Introduction

Due to the rapid growth of industrialization and the human population, the amount of sludge produced has also increased. Every year, the quantity of sludge produced increases, and the sludge handling and transportation costs also increase; these costs often make up 50% of the treatment costs when compared to that of the entire wastewater treatment process (Ealias et al. 2016; Liu et al. 2018). Hence, numerous approaches have been implemented to determine optimal environmental and economical solutions for sludge treatment.

Sewage sludge consists of organic matter in its colloidal form, and it is very difficult to dewater compared to industrially produced sludge. As the particles in the sludge are strongly bonded to each other, settling is prevented and offers resistance to compression and filtration. When the sludge particles are subjected to high compression, deformation of the particles occurs, forming a sludge cake with no voids (Wang et al. 2017). In a sludge cake, the voids are usually closed and lead to low filterability and impermeable layers in the filter channel (Shi et al. 2019). To overcome this, sludge dewatering has been adopted and utilizes physical/chemical or both conditioners.

Chemical conditioners promote particle agglomeration in the sludge, which forms particles of a larger diameter. The involved mechanism is the charge neutralization of anionic particles with the addition of cationic compounds (Duan and Gregory 2003). Ferric chloride, ferrous sulfate, ferric chloride sulfate, aluminum sulfate, poly-aluminum chloride, lime, and cationic polyacrylamide are some of the utilized chemical conditioners (Amokrane et al. 1997). The major limitation of using chemical conditioners is the low compressibility of the sludge, which may eventually lead to low filtration.

Physical conditioners are also implemented as filter aids or skeleton builders. Physical conditioners are normally used to improve the mechanical strength and permeability of sludge solids during compression to reduce the compressibility of sludge. These conditioners remain porous when subjected to relatively high pressures and form a stringent and permeable structure during mechanical dewatering (Zall and Galil 2015). A wide range of carbonaceous materials are usually utilized as skeletal materials, such as rice husks (Zhu et al. 2018), wheat dregs, wood chips, bagasse, sawdust (Guo et al. 2014), and other wastes are also utilized, such as slag, construction and demolition waste (Asakura et al. 2009), fly ash, cement kiln dust (Benitez et al. 1994), lignite (Shi et al. 2015), rice husk biochar (Wu et al. 2016), cement (Liu et al. 2016), and tannery sludge incineration ash (Luo et al. 2013).

The use of physical/chemical conditioners alone cannot enhance sludge dewatering capacity (Jing et al. 1999). Mazaheri et al. (2018) examined sludge conditioning using Moringa peregrina and ferric chloride individually and using a mixture of Moringa peregrina and a small quantity of ferric chloride. The application of both Moringa peregrina and ferric chloride exhibited a synergistic effect and increased the permeability of dewatered sludge. Several investigations on sludge dewatering using both physical and chemical conditioners have been carried out, which, in turn, enhanced sludge dewatering performance by achieving better settleability, the release of trapped water, accelerated sludge filterability, and decreased sludge compressibility (Smollen and Kafaar 1997; Jaafarzadeh et al. 2016; Liang et al. 2019). Hence, it has been inferred that adopting physical and chemical conditioners in combination produces better sludge dewaterability than when implemented individually. Nevertheless, convenient and effective implementation depends upon several factors, including pH, dose, contact time, and time to filter.

Optimization is a real-time control system that makes efficient use of various parameters. The Box-Behnken design (BBD) of response surface methodology (RSM) is widely used for the design of experiments assessing the effects of influencing factors within a system. This method allows for building a model in a time-efficient way and for achieving the optimum conditions for the desired responses.

With this background, the present study attempted to investigate the efficiency of granulated blast furnace slag coupled with quicklime for sludge dewatering by optimizing various system parameters using the Box-Behnken design. 1. An experimental investigation was carried out to assess the initial characteristics of sludge and GBFS. 2. The dose of skeletal materials, pH, and contact time were optimized to achieve maximum sludge dewatering by BBD. 3. Evaluation of sludge dewatering properties (capillary suction time, moisture content, turbidity, and protein, polysaccharide and heavy metal contents) was conducted. 4. Characterization of dewatered sludge was performed via scanning electron microscopy (SEM), X-ray diffractometry (XRD), and Fourier transform infrared spectroscopy (FTIR).

