.38A^{2} \\ & \quad + \,0.96B^{2} + 1.21C^{2} \,{-}\,0.68AB\,{-}\,1.05AC + 0.72BC \\ \end{aligned}$$
where A, B and C refer to parameter value (i.e. − 1, 0 and 1), as shown in Table 6.
Table 8

Observed (OB) and predicted (PR) removal efficiencies in PN–OK and WG–OK experiment

No. of Jar

PN–OK

WG–OK

TUR removal (%)

TSS removal (%)

COD removal (%)

TUR removal (%)

TSS removal (%)

COD removal (%)

OB

PR

OB

PR

OB

PR

OB

PR

OB

PR

OB

PR

1

53.9

56.6

50.0

50.9

21.4

23.2

61.3

60.7

59.9

58.2

20.6

23.5

2

47.9

45.4

45.0

41.7

25.7

23.8

59.3

57.0

50.3

49.0

27.6

21.1

3

79.8

78.5

76.4

73.9

24.4

31.6

76.8

76.1

69.9

69.4

31.7

28.8

4

86.2

84.7

86.4

85.9

37.1

37.3

84.5

85.4

82.9

82.7

32.1

35.8

5

78.5

84.3

85.4

88.3

30.1

30.8

78.5

79.1

79.7

80.8

30.1

29.4

6

79.8

76.6

78.1

73.3

26.1

25.2

75.5

76.5

73.4

73.5

26.4

26.6

7

76.5

78.5

72.6

73.9

33.4

31.6

76.8

76.1

69.3

69.4

20.6

28.8

8

82.9

78.5

77.8

73.9

38.4

31.6

76.8

76.1

69.3

69.4

20.6

28.8

9

88.9

90.7

88.7

87.4

32.9

34.2

89.2

88.7

87.7

86.2

46.8

47.9

10

81.2

78.5

77.4

73.9

36.0

31.6

71.4

76.1

64.3

69.4

26.7

28.8

11

66.7

61.9

59.5

57.6

37.6

37.3

50.9

50.6

48.8

48.1

27.1

29.8

12

73.1

78.5

66.4

73.9

22.6

31.6

76.8

76.1

69.9

69.4

31.7

28.8

13

83.2

82.4

76.8

77.9

33.8

33.0

78.5

76.1

73.7

72.0

36.4

33.5

14

81.8

73.1

76.2

69.1

24.1

24.5

80.8

76.0

74.5

67.5

25.1

25.7

15

35.1

34.3

27.3

29.5

1.5

0.6

59.7

60.5

49.4

51.3

17.4

18.2

16

69.4

78.5

65.1

73.9

31.3

31.6

75.5

76.1

70.3

69.4

30.9

28.8

17

28.0

33.5

32.7

34.8

18.8

20.1

54.3

56.6

47.3

49.1

18.9

21.6

18

92.3

90.6

90.8

90.8

33.8

32.3

87.6

88.5

85.6

87.8

37.9

37.0

19

82.5

78.1

84.5

83.3

38.8

38.0

87.6

85.6

88.5

87.1

38.2

37.5

20

79.5

84.2

74.7

78.1

38.8

36.7

72.8

76.2

64.7

70.1

37.3

33.8

Both PN–OK and WG–OK experiment results fitted the quadratic model, thereby representing the relationship between the independent variables (A, B and C) and dependent responses variables (X, Y and Z), as shown in Tables 9, 10, 11 and 12. Tables 9 and 10 show the statistical significance on the model using the ‘F value’ and ‘P value’. The ‘model (F values)’ of 16.11, 22.73 and 6.75 for TUR, TSS and COD removal from POME, respectively, in PN–OK experiment (29.56, 25.24 and 3.70 for WG–OK experiment) implied that the models were significant and less than 0.01% (except 0.3% and 2.7% for COD removal in PN–OK and WG–OK, respectively) chance that the model (F value) this large could happen due to the presence of noise. The ‘model (P value)’ should always be less than 5% to indicate the statistical significance of the model terms.
Table 9

ANOVA output for PN–OK experiment

Response

 

Sum of squares

Degree of freedom

Mean square

F value

Prob > F (P value)

Remark

TUR (X)

