Lucie Bartoňová*,Lucie Ruppenthalová,Michal Ritz

1V?B-Technical University of Ostrava,Faculty of Metallurgy and Material Engineering,Department of Chemistry,T?.17.listopadu 15,708 33 Ostrava,Poruba,Czech Republic

2V?B-Technical University of Ostrava,Sustainable Development of Centre ENET,T?.17.listopadu 15,708 33 Ostrava,Poruba,Czech Republic

3V?B-Technical University of Ostrava,Regional Materials Science and Technology Centre,T?.17.listopadu 15,708 33 Ostrava,Poruba,Czech Republic

4Institute of Geonics of the CAS,Studentská 1768,708 00 Ostrava,Poruba,Czech Republic

1.Introduction

Dyes are widely used not only in textile or leather manufacturing,but also in the rubber and plastics industry,paper printing,food processing,cosmetics or the pharmaceutical industry.Such an extensive utilization of dyestuff leads to a huge annual production,which is estimated to be 7×105tonnes[1,2].It is assumed that there is about 1%–2%loss in production and 1%–10%loss in use[3].Furthermore,even 1×10—6concentration is highly visible;thus,only a small release of dye can colour a vast amount of the receiving water environment[4].Apart from the undesirable colour,contaminated waters reduce the transmission of sunlight into streams,which consequently reduces the photosynthetic activity of living organisms.

There is a wide range of destructive(biodegradation,oxidation/ozonization)or non-destructive(sedimentation, filtration,coagulation,adsorption)methods designed for the removal of the dye[1].However,since most synthetic dyes are of an aromatic character,their chemical structure is rather stable and resistant to biodegradation.Therefore,the adsorption on suitable adsorbent provides an attractive alternative,which can be very efficient.

There is extensive literature dealing with wastewater cleaning and many adsorbents have been studied in this context,such as wood[5],fruit peel[6],tree fern[7],plant seeds[8],peat[9],coal[10,11],aluminosilicates[12,13],and power station ashes[14–16].

One of themost widely used adsorbents is activatedcarbon which is usednotonly for the cleaning of effluents[17–19]butal so for there tenti on of pollutants from flue gas during coal combustion or municipal solid waste incineration[20–22].

The main drawbacks of activated carbon are its rather high cost in combination with some difficulties associated with its regeneration;therefore,low-cost precursors and alternative materials are being sought,from which unburned carbon(UC)belongs to the most promising ones.Moreover,the utilization of unburned carbon as an adsorbent is beneficial because it is present in all coal combustion ashes and therefore a huge amount of UC is produced annually on a worldwide scale.

Depending on the UC origin,some UCs exhibit a high degree of turbostratic structural order,for whichit has beenstudied asa potential precursor for the preparation of various carbon artefacts(synthetic graphite,electrodesetc.)[23–26].Moreover,UC grains or pyrolysis chars typically provide good adsorption properties[26–29]and for this reason it has been widely used for the retention of pollutants from coal combustion flue gas[30–32]where promising results have been achieved.However,studies based on UC utilization as an adsorbent for the retention of dye from effluents are scarce[33–37]and UCs have been tested predominantly for the retention of the basic(cationic)dyes and so far,there is the only paper dealing with the adsorption of acid(anionic)dye of UC[37].Moreover,all these unburned carbons[33–37]originated from the combustion of coal without any other materials.Despite the trend toward the co-combustion of coal with biomass or various waste materials[38–44],it is quite surprising that the study(in this context)of UC originating from the cocombustion of coal and wastes has not been published yet.

For the reasons described above,this study deals with the adsorption of Naphthol Green B(Acid Green 1)onto two unburned carbons and the parent coal,from which the UCs have been created in the fluidised-bed power station.This study deals with the adsorption equilibrium modelling—experimental data has been analysed using 2-parameter(Langmuir,Freundlich)and 3-parameter(Redlich–Peterson)isotherms.As some recent adsorption studies indicated that linear regression widely used for the analysis of the adsorption process might not always provide reliable results[45–49],both linear and non-linear regressions have been used for the estimation of the isotherm parameters.

