Advertisement

SN Applied Sciences

, 1:250 | Cite as

Adsorption of p-nitrophenol onto acacia glauca saw dust and waste orange peels activated carbon: application of Taguchi’s design of experiment

  • Prashant T. DhorabeEmail author
  • Dilip H. Lataye
  • Ashwini R. Tenpe
  • Ramakant S. Ingole
Research Article
  • 132 Downloads
Part of the following topical collections:
  1. Earth and Environmental Sciences: Waste Reduction, Recycling and Utilization for Value Added Products

Abstract

The study shows the application of Taguchi’s design of experiment method to optimize the parameters for adsorption of p-Nitrophenol (p-NP) on the surface of acacia glauca saw dust activated carbon (AGAC) and waste orange peels activated carbon (ACOP) from aqueous medium. The effect of adsorption parameters viz. adsorbent dose (m), temperature (T), initial concentration (C0) and contact time (t) on response characteristics i.e. adsorption capacity (qt) has been studied at three levels by using L9 orthogonal array (OA) which further analysed by variance analysis (ANOVA) for adsorption data and signal/noise ratio data by using ‘larger the better’ characteristics. The ANOVA shows factor A (m) and B (C0) are the most significant parameters with the contribution of 51.27 and 11.40%, and 40.74 and 80.90% for AGAC and ACOP respectively. Factor C (T) and D (t) are found to be least significant parameters. The percent removal of p-NP at optimum condition was found to be 96.11 and 92.40% with 72.08 and 34.56 mg/g adsorption capacity at m = 2 and 4 g/L, C0 = 150 and 150 mg/L T = 313 and 303 K and t = 90 and 90 min for AGAC and ACOP respectively.

Keywords

Adsorption ANOVA p-Nitrophenol Orthogonal array Taguchi 

List of symbols

µ

Mean of response

ANOVA

Analysis of variance

C0

Initial concentration (mg/L)

Ce

Equilibrium concentration (mg/L)

DOF

Degree of freedom

E

Number of required experiments

fe

DOF error

k

Total numbers of factors and interactions

m

Adsorbent dose (g/L)

N

Total number of results

qt

Adsorption capacity (mg/g)

R

Sample size for confirmation experiment

S/N

Signal to noise ratio (db)

t

Contact time (min)

T

Temperature (K)

Ve

Error variance

1 Introduction

There is a lot of water scarcity and water pollution problems in India and worldwide. Development of civilization and urbanization results in pollution and increasing the level of toxicity in the environment. Among them phenol and its derivatives are also very toxic and hazardous substances and required special attention for removal and treatment from water and wastewater. The p-Nitrophenol (p-NP) is phenol derivative where nitro group and hydroxyl groups attached in benzene ring. p-NP is highly toxic in nature and found in the wastewaters of many agricultural, petrochemicals, pharmaceuticals, pesticides, coke oven, petroleum refineries, steel foundry, herbicides and insecticides industries etc. [1, 2]. It also found in herbicides like nitrofen and bifenox, insecticides ethyl and methyl parathion and fluorodifen. It is also an intermediate synthesized compound formed during the formation of paracetamol drug. The acute and prolong effects of p-NP on human health were observed like irritation of eyes and skin, blood disorder, respiratory disorder and malfunction, nervous disorder etc. It also reduces the oxygen carrying capacity of blood and cause methemoglobinemia blood disease. There are also other health effects like blue lips or finger nails, blue skin, convulsions, confusion, cough, dizziness, nausea, headache, sore throat, unconsciousness, renal malfunctioning etc. USEPA listed p-NP as toxic pollutants. Its water solubility and toxicity destroys aquatic life and the living organisms which depend on water and consumed water for survival. Hence it is priority to remove it from water or wastewater [3, 4, 5].

The chemical oxidation, biodegradation are economical methods but due to lower treatment efficiency and higher time of treatment these are not feasible techniques. The advanced techniques like membrane filtration, advanced oxidation, electrochemical oxidation and photocatalytic removal are more efficient but are not cost effective. Among the available technologies, adsorption is feasible in terms of efficiency, cost and time point of view. Some adsorbent materials like charred saw dust [6], fly ash [7], antimony, zirconium ferrocynide and cadmium [8] activated carbon fibers [9, 10] acid activated jute stick char [11], babool saw dust [12], activated carbon of acacia glauca saw dust [5] and activated carbon prepared from waste orange peels [13], Fenton oxidation process [14], mild steel manufacture [15] and also its collaboration with genetic algorithm as hybrid optimization technique [16] etc. have been reported as effective adsorbents for the removal of p-NP from aqueous medium.

The present study focused on the optimization of adsorption parameters for p-NP sorption onto acacia glauca saw dust activated carbon (AGAC) and waste orange peels activated carbon (ACOP) by Taguchi’s optimization method. The adsorption parameters have been analysed by ANOVA for raw data and S/N data. Taguchi’s optimization is world widely used by many researchers and investigators in the field of manufacturing and industrial production process optimization for best quality products [17, 18, 19, 20, 21, 22]. Wu and Chen [20] used Taguchi’s model for optimization of saw mass sensing device manufacturing. There are other applications of Taguchi’s robust design like polymer optimization [21] optimization of polymer chain reaction in DNA [22], biotech process [23, 24], even in the field of animals and birds social behaviour called swarm optimization [25]. The electrodegradation of 2,4-dichlorophenoxyacetic acid herbicides has also optimised by Taguchi method by Dargahi et al. [26]. But limited contribution of researchers [27, 28, 29, 30] for the use of this method in the field of adsorption for optimization of sorption process parameters impart large scope for its utilization and implementation for resources and energy conservation.

