Potential application of natural coagulant extraction from walnut seeds for water turbidity removal

Water treatment is surely one of the main factors that are involved in the human development considering its direct influence on human lives and its uses in other lifestyles (Antov et al. 2012; Mahanna et al. 2015). Different technologies have been developed for water treatment. Of these technologies, coagulation is considered one of the most widely processes employed for the removal of particles, turbidity, and natural organic matter (Jeon et al. 2009; Mahanna et al. 2018; Ang & Mohammad 2020).

Mainly, the coagulation of suspended particles can be accomplished by four mechanisms: charge neutralization, bridging, double layer compression, and sweep coagulation (Choy et al. 2015). The choice of coagulant is considered to play a major role for pollutant removal. Chemical coagulants, especially inorganic salts and synthetic polymers, are the most widely used coagulants due to their efficiency (Kakoi et al. 2016). Recently, pre-hydrolyzed coagulants such as polyaluminum chlorides and pre-hydrolyzed iron have been widely used to remove suspended solids, colloids, and dissolved solids. Pre-hydrolyzed coagulants are produced by moderately neutralization of the monomeric coagulants to a various basicity ratio for controlling the formation of coagulating particles. This procedure generates highly positive and larger coagulant particles (Zin et al. 2015; Rosińska & Dąbrowska 2021). The advantages of pre-hydrolyzed coagulants over conventional coagulants are the higher effectiveness with less dependence on the pH and temperature (Gumińska & Kłos 2015; Dabrowska 2018).

However, there are some disadvantages induced by using chemical coagulants such as relatively high costs, low effectiveness at low temperature, production of large volume of sludge, affecting the treated water pH, and adverse effects on human health (Bazrafshan et al. 2015; de Oliveira Cardoso Nascimento et al. 2021). Furthermore, the resulting elevated concentration of aluminum residuals in water are linked with neurodegenerative diseases such as Alzheimer's, as well as neurotoxic and carcinogenic effects (Rondeau et al. 2000; Huang et al. 2015). Moreover, aluminum is not biodegradable and can cause environmental problems during the treatment and disposal of the generated sludge (Liew et al. 2006; Kakoi et al. 2016).

Thus, natural coagulants become a potential choice with many advantages over chemical coagulants, especially biodegradability, cost savings, lower toxicity, lower risks to environment and health, lower residual sludge production, and lower coagulant dose (Narasiah et al. 2002). In recent studies, a variety of natural materials has been reported as a source of natural coagulants such as Cactus (Zhang et al. 2006), Moringa oleifera seed (Bhuptawat et al. 2007; Ali et al. 2010; Baptista et al. 2017), Chestnut and Acorn (Šćiban et al. 2009), tannin (Beltrán-Heredia et al. 2010), Phaseolus vulgaris seed (Antov et al. 2010), Plantago ovata (Ramavandi 2014; Dhivya et al. 2017), Pistacia atlantica seed (Bazrafshan et al. 2015), Banana pith (Kakoi et al. 2016), Quercus robur (Antov et al. 2018), Chitosan/CHPATC (Mohammad et al. 2018), and Pine cone (Hussain et al. 2019). Moreover, Jatropha curcas seeds (Abidin et al. 2013), Cassia angustifolia seed (Sanghi et al. 2002), and Nirmali seed and maize (Raghuwanshi et al. 2002) have been investigated as a natural coagulant.

Starch is also getting the attention of many researchers as a source of natural coagulant (Teh et al. 2014). Starches from wheat, rice, potato and corn were used for turbidity removal (Zafar et al. 2015; Choy et al. 2016). Orchis mascula starch was applied as a natural coagulant in synthesized bilge water treatment (Hamidi et al. 2021).