Model

5527.19

9

614.13

16.11

0.0001

Significant

Lack of Fit

248.80

5

49.68

1.87

0.2541

Not significant

TSS (Y)

Model

5925.05

9

658.34

22.73

0.0001

Significant

Lack of Fit

131.62

5

26.32

0.83

0.5770

Not significant

COD (Z)

Model

1353.31

9

150.37

6.75

0.0031

Significant

Lack of Fit

22.55

5

4.51

0.11

0.9842

Not significant

Table 10

ANOVA output for WG–OK experiment

Response

 

Sum of squares

Degree of freedom

Mean square

F value

Prob > F (P value)

Remark

TUR (X)

Model

2259.67

9

251.07

29.56

0.0001

Significant

Lack of Fit

61.56

5

12.31

2.63

0.1558

Not significant

TSS (Y)

Model

2993.51

9

332.61

25.24

0.0001

Significant

Lack of Fit

106.39

5

21.28

4.19

0.0711

Not significant

COD (Z)

Model

856.34

9

95.15

3.70

0.0267

Significant

Lack of Fit

115.88

5

23.18

0.82

0.5831

Not significant

Table 11

Summary of process model in PN–OK experiment

Response

Mean

Std. Dev.

R 2

Adj R2

Pred R2

CV

AP

TUR (X)

72.63

6.17

0.9355

0.8774

0.4221

8.53

13.110

TSS (Y)

69.59

5.38

0.9534

0.9115

0.8121

7.73

16.096

COD (Z)

29.33

4.72

0.8586

0.7314

0.6502

16.09

11.197

Table 12

Summary of process model in WG–OK experiment

Response

Mean

Std. Dev.

R 2

Adj R2

Pred R2

CV

AP

TUR (X)

73.72

2.91

0.9638

0.9312

0.7857

3.95

18.503

TSS (Y)

68.98

3.63

0.9578

0.9199

0.7290

5.26

15.437

COD (Z)

29.21

5.07

0.7691

0.5614

0.1061

17.36

8.297

The ‘lack of fit (F values)’ of 1.87, 0.83 and 0.11 for TUR, TSS and COD removal from POME, respectively, in PN–OK experiment (2.63, 4.19 and 0.82 for WG–OK experiment) implied that the lack of fit was not significantly relative to the pure error. There was a chance of 25.4, 57.7 and 98.4% for TUR, TSS and COD removal from POME, respectively, in PN–OK experiment (15.6, 7.1 and 58.3% for WG–OK experiment) that the lack of fit (F value) this large could occur due to the presence of noise.

The output of ANOVA in Tables 11 and 12 shows high R2 coefficients for TUR (0.9355 and 0.9638, for PN–OK and WG–OK, respectively), TSS (0.9534 and 0.9578, for PN–OK and WG–OK, respectively) and COD (0.8586 and 0.7691, for PN–OK and WG–OK, respectively), thus indicating the reliability in the estimation of TUR, TSS and COD removal efficiencies. All the predicted-R2 (Pred R2) was in reasonable agreement with the adjusted-R2 (Adj R2) except for TUR removal in PN–OK experiment and COD removal in WG–OK experiment. It may indicate a large block effect in the model. The adequate precision (AP) value is a measure of the signal-to-noise ratio. All the AP values in both experiments were greater than 4, since AP value larger than 4 is desirable for the model. It also indicated that an adequate signal and the model can be used to navigate the design space.

The observed and the predicted removal efficiencies were compared to investigate the corresponding model fitting, as shown in Fig. 1. A good distribution of data on the straight line (y = x) was observed for all the models.
Fig. 1

Predicted versus actual values plots of TUR removal using (a) PN–OK, (b) WG–OK; TSS removal using (c) PN–OK, (d) WG–OK and COD removal using (e) PN–OK, (f) WG–OK (Design Expert® Plot)

The residuals from the least square fits are important for analysing model adequacy. By constructing the normal plot of residuals, it provides an indication of the distribution of the residuals. From Fig. 2, the data points were distributed relatively close over the reference line. The lesser the points vary from this line, the greater the indication of normality. Therefore, it indicated an adequate agreement between actual and measured data.
Fig. 2

Normal plot distributions of the residuals for TUR removal using (a) PN–OK, (b) WG–OK; TSS removal using (c) PN–OK, (d) WG–OK and COD removal using (e) PN–OK, (f) WG–OK (Design Expert® Plot)