The main objective of Taguchi’s method is to optimize the effect of the most significant sorption process parameters finalised on the basis of batch adsorption study with minimum variance. In this method, a robust orthogonal array (OA) is used to optimize the raw data obtained from a limited number of experiments. The S/N ratio is also optimized to minimise the effect of uncontrolled factors. The percent contribution of each parameter in the adsorption has been determined by ANOVA.

2 Materials and methods

2.1 Adsorbents

The adsorbents were prepared from acacia glauca saw dust and waste orange peels were collected from local saw mill and orange market of Nagpur (India). The raw materials were cleaned thrice with water for removal of dust, mud and other impurities. The washed material was sun dried for 2 days and then oven dried at 105 ± 5 °C for 5–6 h. The waste orange peels were crushed first and sieved to obtain size in the range of 300–600 μ while the acacia glauca saw dust was directly sieved for required particle size i.e. 300 to 600 μ. To chemically activate the materials for enhancement of its surface and pore properties, it was mixed with 1:0.5 (g: ml) orthophosphoric acid evenly and kept for 24 h at room temperature under mixed condition then after that it was filled in closed container and kept in muffle furnace (450 ± 10 °C for acacia glauca saw dust and 350 ± 10 °C for orange peels) for 1 h where they were converted into char. The developed char of both materials were washed separately with 2 molar liquid ammonia solution and sequentially washed with distilled water to get neutral pH and then were kept in oven at 120 °C for 6 h. The different physical and chemical properties were determined over this prepared activated carbon of acacia glauca saw dust (AGAC) and waste orange peels (ACOP) by standard procedures and then were further used in experiments.

2.2 Adsorbate and other chemicals

The analytical reagent (AR) grade chemicals have been used in the study. p-Nitrophenol (p-NP) was procured from Loba Chemie Pvt. Ltd. Mumbai (India). An accurate weight of 1 g p-Nitrophenol was mixed with doubled-distilled water (DDW) for the preparation 1000 mg L−1 concentration of p-NP as a stock solution. The required concentration solutions were prepared by diluting stock solution with DDW. The other chemicals viz. NaOH, HCl, NaCl, KNO3 etc. were procured from Merck Specialities Pvt. Ltd. (Mumbai, India).

2.3 Taguchi’s design methodology

The followings steps have been implemented for Taguchi experimental methodology: (1) selection of output variable, (2) identification of input parameters and their levels, (3) selection of orthogonal array (OA), (4) assignment of factors to the array columns i.e. Design of Experiment (DOE), (5) performing experiments, (6) statistical analysis by ANOVA and the signal to noise ratio (S/N) and determination of the optimal condition and (7) to perform confirmation experiments, if required [31]. This process is shown in the Fig. 1.
Fig. 1

Taguchi’s methodology Flow sheet

The qt i.e. adsorption capacity (mg g−1) was the output parameter (response characteristic) to be optimized by Taguchi’s method. The four most influencing input variables (factors/parameters) majorly responsible for adsorption on the basis of our previous work [5, 13] viz. adsorbent dose (m), temperature (T), initial concentration of p-NP (C0), and time of contact (t) at three different levels were selected for optimization (shown in Table 1). The other parameters like initial pH and agitating speed which play a major role in sorption process were not considered because in our previous work it was observed that for p-NP adsorption on to AGAC and ACOP not much affected by these parameters [5, 13]. Almost the adsorption capacity was same from varying the initial pH from 2 to 10.
Table 1

Selection of factors and levels for sorption of 4-NP over AGAC and ACOP

Factors

Parameters

Units

Levels

AGAC

ACOP

1

2

3

1

2

3

A

Dose (m)

g/L

2

3

6

4

5

6

B

4NP Initial Concentration (C0)

mg/L

50

100

150

50

100

150

C

Temperature (T)

°C

20

30

40

10

20

30

D

Shaking Time (t)

min

30

60

90

30

60

90

On the basis of partial factorial method of Taguchi’s optimization, the number of experiments, E can be determined by E = 2 k + 1, where k is total factors and interaction if any (i.e. for 4 factors, E = 9) hence, L9 orthogonal array (OA) for three level has selected (shown in Table 2) [22]. The properties of an OA are such that within the columns, each combination of levels (1, 2 or 3) occurs at an equal number of times. The OA reduces the 34 = 81 sets of experiments into 9 runs only.
Table 2

Taguchi’s L9 (34) orthogonal array for sorption of 4-NP over AGAC and ACOP

Run

Factors

A

B

C

D

m

C 0

T

t

1

1

1

1

1

2

1

2

2

2

3

1

3

3

3

4

2

1

2

3

5

2

2

3

1

6

2

3

1

2

7

3

1

3

2

8

3

2

1

3

9

3

3

2

1

As per the selected factors in Table 1, Design of Experiment (DOE) was prepared by putting the factors level value into Table 2 (shown in Table 3).
Table 3

