Introduction

There has been a growing trend in the interest of scientists and researchers to find cost effective alternatives to powdered and granular activated carbons, which are quite expensive. Hence, the quest for cheap activated carbons (ACs) is on the rise (Guo and Rockstraw 2007).

In past times, ACs were synthesized by conventional heating of biomaterials. This led to the formation of ACs with poor surface properties. Apart from this demerit, higher power, longer time, poor heating flux (no uniform interior heating) and ACs with poor surface properties were produced (Guo and Rockstraw 2007; Xin-hui et al. 2011).

In recent times, ACs prepared from microwave irradiated heating has been widely used by researchers. This is due to the fact that ACs prepared from microwave irradiated heating are facile, consume shorter time, make use of less power, excellent heating flux (uniform interior heating), possess higher surface areas and good pores, and produce hierarchical pore structures (Xin-hui et al. 2011). Microwave irradiation also has the merit of improved efficiency, lower time, precision, immediate startup and shutdown, smaller steps, high heating rate, lower activation temperature, improved safety, energy transfer, selective heating, simplicity, smaller equipment size, reduced equipment size and waste, indirect contact between the heating source and heated materials and little or no automation. These advantages have been exploited by researchers over the conventional heating method of producing ACs. Hence, it is currently used as a viable substitute over the conventional heating method for the synthesis of ACs (Ji et al. 2007; Menezes et al. 2007; Zhang et al. 2009; Menéndez et al. 2010).

Over the decades, these commercially activated carbons have found numerous industrial applications in the removal of organic and colored impurities during wine processing, metals finishing, gold purification, purification of electroplating solutions, production of electrodes, gas purification, air purification, metals extraction, sewage treatment, water treatment, oil spill cleanup, decaffeination, molecular sieves (gases adsorbents) due to their good surface areas. They also find therapeutic applications in the treatment of ingested poisons and diarrhea. Agriculturally, they are used as disinfectants, animal feed additives and pesticides in livestock production. The excellent conductivity of ACs make them good materials for super-capacitors fabrication. Recently, ACs are used as catalysts and catalyst supports by researchers due to their high surface area and good pore structure (Yang et al. 2014; Huang et al. 2015). Pine trees are evergreen and resinous conifers. They grow to a height of 3–80 m, with the majority of species growing to a height of 15–45 m. They best survive in temperate and sub-tropical regions of the world. At the various seasons of the year, pine tree shed its bark, as a coping mechanism to adjust to the temperate and sub-tropical conditions. Pine trees are cultivated for their timber values, as gardens, recreation centers and parks ornamental plants. Pine barks (PB) are thick and scaly, though quite a number of species have flaky and thin barks. Their branches comprise very regular tight spiral, which in appearance are ring of branches, which are shot from a singular point. Many pines are uni-nodal, and hence produce a single set of branches per annum. But others are multi-nodal, producing two or more sets of branches annually. Some pine species have been known to be invasive and threaten natural ecological systems (Extension forester 2016; University of Arkansas 2016). Pine barks are shed as wastes in the environment.

This environmental waste (pine bark waste) was dried, pulverized, chemically activated with KOH, and then converted into pine bark activated carbon (PBAC) by microwave irradiation under various experimental conditions. Chemical activation technique for producing activated carbons has drawn the keen interest of scientists in recent times. These chemical activating agents include KOH, K2CO3, NaOH, H3PO4 and ZnCl2. It has been proven that KOH is the most environmentally friendly chemical activator. It also possesses the advantages of the impacting well arrayed and developed pores structures on activated carbons. But these well developed pores begin to break and get destroyed as the impregnation ratio (concentration) of KOH increases, leading to a reduction in the percentage yield of activated carbon. Hence, to optimize the percentage yield for activated carbons impregnated with KOH, a reasonable balance has to be struck in maintaining the production of well developed pores and at the same time increasing the percentage activated carbon yield. This was done by the instrumentality of system designs known as the Central Composite Design (CCD) and the Response Surface Methodology (RSM) that are present in Design Expert Software, Model 6.0.6, Stat-Ease Incorporated, Minneapolis, MN, USA. The Design Expert Software was optimized using the CCD factorial approach that is capable of providing incremental engineered panaceas for percentage yield activated carbons protocols and maximizing pore formation (Wu et al. 2005; Luo 2012; Ahmed and Theydan 2014; Ozdemir et al. 2014; Huang et al. 2015). In this study, the CCD factorial approach took the KOH impregnation ratio (M), power (W) and time (min) as factors (dependent variables), while percentage activated carbon yield, methylene blue number (MBN) and iodine number (IN) were taken as responses (independent variables). This research addressed the synthesis of activated carbon from pine bark powder, synergizing a compromise between well developed pores and percentage yield of activated carbons for scale-up application(s) of activated carbons.