2.Equilibrium Isotherms Description

Due to its simplicity and software availability,the linearization of kinetic and equilibriumadsorption data is often used to estimate the parameters of non-linear models.However,transformations of non-linear equations to linear forms may alter the error distributions and violate the normality assumptions of the least-square method[45–49].Therefore,non-linear regression is more complex and is usually recommended for the estimation of the model parameters[45–49].

In this study,for the description of the adsorption equilibrium,Langmuir,Freundlich and Redlich–Peterson isotherm equations were applied to the experimental data—both non-linear and linearized forms were used for the comparison and for the calculation of the isotherm parameters.The non-linear and linearized forms of the adsorption isotherm models used in this study are summarized in Table 1.

2.1.Statistical calculations and comparison of the models

Non-linear regression and the calculation of the regression characteristics were performed using QC Expert statistical-analysis software(TriloByte)using Gauss–Newton,Mar quardt,Gradient-Cauchy,Simplex and other algorithms(α=0.05)including advanced statistical methods specified by various international standards and regulations,such as ISO 9000,ISO 14000 or ISO 9000.

Two-parameter(Langmuir,Freundlich)and three-parameter(Redlich–Peterson)models were tested in terms of their ability to describe the experimental kinetic data.

The comparison of the accuracy of the fit of the tested model to experimental data was evaluated by means of the coefficient of determination(R2),Chi-square statistic test(χ2)and Akaike's criterion(AIC).Akaike's information criterion is an advantageous statistical approach for the comparison of the kinetic models with different number of parameters[55].Generally speaking,the more parameters used in the model,the better the fit achieved is(resultinge.g.in the higher coefficient of determination).For this reason,AIC has been used with the aim of finding the model with the lowest number of parameters but still providing a good fit(this is achieved by a penalty used for each parameter used).

Values χ2were calculated by MS Excel using the formula:

whereqe,calcandqe,measare the calculated and the measured amounts of NGB adsorbed at equilibrium.

R2and AIC values come from the aforementioned statistical software that offers two different modes of operation.The first(and more common one)is based on further approximation of the input(estimated)parameter values.When iteration procedure is finished,the statistical characteristics are calculated for the actual(improved)parameter values.This procedure has been used for the non-linear regression calculations(for the calculation of theR2and AIC values).The second mode of the operation enables to calculate the statistical characteristics(R2,AIC)for any input values of the isotherm parameters without any other iteration or approximation.This procedure has been used for the testing of the isotherm para meters calculated from the linear regression coefficients(MS Excel).This approach has been preferred due to quite problematic comparison of theR2values if originating from the linear and the non-linear regression.The principle of the aforementioned approach is not to test the closeness of the “l(fā)inearized”experimental data to the equation of the line(in MS Excel)but to test the goodness of the fit of the isotherm(based on these “l(fā)inearized”parameters)to the original experimental data.In other words,e.g.,for the Freundlich isotherm,the two series of the isotherm parameters were obtained(one from the non-linear and one from the linearized model in MSExcel).Then,both calculated Freundlich isotherms were tested for their ability to describe the experimental data using the same procedure.Therefore,suchR2values can be compared and used for the evaluation of the goodness of the fit.

Table 1Equilibrium isotherms used for the description of the adsorption of Naphthol Green B onto coal and two unburned carbons

3.Experimental

3.1.Adsorbents

Three adsorbents were collected at a circulating fluidised-bed power station—parent coal and two unburned carbons originating from its combustion along with waste materials(and limestone due to desulphurization of flue gas)at the temperature of 850°C.Unburned car bong rains were separated manually from bottom ashes(usingtweezers)and ground to a particle size of＜0.09mm.The parent coal was also studied for comparison with both unburned carbons to evaluate the improvement of the adsorption efficiency of the coal during its incomplete combustion in a real power station.