Artificial neural networks (ANNs) are mathematical modeling tools that are mainly used in prediction and forecasting of complex processes. ANNs have been recently employed to solve numerous prediction problems. There are several types of ANNs available, however the most commonly used is the feed-forward back-propagation network. The architecture of a feed-forward network includes input, hidden, and output layers linked to each other by neurons. ANN's modeling has been utilized in the field of water and wastewater treatment (O'Reilly et al. 2018). Moreover, ANNs modeling has been used to predict both turbidity and dissolved organic matter removal during the coagulation process (Kennedy et al. 2015). Modeling jar test experiments, optimum alum dose, and treated water quality using ANNs has been examined (Maier et al. 2004; Haghiri et al. 2018). In addition, prediction of water quality index using ANNs has been applied (Hameed et al. 2017). Until now, there is not a study of using walnut seed extract as a natural coagulant. Moreover, Artificial neural networks (ANNs) modeling of coagulation process using natural coagulants is required to be applied. Therefore, the main objective of the present study is to provide a good understanding on the possibility of using walnut seed extract as a separate natural eco-friendly coagulant or combined with alum as coagulant aid to reduce the amount of alum for turbidity removal. Moreover, ANNs modeling of turbidity removal using extraction of walnut seed was performed.

Walnut seeds were brought locally from the city of Mansura-Egypt. First, they were washed with distilled water, and dried at room temperature for 24 hr. Then, they were ground down to a fine powder using a laboratory mill and then sieved through 0.4 mm sieve. Five grams of prepared powder was added to 1,000 mL of NaCl solution (1M). The suspension was stirred using a magnetic stirrer with 500 rpm for 1 hr at room temperature to get best extraction and coagulation active components. The suspension was filtered with a piece of gauze, gravity filtered through a rugged filter paper, and then through a filter of 45 μm. The filtered solutions, called crude extracts, were kept in a refrigerator at 4 °C for further use.

Different techniques were used for characterization of walnut seeds to establish a thorough understanding of the feasibility of using walnut as natural coagulant in the treatment of turbid water. A Scanning Electron Microscope (SEM) (JEOL JSM 6510 LV, Jaban) was used to identify the surface morphology of samples and give a high-quality and clear stereoscopic image. The elemental analyses of coagulant were conducted using Energy dispersive X-ray spectroscopy (EDX) (Oxford X-MaxN 20, USA). Furthermore, the chemical composition of the investigated walnut seeds and extract was investigated according to the AOAC method. The Fourier transform infrared spectroscopy (FTIR) (Nicolet™ IS10, USA) was used for functional groups investigation with the spectral range of 4,000–400 cm⁻¹. Moreover, a zeta potential analyzer (Malvern Zetasizer Nano-ZS90, UK) was used to measure the zeta potentials of samples.

A turbid water for coagulation tests synthetically was prepared by adding kaolin stock into tab water used for irrigation of a green area discharged from the river Nile directly without treatment. The tab water has an initial average turbidity of 5 NTU and total organic carbon (TOC) of 42 mg/L. The kaolin was added to achieve the high desired levels of turbidity. 15 g of kaolin was added into 1 L of tab water and was stirred for 2 hr at 120 rpm to achieve a uniform dispersion of kaolin particles, and then it was allowed to remain without stirring for 24 h to complete the hydration of the particles (Okuda et al. 2001). Suspensions with different initial turbidity concentrations ranged from 10 to 200 NTU were prepared by dilution using distilled water just before the coagulation experiment. The initial pH of the synthetic water was adjusted to desired values (3–10) using 1 M NaOH or HCl.

Jar test experiments were conducted to evaluate the coagulation activity of the natural coagulant. Six beakers of 1 L capacity were filled with 500 ml of the prepared solution with the required turbidity concentration and placed in a conventional jar tester. Various doses of natural coagulant, walnut seed extract, were added to each beaker and stirred for 5 min at 140 rpm as rapid mixing. Then, the mixing speed was reduced to 40 rpm for 30 min as slow mixing (except for experiments studied the effect of slow mixing time on the turbidity removal efficiency). The stirrer was then stopped, and all suspensions were left for sedimentation for 1 hr (except for experiments investigated the effect of settling time on the turbidity removal efficiency). After that, the samples for the residual measurement were collected carefully from the top of each beaker. The turbidity of each sample was measured using a turbidity meter (TB300 IR Turbid meter, P/N TB300IR – 10, S/N 12/1727). The schematic diagram for the preparation and use of natural coagulant is shown in supplementary materials Fig. S1. All experiments were carried out at room temperature, 25 ± 2 °C.