Interactions between process variables and response variables

The interaction of independent variables such as pH (A), coagulant dosage (B) and flocculant dosage (C) with dependent response variables (TUR, TSS and COD removal) was investigated using the three-dimensional (3D) plots of the regression models. Some interactions among variables were significant, and the curvature of 3D surfaces was obvious, as shown in Figs. 3, 4 and 5.
Fig. 3

TUR removal efficiency (Design Expert® Plot). (a) PN–OK at pH 11, (b) WG–OK at pH 11, (c) PN–OK at pH 12, (d) WG–OK at pH 12

Fig. 4

TSS removal efficiency (Design Expert® Plot). (a) PN–OK at pH 11, (b) WG–OK at pH 11, (c) PN–OK at pH 12, (d) WG–OK at pH 12

Fig. 5

COD removal efficiency (Design Expert® Plot) (a) PN–OK at pH 11, (b) WG–OK at pH 11, (c) PN–OK at pH 12, (d) WG–OK at pH 12

At pH 11 (central level), the effects of coagulant and flocculant dosages on the TUR removal in POME treatment are shown in Fig. 3a, b. When both coagulant and flocculant were of maximum dosages, the TUR removal efficiency was lower; when either of them was of high dosage, the TUR removal efficiency was increased by 3–4%; when both of them were of moderate or low dosage, it resulted in a higher efficiency on TUR removal for PN–OK experiment; however, it resulted in a lower efficiency for WG–OK experiment. This can be explained using coagulation mechanism: higher dosages of coagulant–flocculant may exceed the saturation of intraparticle bridging formed between colloids and bridging molecules; therefore, the excessive coagulant–flocculant has tendency to destroy the intraparticle bridging, which causes an increase in residual TUR. This result also indicated the effect of pH on the turbidity removal. Different optimum operating pH can be observed for different types of coagulant used in POME treatment.

The effect of pH on TUR removal can be noticed by comparing it with Fig. 3c, d. At high pH (maximum level), wheat germ coagulant functioned better in removing TUR of POME as compared to lower pH. With the pH range from 11 to 12, the TUR removal efficiency increased considerably from 77 to 88% for WG–OK experiment. However, the TUR removal efficiency reduced with increasing pH for peanut coagulant. For both types of coagulants, the dosage of coagulant became less dominant to removal efficiency at high pH. This phenomenon can be explained by the effect of charge neutralization, which is one of the mechanisms in coagulation process. A high turbid POME contains hydrophilic colloids which has high affinity for water. It is due to the presence of polar functional groups such as hydroxyl, carboxylic and amine groups (Sincero and Sincero 2003). By properly increasing pH with the addition of NaOH solution, it causes the added OH to neutralize the acid end of the zwitter ion (usually the NH3+), which is the corresponding ion of the colloid. The zwitter ion disappears, and the hydrophilic colloid attains a negative primary charge. The primary charges of the colloids are neutralized by addition of natural coagulant–flocculant in order to initiate destabilization between suspended particles (Sincero and Sincero 2003).

Figure 4a, b shows the effect of coagulant and flocculant dosages on the TSS removal from POME at pH 11 (central point). For PN–OK experiment, the best TSS removal efficiency was observed when both coagulant and flocculant were of low dosage. The TSS removal efficiency decreased with increasing coagulant and flocculant dosages. It may be caused by overdosing of coagulant–flocculant, which is counterproductive. For WG–OK experiment, the best removal efficiency was obtained when low coagulant dosage and high flocculant dosage were applied.

A higher TSS removal efficiency was achieved when pH was increased from 11 to 12 for both types of coagulants, as shown in Fig. 4c, d. For PN–OK experiment, TSS removal efficiency increased with increasing coagulant dosage at pH 12. For WG–OK experiment, a similar interaction between coagulant and flocculant dosages was obtained for different pH (i.e. pH 11 and pH 12). However, TSS removal efficiency increased with increasing coagulant dosage only when the flocculant was of low dosage. By increasing the flocculant dosage, TSS removal efficiency decreased from 87.28 to 80%. From Fig. 4, it is possible to observe the effect of overdosing of coagulant which resulted in a low TSS removal efficiency. The addition of excessive coagulant may cause the reversal of surface charge and re-stabilizing the colloidal particles, thus reducing the removal efficiency in POME treatment.

