Environ Monit Assess (2011) 182:71–84DOI 10.1007/s10661-010-1859-8

Determination of aluminium in groundwater samplesby GF-AAS, ICP-AES, ICP-MS and modellingof inorganic aluminium complexes

Marcin Frankowski · Anetta Zioła-Frankowska ·Iwona Kurzyca · Karel Novotný · Tomas Vaculovic ·Viktor Kanický · Marcin Siepak · Jerzy Siepak

Received: 23 June 2010 / Accepted: 19 December 2011 / Published online: 20 January 2011© The Author(s) 2011. This article is published with open access at Springerlink.com

Abstract The paper presents the results of alu-minium determinations in ground water samplesof the Miocene aquifer from the area of the cityof Poznan (Poland). The determined aluminiumcontent amounted from <0.0001 to 752.7 μg L−1.The aluminium determinations were performedusing three analytical techniques: graphite furnaceatomic absorption spectrometry (GF-AAS), in-ductively coupled plasma atomic emission spec-trometry (ICP-AES) and inductively coupledplasma mass spectrometry (ICP-MS). The resultsof aluminium determinations in groundwater sam-ples for particular analytical techniques were com-pared. The results were used to identify the ascent

M. Frankowski (B) · A. Zioła-Frankowska ·I. Kurzyca · J. SiepakDepartment of Water and Soil Analysis,Faculty of Chemistry, Adam Mickiewicz University,Drzymały 24, 60-613 Poznan, Polande-mail: Marcin.Frankowski@amu.edu.pl

K. Novotný · T. Vaculovic · V. KanickýDepartment of Chemistry, Faculty of Science,Masaryk University, Kotlárská 2,611 37 Brno, Czech Republic

M. SiepakDepartment of Hydrogeology and Water Protection,Institute of Geology, Adam Mickiewicz University,Maków Polnych 16, 61-606 Poznan, Poland

of ground water from the Mesozoic aquifer to theMiocene aquifer in the area of the fault graben.Using the Mineql+ program, the modelling of theoccurrence of aluminium and the following alu-minium complexes: hydroxy, with fluorides andsulphates was performed. The paper presents theresults of aluminium determinations in groundwater using different analytical techniques as wellas the chemical modelling in the Mineql+ pro-gram, which was performed for the first time andwhich enabled the identification of aluminiumcomplexes in the investigated samples. The studyconfirms the occurrence of aluminium hydroxycomplexes and aluminium fluoride complexes inthe analysed groundwater samples. Despite thedominance of sulphates and organic matter inthe sample, major participation of the complexeswith these ligands was not stated based on themodelling.

Keywords Aluminium · Groundwater ·Miocene aquifer · Aluminium complexes ·Modelling · AAS · ICP

Introduction

Aluminium in ground water occurs at the con-centration levels of 60 to 300 μg/L (Kabata-Pendias and Pendias 1999). However, it shouldbe underlined that such low concentrations of

72 Environ Monit Assess (2011) 182:71–84

aluminium occur in unchangeable conditions in-cluding temperature, pH reaction, redox potentialand the level of pollutant concentration. Due tothe aluminium properties, including its ampho-terism and the ability to transform, the changeto any of the components affects the change ofaluminium concentrations level, often by severalranks, which may result in a negative impacton the fauna and flora. The low concentrations ofaluminium in groundwater are connected with thetransformations of aluminosilicates in the activewater exchange zone, where a significant amountof aluminium stays immobile in the structuresof secondary minerals, and only a small part ofthe elements gets to the water. The processesinfluencing the aluminium activation in ground-water include the release of aluminium fromminerals as a result of hydrolysis and dissolu-tion, the reactions of cation exchange, the trans-fer of retained aluminium from pore water togroundwater in the soluble form, and the re-actions of aluminium with other components ofthe solution (Macioszczyk and Dobrzynski 2002).The source of aluminium in groundwater aresoluble complex bonds with dissolved fluoride(AlF2+, AlF+

2 , AlF03, AlF−

4 ), sulphate (AlSO+4 ,

Al(SO4)2−), phosphate (AlHPO2+

4 , AlHPO+4 ) lig-

ands and with low-molecular organic acids. An-other source of aluminium is the so-calledexchangeable aluminium fraction consisting ofAl3+ions and hydroxide complexes: AlOH2+,Al(OH)+

2 , Al(OH)−4 adsorbed on the negatively

charged surfaces of silt mineral particles and hu-mus substances in soils and wastes. The formof aluminium occurring as complexes mainly de-pends on the reaction pH, temperature, the con-centration of organic and inorganic ligands suchas: dissolved organic carbon, fluorides, sulphates,phosphates and suspended particles (Šcancarand Milacic 2006; Frankowski et al. 2010a, b;Frankowski and Zioła-Frankowska 2010). In thewater marked by neutral reaction, aluminiumdominates in the form of hydroxide complexesand organic complexes, while in lower concen-trations it occurs in the form of fluoride andsulphate complexes (Šcancar and Milacic 2006;Zioła-Frankowska et al. 2009; Frankowski et al.2009).

It should be underlined that the investigatedwater of the Miocene aquifer plays an impor-tant complementary role in supplying people withdrinking water as well as with water used inhouseholds and industry. As a result of long-termpumping, a significant lowering of the table of theMiocene aquifer groundwater took place and thenatural flow of this water was disturbed, whichcaused changes in hydrochemical zonality in thewater-bearing layers. The ascent of saline waterfrom the Mesozoic aquifer was observed in thestudy area, and the dislocation of the colouredwater front within the Miocene aquifer was stated.The Miocene in the study area is formed bysands and brown coals which change upwardsinto the mud/silt and silt formations (Górski andPrzybyłek 1996). The groundwater in the Quater-nary and Miocene aquifer is used in householdsand industry (Przybyłek 1986; Dabrowski et al.2007). The water at the Miocene aquifer occursin the study area in the layers of fine-grained andmuddy sands, which locally turn into medium-grained, of the thickness of about several to 80 m.The layers are separated by the discontinuouslayer of muds and brown coals. The discussedlevel occurs at the depth of 50 to 150 m below theground level.