The effects of operational parameters such as initial turbidity, natural coagulant dose, pH, and slow mixing and settling time were investigated. The experiments of natural coagulant dose and initial turbidity effect were conducted using different doses ranged from 1 to 5 ml/L, agitated with different initial turbidities of 13, 54, and 194 NTU at 140 rpm for 5 min then 30 rpm for 30 min then remaining to settle for 60 min. The experiments of pH effect were conducted using the optimum coagulant dose agitated with different initial turbidities of 8, 56, and 190 NTU at various pH values ranging from 3 to 11 at 140 rpm for 5 min then 30 rpm for 30 min then remain to settle for 60 min. Samples were collected for the final turbidity test after 30 and 60 min of settling.

The experiments of slow mix time effect were conducted using the optimum coagulant dose agitated with initial turbidity of 45 NTU at pH 8 at 140 rpm for 5 min then 30 rpm for different slow mixing times ranging from 10 to 35 min then remaining to settle for 60 min. Samples were collected for the final turbidity test after 30 and 60 min of settling.

The experiments of settling time effect were conducted using the optimum coagulant dose agitated with different initial turbidities of 10, 100, and 200 NTU at pH 8 at 140 rpm for 5 min then 30 rpm for 30 min then remain to settle for 90 min. Samples were collected for the final turbidity test after different settling times (15, 30, 45, 60, 75, and 90 min).

Moreover, to determine the effect of combination of walnut extract with alum, different walnut extract doses (0, 1, 3, and 5 ml/L) combined with alum doses ranged from 13 to 20 mg/L were agitated with initial turbidity 11 NTU at pH 8 at 140 rpm for 5 min then 30 rpm for 30 min then remaining to settle for 30 min. Samples were collected for the final turbidity test after 30 min settling time. In addition, the effect of initial turbidity (11 to 152 NTU) was studied at optimum dose of alum combined with walnut extract.

The Final and initial turbidity of the sample are corresponded to the values of turbidity after and before treatment, respectively. A turbid sample without the addition of any coagulant was prepared and used in a jar test as a blank where the residual turbidity measured after settling was termed the ‘Final turbidity of blank’ in Equation (2).

Organic load in treated water experiments

The use of some natural coagulants contributes to increase of organic matter in the treated water (Camacho et al. 2017). In this work, organic matter as total organic carbon (TOC) in water after and before the coagulation process was tested to examine the increase in organic load using walnut natural coagulant. Five beakers of 1 L were prepared with a fixed initial turbidity of 74 NTU and TOC = 42 mg/L with the same previous method at pH equal to 8. Different doses of walnut extract ranged from 2 to 5 ml/L were added to the beakers. In addition, in one beaker, a turbid sample without the addition of any coagulant was prepared and used in jar test as a blank where the TOC value measured after settling was termed as the ‘Final TOC of blank’. TOC was measured for each beaker using TOC analyzer (AJ-Analyzer multi-N/C 3100). Moreover, another experiment was conducted to study the effect of using alum on the TOC value compared to using walnut extract. For a fixed turbidity of 73 NTU and pH equal to 8, walnut extract dose of 3 ml/L and alum dose of 16 mg/L were used separately and combined.

SEM microphotograph for the walnut powder showed the presence of platelet networks with rough and porous surface as shown in Figure 1(a) and 1(b), which imposed an influence on the coagulation–flocculation process (Shak & Wu 2015). The porous and rough surfaces help to improve the bridging mechanism due to the presence of a larger surface area, which provides more adsorption sites. Energy dispersive X-ray spectroscopy (EDX) of the walnut powder confirmed the presence of C, N, O, Na, Mg, P, S, K, Ca, Fe and Cu compounds as shown in Figure 1(c). The presence of O and C is related to the protein content in walnut seeds, which contributes to achieving the high performance of this natural material in the coagulation process. Furthermore, the presence of minerals in the walnut seed composition suggests the availability of free active sites for chemical adsorption of anions and cations (Mateus et al. 2018).

Table 1 presents the chemical characterization of the walnut seed powder and the walnut seed extraction. It was shown that the extraction resulted in a reduction of the lipid of the seeds. Walnut seed lipids were reduced by ∼99% with extraction. The results also showed that the crude fibers of the walnut seeds vanished with extraction. So, this reduces the water turbidity and the treatment level required. Moreover, the protein content of the walnut seeds is nearly like the walnut extract, so the extraction has low effect on the protein content of the seeds, which increases coagulation and turbidity removal efficiency (Hussain et al. 2019).