Materials and methodology

Materials

Sewage sludge was obtained from the treatment plant of the National Institute of Technology Karnataka (NITK), Karnataka, India. The sludge sample was collected in an airtight container, and it was stored in a refrigerator at 4 °C to minimize microbiological decomposition. Before proceeding with the experiments, the sample was placed in a water bath at 20 to 30 °C. Calcium oxide (Loba Chemicals) was used for pH control and conditioning. Granulated blast furnace slag (GBFS) is an industrial waste material and was used as a physical conditioner in the present study. The chemical composition of GBFS is given in Table 1.

Table 1 Chemical composition of granulated blast furnace slag
Full size table

The sludge characteristics, such as pH, moisture content, electrical conductivity (EC), temperature, total solids (TS), total dissolved solids (TDS), volatile solids (VS), turbidity, chemical oxygen demand (COD), and capillary suction time (CST), were determined experimentally. The obtained pH value of the sludge was 6.9 using a digital pH meter (Henna). The moisture content of the raw sludge was 96.4%. The TS, TDS, VS, and COD were determined as per IS: 3025 Bureau of Indian Standards (BIS) methods. The characteristics of the raw sludge were determined and are presented in Table 2.

Table 2 Raw sewage sludge characteristics
Full size table

Methodology

The skeletal material (GBFS) and CaO (quicklime) were used to enhance the dewaterability of secondary sludge. A set of 500 mL conical flasks was used to perform batch-mode experiments. To study the effect of pH on sludge dewatering, quicklime was added to the sludge sample to achieve a target pH (8.0, 9.0, 10, and 11) for maximum dewatering. The mixture was then subjected to 5 min of rapid mixing at 300 rpm and 5 min of slow mixing at 60 rpm to ensure CaO dispersion with 200 mL of secondary sludge. Later, the effect of the skeletal material on sludge dewatering was studied by varying the dose of slag (0.12, 0.25, 0.37, and 0.5 g/g DS) and mixing for 2 min at 300 rpm and 3 min at 60 rpm. After agitation, the resulting sludge was evaluated for CST and moisture content. Other significant properties of the raw and dewatered sludge, such as the zeta potential, turbidity, and heavy metal and biopolymer contents, were studied. The raw and dewatered sludge samples were dried and then analyzed by SEM in conjunction with EDS, XRD, and FTIR. The detailed methodology is represented in Fig. 1.

Fig. 1
figure 1

The schematic representation of the study

Full size image

Analytical methods

The moisture content and filter volume were determined using a vacuum filter. Fifty milliliters sludge was poured onto a vacuum filter, and a pressure of 50 kPa was applied. Then, the liquid was collected in the bottom chamber. The sludge remaining in the top chamber was collected, and the moisture content was determined as per a standard method (APHA 2012).

The CST analysis test was carried out in accordance with Lee and Liu (2000). The zeta potential was measured for the raw and conditioned sludge with varying doses of the skeletal material (Ma et al. 2017). The sludge was mixed with quicklime and skeletal material in a jar. The mixed sludge was then centrifuged, and the supernatant was used to measure the zeta potential using a HORIBA Scientific SZ-100 zeta potential analyzer. The turbidity was measured with a digital Nephelo-turbidity meter 132. Experiments were conducted in triplicate and reported as average values.

The polysaccharide and protein contents in the raw and treated sludge were determined according to the Anthrone and Lowry method (Felz et al. 2019). Heavy metal detection was performed with the sludge filtrate using an atomic absorption spectrophotometer (TIFAC, GBC 932 plus). The raw and dewatered sludge were collected and dried in a hot air oven at 105 °C for 24 h. The morphology of the raw and dewatered samples was evaluated using a JEOL scanning electron microscope along with energy dispersive X-ray analysis (EDAX). An XRD study was carried out to determine the crystal phase of the dewatered sludge. This was conducted at a 2θ range from 10° to 60° using Cu-Kἀ radiation. The chemical modifications in the sludge were verified by the presence/absence of functional groups as determined by FTIR analysis (Thermo Nicolet).