Theoretical

RSM and CCD statistical optimization

RSM is an applied statistical technique that is used for the optimization of the effects of process variables on the properties of prepared products using a combination of mathematical and statistical approach (Huang et al. 2015). In this research, CCD which was a sub-component of RSM used to study the variables for synthesizing KOH-activated carbon from pine bark powder via microwave irradiated technique. The CCD factorial approach took the KOH impregnation ratio (M), power (W) and time (min) as factors (dependent variables), while percentage activated carbon yield, methylene blue number (MBN) and iodine number (IN) were taken as responses (independent variables) which were considered \({x_1}\), \({x_2}\) and \({x_3}\), respectively. A CCD factorial design with twenty (20) experimental runs was constructed (see Table 1). By doing this, the approximate empirical equation was developed below.

Table 1 CCD factorial design matrices for experimental factors and the responses
Full size table
$$Y={b_0}+\sum\limits_{i=1}^3 {{b_i}{x_i}+\sum\limits_{i=1}^3 {{b_i}x_{i}^{2}+\sum\limits_{i=1}^3 {\sum\limits_{j=i1}^3 {{b_{ij}}{x_i}{x_j}} } } },$$
(1)

where Y is the response variable, \({x_i}\) and \({x_j}\) are the coded variables, that determine \(Y\), \({b_o}\) is the constant coefficient, \({b_i}\) is the linear coefficient, \({b_{ij}}\) is the interaction coefficient and \({b_{ii}}\) is the quadratic coefficient. The experimental data were analysed using Design Expert Software, Model 6.0.6, Stat-Ease Incorporated, Minneapolis, MN, USA for regression analysis and evaluation of the statistical significance of the equation.

Experimental

The pine bark lumps used for this research work were got from a pine plantation in Vanderbijlpark (26°41′57″ South, 27°50′8″ East), Gauteng Province, South Africa in December 2015. The pine barks were washed to remove impurities like sand and leaves, then dried in an oven at 100 °C for 72 h. The dried pine barks were pulverized into powder with particle size ≤300 μm.

Chemical activation of 5 g pine bark powder was impregnated with 0.2, 1.1 and 2.0 M KOH solutions. The KOH impregnated pine bark powder stood in a heating mantle at 30 °C for 24 h and agitated at 250 rpm. This was done to homogenize the pine bark powder in the KOH solution. Thereafter, the slurry of the impregnated materials was placed in an oven at 100 °C overnight. The pyrolysis of the dried impregnated material took place in a Kelvinator microwave oven, KML62B model, 230 V, 50 Hz, with maximum power of 1200 W. The dried impregnated samples were initially subjected to nitrogen atmosphere, before being introduced into the microwave oven at different power ratings of 600–1080 W, with impregnation ratios of 0.2–2.0 for 10–30 min. The pine bark activated carbon (PBAC) produced after the microwave irradiation was cooled in a desiccator to room temperature and thereafter subjected to mild washing with 0.1 M HCl and hot distilled water to remove impurities until the pHs of the washed PBAC were found to be 6.5 and 7.5. The percentage yield was calculated as the ratio of the dry weight of produced PBAC to weight of the dried raw pine bark.

$$\% \;dry\;weight\;PBAC\;produced\;=\;\frac{{dry\;weight\;PBAC\;produced\;(\text{g})}}{{dry\;weight\;of\;pine\;bark\;powder\;(\text{g})}} \times 100.$$
(2)

Methylene blue number (MBN)

MBN is defined as the maximum amount of dye (mg) adsorbed per g of an adsorbent. The MBN experiment was conducted according to the published protocol of Nunes and Guerreiro (2011). In a typical experiment, 10 mg of PBAC was placed in contact with 10 mL of a methylene blue solution at different concentrations of 50–1000 mg L−1 at pH 12.0 for 2 h at room temperature (approximately 30 °C). Thereafter, the residual concentration of methylene blue was analysed by a UV/vis spectrophotometer at 665 nm. The amount of methylene blue adsorbed (mg g−1) from each solution was calculated by difference using;

$${q_{}}=\frac{V}{W}\int\limits_{i=1}^f {({C_i} - {C_f})},$$
(3)

where \({C_i}\) (mg L−1) was the initial concentration of methylene \({C_f}\) (mg L−1) was the final or equilibrium concentration of methylene blue solution, \(V\) (L) was the volume of the solution and \(W\) (g) is the mass of PBAC. To determine the MBN for the Langmuir isotherm, a plot of \({q_{}}\) (mg g−1) was plotted against \({C_i}\) (mg g−1). The Langmuir parameters (\({q_{{{\max }_L}}}\) and \({K_L}\)) were calculated by a least squares fitting regression, KyPlot 2.0 software, with a Quasi–Newton algorithm for finding the parameter values. Also, the specific surface areas (SSA) of PBAC (Karimifard and Moghaddam 2016; Palma et al. 2016) were calculated from MBN values using;

$$SSA\;\;=\;{q_{{{\max }_L}}} \times N \times \frac{{{A_{MB}}}}{{M{W_{MB}}}}$$
(4)

where \(SSA\), \(N\), \({A_{MB}}\) and \(M{W_{MB}}\) are specific surface area (m2 g−1), Avogadro’s constant (6.022 × 1023 mol−1), area per methylene blue molecule (130 × 10−20 m2) and molecular weight of methylene blue dye (319.85 × 103 mg mol−1).