For COD removal from POME, both experiments showed a similar correlation between COD removal efficiency and coagulant–flocculant dosages at pH 11, as shown in Fig. 5a, b. The lowest COD removal efficiency was caused by the high dosages of both coagulant and flocculant.

When coagulant and flocculant were of low dosage, it resulted in a highest COD removal from POME. This phenomenon can be explained by the overdosing effect, which causes the saturation of intraparticle bridging formed between organic matters and bridging molecules; therefore, the excessive coagulant–flocculant has tendency to destroy the intraparticle bridging, which causes an increase in residual COD.

At high pH (maximum level), the COD removal efficiency of both PN–OK and WG–OK experiments increased as compared to lower pH. For PN–OK experiment, the highest COD removal efficiency can be achieved by low coagulant dosage. However, it was highly dependent on the flocculant dosage. For WG–OK experiment, a similar interaction between coagulant and flocculant dosages was obtained. However, with the pH range from 11 to 12, the highest COD removal efficiency increased significantly from 39.0 to 47.4%. A low COD removal efficiency range from 40 to 60% can be observed by the previous studies performed using natural coagulant–flocculant alone, which is previously shown in Table 2. With the addition of chemical coagulant–flocculant, the COD removal in treatment can increase significantly. The dosage of chemical coagulant (Fe3+) was reduced when it combined with natural flocculant (okra) in treatment for achieving maximum COD removal (Freitas et al. 2015). In order to comply with the regulatory discharge limits, as previously shown in Table 1, it is recommended that the moderate COD removal by applying natural coagulant–flocculant requires further treatment such as filtration process and reverse osmosis. The treated POME is shown in Fig. 6.
Fig. 6

Treated POME with flocs settled at the bottom of the beaker

Optimization and validation experiment

The high desirability option in the numerical optimization of RSM was applied to determine the optimum process parameters such as pH, coagulant and flocculant dosages in order to maximize the removal efficiencies of TUR, TSS and COD from POME. The optimum operating parameters are tabulated in Table 13. The validation experiment was performed by using the optimum operating parameters in order to obtain the observed values of the removal efficiencies of TUR, TSS and COD in POME treatment. The obtained and predicted removal efficiencies were compared for validation purpose, as shown in Table 14. By comparing the result between simulated values and laboratory experiment, a small variation was observed for both PN–OK and WG–OK. It clearly indicates that the RSM using CCD model is an excellent tool to obtain the optimum operating parameters of coagulation–flocculation of POME treatment for maximum TUR, TSS and COD removal.
Table 13

Optimum operating parameters (Design Expert® numerical optimization)

Variables

PN–OK

WG–OK

pH

11.6

12.0

Coagulant dosage (mg/L)

1000.1

1170.5

Flocculant dosage (mg/L)

135.5

100.0

Desirability

0.957

0.978

Table 14

Observed and predicted result based on validation experiment

Response variables

PN–OK

WG–OK

Observed

Predicted

Observed

Predicted

TUR removal (%)

92.5

92.3

86.6

88.7

TSS removal (%)

86.6

86.8

87.5

86.4

COD removal (%)

34.8

36.3

43.6

46.8

The optimum results for both PN–OK and WG–OK, as shown in Table 14, are in agreement with the study by Saifuddin and Dinara (2011), which reported that chitosan was able to achieve the best TUR, TSS and COD reductions in POME treatment, which was 97.7, 91.7 and 42.7%, respectively, at dosage of 370 mg/L and pH 6. Furthermore, the optimum result using either PN–OK or WG–OK as coagulant–flocculant is also comparable to other studies on coagulation–flocculation, as previously shown in Table 2.