Results and discussion

Batch test results

The GBFS and CaO were added to sludge to examine the effect of their doses on sludge dewaterability. At the initial pH of the raw sludge (6.9), the dewatering efficiency was measured in terms of CST and was found to be low. Hence, the addition of CaO was performed. The maximum dewatering efficiency was confirmed at pH 10, as shown in Fig. 2 a. When the pH was further increased, the absorption of water was high because the efficiency of dewatering was reduced. For better dewaterability, the CST should be low.

Fig. 2
figure 2

a Effect of pH on sludge dewaterability, CST vs. pH. b Effect of pH and dosage on sludge dewaterability, CST vs. slag dosage

Full size image

It was observed that the dose of 0.37 g/g DS of slag was sufficient for obtaining satisfactory dewatering. At the abovementioned dose, the obtained CST value was 38 s. It can be noted that the addition of CaO and slag is able to achieve a CST of 38 s, as shown in Fig. 2 b. Thus, the addition of GBFS and CaO helped improve the dewatering capability.

Effect of conditioner on sludge dewatering properties

Moisture content and turbidity

The distribution of moisture content with varying slag dose is shown in Fig. 3 a. The moisture content decreased to 78% at pH 10 due to the reactivity of CaO. At pH 10, the addition of slag lowered the moisture content to 68% at a 0.37 g/g DS dose and resulted in a rigid lattice structure with more voids than in the raw sludge structure. The addition of slag resulted in a bridging effect of the sludge particles for the passage of water and reduced the compressibility of the sludge cake.

Fig. 3
figure 3

a Distribution of moisture content with a varying slag dosage. b Responses of turbidity with respect to varying slag dosage

Full size image

After the addition of slag to the sludge, more voids and pores were produced (Dinçer et al. 2007), which enhanced the reduction in moisture content. The addition of a large quantity of slag also produced a higher solids content in the sludge during filtration, the deformation of fine particles, and a decrease in the porosity of the sludge cake, resulting in a reduction of permeability. Hence, the sludge dewatering decreased with increasing fine particle deformation. In the present study, the moisture content did not decrease much with the increasing slag concentration beyond 0.37 g/g DS due to the presence of fine particles. Hence, it was inferred that a dose of 0.37 g/g DS was sufficient for a significant reduction of the moisture content. Our results are in agreement with those found for sludge dewatering using wood chips (Ding et al. 2014) and gypsum as skeletal materials (Zhao 2002).

The initial turbidity (24.6 NTU) was measured in the supernatant of the centrifuged sludge. Later, various doses of slag were added to the sludge, the centrifuged supernatant was measured, and the results are portrayed in Fig. 3 b. Reduction in the turbidity happened with respect to the increasing slag dose, and a turbidity of 8 NTU was obtained at the slag dose of 0.37 g/g DS. A further increase in the slag dose resulted in an increase in turbidity.

Zeta potential

The zeta potential is a key factor in understanding the surface properties of sludge flocs with respect to flocculation and dewatering. Typically, the zeta potential of sludge obtained from wastewater treatment plants is negative (To et al. 2018). The observed decrease in zeta potential through charge neutralization using quicklime (Ca2+) and slag was efficient (Fig. 4). The zeta values changed from − 7.8 mV for the raw sludge to − 1.7 mV for the dewatered sludge at a dose of 1.5 g. The addition of quicklime increased the pH value of sludge. pH and electrical conductivity are related to the structure of the sludge floc. The formation of an alkaline environment in the sludge could increase the dissolved extracellular polymeric substances (EPSs) and destroy the matrix structure of the EPSs. Thus, the obtained sludge dewatering performance was satisfactory and occurred by releasing bound water and small particles from the sludge floc.