Design of experiment (DOE) for sorption of 4-NP over AGAC and ACOP

Run

Factors

AGAC

ACOP

A

B

C

D

A

B

C

D

m

C 0

T

t

m

C 0

T

t

1

2

50

20

30

4

50

10

30

2

2

100

30

60

4

100

20

60

3

2

150

40

90

4

150

30

90

4

3

50

30

90

5

50

20

90

5

3

100

40

30

5

100

30

30

6

3

150

20

60

5

150

10

60

7

6

50

40

60

6

50

30

60

8

6

100

20

90

6

100

10

90

9

6

150

30

10

6

150

20

10

2.4 Experimental programme

As per DOE (Table 3a, b), each run of experiment among 9 runs were repeated three times to eliminate the experimental and analysis errors. All the experiments were performed in 250 mL glass conical flasks with fixed (50 mL) volume of p-NP adsorbate solution with varying ‘C0’ and ‘m’ as per DOE run in triplicate and kept in an orbital shaking incubator (Remi Instruments, Mumbai) at set temperature ‘T’ and constant shaking speed of 150 RPM for required time ‘t’. The residual concentration of p-NP was determined by taking absorbance on SHIMADZU double beam UV–Vis Spectrophotometer (model UV-2450) at 318 nm wavelength. Adsorption capacity qt was calculated by formula:
$$q_{t} = \frac{{(C_{0} - C_{e} )}}{m}$$
(1)
where, qt = amount of p-NP adsorbed by the adsorbent at equilibrium (mg g−1), Co = initial concentration (mg L−1) and Ce = equilibrium concentration (mg L−1) of p-NP solution and m = mass of the adsorbent (g L−1).

2.5 Experimental data analysis

The data obtained from experimental programme was analyzed by ANOVA for raw data, and S/N. S/N ratio reduced the variability of uncontrollable factors (or factors of noise) [22]. The S/N ratio is defined as signal to noise ratio which shows the quality of functions if the ratio is more, then the function quality will be better. It is also defined as the good to error ratio. There are three conditional types of S/N ratio are as follows:

Smaller the better
$$\frac{S}{N} = - { 10 }\log \left[ {\frac{1}{n}\sum\limits_{i = 1}^{n} {y_{i}^{2} } } \right]$$
(2)
Larger the better
$$\frac{S}{N} = - { 10 }\log \left[ {\frac{1}{n}\sum\limits_{i = 1}^{n} {\frac{1}{{y_{{_{i} }}^{2} }}} } \right]$$
(3)
and Nominal the better
$$\frac{S}{N} = 10 \, \log \frac{{\mu^{2} }}{{\sigma^{2} }}$$
(4)
where µ2 = square of mean, σ2 = variance of observations and yi = observation of response variables for ‘n’ trials.
Equation (3) i.e. larger the better S/N condition was used to calculate the variance of data. The S/N data were further optimized by ANOVA. Each column of OA is associated with two DOF (number of levels minus one) and can be assigned one factor. The array should satisfy the condition, total degree of freedom for OA should always greater or equal to the total DOF of experiment. The raw data were used for the results analysis and to calculate the adsorption performance. After determination of optimum condition the mean of response (µ) is predicted. ANOVA gives the optimized factors at significant level. The detailed methodology for the analysis of experimental data is given elsewhere [27, 29]. Suppose if A1, B2, C3 and D3 are the factors optimized at level second then the mean of the response (µ) at the optimal condition was estimated as:
$$\mu = \bar{T} + (\bar{A}_{1} - \bar{T}) + (\bar{B}_{2} - \bar{T}) + (\overline{C}_{3} - \bar{T}) + (\overline{D}_{3} - \bar{T}) = \bar{A}_{1} + \bar{B}_{2} + \overline{{C_{3} }} + \overline{{D_{3} }} - \bar{T}$$
(5)
where \(\bar{T}\) is the overall response mean, and \(\bar{A}_{1}\), \(\bar{B}_{2}\), \(\overline{{C_{3} }}\) and \(\overline{{D_{3} }}\) represent average response values at the first level of factor A, second level of factor B, third level of factors C and D respectively.
The determination of µ is based on the average results obtained from the experiments. The optimized result was checked at confidence interval (CI) which represents the value of statistical parameter lies between true mean, \(\mu\) should fall at some stated confidence level. The CI characterized in two types, CIPOP which is confidence interval for entire population and CICE which is for only a sample group or group of experiments at specified conditions. These confidence intervals are defined as:
$$CI_{POP} = \sqrt {\frac{{F_{\alpha } (1,f_{e} )V_{e} }}{{n_{eff} }}}$$
(6)
$$CI_{CE} = \sqrt {F_{\alpha } (1,f_{e} )V_{e} \left[ {\frac{1}{{n_{eff} }} + \frac{1}{R}} \right]}$$
(7)
where \({\text{F}}_{\upalpha} (1,{\text{f}}_{\text{e}} )\) = the F-ratio at a confidence level of (\(1 - \alpha\)) against DOF 1 and error of DOF (fe) its value can find in F table, \(V_{e}\) = error variance (from pooled ANOVA)
$$n_{eff} = \frac{N}{1 + [Total\,DOF\,associated\,in\,the\,estimate\,of\,the\,mean]}$$
(8)
where N = total number of results and R = sample size for confirmation experimentIt can be seen from Eq. (7), that as R approaches infinity, i.e. the entire population, the value 1/R approaches zero and \({\text{CI}}_{{\text{CE}}} = {\text{CI}}_{{\text{POP}}}\). As R approaches 1, \({\text{CI}}_{{\text{CE}}}\) becomes wider [28, 29]. The confirmation experiment will be carried out at optimized condition for parameters and the average of result checked at 95% CI. If the Taguchi’s model value and confirmation experiments value lies in 95% CI confirm the validity of optimal parametric values determined by Taguchi’s design.