Characteristics of coagulants and flocculant

FTIR spectra of okra, peanut and wheat grain are shown in Fig. 7a, b. The FTIR spectra of both coagulants are similar in terms of the location of the peak. For both coagulants, at 3300–3250 cm−1, there is a broad band due to the stretching of O–H stretching, thereby indicating that the presence of OH functional group. Furthermore, C–H stretching of aliphatic structures assigned to fatty acids and lipids was detected at about 2854 and 2923 cm−1 (Araújo et al. 2010). The presence of O–H and C–H stretching could suggest two main functional groups in the coagulant, which are fatty acids. C=O stretch was found at the peak from 1800 to 1600 cm−1 (Shak and Wu 2014). In addition, at 1047 cm−1, peak was detected due to the stretching of SO42− (Teh et al. 2014a). The FTIR spectrum of okra is similar to peanut and wheat germ. But when it comes to the intensity, it is seen that PN has greater intensity when compared to WG, hence the better its performance. PN had greater concentration of active groups when compared to WG, as seen from the FTIR results. The performance of PN was better as the active group concentration was higher. Thus, a lower dosage of PN was needed when compared to WG, to achieve the same level of performance. The FTIR spectra of okra, peanut and wheat grain are in agreement with the study performed by Shak and Wu (2014), which confirmed the presence of fatty acids in C. obtusifolia seed coagulant through the analysis of FTIR spectrum.
Fig. 7

FTIR spectra of (a) peanut (yellow) and wheat germ (red), (b) okra. (Frontier FTIR plot)

The elemental compositions of okra, peanut and wheat germ were determined by EDX analysis. The result is tabulated in Table 15. The most abundant elements in both coagulant–flocculant are carbon and oxygen. The other elements such as aluminium and potassium were found trace in concentration. FESEM images of okra, peanut and wheat grain were generated and investigated for studying their surface morphology, as shown in Fig. 8. It shows the presence of fibrous network with rough and high porosity on the surfaces of coagulant–flocculant, which can be related to their coagulation–flocculation performance. The rough and highly porous surfaces of natural coagulant–flocculant increase the overall surface area, which effectively increases the adsorption sites for bridging mechanism (Shak and Wu 2015).
Table 15

EDX elemental analysis (in wt%)

Element

Okra

Peanut

Wheat germ

C

53.32

78.31

62.93

O

41.72

20.63

36.62

Mg

0.57

0.00

0.00

Al

0.97

0.20

0.58

P

0.50

0.30

0.00

S

0.11

0.18

0.00

Cl

0.63

0.00

0.00

K

1.60

0.52

0.26

Ca

0.59

0.00

0.00

Fig. 8

FESEM images of okra (left), peanut (middle) and wheat germ (right)

Natural coagulants consist mainly of polysaccharides, fatty acids and proteins, which contribute to the coagulating ability in water treatment (Antov et al. 2010). Based on the EDX results, the peanut and wheat germ do not consist of polysaccharides on their surface as lower amount of oxygen element present as compared to carbon element. In addition, the phosphorus element present in peanut, as shown in Table 15, may indicate the presence of phospholipids acids in the peanut coagulant (Ramavandi 2014). However, the phosphorus element was absent in wheat germ coagulant, which may contribute to a lower efficiency in removal efficiencies of TUR, TSS and COD in POME treatment as compared to peanut coagulant.

Based on the EDX analysis of okra flocculant, it was found that the carbon and oxygen element are similar in terms of weight percentage, which could be attributed to the empirical formula of polysaccharides. Therefore, okra flocculant could be made from polysaccharides. Based on the reported study, mucilage found in okra flocculant is a highly branched carbohydrate polymer, which is also known as polysaccharides (Ramavandi 2014). The mechanism of okra mucilage was deducted to be adsorption and bridging, which take place at the surface of okra. The presence of polysaccharides in okra serves as bridges to link colloidal particles together, resulting in the increment of floc size which improves the coagulation–flocculation process (Choy et al. 2014).

Characterization of sludge

FTIR spectra of PN sludge and WG sludge are shown in Fig. 9. The FTIR spectra of two types of sludge are similar in terms of the location of the peak. For both types of sludge, there is a strong peak at 1380 cm−1 due to the stretching of C–O bond. O–H bend was detected at the medium peak at 1000 cm−1 due to the presence of carboxylic groups. The elemental compositions of PN sludge and WG sludge were determined by EDX analysis, as tabulated in Table 16. The most abundant elements in both types of sludge are carbon and oxygen. The other significant elements such as sodium, magnesium, chloride and potassium were found trace in concentration. The trace metal present in sludge may be caused by the sediments of POME. As compared to elemental compositions of OK, PN and WG, as shown in Table 15, the elements of Na and Cl were found present in the sludge due to the addition of NaCl solution for the extraction of active agents in natural coagulant–flocculant. FESEM images of PN sludge and WG sludge were generated and investigated for studying their surface morphology, as shown in Fig. 10. The surfaces of solid sludge appeared relatively compact and clustered compared to their corresponding natural coagulant–flocculant, as shown in Fig. 7. It is due to the agglomeration between the colloid particles formed by adsorption and intraparticle bridging.
Fig. 9