Ash content(determined gravimetrically in muffle furnace at 815°C)in UC1,UC2 and the coal was66%,62%and22%.The major mineral phases(identified on the PANalytical X'Pert3Powder X-ray diffraction spectrometer)in both UCs were quartz,anatase and magnetite and the dominant minerals in the coal were quartz,kaolinite,anatase,siderite and pyrite.Specific surface areas(determined by the adsorption of nitrogen and the BET method on Sorptomatic 1990)of UC1,UC2 and the coal were 183,123 and 21 m2·g-1.The most abundant pores in UC1 and UC2(determined by the BJH method,Sorptomatic 1990)are mesopores falling approximately within the range of 10–16 nm and in the case of the coal it was 8–12 nm(but with a somewhat lower pore volume).The degree of graphitization and carbon disorder studied by Raman spectroscopy(DXR Smart Raman Spectrometer)revealed the presence of both G(graphite)and D(disorder)bands indicating some disorder of the crystallites and a relatively low degree of the graphitization of both UCs.

3.2.Adsorbate

The adsorbate used in this study was Naphthol Green B(Sigma Aldrich);the molecular formula is given in Fig.1(M=878.46 g?mol-1,λmax=714 nm).Stock solutions were prepared by dissolution of Naphthol Green B in demineralised water.The spectrogram of Naphthol Green B is shown in Fig.2.

Fig.1.Molecular formula of Naphthol Green B.

Fig.2.Spectrogram of Naphthol Green B(the visible region of the spectrum).

3.3.Adsorption procedure

Batch adsorption experiments were performed in 150 ml of the studied dye solution(in demineralized water)in 250 ml PP bottles with 0.0375 g of unburned carbon.

All adsorption tests were carried out at laboratory temperature using Multi-shaker PSU 20(Biosan)at 200 r·min-1.Then the solutions were centrifuged in a laboratory centrifuge for 30 min at 3500gand the concentration of the dye in the supernatant solution was measured spectrophotometrically with UV–VIS spectrophotometry (UV-1601,Shimadzu).

The amount of adsorbed dye was calculated using Eq.(1):

where:

qethe adsorbed amount at equilibrium,mol·g-1

c0,cethe initial and equilibrium concentration of Naphthol Green B solution,mol·L-1

Vvolume of the solution,L

mmass of the unburned carbon,g

4.Results and Discussion

4.1.Non-linear regression of equilibrium isotherm models

Equilibrium data were analysed using 2-parameter(Langmuir,Freundlich)and 3-parameter(Redlich–Peterson)adsorption isotherm models.The parameter values were calculated by means of the nonlinear regression(which provides a mathematically rigorous method for determining isotherm parameters using the original form of the isotherm equation).The comparison of the non-linear isotherm models and the fit of the experimental data were performed byR2and namely by χ2and Akaike's information criterion(AIC statistic test).

The calculated parameters as well as the values ofR2,AIC and χ2for the two unburned carbons and the parent coal are summarized in Table 2.The fit of the experimental data can also be checked visually in Figs.3–5.

The data given in Table 2 and the comparison of the applied models suggest that the highestR2and the lowest χ2and AIC values have been calculated for both UCs using the Redlich–Peterson isotherm model.Hence,the best fit of the experimental data of the adsorption of Naphthol Green B on UCs has been achieved by the Redlich–Peterson isotherm equation—the AIC values for this model were the lowest ones even if some penalty has been used due to higher number of the parameters(when compared to 2-parameter models).Nevertheless,some caution is needed due to quite broad confidence interval of some parameters calculated by Redlich–Peterson(R–P)model.

Table 2Parameters and statistics of nonlinear isotherm models

Fig.3.Langmuir,Freundlich and Redlich–Peterson equation isotherms for the adsorption of Naphthol Green B on UC1(based on nonlinear regression).

Fig.4.Langmuir,Freundlich and Redlich–Peterson equation isotherms for the adsorption of Naphthol Green B on UC2(based on nonlinear regression).

In the case of the parent coal,the highestR2and the lowest χ2were calculated for the Red lich–Peterson and the Langmuir models indicating comparable results with slightly better fit achieved by the R–P model.However,the AIC value for the R–P isotherm is higher than that of the Langmuir model.It can be explained by a similar closeness of the fit by the Langmuir and the R–P model but in the latter case,the penalty applied for the third parameter increased the final AIC value suggesting that there is no justification of using the3-para meter model in this case.Moreover,the observation is consistent with a very high standard deviation calculated for the para meter B that is even higher than the parameter itself.The conclusion can also been checked in Fig.5 where the R–P and the Langmuir isotherms exhibit nearly the same shape—therefore,there is no justification for using the R–P isotherm for the description of the NGB adsorption on the coal(using the Langmuir isotherm is absolutely sufficient in this case).