Iodine number (IN)

The iodine number was determined by the published protocol of Nunes and Guerreiro Guerreiro (2011). The iodine number is defined as an amount of iodine (mg) adsorbed per g of an adsorbent when the iodine concentration of the filtrate was 0.02 N. This experiment comprise treating PBAC with 10 mL of 5% HCl. This mixture was boiled for 30 s and then cooled. Thereafter, 100 mL of 0.1 N (0.1 mol L−1) iodine solution was added to the mixture and stirred for 30 s. The resulting solution was filtered and 50 mL of the filtrate was titrated with 0.1 N Na2S2O3·5H2O, using starch as indicator. The iodine number (mg g−1) and C values were calculated using;

$$IN\;(\text{mg/g})=\{ ({N_1} \times 126.93 \times {V_1}) - [({V_1}+{V_{HCl}})/{V_F}] \times ({N_{N{a_2}{S_2}{O_3}}} \times 126.93) \times {V_{N{a_2}{S_2}{O_3}}}\} \frac{1}{{{M_C}}}.$$
(5)

Surface microstructures of PBAC

The Fourier transform infra red (FTIR) spectra of the raw pine bark and the KOH-activated PBAC were recorded on a Perkin–Elmer (USA) FTIR Spectra 400 spectrometer in the range 650–4000 cm−1 to elucidate the functional moieties of these materials. The thermal analysis of PBAC was conducted on Perkin Elmer Thermogravimetric Analyzer (TGA-4000 model), the samples were heated between 30 and 900 °C at a heating rate of 10 °C min−1 and nitrogen flow of 20 mL min−1. Percentage weight loss against temperature was applied for thermo gravimetric analysis (TGA). Also, field emission scanning electron microscopy (FESEM), Carl Zeiss, Sigma model was used to study the surface microstructures of PBAC. The energy dispersive X-ray (EDAX) analyzer which was coupled to the FESEM was used to detect the carbon peaks for PBAC. The surface area and pore structure of PBAC were studied using Tristar II nitrogen sorptometer, 3020 V1.02 model. Samples were outgassed at 10–20 Torr vacuum, at 150 °C for 24 h before the N2 adsorption by multipoint technique at 77 K, between 10−3 and 10° atm pressure.

Results and discussion

Regression and statistical analyses of experimental data

The CCD was utilized for the development of correlations between the dependent variables (factors) and independent variables (responses). This was directly applied in the identification of significant factors contributing to the regression model. The response value \(Y\) obtained from 20 runs at microwave power ratings of 600–1080 W, activation time of 10–30 min and impregnation ratios of 0.2–2.0, with six (6) repetitions at centroid was used to determine the experimental error. Table 2 depicts the detailed summary of experimental variables for the synthesis of PBAC. From the experimental design, it was observed that yield of PBAC, MBN and IN were obtained as 2.27–34.82%, 212.31–1736 mg g−1 (SSA = 10.44–85.37 m2 g−1) and 29.19–2919.39 mg g−1, respectively. This software depicted from Table 3, the proposed model, which were the cubic (aliased) and quadratic (suggested) had probabilities of prob > F = 0.47 and prob > F = 0.08 respectively. Though, both suggested models are not significant (with probability >0.05 degree of confident limit), but the quadratic model explained the experimental data better. On this ground, Table 4 which is an overall summary of lack of fit test suggested the quadratic model to be prob > F = 0.63 and cubic model to be prob > F = 0.69 being aliased. This axiom simply points the fact that the quadratic model slightly explained the factorial design than both mean and cubic models. Therefore, the final equation in terms of coded factors (see Table 4) for % PBAC yield according to the factorial design is given as:

$$\% {\text{ PBAC yield }} = + \,13.34 - 5.90x_{1} - 0.12x_{2} + 3.20x_{3} + 12.62x_{1} ^{2} - 3.51x_{2} ^{2} - 8.91x_{3} ^{2} + 1.48x_{1} x_{2} - 4.12x_{1} x_{3} - 2.09x_{2} x_{3} .$$
(6)
Table 2 Detailed design summary for experimental variables for the synthesis of PBAC
Full size table
Table 3 Suggested models appropriate for the CCD optimization of experimental data
Full size table
Table 4 ANOVA for CCD factorial design in response surface quadratic model for the optimization of % PBAC yield
Full size table