The functional groups present in walnut seed powder were studied by FTIR technique as shown in Figure 2. The wide band at 3,422 cm⁻¹ and bands from 3,653 to 3,754 cm⁻¹ are attributed to O-H stretching that can be found in the lignin, carbohydrates, protein, and fatty acids (Baptista et al. 2017; Younes et al. 2019). The peaks from 2,856 to 3,010 cm⁻¹ corresponded to C–H (Jadhav & Mahajan 2014; El-bendary et al. 2021b). The characteristic peaks of the COO- double bond of deprotonated carboxylate groups were observed at 1,655 and 1,746 cm⁻¹ (Kakoi et al. 2016). The 1,545 cm⁻¹ was ascribed to the C-N elongation vibration due to the amide group (Mateus et al. 2018). Peaks from 1,238 to 1,461 cm⁻¹ corresponded to the C-O stretching vibration (Wan et al. 2019; El-Bendary et al. 2021a). The characteristic vibrations of the elongation of the C-OH hydroxyl bond were observed at bands 1,100 and 1,163 cm⁻¹ (Mateus et al. 2018). Peaks at 914 and 994 cm⁻¹ are assigned to C-O-C skeletal mode vibration. While the peaks located at 722 and 847 cm⁻¹ suggested the presence of CH₂ deformation (Choy et al. 2016). The other peaks at 518 and 593 cm⁻¹ could be associated with metal oxygen and metal hydroxyl (Mahanna & Azab 2020; Younes et al. 2021) Thus, the presence of different functional groups at various wave numbers for walnut natural coagulant should be the principal reasons for the effective coagulation activity.

Figure 3 shows the zeta potential of walnut natural coagulant as a function of pH. It shows that the walnut extract coagulant has a wide negative charge potential by increasing pH value. The presence of hydroxyl groups and carboxyl groups of proteins on the coagulant surface (as demonstrated in the previous FTIR analysis), which varies with protein content ratio, is associated with the negative charge potential (Choy et al. 2016). The value of zeta potential reached −24.73 mV at pH 10. The higher intense charge variation for walnut extract implies that there were higher numbers of dissociable groups caused by the pH change (Mateus et al. 2018). High zeta potential values indicate high coagulation activity performance of the natural coagulant (Baptista et al. 2017).

Coagulant dose is a major parameter that has been considered to assess the performance of coagulation and flocculation process. Mostly, overdosing or insufficient doses may weaken the flocculation performance. Therefore, the effect of walnut coagulant dose on the coagulation activity and turbidity removal efficiency was investigated at different initial turbidities as shown in Figure 4. Figure 4(a) shows that coagulation activity changed by increasing walnut seed extract dose from 1.0 to 5.0 ml/L. The optimum dose of walnut seed extract at various water turbidity was 3.0 ml/L. The coagulation activity was found to be 70.6, 90.2, and 94.9% at initial turbidity of 13, 54, and 194 NTU, respectively. In addition, as shown in Figure 4(b), the highest removal efficiency of turbidity was found to be 84.6, 95.2, and 97.8% at natural coagulant doses equal to 3.0 ml/L, at turbidity of 13, 54, and 194 NTU, respectively. The high turbidity removal was achieved by water-soluble cationic proteins in the walnut seed extract, which improve both coagulation and flocculation. Thus, coagulant dose, 3.0 ml/L, was selected as the optimum dose of walnut seed extract. Conversely, over the optimum dose, the excess of the flocculating agent would disrupt sedimentation and cause the resuspension of the accumulated particles (Bazrafshan et al. 2015). This was in agreement with results obtained by other researches (Ghebremichael et al. 2006; Bodlund et al. 2014).

Solution pH is a vital factor regarding charges on protein molecules and coagulation process. Figure 5 shows the effect of initial pH of solution on the turbidity removal efficiency at various solution turbidities. The highest removal was achieved at pH values from 7 to 11. The removal efficiency was 80.3, 89.6, and 99.5% at turbidities of 8, 56, and 190 NTU, respectively at pH equal to 11. It was clear that as the pH decreased the removal efficiency decreased. High turbidity removal efficiency was achieved at natural and alkaline pH ranges. This may be due to precipitation and the adsorption onto hydroxide flocs. The results of walnut seed extract were in agreement with other natural coagulants which are most effective at basic water (Okuda et al. 2001; Sanghi et al. 2002).