3 Results and discussion

3.1 Characterisation of adsorbents

The BET surface area for AGAC and ACOP were determined by adsorption of N2 gas on their surface at controlled condition. For AGAC and ACOP adsorbents, BET surface area was found to be 788.93 m2 g−1 and 540.61 m2 g−1 before adsorption which were decreased after adsorption of p-NP and was found to be 223.83 m2 g−1 and 84.81 m2 g−1 respectively. There was decreased in 565.10 m2 g−1 (i.e. 71.63%) and 455.80 m2 g−1 (i.e. 84.31%) surface area of AGAC and ACOP after adsorption. The pore surface area of AGAC and ACOP i.e. iodine number were found 755.10 mg g−1 and 420.14 mg g−1. As per BIS 1350-I (1984) [32] the proximate analysis was done and the results are given in Table 4. The ash content i.e. non-carbonate contents and fixed carbon contents of AGAC and ACOP were found to be 20.23 and 23.60% and 55.38 and 48.20% respectively. The presence of fixed carbon indicates lignocellulose in organic materials which shows that raw form of AGAC and ACOP are rich of lignocellulose. The inversely affecting parameters which reduce the adsorption capacity like moisture and volatile content were observed to be 10.07 and 9.67% and 14.32 and 18.87% for AGAC and ACOP. Figure 2a, b are the SEMs of AGAC and ACOP. The rough surface texture and irregular size micropores were observed before adsorption of p-NP which gets clogged after the adsorption of p-NP. The point of zero charge for the AGAC and ACOP were found to be 6.15 and 6.50 (Fig. 3) by solid addition method [5, 13, 33].
Table 4

Proximate analysis of AGAC and ACOP

Sr. no.

Characteristics

AGAC (%)

ACOP (%)

1

Moisture content

10.07

9.67

2

Ash content

20.23

23.26

3

Volatile matter

14.32

18.87

4

Fixed carbon

55.38

48.20

Fig. 2

SEMs picture of A AGAC and B ACOP

Fig. 3

Point of Zero Charge of AGAC and ACOP

3.2 L9 OA experimental results

Experiments were performed for p-NP adsorption onto AGAC and ACOP at specified conditions as mentioned in Table 3. The experiments were repeated three times and average of qt was used for Taguchi’s design. The experimental output for runs and S/N calculated for larger the better condition shows in Table 5.
Table 5

Taguchi L9 OA with experimental qt values and S/N ratio data for sorption of 4-NP by ACOP

Adsorbent

Run

Factors

Experimental results

S/N ratio

A

B

C

D

qt (mg/g)

m

C 0

T

t

R1

R2

R3

AGAC

1

2

50

20

30

37.19

37.18

37.18

31.41

2

2

100

30

60

24.67

24.66

24.66

27.84

3

2

150

40

90

8.25

8.25

8.24

18.33

4

3

50

30

90

48.95

48.99

48.99

33.80

5

3

100

40

30

16.58

16.58

16.57

24.39

6

3

150

20

60

24.79

24.80

24.80

27.89

7

6

50

40

60

72.09

72.07

72.09

37.16

8

6

100

20

90

12.41

12.40

12.41

21.87

9

6

150

30

10

24.56

24.50

24.56

27.80

ACOP

1

4

50

10

30

10.61

10.62

10.59

20.51

2

4

100

20

60

22.32

22.16

22.28

26.95

3

4

150

30

90

34.62

34.54

34.50

30.77

4

5

50

20

90

10.47

10.47

10.47

20.40

5

5

100

30

30

17.96

18.02

17.96

25.09

6

5

150

10

60

26.36

26.33

26.42

28.42

7

6

50

30

60

10.54

10.54

10.53

20.45

8

6

100

10

90

16.28

16.29

16.28

24.24

9

6

150

20

10

21.81

21.81

21.78

26.77

3.3 Effect of adsorption parameters

Factors m, T, C0, and t at various levels significantly affect the response values qt (shown in Table 5). The qt considered as final response characteristic for optimization and the factor A (m) shows highest influence at level 1, factor B (C0) shows highest influence at level 3, factor C (T) shows highest influence at level 3 and factor D (t) shows highest influence at level 3 (shows in Fig. 3a, b and Table 6) for both adsorbents AGAC and ACOP. The difference between the level 1 and level 2 (i.e. L2–L1) and level 2 and level 3 (i.e. L3–L2) shows the influence of one level over another. Higher the difference indicate greater will be the impact of level. Table 6 indicates that factor B (C0) shows the highest impact on qt values. qt values increases with increase in C0 because of more mass driving force and less resistance to the sorption uptake of p-NP over AGAC and ACOP surface [28].
Table 6

Average and main effects of qt values for AGAC and ACOP: Raw and S/N data

Adsorbent

Factor

Raw data, average value

Main effects (raw data)

S/N data, average value

Main effects (S/N data)