The significant (with probability <0.05 degree of confident limit) segment of equation above is given as:

$$\% {\text{ PBAC yield }}= +{{13}}.{{34}} - {{5}}.{{9}}0{x_1}+{{12}}.{{62}}{x_1}^{2}.$$
(7)

It is understood that the negative and positive signs depict antagonistic and synergistic effects. From Eq. (7), which is the significant equation, it was shown that the percentage PBAC yield had ca. six times first order antagonistic effect, and also had ca. 13 times second order synergistic effect. The lack of fit test value was 0.74, showing that this test was not significant when compared to the pure error. It is good to note that MBN gave the least values for the sum of squares, mean squares and F value. The implication of this is that the MBN is largely controlled by the percentage PBAC yield. The correlation co-efficient (\({R^2}\)) is 0.6723. For MBN, the suggested model was linear, while the cubic model was aliased. Moreover, the model for MBN is not significant (with probability >0.05 degree of confident limit), and the lack of fit test was also not significant. The final equation in terms of coded factors (see Table 5) for MBN according to the factorial design is given as:

Table 5 ANOVA for CCD factorial design in response surface quadratic model for the optimization of MBN
Full size table
$${\text{MBN}}={{+1085.50}} - {{14}}.{{23}}{x_1}+{{19}}0.{{86}}{x_2} - {{169}}.{{65}}{x_3}+{{42}}.{{83}}{x_1}^{2}+{{82}}.{{11}}{x_2}^{2} - {{165}}.{{76}}{x_3}^{2} - {{213}}.{{37}}{x_1}{x_2} - {{145}}.{{19}}{x_1}{x_3} - {{7}}0{x_2}{x_3}.$$
(8)

On the contrary, the suggested model is quadratic, while the cubic model is aliased. For IN, the suggested model and the lack of fit test are both significant. The lack of fit test showed gave prob > F = 0.0067 for the quadratic model and prob > F = 0.0045 for the aliased model. The lack of fit test has the probability of prob > F = 0.0037. The adjusted \({R^2}\) and \({R^2}\) are 0.7806 and 0.8845 respectively. It is interesting to note that the model for IN is significant (with probability <0.05 degree of confident limit). The final equation in coded factors (see Table 6) for IN according to the factorial design is:

Table 6 ANOVA for CCD factorial design in response surface quadratic model for the optimization of IN
Full size table
$$\text{IN}=+82.90+463.55x_{1}-501.25x_{2}-199.86x_{3}+317.35x_{1}^2+338.29x_{2}^2+227.60x_{3}^2-486.94x_{1}x_{2}+57.59x_{1}x_{3}+64.58x_{1}x_{3}$$
(9)

The significant (with probability <0.05 degree of confident limit) segment of equation above is given as:

$$\text{IN}=+82.90+463.55x_{1}-501.25x_{2}-486.94x_{1}x_{2}.$$
(10)

From Eq. (9), it is shown that IN had ca. 464 times first order synergistic effect, ca. 501 times second order antagonistic effect and ca. 487 times influence on the factorial design. This simply implies that IN rapidly increased due to ca. 464 times increase in % PBAC yield (+463.55\({x_1})\), ca. 501 times decrease in MBN (−501.25\({x_2}\)) and ca. 487 times simultaneous decrease in % PBAC yield and MBN (−486.94\({x_1}{x_2}\)).

Figure 1a–c show the normal plot of residual of studentized residuals against normal percentage probability, the experimental data are more concentrated at >50% of the normal plot of residual for % PBAC yield, MBN and IN respectively.

Fig. 1
figure 1

The plot of normal % probability against studentized residuals for a % yield PBAC, b MBN, c IN

Full size image

2D contour plots; effects of the independent variables on % PBAC yield

The nature of the model, the relationship and interaction between the dependent variables (factors) and the independent variables (responses) can be seen from the contour lines of the 2D plots of three independent variables with a specified response (such as % PBAC yield). It is understood that elliptical contours show a strong interaction between two independent variables (factors) with respect to a dependent variable (response). This interaction reduces as the contour lines get less perfectly elliptical (Danish et al. 2014; Zhao-qiang et al. 2014; Karimifard and Moghaddam 2016). The linear contour lines show that there is a direct relationship between two dependent variables (factors) with respect to a independent variable (response). Contrariwise, the circular contour lines show that there is neither interaction between two dependent variables (factors) nor a independent variable (response). Figure 2a–i depict the 2D contour plots of the relationship between impregnation ratio and power, impregnation ratio and time, then power and time as they influence % PBAC yield, impregnation ratio and power, impregnation ratio and time, then power and time as they influence MBN, impregnation ratio and power, impregnation ratio and time, then power and time ratio as they influence IN.