FTIR spectra of PN sludge (purple) and WG sludge (green). (Frontier FTIR plot)

Table 16

Elemental analysis of PN sludge and WG sludge (in wt%)

Element

PN sludge

WG sludge

C

31.46

35.59

O

34.40

23.01

Na

10.64

9.62

Mg

5.92

4.00

Al

0.36

0.23

Si

1.32

0.81

P

0.57

0.31

S

0.20

0.00

Cl

8.53

16.26

K

5.00

8.92

Ca

1.59

1.24

Fig. 10

FESEM images of PN sludge (left) and WG sludge (right)

After coagulation–flocculation, a significant amount of sludge is produced. Sludge treatment is commonly required after wastewater treatment. Recently, Zahrim et al. (2017) showed that the coagulated palm oil mill sludge has germination index greater than 80% indicating its potential as soil conditioner. Therefore, a high sludge dewatering capability is important in volume reduction, which can effectively reduce the sludge treatment cost. Table 17 shows the recovery of water by sludge dewatering. The water recovery from PN sludge using centrifuge method was 72.8%, which was higher than that from WG sludge (61.38%). However, after drying process, it was found that the solid content of PN concentrated sludge (5.2%) was higher than WG concentrated sludge (4.1%). By using centrifuge and drying methods, more than 98% of water can be removed from sludge produced in POME treatment for both types of natural coagulant–flocculant. Therefore, only 2% of solid remained in POME sludge requires further treatment and handling after coagulation–flocculation process.
Table 17

Sludge dewatering analysis

 

PN–OK

WG–OK

Mass (g)

Mass fraction (%)

Mass (g)

Mass fraction (%)

100 mL sludge (before centrifuge)

94.0

100

83.8

100

Water recovered by centrifuge

68.5

72.8

51.5

61.4

Concentrated sludge (after centrifuge)

25.4

27.2

32.3

38.6

Dried solid sludge

1.3

5.2

1.3

4.1

Water loss by drying

24.1

94.8

31.0

95.9

Overall mass reduction

92.7

98.6

82.5

98.4

Conclusion

An evaluation of two pairs of coagulant–flocculant (peanut–okra and wheat germ–okra) was conducted for the removal of turbidity (TUR), total suspended solid (TSS) and chemical oxygen demand (COD) in palm oil mill effluent (POME) treatment. The application of response surface methodology (RSM) using central composite design (CCD) model by Design Expert® was performed in this research to determine the maximum removal efficiencies of TUR, TSS and COD and the optimum operating parameters for coagulation–flocculation process of POME treatment. The performance of peanut–okra (PN–OK) was found comparable to wheat germ–okra (WG–OK) as natural coagulant–flocculant in POME treatment. The optimum operating parameters of coagulation–flocculation process by using PN–OK as coagulant–flocculant were pH 11.6, 1000.1 mg/L PN dosage and 135.5 mg/L OK dosage in order to yield the maximum TUR, TSS and COD removal efficiencies of 92.5, 86.6 and 34.8%, respectively. The optimum operating parameters using WG–OK as coagulant–flocculant were pH 12, 1170.5 mg/L WG dosage and 100 mg/L OK dosage in order to yield the maximum TUR, TSS and COD removal efficiencies of 86.6, 87.5 and 43.6%, respectively. The high percentage of TUR and TSS removal showed reliable results for POME treatment in comparison with the previous studies. However, the moderate COD removal requires further studies. It could have happened due to the presence of oil in both the PN and WG, which added on to the COD load. The removal of oil prior to usage as coagulants will be studied in the future. Sludge dewatering by centrifuge and drying achieved 98% water removal, hence effectively reducing the volume of sludge to be treated. In conclusion, peanut–okra and wheat germ–okra have the potential as a natural coagulant–flocculant in POME treatment.