L1

L2

L3

L2–L1

L3–L2

L1

L2

L3

L2–L1

L3–L2

AGAC

A

48.53

24.75

16.54

− 23.78

− 8.21

32.92

27.04

23.53

− 5.88

− 3.51

B

15.07

30.07

44.69

15.01

14.62

22.67

28.68

32.15

6.01

3.47

C

26.10

28.73

35.00

2.63

6.27

27.86

27.85

27.77

− 0.01

− 0.08

D

24.67

31.47

33.69

6.80

2.22

27.84

27.84

27.81

0.00

− 0.04

ACOP

A

22.47

18.27

16.20

− 4.20

− 2.07

26.08

24.64

23.82

− 1.44

− 0.82

B

10.54

18.84

27.57

8.30

8.73

20.45

25.43

28.65

4.97

3.23

C

17.75

18.17

21.02

0.42

2.85

24.39

24.71

25.44

0.32

0.73

D

16.79

19.72

20.44

2.93

0.72

24.12

25.27

25.14

1.15

− 0.14

Figure 4a, b show the effect of factors on adsorption capacity. It was observed that with increase of factor A (m) from level 1 to 3 there was decrease in qt values from 48.53 to 16.54 mg g−1 and 22.47 to 16.20 mg g−1 for AGAC and ACOP respectively. This is because of decrease in adsorbate to adsorbent ratio with increase in m values. As the value of factor B (C0) increases from level 1 to level 3 the qt values also increases from 15.07 to 44.69 and from 10.54 to 27.57 mg g−1 for AGAC and ACOP respectively, due to increase in mass transfer driving force and decreased in uptake resistance of p-NP over ACOP surface with increased in C0. It is also observed that adsorption of p-NP increases with the increase in factor C (T) for both adsorbents from level 1 to level 3. The higher temperature reduces the viscosity of p-NP and increases mobility of molecules and kinetic energy which increases the probability of p-NP molecule to adsorb over AGAC and ACOP which ultimately increased the diffusion rate [34]. It also confirmed that the adsorption of p-NP over AGAC and ACOP is endothermic in nature. With the increase in time of contact (t) from level 1 to 3, initially it was found that from level 1 to 2 rapid increase in qt from 24.67 to 28.73 mg g−1 for AGAC and 16.79 to 19.72 mg g−1 for ACOP due to availability of large vacant sites for sorption of p-NP but in later phase i.e. from level 2 to level 3 number of vacant sites decrease with time of contact which retard the uptake capacity (shown in Fig. 4a, b) and it were found that only 2.22 and 0.72 mg g−1 increase in adsorption capacity from level 2 to 3 for AGAC and ACOP respectively.
Fig. 4

Effect of adsorption parameters for sorption of p-NP on A AGAC and B ACOP

Figure 5a, b and Table 7 show the percentage contribution of various factors in overall sorption of p-NP over AGAC and ACOP. The first most affecting factor on sorption of p-NP over AGAC is factor A (m i.e. adsorbent dose) with 51.3% contribution while factor B (C0 i.e. initial concentration) with 80.9% contribution in case of ACOP were observed. The second most influencing factor is B (C0 i.e. initial concentration) in case of AGAC and factor A (m i.e. adsorbent dose) in case of ACOP with 40.70% and 11.40% contribution. After that factor D (t i.e. time) which shows 4.1% and 4.2% contribution for AGAC and ACOP materials. Factor C (T i.e. temperature) shows least effect in sorption process with only 3.9 and 3.5% contribution for AGAC and ACOP respectively.
Fig. 5

Percent contribution of each parameter for sorption of p-NP on A AGAC and B ACOP

Table 7

ANOVA for 4-NP adsorption on AGAC and ACOP

Adsorbent

Factors

ANOVA qt

Pooled ANOVA qt

S

DOF

V

F

P

S

DOF

V

F

P

AGAC

A

4970.08

2

2485.04

10,376,110.94

51.27

4970.08

2

2485.04

132.07

50.88

B

3949.03

2

1974.52

8,244,449.42

40.74

3949.03

2

1974.52

104.94

40.35

C

376.31

2

188.15

785,617.54

3.88

0.00

0

0.00

0.00

0.00

D

397.92

2

198.96

830,738.83

4.11

397.92

2

198.96

10.57

3.72

Error

0.004

18

0.002

1

0.002

376.31

20

18.82

1.00

5.05

Totals

9693.34

26

4846.67

 

100

9693.34

26

4677.33

 

100.00

ACOP

A

183.74

2

91.87

55,199.5

11.38

183.74

2

91.87

32.26

11.03

B

1306.46

2

653.23

392,485

80.93

1306.46

2

653.23

229.40

80.58

C

56.92

2

28.46

17,100.3

3.53

0.00

0

0.00

0.00

0.00

D

67.07

2

33.53

20,147.9

4.15

67.07

2

33.53

11.78

3.80

Error

0.03

18

0

1

0

56.95

20

2.85

1.00

4.59

Totals

1614.22

26

807.09

 

100

1614.22

26

781.48

 

100.00

S∇sum of squares, DOF degree of freedom, V variance, F variance ratio, and P  % contribution

3.4 Selection of optimized level and estimation of optimized response characteristics

To optimize the response value i.e. qt, the maximum values of qt at certain level for particular factor has been selected. From response curve (Fig. 3a, b) and Table 6 it was observed that first level of factor A (m) and third level of factor B (C0), C (T) and D (t) have higher values of qt for AGAC and ACOP both. The average values of qt (mg g−1) for optimal level of factors are as given:
$$\begin{aligned} & \overline{A}_{1} = 48.53,\,\overline{B}_{3} = 44.69,\,\overline{C}_{3} = 35.00\,{\text{and}}\,\overline{D}_{3} = 33.69\,{\text{for}}\,{\text{AGAC}}\,{\text{and}} \\ & \overline{A}_{1} = 22.47,\,\overline{B}_{3} = 21.02\,\overline{C}_{3} = 21.02\,{\text{and}}\,\overline{D}_{3} = 20.44\,{\text{and}}\,{\text{for}}\,{\text{ACOP}}\, ( {\text{from}}\,{\text{Table}}\, 6 )\\ \end{aligned}$$
Grand total of all results T = 808.43 and 512.56 mg g−1 for AGAC and ACOP respectively, and total number of results (N) = 27 for both adsorbents.