Fig. 2
figure 2

2D contour plots of the relationship between a impregnation ratio and power, b impregnation ratio and time, c power and time as they influence % PBAC yield, d impregnation ratio and power, e impregnation ratio and time f power and time as they influence MBN, g impregnation ratio and power, h impregnation ratio and time, i power and time ratio as they influence IN

Full size image

From Fig. 2a, the contour plot showed that there is a strong interaction between the impregnation ratio and power due to the increase in the curvature of the elliptical contours. This interaction between impregnation ratio and power tends to increase as impregnation ratio values increased from 0.20 to about 1.10 when power is held constant at 840 W. This interaction got stronger when the impregnation ratio was >1.10 to <2.00. At ca. 2.00, the interaction between the impregnation ratio and power decreased drastically. Conversely, increase in the impregnation ratio from 0.20 to 2.00 was accompanied with a corresponding decrease in % yield of PBAC from 28.18 to 12.73%, and at ca. 2.00, there was a slight increase in the % yield of PBAC from 12.73 to 16.59%. In the same vein, when time was held constant at 20 min, it was observed from Fig. 2b that a strong interaction exists between the impregnation ratio and time due to the elliptical nature of the contours. The curvature of the elliptical contours tends to increase with increase in the impregnation ratio from 0.20 to ca. 1.10, coupled with a corresponding decrease in % yield of PBAC from 28.18 to 12.73%. The curvature of the elliptical contours increased as the impregnation ratio approached ca. 2.00, coupled with a slight increase in the % yield of PBAC from 12.73 to 16.59%.

It is interesting to note from Fig. 2c that there was no interaction between the power and time. This simply implies that the % yield of PBAC was not influenced by power and time.

Figure 2d showed that there is a strong interaction between impregnation ratio and power, as they influence the MBN values of PBAC. The interaction became more pronounced as power decreased from 1080 W to ca. 600 W, with decrease in MBN values from 1494.16 to 955.18 mg g−1 as impregnation ratio approached ca. 2.00. This supports the fact that the formation of mesopores decreased with decreasing power. The interaction became stronger at ca. 840 W with impregnation ratio value of ca. 1.55. This also shows a sharp decrease in the formation of mesopores below 840 W. Hence, a sharp decrease in the formation of mesopores of PBAC is accelerated decrease in power. Figure 2e depicted that there is a strong interaction between impregnation ratio and time. The interaction became stronger as time decreased with increase in MBN value at impregnation ratio of 1.10 and ca. 25 min. The MBN value was constant <15 min at impregnation ratio value of 1.10, and the MBN value steeply increased to 1224.67 mg g−1 as the impregnation ratio approached 2.00. For Fig. 2f, increase in power from 600 to 1080 W and decrease in time from 30 to 10 min led to less elliptical nature of the contours. The implication of this lies in the fact that the formation of the mesopores of PBAC increased from 955.18 to 1359.41 mg g−1 with decreasing time and increasing power. The values of MBN increased more as power increased from 840 to 1080 W.

It is understood that KOH activating agent performs the function of a reductant for lignocellulosic precursors and other materials as it produces activated carbons with high specific areas and tunable pores with a low yield of carbon that is ascribed to the development of pores that accelerates carbon due to the intercalation of K+ into the carbon matrix.

The KOH activation process has been established to follow the reaction trend below:

$${\text{C}}+2{\text{KOH}} \to 2{\text{K}}+{{\text{H}}_2}+{\text{CO}}_2,$$
(11)
$${\text{C}}+2{\text{KOH}} \to 2{\text{K}}+{{\text{H}}_2}{\text{O}}+{\text{CO}},$$
(12)
$${\text{CO}}_2+2{\text{KOH}} \to {{\text{K}}_2}{\text{CO}}_3+{{\text{H}}_2}{\text{O}},$$
(13)
$${{\text{H}}_2}{\text{O}}+{\text{C}}+2{\text{KOH}} \to {{\text{K}}_2}{\text{CO}}_3+{{\text{H}}_2}.$$
(14)

Activated carbon yield is low at high power ratings or activation temperatures due to the fact that KOH may catalyze oxidation reactions that will oxidize the outer surface carbon atoms, leading to pores formation. The intercalation of K into the carbon matrix during activation results in the formation of both mesopores and micropores. This K diffuses into the carbon matrix, producing a significant number of micropores (Prahas et al. 2008; Hui and Zaini 2015).

Hence, high power rating and low impregnation ratio will lead to the collapse of micropores as a result of carbon burn off that widens and increases the pore sizes of the micropores, thereby converting these micropores to mesopores. Therefore, increase in power rating will tend to generate activated carbon that is more mesoporous in nature than micropores. Abechi et al. (2013) expounded that K2CO3 produced from KOH activation of lignocelluloses might avert their excessive burn-off, and evidently lead to a high product yield and well developed porosity.