Notes

Acknowledgements

We want to express our gratitude to Seri Ulu Langat Palm Oil Mill Sdn Bhd for providing the POME samples throughout the research studies.

References

  1. Ahmad A, Chong M, Bhatia S, Ismail S (2006) Drinking water reclamation from palm oil mill effluent (POME) using membrane technology. Desalination 191(1–3):35–44CrossRefGoogle Scholar
  2. Ahmad A, Chong M, Bhatia S (2009) A comparative study on the membrane based palm oil mill effluent (POME) treatment plant. J Hazard Mater 171(1–3):166–174CrossRefGoogle Scholar
  3. Anastasakis K, Kalderis D, Diamadopoulos E (2009) Flocculation behaviour of mallow and okra mucilage in treating wastewater. Desalination 249(2):786–791CrossRefGoogle Scholar
  4. Antov M, Šćiban M, Petrović N (2010) Proteins from common bean (Phaseolus vulgaris) seed as a natural coagulant for potential application in water turbidity removal. Biores Technol 101(7):2167–2172CrossRefGoogle Scholar
  5. Araújo C, Alves V, Rezende H, Almeida I, de Assunção R, Tarley C, Segatelli M, Coelho N (2010) Characterization and use of Moringa oleifera seeds as biosorbent for removing metal ions from aqueous effluents. Water Sci Technol 62(9):2198CrossRefGoogle Scholar
  6. Bhatia S, Othman Z, Ahmad A (2007) Pretreatment of palm oil mill effluent (POME) using Moringa oleifera seeds as natural coagulant. J Hazard Mater 145(1–2):120–126CrossRefGoogle Scholar
  7. Birima A, Hammad H, Desa M, Muda Z (2013) Extraction of natural coagulant from peanut seeds for treatment of turbid water. IOP Conf Ser Earth Environ Sci 16:012065CrossRefGoogle Scholar
  8. Birima A, Ahmed A, Noor M, Sidek L, Muda Z, Wong L (2014) Application of salt extracted peanut seeds in the pretreatment of palm oil mill effluent (POME). Desalin Water Treat 55(8):2196–2200CrossRefGoogle Scholar
  9. Choy S, Prasad K, Wu T, Raghunandan M, Ramanan R (2014) Utilization of plant-based natural coagulants as future alternatives towards sustainable water clarification. J Environ Sci 26(11):2178–2189CrossRefGoogle Scholar
  10. Daud NS, Ghazi TIM, Ahamad IS (2014) Wheat germ as natural coagulant for treatment of palm oil mill effluent (POME). Int J Chem Environ Eng 5:111–114Google Scholar
  11. Flaten T (2001) Aluminium as a risk factor in Alzheimer’s disease, with emphasis on drinking water. Brain Res Bull 55(2):187–196CrossRefGoogle Scholar
  12. Freitas T, Oliveira V, de Souza M, Geraldino H, Almeida V, Fávaro S, Garcia J (2015) Optimization of coagulation–flocculation process for treatment of industrial textile wastewater using okra (A. esculentus) mucilage as natural coagulant. Ind Crops Prod 76:538–544CrossRefGoogle Scholar
  13. Garde W, Buchberger S, Wendell D, Kupferle M (2017) Application of Moringa Oleifera seed extract to treat coffee fermentation wastewater. J Hazard Mater 329:102–109CrossRefGoogle Scholar
  14. Latif Ahmad A, Ismail S, Bhatia S (2003) Water recycling from palm oil mill effluent (POME) using membrane technology. Desalination 157(1–3):87–95CrossRefGoogle Scholar
  15. Lee C, Robinson J, Chong M (2014) A review on application of flocculants in wastewater treatment. Process Saf Environ Prot 92(6):489–508CrossRefGoogle Scholar
  16. Ma AN (2000) Environmental management for the oil palm industry. Palm Oil Dev 30:1–10Google Scholar
  17. Meraz K, Vargas S, Maldonado J, Bravo J, Guzman M, Maldonado E (2016) Eco-friendly innovation for nejayote coagulation–flocculation process using chitosan: evaluation through zeta potential measurements. Chem Eng J 284:536–542CrossRefGoogle Scholar