Therefore \(\overline{T}\) = T/N = 29.94 and 18.98 for AGAC and ACOP respectively

Hence,
$$\mu_{AGAC} = \bar{T} + (\bar{A}_{1} - \bar{T}) + (\bar{B}_{3} - \bar{T}) + (\overline{C}_{3} - \bar{T}) + (\overline{D}_{3} - \bar{T}) = 72.08\,{\text{mg}}\,{\text{g}}^{ - 1}$$
and
$$\mu_{ACOP} = \bar{T} + (\bar{A}_{1} - \bar{T}) + (\bar{B}_{3} - \bar{T}) + (\overline{C}_{3} - \bar{T}) + (\overline{D}_{3} - \bar{T}) = 34.56\,{\text{mg}}\,{\text{g}}^{ - 1}$$
were computed.
The confidence interval of 95% for the population mean and three confirmation experiments (CIPOP and CICE) are calculated by substituting N = total number of results = 9 × 3 = 27, fe (DOF error) = 26 − 8 = 18, Ve (recalculated error variance after pooling ANOVA) = 18.82 for AGAC and 2.85 for ACOP (from Table 7) in Eqs. 68.
$$n_{eff} = \frac{N}{1 + [Total\,DOF\,associated\,in\,the\,estimate\,of\,the\,mean]} = 3$$
F0.05 (1, 18) = 4.4139 (from standard F-distribution table)Therefore,
$$CI_{POP(AGAC)} = \sqrt {\frac{{F_{\alpha } (1,f_{e} )V_{e} }}{{n_{eff} }}} = \pm 5.26$$
$$CI_{CE(AGAC)} = \sqrt {F_{\alpha } (1,f_{e} )V_{e} \left[ {\frac{1}{{n_{eff} }} + \frac{1}{R}} \right]} = \pm 5.546$$
Similarly for ACOP,
$$CI_{POP(ACOP)} = \sqrt {\frac{{F_{\alpha } (1,f_{e} )V_{e} }}{{n_{eff} }}} = \pm 2.046$$
$$CI_{CE(ACOP)} = \sqrt {F_{\alpha } (1,f_{e} )V_{e} \left[ {\frac{1}{{n_{eff} }} + \frac{1}{R}} \right]} = \pm 2.157$$
The predicted range of qt for sorption of p-NP onto ACOP at 95% confidence interval for CIPOP and CICE for both AGAC and ACOP are shown in Table 8.
Table 8

Predicted optimal qt values, confidence interval and result of confirmation experiment

Adsorbent

Optimal level process parameters

Predicted optimal value (mg/g)

Confidence interval 95%

Average of confirmation (mg/g)

AGAC

A1B3C3D3

72.08

CIPOP(AGAC): 66.82 < µAGAC < 77.35

72.06

CICE (AGAC): 66.54 < µAGAC < 77.63

ACOP

A1B3C3D3

34.56

CIPOP(ACOP): 32.51 < µACOP < 36.60

34.55

CICE(ACOP): 32.40 < µACOP < 36.72

3.5 Confirmation experiment

A1B3C3D3 run already exists in OA i.e. run number 3 and the average of qt of the three repetitions is 72.08 mg g−1 for AGAC and 34.55 mg g−1 for ACOP (from Table 2). This average value lies in the range of 95% CI for CICE (shown in Table 8). This shows the applicability of Taguchi’s process parameters optimization method for sorption.

3.6 Mechanism of adsorption

The molecular structure of p-NP shows aromatic benzene ring with nitro and hydroxyl group. The π electron present in the benzene ring have strong affinity towards the π electron of AGAC and ACOP resulted to π- π bonding. The nitro group which is electron withdrawing groups boosted the π- π interaction and also reduces the repulsive forces between π electrons of two consecutive benzene rings and resulted into enhancement of adsorption of p-NP on AGAC and ACOP surface [5, 13, 35].

4 Conclusions

The surface area by BET for AGAC and ACOP were found 788.93 and 540.61 m2 g−1 respectively, which is decreased by 71.63 and 88.93% after p-NP adsorption showing good removal capacity of p-NP. Internal pores area represented by iodine value 755.10 and 420.14 mg g−1 for AGAC and ACOP respectively indicated the availability of pore surface area and pore volume space for diffusion. SEM images of ACOP showed rough surface with vacant internal pores of varying sizes and the proximate characteristics reflects the good adsorption capacity of AGAC and ACOP for adsorption of p-NP. The higher value of fixed carbon (i.e. 55.38 and 48.20%) shows the better sorption ability of AGAC and ACOP. Application of the Taguchi orthogonal array design methodology on adsorption of p-NP by AGAC and ACOP is very economical in cost, time and resources point of view. The 81 sets of experiment are reduced to 9 sets only. The major responsible factors for sorption such as adsorbent dose (m), initial concentration of p-NP (C0), temperature (T) and time of contact (t) at three levels are optimized in only nine runs. The parameter adsorbent dose (m) and initial concentration of p-NP (C0) were found to be most significant factors with 51.27 and 40.74% for AGAC and 80.9 and 11.4% contribution for ACOP. The time of contact (t) and temperature (T) found to be least significant factor in overall sorption. The optimized factors at different levels found to be A1, B3, C3, D3 for AGAC and ACOP both. The predicted value of qt by Taguchi’s optimization method is found to be 72.08 and 34.56 mg/g for AGAC and ACOP respectively. The confirmatory experiment i.e. run number 3, average value of sorption capacity lies in the range of 95% CICE indicated the applicability of Taguchi’s optimization method for p-NP sorption onto AGAC and ACOP. The percentage removal at optimum condition is found to be 96.11% for AGAC and 92.15% for ACOP.

Notes

Compliance with ethical standards

Conflict of interest

The authors declared that they have no conflict of interest.