From Fig. 2 g, the interaction between the impregnation ratio and power was not strong due to the less elliptical nature of the contours. Decrease in power and increase in the impregnation ratio led to decrease in the curvature of the elliptical contours. A reduction in this interaction had led to an increase in the IN values from 263.94 to 1805.01 mg g−1. This means an increase in the formation of the micropores of PBAC was controlled by decrease in power from 1080 to 600 W and increase in values of impregnation ratio from 0.20 to 2.00.

The interactions between the impregnation ratio and time, than power and time (Fig. 2h, i) are stronger than that of impregnation ratio and power that was discussed above. These contours became less elliptical, thereby reducing the interactions with corresponding increase in the values of IN from 263.94 to 1034.47 mg g−1. For Fig. 2h, the values of IN increased with increase in the impregnation ratio values from 0.20 to 2.00. Conversely, for Fig. 2i, the IN values decreased from 1034.47 to 263.94 mg g−1 as power increased from 600 to 1080 W. This implies that increase in the formation of the micropores of PBAC is controlled by a corresponding increase in the impregnation ratio and decrease in power. It is known in literature that KOH activation of lignocellulosic materials yields of microporous activated carbons, and that mesoporous activated carbons produced by KOH activation might be due to the manifestation of the lower carbon content of the precursor, and the high impregnation ratio used in the activation.

As K+ is formed, it diffuses into the carbon layer producing a large number of micropores. If the mesopores content is kept constant and low, the formation of the micropores increase at an elevated rate, in turn widen and burn off, leading to a decrease in the % of PBAC yield. The occurrence known as pore coalescence or collapse involves the widening of micropores into mesopores and macropores, have been identified as the leading cause for mesopore formation during KOH activation. It is also cogent to state here that low power rating and high impregnation ratio will increase the intercalation and migration of K in the lignocelluloses, thereby increasing their microporous contents (Rodriguez-Reinoso and Molina-Sabio 1992; Sudaryyanto et al. 2006; Li et al. 2008; Khalil et al. 2010; Zhang and Chen 2015).

3D response surface plots; effects of the dependent variables on % PBAC yield

The 3D response surface plots have been recommended and recognized to be highly informative and explicit for the interpretation of the interaction behaviour among the system variables. Figure 3a–i depict 3D response surface plots for the relationship between impregnation ratio and power, impregnation ratio and time, then power and time as they influence % PBAC yield, time and power, power and impregnation ratio, then time and impregnation ratio as they influence MBN, timemes and pore width characteristics of PBAC synthesized by CCD factorial design and power, power and impregnation ratio, then time and impregnation ratio as they influence IN.

Fig. 3
figure 3

3D response surface plots for the relationship between a impregnation ratio and power, b impregnation ratio and time c power and time as they influence % PBAC yield, d time and power, e power and impregnation ratio, f time and impregnation ratio as they influence MBN, g time and power, h power and impregnation ratio, i time and impregnation ratio as they influence IN