Fig. 4
figure 4

Responses of zeta potential and electrical conductivity with respect to varying conditioner

Full size image

Heavy metal content

The concentrations of heavy metals in the raw sludge filtrate and treated sludge filtrate are shown in Fig. 5 a. The slag mainly consisted of calcium oxide and aluminum oxide. These oxides adsorb heavy metals, thereby reducing the concentration of heavy metals in the filtrate (Nguyen et al. 2018).The reduction of Pb, Cd, Cr, and Zn was 76, 42, 41, and 70%, respectively. The soluble fraction of heavy metals observed in the raw and treated sludge cakes by EDAX is shown in Fig. 5 b. A major reduction in the concentration of elements, such as lead (Pb), cadmium (Cd), chromium (Cr), and zinc (Zn), was observed in the solid phase after conditioning with GBFS. The findings showed that there was a significant reduction in the heavy metal content compared to that in the raw filtrate. The heavy metal concentration in the filtrate after reduction with the addition of GBFS was within the permissible limit for effluent discharge into inland surface water (CPCB 2000).

Fig. 5
figure 5

a Soluble fraction of heavy metals in the raw sludge cake and treated sludge cake. b Concentration of heavy metals in the raw sludge filtrate and treated sludge filtrate

Full size image

Protein and polysaccharide determination

The major factors considered in the sludge dewatering process are EPSs, proteins, and polysaccharides (Xiao et al. 2019). The presence of EPSs was measured for both the raw and treated samples (Fig. 6). The results show that the presence of polysaccharides was greater than the protein component in the sludge. The initial concentrations of protein and polysaccharides in the raw sludge were 108 and 1249 g/L, respectively, which significantly increased to twice the concentrations in treated sludge (826 and 2316 g/L). Hence, the results indicate that the destruction of EPSs occurred and that biopolymers were released into the dewatered liquor, whereas protein and polysaccharides remained in the sludge filtrate. This may be due to the rough surface of slag resulting in the stripping of sludge EPSs and the breakage of cells.

Fig. 6
figure 6

The concentration of protein and polysaccharide in raw sludge and treated sludge

Full size image

Characterization study

Surface morphology

The SEM images of the raw and dewatered sludge are depicted in Fig. 7 a and b. The morphological analysis of the raw sludge indicated its non-porosity and that it has a dense structure. However, in the case of the treated sludge, the morphology had a higher porosity than that of the raw sludge, and the destruction of the particles and the surface tended to be convoluted. This may be due to reactions that occurred when CaO was added. The slag helped increase the number of water pore channels, which leads to an increase in the dewaterability of the conditioned sludge. The sludge cake obtained after dewatering had unconnected surfaces with voids; therefore, during mechanical dewatering, the formation of impermeable thin layers in the upper layer of the filter medium can be prevented (Ning et al. 2013).

Fig. 7
figure 7

a Surface morphology of raw sludge. b Surface morphology of treated sludge. c XRD of treated sludge

Full size image

XRD was used to identify the mineralogical compounds and crystalline phase (Fig. 7c) present in the samples. The major peaks obtained were identified as quartz, calcium hydroxide, and calcite.

Surface analysis

The information regarding the functional groups present in the raw and treated sludge was obtained from FTIR spectra in the range of 500 to 4000 cm−1. The FTIR spectra of the raw and treated sludge are shown in Fig. 8a and b, respectively. FTIR analysis allows for the monitoring of atoms experiencing stretching vibrations in which the distance between the atoms increases or decreases. The broad peak obtained in the range of 3800 to 3600 cm−1 was attributed to the O-H and N-H vibrational stretching overlap. The peaks observed in the region below 700 cm−1 were present due to halogenated stretching (Betatache et al. 2014). The band located at 1647.26 cm−1 was related to the stretching and deformational vibration of the peptic bonds of proteins, such as C=O, C-N, and N-H. The peak at 1037.30 cm−1 and a shoulder at 1067 cm−1 were attributed to CO32+ splitting due to the presence of hydrated monocalcium aluminate. The bands observed in the range between 780 and 560 cm−1 indicated the generation of CO2 and H2O due to the reaction of organic acids and saccharides.