References

  1. 1.
    Kumar A, Kumar S, Kumar S, Gupta DV (2007) Adsorption of phenol and 4-nitrophenol on granular activated carbon in basal salt medium: equilibrium and kinetics. J Hazard Mater 147:155–166.  https://doi.org/10.1016/j.jhazmat.2006.12.062 CrossRefGoogle Scholar
  2. 2.
    Ahmad F, Daud W, Ahmad MA, Radzi R (2011) Using cocoa (Theobroma cacao) shell-based activated carbon to remove 4-nitrophenol from aqueous solution: kinetics and equilibrium studies. Chem Eng J 178:461–467.  https://doi.org/10.1016/j.cej.2011.10.044 CrossRefGoogle Scholar
  3. 3.
    Sabio E, Gonalez E, Gonalez JF, Gonalez-Garcia CM, Ramiro A, Ganan J (2004) Thermal regeneration of activated carbon saturated with p-nitrophenol. Carbon 42:2285–2293.  https://doi.org/10.1016/j.carbon.2004.05.007 CrossRefGoogle Scholar
  4. 4.
    Marais E, Nyokong T (2008) Adsorption of 4-nitrophenol onto amberlite IRA-900 modified with metallopthalocyanines. J Hazard Mater 152:293–301.  https://doi.org/10.1016/j.jhazmat.2007.06.096 CrossRefGoogle Scholar
  5. 5.
    Dhorabe PT, Lataye DH, Ingole RS (2016) Removal of 4-nitrophenol from aqueous solution by adsorption onto activated carbon prepared from acacia glauca saw dust. Water Sci Technol 73(4):955–966.  https://doi.org/10.2166/wst.2015.575 CrossRefGoogle Scholar
  6. 6.
    Dutta S, Basu JK, Ghar RN (2001) Studies on adsorption of p-nitrophenol on charred saw dust. Sep Purif Technol 21:227–235.  https://doi.org/10.1016/S1383-5866(00)00205-7 CrossRefGoogle Scholar
  7. 7.
    Singh BK, Nayak PS (2004) Sorption equilibrium studies of toxic nitro-substituted phenols on fly ash. Adsorpt Sci Technol 22:295–310.  https://doi.org/10.1260/0263617041514901 CrossRefGoogle Scholar
  8. 8.
    Tewari BB, Boodhoo M (2005) Removal of p-nitrophenol from aqueous solution through adsorption on antimony, cadmium and zirconium ferrocyanides. J Colloid Interface Sci 289:328–332.  https://doi.org/10.1016/j.jcis.2005.04.032 CrossRefGoogle Scholar
  9. 9.
    Tang D, Zheng Z, Lin K, Luan J, Zhang J (2007) Adsorption of p-nitrophenol from aqueous solutions onto activated carbon fiber. J Hazard Mater 143:49–56.  https://doi.org/10.1016/j.jhazmat.2006.08.066 CrossRefGoogle Scholar
  10. 10.
    Liu QS, Zheng T, Wang P, Jiang JP, Li N (2010) Adsorption isotherm, kinetics and mechanism studies of some substituted phenols on activated carbon fibres. Chem Eng J 157:348–356.  https://doi.org/10.1016/j.cej.2009.11.013 CrossRefGoogle Scholar
  11. 11.
    Ahmaruzzaman M, Gayatri SL (2010) Batch adsorption of 4-nitrophenol by acid activated jute stick char: equilibrium, kinetics and thermodyanamic studies. Chem Eng J 158:173–180.  https://doi.org/10.1016/j.cej.2009.12.027 CrossRefGoogle Scholar
  12. 12.
    Ingole RS, Lataye DH (2015) Adsorptive removal of phenol from aqueous solution using activated carbon prepared from babul sawdust. J Hazard Toxic Radioact Waste 25:30.  https://doi.org/10.1061/(ASCE)HZ.2153-5515.0000271,19(4):04015002 CrossRefGoogle Scholar
  13. 13.
    Dhorabe PT, Lataye DH, Ingole RS (2016) Adsorptive removal of 4-Nitrophenol from aqueous solution by activated carbon prepared from waste orange peels. J Hazard Toxic Radioact Waste.  https://doi.org/10.1061/(ASCE)HZ.2153-5515.0000332 CrossRefGoogle Scholar
  14. 14.
    Khorsandi H, Mohammadi A, Kariminejad F, Haghighi M, Alizadeh R (2015) Determination of optimum conditions for Fenton Oxidation to remove linear alkyl benzene sulfonate using taghuchi Method and requirement evaluation to it for additional treatment of biolac effluent. IJHE 7(4):557–568Google Scholar
  15. 15.
    Rao MS, Ravinder S, Kumar AS (2014) Parametric optimization of abrasive waterjet machining for mild steel: Taguchi approach. Int J Curr Eng Technol 2:1–3.  https://doi.org/10.14741/ijcet/spl.2.2014.06 CrossRefGoogle Scholar
  16. 16.
    Vaghela PA, Prajapati JM (2019) Hybridization of Taguchi and Genetic Algorithm to minimize iteration for optimization of solution. MethodsX.  https://doi.org/10.1016/j.mex.2019.01.002. (in Press)
  17. 17.
    Antony J (2001) Simultaneous optimisation of multiple quality characteristics in manufacturing processes using Taguchi’s quality loss function. Int J Adv Manuf Technol 17:134–138.  https://doi.org/10.1007/s001700170201 CrossRefGoogle Scholar
  18. 18.
    Yeniay O (2001) A comparision of the performances between a genetic algorith and the Taguchi method over artificial problems. Turk J Eng Environ Sci 25:561–568Google Scholar
  19. 19.
    Takeshita S, Hosokawa T (2007) Achieving robust design through quality engineering: Taguchi method. FUJITSU Sci Tech J 43(1):105–112Google Scholar
  20. 20.
    Wu D, Chen HH (2005) Application of Taguchi robust design method to SAW mass sensing devices. IEEE Trans Ultrason Ferroelectr Freq Control 52(12):2403–2410.  https://doi.org/10.1109/TUFFC.2005.1563284 CrossRefGoogle Scholar