Full size image

From Fig. 3a, when time was held constant at 20 min, the % PBAC yield decreased with decrease in power from 1080 to 600 W and increase in impregnation ratio from 0.20 to ca. 1.10. A slight increase in the % PBAC was observed with an increase in impregnation ratio from >1.10–2.00. This figure showed that there is an interaction between power and impregnation ratio. This interaction tends to increase with increase in impregnation ratio, and became more significant at ca. 1.00–1.80. This exposed the fact that the % PBAC yield was more influenced at impregnation ratio ca. 1.00–1.80. When power was constant at 840 W (Fig. 3b), the % PBAC yield decreased with decrease in time from 30 to 10 min, and decrease in impregnation ratio from ca. 1.10–0.20. Moreover, a slight increase in the % PBAC yield was observed at impregnation ratio of >1.10–2.00. The interaction between time and impregnation ratio was more pronounced for impregnation ratio of 0.20–2.00. It shows that impregnation ratio played more role than time in ascertaining the % PBAC yield obtained. Figure 3c has the impregnation ratio held constant at 1.10, as the % PBAC yield underwent a sharp decrease with decrease in time from 30 to 10 min, and increase in power from 600 to 1080 W. It is lucid here that there is not interaction between power and time, as it % PBAC yield that was obtained. From Fig. 3d, as the time was held constant at 20 min, there was a drastic decrease in MBN value from 1628.90 to 820.43 with corresponding decrease in power from 1080 to 600 W and increase in impregnation ratio from 0.20 to 2.00. The interaction between power and impregnation ratio is more significant at power <840 W and impregnation ratio of ca. 1.10–1.55. This suggests that power and impregnation ratio play a cogent role in determining the MBN value of PBAC. Figure 3e depicts a gradual decrease in the MBN value from 1263.55 to 633.50 with a corresponding decrease in time from 30 to 10 min, and increase in impregnation ratio from 0.20 to 2.00. The interaction between time and impregnation ratio is less significant when compared to that between power and impregnation ratio for the determination of MBN values. For Fig. 3f shows that the MBN value slightly decreased with decrease in time from 30 to 10 min and increase in power from 600 to 1080 W. Here, the interaction between time and power for the determination of MBN value of PBAC was more pronounced at power ca. 1080–840 W, showing that the change in MBN value was most influenced around this power region. From Fig. 3g, it is interesting to note that there was a very slight increase in the IN value with decrease in power from 1080 to 600 W, with increase in impregnation ratio from 0.20 to 2.00 when time was held constant at 20 min. The interaction between power and impregnation ratio increased and became more pronounced as the impregnation ratio decreased from 2.00 to 0.20. Figure 3h shows a gradual increase in IN value with corresponding decrease in time from 30 to 10 min and increase in impregnation ratio from 0.20 to 2.00, when power was held constant at 840 W. The interaction between time and impregnation ratio was very pronounced, showing that these factors significantly influence the IN value for the PBAC. Holding impregnation ratio constant at 1.10 in Fig. 3i, the IN value slightly decreased, and then increased as time decreased from 30 to 10 min. Also, the IN value decreased with increase in power from 600 to 1080 W. Hence, the interaction between time and power significantly increased from 600 to 1080 W, showing that the influence of power in determining the IN value for PBAC is greater than time.

Surface microstructures of PBAC

FTIR, TGA, FESEM, EDAX and nitrogen sorptometry–desorptometry analyses

The FTIR spectra in Fig. 4a were taken as representatives of all the PBAC samples that were synthesized and PB. These samples had the best BET surface areas from nitrogen adsorption–desorption. For PBAC synthesized with 2 M KOH, 1080 W, 10 min, peaks were observed at 3308, 1877, 1534, 980 and 677 cm−1, which were O–H stretch, C=O stretch associated with lignocelluloses precursor, C=O stretch of aromatics, olefinic C=C stretch and two out of plane C–H bend of aromatic ring respectively. For PBAC synthesized with 0.2 M KOH, 600 W, 30 min, peaks were found at the 3336, 2916, 1561, 1014 and 779 cm−1, which were O–H stretch, C–H stretch of CH2 and CH3, olefinic C=C stretch, C–O stretch and out of plane C–H bend of aromatic ring respectively. For PBAC synthesized with 1.1 M KOH, 840 W, 30 min, peaks were observed at 3315, 2942, 1595, 1045 and 736 cm−1, which were also O–H stretch, C–H stretch of CH2 and CH3, olefinic C=C stretch, C–O stretch and out of plane C–H bend of aromatic ring respectively. The C–H stretch of CH2 and CH3 disappeared in PBAC synthesized with 2 M KOH, 1080 W, 10 min. An evidence of the disappearance of this peak is the formation of the C=O stretch of lignocelluloses at 1877 cm−1. This particular C=O stretch were not found in PBACs synthesized with 0.2 M KOH, 600 W, 30 min and 1.1 M KOH, 840 W. The reason for this is ascribed to the fact that the very high power rating under which the PBAC synthesized with 2 M KOH, 1080 W, 10 min was made decomposed its C–H stretch of CH2 and CH3. The PB had peaks at 3335, 2927, 1744, 1601, 1285, 1051 and 796 cm−1, which were O–H stretch, C–H stretch of CH2 and CH3, C=O stretch from lignocelluloses precursor, olefinic C=C stretch, C=O bend, C–O stretch and out of plane C–H bend of aromatic ring respectively (Rodriguez-Reinoso and Molina-Sabio 1992; Sudaryyanto et al. 2006; Li et al. 2008; Khalil et al. 2010).

Fig. 4
figure 4

a The thermograms of 1.1 M KOH, 840 W, 30 min, 0.2 M KOH, 600 W, 30 min, 2 M KOH, 1080 W, 10 min and PB, b the FTIR spectra of 1.1 M KOH, 840 W, 30 min, 0.2 M KOH, 600 W, 30 min, 2 M KOH, 1080 W, 10 min and PB