The visual check of the quality of the fit of the experimental data in the case of the unburned carbons(Figs.3 and 4)is consistent with the worst fit by the Langmuir isotherm,whereas 2-parameter Freundlich and 3-parameter R–P are both good.

An extensive literature survey on the Redlich–Peterson(R–P)isotherm was presented by Wuet al.[56].In thirty evaluated papers,R–P isotherm was more accurate than the Langmuir and the Freundlich ones;twelve papers reported that both R–P and Langmuir isotherms had equally high accuracy.It has been mentioned in this context that the R–P isotherm is unlikely to provide worse fit than Langmuir or Freundlich isotherms because if exponentgis adjusted to 1 it is the same as the Langmuir isotherm and if 1/A→0 it is the same form as the Freundlich isotherm[56].Therefore,the question is whether using the 3-parameter model instead of the 2-parameter ones is justified.For this reason,in the case of the unburned carbons,AIC has been used for such evaluation(despite the penalty for the third parameter,R–P model provided the lowest AIC values for both unburned carbons).

The comparison of the values of the exponentg(from the R–P model)is interesting as well.In the aforementioned study[56](based on the evaluation of35studies of the dye adsorption systems),it was reported that 31 exponent values were lower than 1(with most values falling within the range of 0.82–0.98).Suchgvalues were explained as the existence of a solid impediment between pores and a large molecule of the adsorbate(dye).In the case of the unburned carbons,gvalues of 0.84 and 0.94 have been calculated,which is in an agreement with the adsorption of a large-molecule adsorbent(the molecular weight of Naphthol Green B is 878.46 g·mol-1).

Fig.5.Langmuir,Freundlich and Redlich–Peterson equation isotherms for the adsorption of Naphthol Green B on coal(based on nonlinear regression).

Moreover,the value of the exponentg(from the R–P isotherm)affects the shape of the isotherm;according to Wuet al.[56],the higher thegvalue(up tog=1),the more significant the curvature of the isotherm(i.e.,the sharper the increase of the adsorbed amount at the beginning with more horizontal end of the curve).Then,ifgvalue in the R–P equation reaches 1,its form is the same as the Langmuir equation exhibiting the nearly horizontal end of the curve[56].In the case of the unburned carbons,thegvalues were calculated as 0.94 and 0.84 indicating the more obtuse(open)bend angle(the less significant curvature)in relation to the Langmuir isotherm.At any case,the more obtuse bend angle(brought about byg＜1)helped to achieve the better fit of the experimental data.

As far as the adsorbed amount of Naphthol Green Bis concerned,the papers in literature are quite scarce.For example,metal hydroxides sludge from hot dipping galvanizing plant(low-cost adsorbent)have been used for this purpose— the adsorption capacity was 10 mg·g-1[57].The polymeric xerogel(with amino groups)was tested as well and the retention of Naphthol Green B was 15.9 mg·g-1[58].Both these values are comparable with the adsorption capacity on the unburned carbons received in this study — 25.5 and 15.2 mg·g-1.It is worth mentioning in this context that the parent coal adsorbed only 3 mg·g-1and that the increased adsorption on the unburned carbons corresponds preferentially with the(natural)release of the volatile combustibles out of the coal particles during the incomplete combustion at the power station(creating the porous texture).Higher adsorbed amount has been presented by Bezak-Mazuk and Adamczyk[59]for a commercial activated carbon(130 mg·g-1);however,the cost of such adsorbent is rather high.

4.2.Calculation of isotherm parameters by linearization

Due to its simplicity(and often also due to software availability),the linear regression is widely used for the evaluation of the non-linear models.In the case of Langmuir and Freundlich isotherm,the linearized models(Table 1)are generally known and easy to use.However,asR–P isotherm is a 3-parameter model,the trial-and-error method is to be used[54]because the linear equation still contains parameterA.The principle of this optimization is to plot

and to find the optimized value ofA(characterized by the highest correlation coefficient value).This linear equation(for the optimized parameterA)is then used for the calculation of the other two parameters of the R–P isotherm.In this study,due to its good availability,it has been done in an MS Excel spreadsheet.