  21. 21.
    Tripathi AD, Srivastava SK (2011) Novel approach for optimization of fermentive condition for polyhydroxybutyrate (PHB) production by Alcaligenes sp. using Taguchi (DOE) methodology. Afr J Biotech 10(37):7219–7224Google Scholar
  22. 22.
    Cobb BD, Clarkson J (1994) A simple procedure for optimising the polymerase chain reaction (PCR) using modified Taguchi method. Nucleic Acids Res 22(18):3801–3805.  https://doi.org/10.1093/nar/22.18.3801 CrossRefGoogle Scholar
  23. 23.
    Dasu VV, Panda T, Chidambaram M (2003) Determination of significant parameters for improved griseofulvin production in a batch bioreactor by Taguchi method. Process Biochem 38:877–880.  https://doi.org/10.1016/S0032-9592(02)00068-7 CrossRefGoogle Scholar
  24. 24.
    Prasad KK, Mohan SV, Rao RS, Pati BR, Sarma PN (2005) Laccase production by Pleurotus ostreatus 1804: optimization of submerged culture conditions by Taguchi DOE methodology”. Biochem Eng J 24:17–26.  https://doi.org/10.1016/j.bej.2005.01.019 CrossRefGoogle Scholar
  25. 25.
    Arun K, Shakti K, Aggarwal KK (2006) Identification of strategy parameters for particle swarm optimizer through Taguchi method. J Zhejiang Univ Sci A 7(12):1989–1994.  https://doi.org/10.1631/jzus.2006.A1989 CrossRefzbMATHGoogle Scholar
  26. 26.
    Dargahi A, Nematollahi D, Asgari G, Shokoohi R, Ansari A, Samarghandi MR (2018) Electrodegradation of 2, 4-dichlorophenoxyacetic acid herbicide from aqueous solution using three-dimensional electrode reactor with G/β-PbO 2 anode: Taguchi optimization and degradation mechanism determination. RSC Adv 8(69):39256–39268.  https://doi.org/10.1039/C8RA08471H CrossRefGoogle Scholar
  27. 27.
    Srivastava VC, Mall ID, Mishra IM (2007) Multicomponent adsorption study of metal ions onto bagasse fly ash using Taguchi’s design of experimental methodology. Ind Eng Chem Res 46(17):5697–5706.  https://doi.org/10.1021/ie0609822 CrossRefGoogle Scholar
  28. 28.
    Lataye DH, Mishra IM, Mall ID (2008) Pyridine sorption from aqueous solution by rice husk ash (RHA) and granular activated carbon (GAC): parametric, kinetic, equilibrium and thermodynamic aspects. J Hazard Mater 154:858–870.  https://doi.org/10.1016/j.jhazmat.2007.10.111 CrossRefGoogle Scholar
  29. 29.
    Lataye DH, Mishra IM, Mall ID (2009) Multicomponent sorption of pyridine and its derivatives from aqueous solution on to rice husk ash and granular activated carbon. Pract Period Hazard Toxic Radioact Waste Manag 13(4):218–229.  https://doi.org/10.1061/(ASCE)HZ.1944-8376.0000013 CrossRefGoogle Scholar
  30. 30.
    Singh K, Lataye DH, Wasewa KL (2015) Adsorption of fluoride onto sugarcane bagasse (Saccharum Officinarum): an application of Taguchi’s design experimental methodology. J Indian Water Works Assoc 47(1):285–294Google Scholar
  31. 31.
    Silva MB, Carneiro LM, Silva JPA, Oliveira IDS, Filho HJI, Almeida CRO (2014) An application of Taguchi method (Robust Design) to environmental engineering: evaluating advanced oxidation process in polyester-resin wastewater treatment. Am J Anal Chem 5:828–837.  https://doi.org/10.4236/ajac.2014.513092 CrossRefGoogle Scholar
  32. 32.
    Bureau of Indian Standards (1984) Methods of test for coal and proximate analysis, IS 1350 (I), Manak Bhawan, New Delhi, IndiaGoogle Scholar
  33. 33.
    Lataye DH, Mishra IM, Mall ID (2006) Removal of pyridine from aqueous solution by adsorption on bagasse fly ash. Ind Eng Chem Res 45(11):3934–3943.  https://doi.org/10.1021/ie051315w CrossRefGoogle Scholar
  34. 34.
    Chowdhary S, Mishra R, Saha P, Kushwaha P (2011) Adsorption thermodynamics, kinetics and isosteric heat of adsorption of malachite green onto chemically modified rice husk. Desalination 265:159–168.  https://doi.org/10.1016/j.desal.2010.07.047 CrossRefGoogle Scholar
  35. 35.
    Shokoohi R, Roya AG, Mohammad MM, Abdollah D (2018) Investigation of the efficiency of heterogeneous Fenton-like process using modified magnetic nanoparticles with sodium alginate in removing Bisphenol A from aquatic environments: kinetic studies. Desalin Water Treat 101:185–192.  https://doi.org/10.5004/dwt.2018.21736 CrossRefGoogle Scholar

Copyright information

© Springer Nature Switzerland AG 2019

Authors and Affiliations

  • Prashant T. Dhorabe
    • 1
    Email author
  • Dilip H. Lataye
    • 2
  • Ashwini R. Tenpe
    • 2
  • Ramakant S. Ingole
    • 3
  1. 1.Department of Civil EngineeringPriyadarshini College of EngineeringNagpurIndia
  2. 2.Department of Civil EngineeringVisvesvaraya National Institute of TechnologyNagpurIndia
  3. 3.Department of Civil EngineeringAmrutvahini College of EngineeringSangamnerIndia

Personalised recommendations