Fig. 8
figure 8

a FTIR spectra of raw sludge. b FTIR spectra of treated sludge

Full size image

FTIR analysis of the treated sludge showed a broad peak at 3638.17 cm−1, attributed to the O-H stretching and N-H stretching in alcohols, acids, and alkyl structures. The S=O stretching vibration was assigned to the peak at 866.13 cm−1. The peaks ranging from 1300 to 850 cm−1 were attributed to the O-H stretching due to the decomposition of minerals present in the treated sludge. When the raw sludge and treated sludge spectra were compared, a peak at 1401.28 cm−1 was observed, confirming the presence of C=C double bonds and therefore aromatic rings in the treated sludge. Overall, there was a decrease in the number of peaks in the treated sludge spectra when compared to that in the raw spectra.

Optimization

Response surface methodology is used to determine optimal performance parameters of a system by optimizing variables at different levels. The Box-Behnken design was used to approximate a response function using Design Expert v11.1.2.0.

For the current study, the effects of slag dose, pH, and contact time were investigated. Every independent variable was serially coded at three levels: low (− 1), medium (0), and high (1). Table 3 shows the independent variables that were selected, which all interacted with the sludge and had a significant impact on the system. The Box-Behnken design is more efficient when utilized as a spherical design and when factors are run at three levels. Experimental studies were performed in the laboratory to confirm the model-obtained results; the experiments were conducted at least three times and are reported as the average values of filtrate. A quadratic model was fitted to the experimental data, as the F value was higher and the p value was lower than those of the 2FI, cubic, and linear models from the sequential model sum of squares. Normally, “The lack of fit” is an inadmissible factor for the design; hence, it has to be insignificant, i.e., when the given data properly fit the filtrate volume.

Table 3 Experimental responses of the liquid separation
Full size table

The dewatering efficiency was illustrated by a second-order polynomial equation in coded form using the shown responses and by implementing multiple regression analysis in the design matrix:

Filtrate (mL) = 80.48 + 3.54A − 0.7375B + 0.5250C − 0.8500AB − 0.6750 AC − 0.000 BC − 4.32A2 − 4.02B2 − 1.19C2, where the dewatering efficiency is expressed in terms of filtrate and A, B, and C are the coded terms for the three independent test variables dose, pH, and contact time, respectively. From the above equation, the optimum values of the selected variables were obtained. ANOVA was carried out to determine the significance of the second-order polynomial equation (Table 4).

Table 4 Statistical and ANOVA analysis of the model
Full size table

The determination coefficient (R2) is one way to determine the model’s goodness of fit. The obtained determination coefficient (R2 = 0.9945) expresses that more than 95% of the experimental responses were well fitted to our model. To analyze the model fitness and adequacy, prediction of the adjusted R2 was necessary. The model showed high significance, as the value of the adjusted determination coefficient was R2adj = 0.9875. In our model, the R2adj value was very close to the predicted R2 value, and the difference between the adjusted R2 (0.9875) and predicted R2 (0.9561) was less than 0.2. Hence, our experimental results and the predicted values show a significant correlation coefficient (R). The coefficient of variation was low (0.615), which implies that the experiments were carried with a high degree of precision.

The “Adeq precision” value was determined by the signal-to-noise ratio. In our model, the Adeq precision value was 33.40, indicating that the model can be used effectively (Wang and Guo 2019; Behera et al. 2018). The variation inflation factor (VIF) describes that the model variance in the design is magnified by the absence of originality. For all the independent and dependent factors in the current work, the VIF was found to be 1, which denoted that all the factors in the model and the desired factor were orthogonal. The results indicate that the model utilized for the liquid separation process was capable of identifying the optimal operational conditions for the separation of liquid from the sludge.