The effect of parameterAon the correlation coefficient values in the dependence of

can be visually evaluated in Figs.6–8.The regression equations of the linearized R–P model for the given values of parameterAare documented in Figs.9–11.The regression equations with the highest correlation coefficients are highlighted.

Isotherm parameters calculated by linearized models(along with the values of the given statistic characteristics)are summarized in Table 3.

Fig.6.Linearized Redlich–Peterson isotherm model for UC1—shift of the determination coefficient R2in dependence of the actual value of parameter A.

Fig.7.Linearized Redlich–Peterson isotherm model for UC2—shift of the determination coefficient R2in dependence of the actual value of parameter A.

Fig.8.Linearized Redlich–Peterson isotherm model for the coal— shift of the determination coefficient R2in dependence of the actual value of parameter A.

Fig.9.Linearized Redlich–Peterson isotherm model for UC1— regression equations and the determination coefficients R2for selected values of parameter A.

The direct comparison of theR2values coming from the linear and the non-linear regression is quite problematic becauseR2values in the linear regression correspond with the closeness of the “l(fā)inearized”experimental data to the regression line and in the non-linear regression they relate to the fit of the original measured data to the isotherm(which is not a line).For this reason,the isotherm parameters calculated by the linear regression(in MS Excel)have been introduced into the statistical software(where all the non-linear regressions have been calculated)and theR2and AIC values were calculated directly for these input parameters using the original non-linear isotherm model(i.e.,without any other iteration or further approximation).SuchR2values were calculated in the same way as in the case of the nonlinear isotherm parameters;therefore,the mutual comparison is possible.

In the case of the Langmuir and Freundlich isotherm,the data summarized in Tables2and3suggest that the parameters and the statistical characteristics calculated by the non-linear and the linear regression are roughly comparable.In other words,in this case,the linearized models of the Langmuir and Freundlich isotherm provide relatively good results and can also be used for the description of the adsorption process(the only exception is the Langmuir linearization#2 if calculated for the coal).

Fig.10.Linearized Redlich–Peterson isotherm model for UC2—regression equations and the determination coefficients R2for selected values of parameter A.

Fig.11.Linearized Redlich–Peterson isotherm model for the coal— regression equations and the determination coefficients R2for selected values of parameter A.

Moreover,even linearization of the 3-parameter R–P model provides good results if calculated for the UCs(for which the R–P equation is a suitable model).In the case of the parent coal,the R–P linearization provides somewhat worse results,which is also the case of its original non-linear form,where the Langmuir model exhibited better results.Hence,the results indicate that the linearized R–P isotherm can be used in its linearized form with good results as well;at least in the cases where this(non-linear)model is suitable for the description of the adsorption process on a given adsorbent.

5.Conclusions

The study deals with the adsorption of Naphthol Green B on two unburned carbons and the parent coal,from which the UCs have been created in a fluidised-bed power station.Particular attention has been paid to the adsorption equilibrium modelling—experimental data has been analysed using 2-parameter(Langmuir,Freundlich)and 3-parameter(Redlich–Peterson)isotherms.As some recent adsorption studies indicated that linear regression might not always provide reliable results,both linear and non-linear regressions have been used for the estimation of the isotherm parameters.

In the case of both UCs,the Langmuir isotherm model provides the worst fit,whereas 2-parameter Freundlich and 3-parameter Redlich–Peterson models are both good,from which 3-parameter Redlich–Peterson isotherm provides slightly better results despite the penalty used for the higher number of parameters(applied in the Akaike's in formation criterion).The comparison of the linear regression of the Freundlich and Redlich–Peterson models with the non-linear calculation procedures documents that(at least in this case)the linearized models provide good results and can be used for the calculation of the isotherm parameters(e.g.,if the software needed for the non-linear calculations is not available).

Unlike both UCs,the best fit of the experimental data from the adsorption of Naphthol Green B on the coal has been achieved by the Langmuir isotherm model.The results based on the Freundlich or Redlich–Peterson model were(in this case)somewhat worse.