The time of slow mixing is a main contributor in achieving optimal coagulation performance. Therefore, the effect of slow mix time on the turbidity removal efficiency was also investigated. Figure 6 shows an increase in removal efficiency when slow mixing time was raised from 10 to 20 min. After 20 min of slow mixing, the removal efficiency reached 91.4 and 94.3% at 30 and 60 min of settling time, respectively. Increasing the time of slow mixing increases the opportunity for the extract to coagulate and form flocs with a good size, which leads to faster sedimentation (Ernest et al. 2017). Moreover, a slight increase in the removal efficiency was observed at slow mixing from 20 to 30 min. While insignificant removal was observed beyond 30 min of slow mixing. It was observed from Figure 6 that the settling time is an effective factor only for slow mixing duration less than 20 min.

Settling time has also a considerable and quantitative effect on coagulation efficiency. Suspended and other materials present in water differ in their size, weight, and charge after coagulation. Therefore, the effect of settling time after coagulation by walnut seed extract was investigated. Figure 7 shows that the removal efficiency after 30 min of settling time was 81.98, 94.73, and 97.69% at initial turbidity of 10, 100, and 200 NTU, respectively. While the removal efficiency after 60 min of settling time was 88.65, 96.58, and 98.43% at initial turbidity of 10, 100, and 200 NTU, respectively. High removal achieved within 30 min for high turbidity and 60 min for low turbidity as shown in Figure 7. This means that the amount of active protein present in the extract dose was active and have high reaction rates within 30 to 60 min of sedimentation according to initial turbidity then started to decrease with time gradually. Also, the amount of protein with weak activity needed longer time in order to make connections between suspended particles (Shak & Wu 2014).

Natural coagulant could be used as a coagulant aid with alum or used to reduce the alum dose. Thus, the effect of combination of natural walnut extract with alum was investigated as shown in Figure 8. Figure 8(a) shows the effluent turbidity at initial turbidity of 11 NTU and pH 8 by using various doses of natural coagulant with alum. It was found that the effluent turbidity decreases by increasing alum dose till 16 mg/l then the effluent turbidity started to increase. The effluent turbidity was 0.68 NTU using alum dose of 16 mg/L in addition to 3 ml/L of walnut extract for initial turbidity of 11 NTU (93.8% removal efficiency) after 30 min of sedimentation. While by using alum only, the effluent turbidity was 2.1 NTU (80.9% removal efficiency) using an alum dose of 16 mg/L. The effect of initial turbidity on the removal efficiency by using a combination of alum with natural walnut extract was shown in Figure 8(b). The turbidity removal efficiency increases from 93.8 to 99.5% by increasing the initial turbidity from 11 to 152 NTU, respectively. Therefore, alum dose of 16 mg/l is an optimum dose in addition to 3 ml/L of walnut extract. It was noticed that using natural walnut extract as a coagulant aid was able to improve the coagulation process performance. In addition, it can be used to decrease the dose of alum required for the coagulation process and achieve higher removal efficiency.

To study the effect of using walnut extract on the organic matter, the TOC value after using different doses of walnut extract (2–5 ml/L) was tested and the results are shown in Figure 9(a). The TOC value of the blank beaker was 30.76 mg/L after 60 min of settling where some of the organic suspended particles were settled by gravity without coagulation. However, the values of TOC were 40.21, 37.2, 36.09, 37.08 mg/L at walnut doses of 2, 3, 4, and 5 ml/L, respectively after 30 min sedimentation. While, after 60 min of settling, the results were 40.66, 37.11, 37.14, 35.81 mg/L, respectively as shown in Figure 9(a). This result shows that the measured TOC values with natural walnut extract slightly increased and were within the required limits (Herschy, 2012). Moreover, extract dose of 3 ml/L was the optimum dose to reduce the organic matter in coagulated water.