Figure 9 shows the 3D response surface contour plots, illustrating the effects of dose, pH, and contact time on the liquid separation from the sludge. In all 3D surface plots, the effect of two variables was shown while keeping the other variables constant. The 3D response surfaces were generated with respect to the equation formed in the model. The interaction effects between pH and dose and pH and contact time showed positive effects. The interaction effect of every independent variable, such as that of pH and dose, is shown in the response surface plots, which exhibits a strong positive quadratic effect on the liquid separation of the sludge. The curvature formed in the 3D response surface for all variables indicates that the removal of water from the sludge was high.

Fig. 9
figure 9

3D surface graphs and contour plots of liquid separation with two variables

Full size image

Derringer’s desirability function was used for every independent variable in the optimization process. The desirability function ranged between 0 and 1, where the former specifies an undesirable response and the latter specifies a fully desirable response (Awotwe-Otoo et al. 2012; Ferreira et al. 2007). For individual desirability, a weight factor of 1 was preferred in the current work. Figure 10 shows the ramp desirability for liquid separation. The optimized process variables were achieved by implementing the desirability function. This indicated that liquid separation at pH = 10.2, dose = 0.34 g/g DS, and contact time = 14 min would yield 81.1 mL of collected filtrate from sludge with an overall desirability value of 1.

Fig. 10
figure 10

Ramp desirability for liquid separation

Full size image

The addition of quicklime was carried out carefully; otherwise, it would have caused a large increase in pH even with a minor variation in dose. If the quantity of quicklime was low or high, it led to a reduction in sludge dewatering efficiency. The relatively high-water absorption capacity at a high pH is one of the properties of quicklime, which, in turn, affects sludge dewatering. Hence, the amount of quicklime was limited to that resulting in a pH of 10, so that the effect of the skeletal material could be established. Quicklime addition alone would not contribute to any void formation; hence, the addition of GBFS was introduced to reduce the compressibility of the sludge when a high pressure was applied. Using quicklime, the pH of the sludge was increased to more than 12. GBFS influences sludge dewatering by reducing the maximum moisture content. The presence of oxides of calcium, magnesium, and aluminum is high in GBFS, and these oxides facilitate the action of GBFS as a conditioner in sludge dewatering. A high solid content was produced during conditioning with the skeletal material, which was beneficial for developing a more stringent lattice structure. Surface morphology analysis indicated that the slag, as a skeletal material, exhibited a rigid structure with more voids than those in the raw sludge when pressure was applied to the sludge. Further reduction in the compressibility of the sludge led to the development of channels for the removal of excess water from the sludge. Hence, the adopted combination of the conditioners may also serve as skeletal additives in landfill cover materials and construction materials and for mining site reclamation (Li et al. 2014). Furthermore, GBFS is not a flocculent, as it does not have the ability to flocculate sludge particles. Hence, it is worth noting that even GBFS alone does not yield better results of dewatering. To achieve flocculation, CaO was used as a flocculent, thereby establishing GBFS as a promising skeletal material in terms of economy and scalability. The slag acted as a skeletal material by achieving good dewatering efficiency with the reduction of environmental risk and treatment cost.

Conclusions

In this study, GBFS was used for sludge dewatering and was found to be a significant skeletal material for dewatering secondary sludge. The Box-Behnken design was used to successfully determine the dewatering efficiency with various influencing parameters. The optimum values obtained were a pH of 10.2, a dose of 0.34 g/g DS, and contact time of 14 min, resulting in a liquid separation efficiency from sludge that yielded 81.1 mL filtrate. The dewatering ability of the sludge was successfully evaluated by studying various properties of the dewatering process. Surface morphology analysis indicated that the slag, as a skeleton material, exhibited a rigid structure that could act as a channel for the removal of excess water from the sludge. Furthermore, the treated sludge can be used in various industrial applications, such as in construction, agriculture, etc., as the heavy metal contents in the filtrate were reduced to below the desired levels. The present study demonstrates that slag can act as a skeletal material for effective sludge dewatering in waste recycling to establish sustainable waste management practices.