Table 3Parameters of linearized isotherm models(parameters 1–3 are the same as in Table 2)

[1]C.Fernández,M.S.Larrechi,M.P.Callao,An analytical overview of processes for removing organic dyes from wastewater effluents,Trac Trends Anal.Chem.29(10)(2010)1202–2011.

[2]T.Robinson,G.McMullan,R.Marchant,P.Nigam,Remediation of dyes in textile effluent:A critical review on current treatment technologies with a proposed alternative,Bioresour.Technol.77(3)(2001)247–255.

[3]E.Forgacs,T.Cserháti,R.Oros,Removal of synthetic dyes from wastewaters:A review,Environ.Int.30(7)(2004)953–971.

[4]R.M.Gong,Y.Z.Sun,J.Chen,H.J.Liu,C.Yang,Effect of chemical modification on dye adsorption capacity of peanut hull,Dyes Pigments67(3)(2005)175–181.

[5]Y.S.Ho,G.McKay,Kinetic models for the sorption of dye from aqueous solution by wood,Process Saf.Environ.76(B2)(1998)183–191.

[6]C.Palma,L.Lloret,A.Puen,M.Tobar,E.Contreras,Production of carbonaceous material from avocado peel for its application as alternative adsorbent for dyes removal,Chin.J.Chem.Eng.24(4)(2016)521–528.

[7]Y.S.Ho,T.H.Chiang,Y.M.Hsueh,Removal of basic dye from aqueous solution using tree fern as a biosorbent,Process Biochem.40(1)(2005)119–124.

[8]N.Sivarajasekar,R.Baskar,Biosorption of basic violet 10 onto activatedGossypium hirsutumseeds:Batch and fixed-bed column studies,Chin.J.Chem.Eng.23(10)(2015)1614–1619.

[9]Y.S.Ho,G.McKay,Sorption of dye from aqueous solution by peat,Chem.Eng.J.70(2)(1998)115–124.

[10]A.Hassani,L.Alidokht,A.R.Khatae,S.Karaca,Optimization of comparative removal of two structurally different basic dyes using coal as a low-cost and available adsorbent,J.Taiwan Inst.Chem.E45(4)(2014)1597–1607.

[11]A.Hassani,F.Vafaei,S.Karaca,A.R.Khataee,Adsorption of cationic dye from aqueous solution using Turkish lignite:Kinetic,isotherm,thermodynamic studies and neural network,J.Ind.Eng.Chem.20(4)(2014)2615–2624.

[12]N.Mirzaei,M.Hadi,M.Gholami,R.F.Fard,M.S.Aminabad,Sorption of acid dye by surfactant modi ficated natural zeolites,J.Taiwan Inst.Chem.E59(2016)186–194.

[13]D.Plachá,G.S.Martynková,J.Kukutschova,Sorpce par naftalenu na organicky modi fikovany vermikulit,Chem.List.105(3)(2011)186–192.

[14]P.Jano?,H.Buchtová,M.Ryznarová,Sorption of dyes from aqueous solutions onto lf y ash,Water Res.37(20)(2003)4938–4944.

[15]S.Wang,Q.Ma,Z.H.Zhu,Characteristic of coal fly ash and adsorption application,Fuel87(15–16)(2008)3469–3473.

[16]Z.Liu,Y.Liu,Structure andproperties of forming adsorbents prepared from different particle sizes of coal fly ash,Chin.J.Chem.Eng.23(1)(2015)290–295.

[17]Y.Gao,S.Xu,Q.Yue,Y.Wu,B.Gao,Chemical preparation of crab shell-based activated carbon with superior adsorption performance for dye removal from wastewater,J.Taiwan Inst.Chem.E61(2016)327–335.

[18]D.A.Giannakoudakis,G.Z.Kyzas,A.Avranas,N.K.Lazaridis,Multi-parametric adsorption effects of the reactive dye removal with commercial activated carbons,J.Mol.Liq.213(2016)381–389.

[19]R.Baccar,P.Blánquez,J.Bouzid,M.Feki,H.Attiya,M.Sarra,Modeling of adsorption isotherms and kinetics of a tannery dye onto an activated carbon prepared from an agricultural by-product,Fuel Process.Technol.106(2013)408–415.