To study the effect of using alum on the organic matter compared to using walnut extract, the TOC value after using alum, natural extract, and combination of them was tested and the results are shown in Figure 9(b). The TOC value of the blank beaker was 32.25 mg/L after 60 min of settling. However, the TOC value was 33.07, 31.91 and 30.72 mg/L using walnut extract only, walnut and alum, and alum only, respectively after settling time of 30 min. While after 60 min of settling, the TOC values were 34.07, 33.41, and 31.19 mg/L, respectively as shown in Figure 9(b). The results showed that using alum only had no significant impact (slight decrease) on the TOC value. While using walnut extract only or combined with alum led to a slight increase in TOC value. It was concluded that using either walnut extract only or combined with alum results in the organic matter of the treated samples being within the accepted limits (Herschy 2012).

Chemical coagulants used in water and wastewater treatment are costly and cause side effects to human health and the environment. In this study, walnut seed extract was used as a natural coagulant. Water treatment using natural coagulant from walnut seeds is safe and non-hazardous sludge is produced. Moreover, the amount of sludge generated by seeds extract coagulation is lower compared with chemical coagulants. Furthermore, walnut seed extract was effective in turbidity removal alone or combined with alum. Compared with the turbidity removal efficiency by different natural coagulants, walnut seed extract exhibited the highest removal efficiency with small dose as shown in Table 2. The turbidity removal efficiency reached 91–98% using walnut extract with dose of 3 ml/L and 99.5% using walnut extract with dose of 3 ml/L combined with alum dose of 16 mg/L. So, walnut extract can therefore be effectively used in the removal of turbidity from aqueous solutions as natural, eco-friendly coagulant.

The artificial neural networks using activation function of Levenberg–Marquardt algorithm was used in removal efficiency model development. The parameters used in the ANNs model development were coagulant dose, initial turbidity, pH, and slow mix time. The model network structure is shown in Figure 10(a). Trail (4-30-10-10-1) (four inputs, three hidden layers with 30,10, and 10 neurons, and one output), yielded the highest coefficient of determination (R² of 0.92). The goodness of fit measures MSE, MAD, and MPB were 31.12, 3.08, and −0.55, respectively. It is noted that the output tracks the targets well for training (R-value = 0.99) and testing (R-value = 0.91). These values can be corresponding to a total response of R-value = 0.96. So, the network response is suitable, and simulation can be used for new inputs. Figure 10(b)–10(d) shows the regression plots for training, testing, and the total response for turbidity removal efficiency in the ANN model.

In this work, walnut seed extract was used for water turbidity removal as a natural ecofriendly coagulant. The results revealed that the highest removal efficiency was found to be 84.6, 95.2, and 97.8% at walnut seed extract doses equal to 3.0 ml/L, at turbidity of 13, 54, and 194 NTU, respectively. The highest turbidity removal efficiency was achieved at natural and alkaline pH. Walnut seed extract characterization analysis identifies the main properties of the material that contribute to achieving high performance in the coagulation process. Using walnut seed extract with alum as coagulant aid was able to enhance the coagulation process performance. Walnut seed extract can be used to decrease the alum dose required in the coagulation process, in addition to increasing the turbidity removal efficiency. The results showed that alum dose of 16 mg/l is an optimum dose added with 3 ml/L of walnut extract, which gives 99.5% removal efficiency. The TOC values of treated water by walnut extract doses of 2, 3, 4, and 5 ml/L were 40.66, 37.11, 37.14, 35.81 mg/L, respectively after 60 min of settling. While the TOC value of the settled water without coagulation was 30.76 mg/L where some of the organic suspended particles were settled by gravity without coagulation. It was concluded that insignificant increase of the organic matter content occurred using walnut seed extract and the TOC value within the accepted limits by using either walnut extract only or combined with alum. An ANNs model was developed for turbidity removal using walnut seed extract with high coefficient of determination (R² of 0.96). It was proved that walnut seed extract is an excellent environmental friendly coagulant compared to other natural coagulants.

The authors declare that they have no competing interests, and they are not affiliated with or involved with any organization or entity with any financial interest or non-financial interest in the subject matter or materials discussed in this paper.

Not applicable

Not applicable

TZ carried out the experimental tests and collected the required data. HM performed the analysis of data, MM helped in writing the manuscript. HM and MM revised the analysis of data. MF revised the manuscript writing. All authors read and approved the final manuscript.

All relevant data are included in the paper or its Supplementary Information.