Full size image

The TGA thermograms in Fig. 4b showed three regions of weight losses. The first region of weight losses from 30 to 230 °C consists of 6% weight loss for PBAC prepared from 0.2 M KOH, 600 W, 30 min, 7.5% weight loss for PBAC prepared from 1.1 M KOH, 840 W, 30 min, 8% weight loss for PBAC prepared from 2 M KOH, 1080 W, 10 min and 10% weight loss for PB. These weight losses were due to the decomposition of hemicelluloses. The second region of weight losses from 230 to 440 °C consists of 14% weight loss for PBAC prepared from 0.2 M KOH, 600 W, 30 min, 12% weight loss for PBAC prepared from 1.1 M KOH, 840 W, 30 min, 11% weight loss for PBAC prepared from 2 M KOH, 1080 W, 10 min and 37% weight loss for PB. These weight losses represent tightly held lignin volatiles in the various PBAC samples. The third region of weight losses from >440 °C consists of 10% weight loss for PBAC prepared from 0.2 M KOH, 600 W, 30 min, 9.5% weight loss for PBAC prepared from 1.1 M KOH, 840 W, 30 min, 8% weight loss for PBAC prepared from 2 M KOH, 1080 W, 10 min and 22% weight loss for PB. These weight losses represent the decomposition tightly held lignin volatiles in the various PBAC samples. These weight losses amounted from the decomposition other tightly held organics on the PBAC samples (Sudaryyanto et al. 2006; Li et al. 2008; Khalil et al. 2010; Zhang and Chen 2015).

Figure 5 showed nitrogen adsorption–desorption plots [the plots of quantity adsorbed/desorbed in cm3 g−1 at s.t.p against relative pressure, P/Po (atm)] for the various PBAC synthesized under different conditions. The responses of all the synthesized PBAC samples to nitrogen gas adsorbed or desorbed at s.t.p at <1 atm was between 2 and 140 cm3 g−1. From Table 7, the surface areas, pore structure characteristics of the PBAC synthesized are shown. The BET surface areas of the PBAC samples span from 19.11 to 411.96 m2 g−1. The BET surface areas and the pore structures of the PBAC samples varied due to the different synthesis conditions. Table 7 showed that most of the PBAC synthesized have larger micropore surface areas, rather than mesopore surface areas. Only four PBAC samples have more mesopore surface areas than micropore surface areas. The PBAC samples synthesized with 0.2 M KOH, 1080 W, 10 min and 2 M KOH, 1080 W, 30 min formed the largest percentage of micropore surface areas and mesopore surface areas respectively. Similarly, more micropore volumes were formed by the PBAC samples rather than mesopore volumes. Only five PBAC samples have more mesopore volumes than micropore volumes. This substantiates the fact that KOH impregnating agent impacts high micro-porosity in activated carbons (Zhang and Chen 2015).

Fig. 5
figure 5

Nitrogen adsorption–desorption plots for the various PBAC synthesized under different conditions

Full size image
Table 7 Surface areas, pore volumes and pore width characteristics of PBAC synthesized by CCD factorial design
Full size table

The SEM images of three PBAC samples and PB with the highest surface areas are shown in Fig. 6a–h. Figure 6a, b showed a crispy surface with particles of irregular shapes, and scattered unevenly for PBAC sample synthesized with 1.1 M KOH, 840 W, 30 min. For Fig. 6c, d, the surface of PBAC sample synthesized with 0.2 M KOH, 600 W, 30 min showed a porous surface with mono-dispersed particles arrayed in small heaps and intertwined connectivity. For Fig. 6e, f, this porous surface showed arrayed and mono-dispersed pores and particles, gathered in layers. The pores in Fig. 6e, f are larger and prominent than the pores in Fig. 6c, d. For Fig. 6g, h, its surface showed large irregular particles that are scattered in large flaky lumps on the surface of PB. The EDAX spectra, the peak intensities of carbon were greater than the peak intensities of the other elements that were found in the PBAC samples in Fig. 7a–c.

Fig. 6
figure 6

The scanning electron microscopy images for a, b 1.1 M KOH, 840 W, 30 min, c, d 0.2 M KOH, 600 W, 30 min and e, f 2 M KOH, 1080 W, 10 min and g, h PB

Full size image
Fig. 7
figure 7

The energy dispersive analysis of X-ray (EDAX) spectra, showing carbon peaks for a 1.1 M KOH, 840 W, 30 min, b 0.2 M KOH, 600 W, 30 min and c 2 M KOH, 1080 W, 10 min

Full size image

Conclusion

Potassium hydroxide impregnated activated carbons were synthesized from a biogenic waste (pine bark) by microwave carbonization technique. The central composite design of the response surface methodology was used as a scientific tool to design the various experimental conditions under which these activated carbons were synthesized. These activated carbons were characterized using Fourier transform infra red (FTIR) spectrometry, thermogravimetric analysis (TGA), field emission scanning electron microscopy (FESEM), energy dispersive analysis of X-ray (EDAX) and nitrogen sorptometry–desorptometry. This research finding depicts activated carbons with high microporosity and relatively large surface areas, as well as provides a synthesis route for cost effective activated carbons for various industrial applications.