[20]F.Scala,R.Chirone,A.Lancia,In-duct removal of mercury from coal- fired power plant flue gas by activated carbon:Assessment of entrained flow versus wall surface contributions,Environ.Eng.Sci.25(10)(2008)1423–1428.

[21]C.Senior,M.Denison,M.Bockelie,A.Saro fim,J.Siperstein,Q.He,Modelling of thermal desorption of Hg from activated carbon,Fuel Process.Technol.91(10)(2010)1282–1287.

[22]F.Scala,R.Chirone,A.Lancia,Elemental mercury vapor capture by powdered activated carbon in a fluidized bed reactor,Fuel90(6)(2011)2077–2082.

[23]M.Cabielles,M.A.Montes-Motán,A.B.Garcia,Structural study of graphite materials prepared by HTT of unburned carbon concentrates from coal combustion fly ashes,Energy Fuel22(2)(2008)1239–1243.

[24]T.Navrátil,Z.?enholdová,K.Shanmugam,J.Barek,Voltametric determination of phenylglyoxylic acid in urine using graphite composite electrode,Electroanalysis18(2)(2006)201–206.

[25]M.Cabielles,J.-N.Rouzaud,A.B.Garcia,High-resolution transmission electron microscopy studies of graphite materials prepared by high-temperature treatment of unburned carbon concentrates from combustion fly ashes,Energy Fuel23(2)(2009)942–950.

[26]L.Bartoňová,Unburnedcarbonfrom coal combustion ash:Anoverview,FuelProcess.Technol.134(2015)136–158.

[27]N.J.Wagner,R.H.Matjie,J.H.Slaghuis,J.H.P.van Heerden,Characterization of unburned carbon present in coarse gasification ash,Fuel87(6)(2008)683–691.

[28]N.Malumbazo,N.J.Wagner,J.R.Bunt,D.van Niekerk,H.Assumption,Structural analysis of chars generated from South African inertinite coals in a pipe-reactor combustion unit,Fuel Process.Technol.92(4)(2011)743–749.

[29]V.P.Chabalala,N.J.Wagner,S.Potgieter-Vermaak,Investigation into the evolution of char structure using Raman spectroscopy in conjunction with coal petrography:Part 1,Fuel Process.Technol.92(4)(2011)750–756.

[30]L.Bartoňová,Z.Klika,D.A.Spears,Characterization of unburned carbon from ash after bituminous coal and lignite combustion in CFBs,Fuel86(3)(2007)455–463.

[31]J.C.Hower,C.L.Senior,E.M.Suuberg,R.H.Hurt,J.L.Wilcox,E.S.Olson,Mercury capture by native fly ash carbons in coal- fired power plants,Prog.Energy Combust.36(4)(2010)510–529.

[32]L.Bartoňová,B.?ech,L.Ruppenthalová,V.Majvelderová,D.Juchelková,Z.Klika,Effect of unburned carbon content in fly ash on the retention of 12 elements out of coal-combustion flue gas,J.Environ.Sci.24(9)(2012)1624–1629.

[33]S.Wang,L.Li,H.Wu,Z.H.Zhu,Unburned carbon as a low-cost adsorbent for treatment of methylene blue-containing wastewater,J.Colloid Interface Sci.292(2)(2005)336–343.

[34]S.Wang,H.Li,Dye adsorption on unburned carbon:Kinetics and equilibrium,J.Hazard.Mater.126(1–3)(2005)71–77.

[35]F.Montagnaro,L.Santoro,Reuse of coal combustion ashes as dyes and heavy metal adsorbents:Effect of sieving and demineralization on waste properties and adsorption capacity,Chem.Eng.J.150(1)(2009)174–180.

[36]F.C.Wu,P.H.Wu,R.L.Tseng,R.S.Juang,Preparation of activated carbons from unburnt coal in bottom ash with KOH activation for liquid-phase adsorption,J.Environ.Manag.91(5)(2010)1097–1102.

[37]S.Wang,H.Li,Kinetic modelling and mechanism of dye adsorption on unburned carbon,Dyes Pigments72(3)(2007)308–314.

[38]J.C.Hower,J.D.Robertson,Chemistry and petrology of fly ash derived from the co-combustion of western United States coal and tire-derived fuel,Fuel Process.Technol.85(5)(2004)359–377.