  18. Miller S, Fugate E, Craver V, Smith J, Zimmerman J (2008) Toward understanding the efficacy and mechanism of opuntia spp. as a natural coagulant for potential application in water treatment. Environ Sci Technol 42(12):4274–4279CrossRefGoogle Scholar
  19. Mohajeri L, Aziz H, Isa M, Zahed M (2010) A statistical experiment design approach for optimizing biodegradation of weathered crude oil in coastal sediments. Biores Technol 101(3):893–900CrossRefGoogle Scholar
  20. Ndabigengesere A, Narasiah K, Talbot B (1995) Active agents and mechanism of coagulation of turbid waters using Moringa oleifera. Water Res 29(2):703–710CrossRefGoogle Scholar
  21. Oladoja N (2015) Headway on natural polymeric coagulants in water and wastewater treatment operations. J Water Process Eng 6:174–192CrossRefGoogle Scholar
  22. Quah SK, Lim KH, Gillies D, Wood BJ, Kanagaratnam J (1982) Proc. regional workshop on. Palm oil mill technology and effluent treatment. PORIM, Kuala Lumpur, pp 193–200Google Scholar
  23. Ramavandi B (2014) Treatment of water turbidity and bacteria by using a coagulant extracted from Plantago ovata. Water Resour Ind 6:36–50CrossRefGoogle Scholar
  24. Rupani PF, Sigh RP, Ibrahim MH, Esa N (2010) Review of current palm oil mill effluent (POME) treatment methods: vermicomposting as a sustainable practice. World Appl Sci J 10(10):1190–1201Google Scholar
  25. Saifuddin N, Dinara S (2011) Pretreatment of palm oil mill effluent (POME) using magnetic chitosan. E J Chem 8(s1):S67–S78CrossRefGoogle Scholar
  26. Sanghi R, Bhattacharya B, Dixit A, Singh V (2006) Ipomoea dasysperma seed gum: an effective natural coagulant for the decolorization of textile dye solutions. J Environ Manag 81(1):36–41CrossRefGoogle Scholar
  27. Shak K, Wu T (2014) Coagulation–flocculation treatment of high-strength agro-industrial wastewater using natural Cassia obtusifolia seed gum: treatment efficiencies and flocs characterization. Chem Eng J 256:293–305CrossRefGoogle Scholar
  28. Shak K, Wu T (2015) Optimized use of alum together with unmodified Cassia obtusifolia seed gum as a coagulant aid in treatment of palm oil mill effluent under natural pH of wastewater. Ind Crops Prod 76:1169–1178CrossRefGoogle Scholar
  29. Sincero A, Sincero G (2003) Physical–chemical treatment of water and wastewater, 1st edn. CRC Press, Boca RatonGoogle Scholar
  30. Syafalni, Lim H, Ismail N, Abustan I, Murshed M, Ahmad A (2012) Treatment of landfill leachate by using lateritic soil as a natural coagulant. J Environ Manag 112:353–359CrossRefGoogle Scholar
  31. Tajuddin HA, Luqman Chuah A, Choong Thomas S Y (2015) Microbial community analysis in anaerobic palm oil mill effluent (POME) wastewater by denaturing gradient gel electrophoresis (DGGE). Int J Res Eng Technol 4(8):1–9CrossRefGoogle Scholar
  32. Teh C, Wu T, Juan J (2014a) Potential use of rice starch in coagulation–flocculation process of agro-industrial wastewater: treatment performance and flocs characterization. Ecol Eng 71:509–519CrossRefGoogle Scholar
  33. Teh C, Wu T, Juan J (2014b) Optimization of agro-industrial wastewater treatment using unmodified rice starch as a natural coagulant. Ind Crops Prod 56:17–26CrossRefGoogle Scholar
  34. Zahrim AY, Dexter ZD, Collin JG, Hilal N (2017) Effective coagulation–flocculation treatment of highly polluted palm oil mill biogas plant wastewater using dual coagulants: decolourisation, kinetics and phytotoxicity studies. J Water Process Eng 16:258–269CrossRefGoogle Scholar

Copyright information

© Springer-Verlag GmbH Germany, part of Springer Nature 2018

Authors and Affiliations

  1. 1.Department of Chemical and Environmental Engineering, Faculty of EngineeringThe University of Nottingham Malaysia CampusSemenyihMalaysia

Personalised recommendations