[39]H.Raclavská,D.Juchelková,V.Roubí?ek,D.Matysek,Energy utilization of biowaste—Sun flower-seed hulls for co- firing with coal,Fuel Process.Technol.92(1)(2011)13–20.

[40]D.O.Glushkov,N.E.Schlegel,P.A.Strizhak,K.Y.Vershinina,Heat transfer under ignition of droplet of composite liquid fuel made of coal,water and oil in an oxidant flow,Adv.Appl.Fluid Mech.19(1)(2016)157–168.

[41]F.Scala,R.Chirone,Fluidised bed combustion of alternative solid fuels,Exp.Thermal Fluid Sci.28(7)(2004)691–699.

[42]H.Raclavská,D.Juchelková,H.?krobánková,T.Wiltowski,A.Campen,Conditionsfor energy generation as an alternative approach to compost utilization,Environ.Technol.32(4)(2011)407–417.

[43]D.O.Glushkov,P.A.Strizhak,K.Y.Vershinina,Minimum temperatures for sustainable ignition of coal water slurry containing petrochemicals,Appl.Therm.Eng.96(2016)534–546.

[44]D.Juchelková,A.Corsaro,A.Hlavsová,H.Raclavská,Effect of composting on the production of syngas during pyrolysis of perennial grasses,Fuel154(2015)380–390.

[45]K.V.Kumar,Optimum sorption isotherm by linear and non-linear methods for malachite green lemon peel,Dyes Pigments74(3)(2007)595–597.

[46]M.I.El-Khaiary,Least-squares regression of adsorption equilibrium data:Comparing the options,J.Hazard.Mater.158(1)(2008)73–87.

[47]Y.S.Ho,Selection of optimum sorption isotherm,Carbon42(10)(2004)2115–2116.

[48]K.V.Kumar,S.Sivanesan,Prediction of optimum sorption isotherm:Comparison of linear and non-linear method,J.Hazard.Mater.B126(2005)198–201.

[49]K.Y.Foo,B.H.Hameed,Insights into the modeling of adsorption isotherm systems,Chem.Eng.J.156(1)(2010)2–10.

[50]I.Langmuir,The adsorption of gases on plane surface of glass,mica and platinum,J.Am.Chem.Soc.40(1916)1361–1368.

[51]Y.S.Ho,J.F.Porter,G.McKay,Equilibrium isotherm studies for the sorption of divalent metal ions onto peat:Copper,nickel and lead single component systems,Water Air Soil Pollut.141(1–4)(2002)1–33.

[52]H.M.F.Freundlich,Over the adsorption in solution,J.Phys.Chem.57(1906)385–470.

[53]O.Redlich,D.L.Peterson,A useful adsorption isotherm,J.Phys.Chem.63(1959)1024–1026.

[54]G.McKay,S.J.Allen,I.F.McConvey,The adsorption of dyes from solution—Equilibrium and column studies,Water Air Soil Pollut.21(1–4)(1984)127–129.

[55]M.I.El-Khaiary,G.F.Malash,Common data analysis errors in batch adsorption studies,Hydro metallurgy105(2011)314–320.57.

[56]F.-C.Wu,B.-L.Liu,K.-T.Wu,R.-L.Tseng,A new linear form analysis of Redlich–Peterson isotherm equation for the adsorptions of dyes,Chem.Eng.J.162(2010)21–27.

[57]M.F.Attallah,I.M.Ahmed,M.M.Hamed,Treatment of industrial wastewater containing Congo Red and Naphthol Green B using low-cost adsorbent,Environ.Sci.Pollut.Res.20(2013)1106–1116.

[58]Y.Cheng,Q.Feng,X.Ren,M.Yin,Y.Zhou,Z.Xue,Adsorption and removal of sulfonic dyes from aqueous solution onto a coordination polymeric xerogel with amino groups,Colloids Surf.A Physicochem.Eng.Asp.485(2015)125–135.

[59]E.Bezak-Mazur,D.Adamczyk,Changes in the chemistry of wd-extra activated carbon surface after Fenton's reagent regeneration used for adsorption of Naphthol Green B,Rocz.Ochrona Srodowiska15(1)(2013)966–980.