By forming complexes with aluminium, ligandssuch as humic acids, fulvic acids, fluorides, phos-phates and silicates decrease its toxicity, whichdepends on the form of aluminium occurrence inthe environment and only appears in the case ofhydroxy complexes, in particular Al(OH)+

2 . Inor-ganic complexes of aluminium are more toxic thanits organic complexes. The toxicity of aluminiumusually decreases in the following order: polymerAl13, Al3+, Al(OH)2+, Al(OH)+

2 , Al(OH)−4 and

AlSO+4 . In connection with fluorides, aluminium

and its organic complexes, as well as Al(OH)3 areconsidered harmless, while its labile forms such asAl3+,AlOH2+, Al(OH)+

2 penetrate into the foodchain and may be toxic for all living organisms(Boudot et al. 1994; Driscoll and Schecher 1990).

The analysis of aluminium in different com-ponents of the environment is a difficult taskdue to the physicochemical properties of this el-ement. Aluminium determinations can then beperformed by numerous methods, such as by

Environ Monit Assess (2011) 182:71–84 73

Inductively Coupled Plasma Mass Spectrometry(ICP-MS), Graphite Furnace Atomic AbsorptionSpectrometry (GF-AAS) or Inductively CoupledPlasma Atomic Emission Spectrometry (ICP-AES). GF-AAS has successfully been used todetermine trace aluminium levels natural waters,seawater, river, soil and water (Danielsson andSparén 1995; Mitrovic et al. 1998; Salomon et al.2000; Narin et al. 2004; Rezaee et al. 2010).Also ICP-MS method has been used in deter-mination of aluminium in aqueous medium (wa-ter, lake water, limed lake, seawater, spring andforest soil waters; Bérubé and Brûlé 1999; Xiaet al. 2005; Préndez and Carrasco 2003; Sjöstedtet al. 2009; Hills et al. 1999; Fairman et al. 1998;Rodushkin and Ruth 1997) and ICP-OES in de-termination aluminium in water samples (Rezaeeet al. 2010) drinking water (Rossiter et al. 2010)or in boreal streams (Cory et al. 2009). Thedetermination of aluminium by the ICP-MS tech-nique is limited by the occurrence of interfer-ences caused by other elements in the sample(matrix effects), due to which the aluminiumdetermination using this technique is not fullyjustified. Therefore, the aluminium determina-tions and speciation analysis are commonly per-formed using the technique of graphite furnaceatomic absorption spectrometry. Moreover, thealuminium speciation schemes are applied. Theschemes are based on the speed of reaction witha complexing agent, the use of ionoselective elec-trode for the determination of fluorides, the ionexchange on ion-exchangeable resins and other(Bloom and Erich 1996).

The aims of the study were: (1) the comparisonof analytical techniques (AAS and ICP) in alu-minium determinations in groundwater samplesof the Miocene aquifer, (2) defining the variabilityof aluminium occurrence in groundwater and thefactors affecting the concentration of aluminiumin this water, (3) applying the results of aluminiumdeterminations in the identification of water as-cent from the Mesozoic to the Miocene aquiferin the region of the fault graben, (4) modellingthe occurrence of aluminium and aluminium com-plexes in the form of labile inorganic hydroxy,fluoride and sulphate complexes in the groundwa-ter samples of the Miocene aquifer.

Materials and methods

Sample collection

The groundwater samples of the Miocene aquiferin the area of the city of Poznan were collected forthe physicochemical analysis in the year 2006 fromthe exploited drilled wells (Fig. 1).

The non treated water for the analysis wascollected after stabilising such parameters as:temperature, pH reaction, electrolytic conductiv-ity. The measurements of temperature, pH re-action and electrolytic conductivity were takenusing a multi-function Multi 197i meter pro-duced by WTW (Weilheim, Germany), equippedwith the following electrodes: pH-ElectrodeSenTix 41 and TetraCon® 325 produced by WTW(Weilheim, Germany). The measurement elec-trodes were placed in the flow-through cellproduced by Eijkelkamp Agrisearch EquipmentBV (Giesbeek, The Netherlands). The sampleswere collected and transported anaerobically inNalgene® polyethylene bottles.

Groundwater conditions in the study area

Poznan is a city in west-central Poland with over550,000 inhabitants. Located by the Warta River,it is one of the oldest Polish Cities, an importanthistorical center and capital of the WielkopolskaDistrict. There is a storeyed arrangement ofgroundwaters in the Poznan area within its bound-aries, with the following aquifers (from top to bot-tom): the shallow aquifer, the Upper Quaternaryaquifer (sandwiched between the glacial tills),the Wielkopolska Buried Valley aquifer (or theMiddle Quaternary aquifer between the glacialtills), the Miocene aquifer with its Upper, Middleand Lower Zones), and the Oligocene aquifer(Dabrowski et al. 2007; Przybyłek 1986; Górski1989; Górski and Przybyłek 1996). Groundwaterin the Miocene aquifer in the study area is storedin layers of fine-grained and muddy sands whichlocally turn into medium-grained. The thicknessof sandy layers is changeable, from several to80 m. They are interlayered with the muddyseams and brown coals of discontinuous character.The Poznan Miocene aquifer contains confined

74 Environ Monit Assess (2011) 182:71–84

Fig. 1 Hydroizohypse map of ground water of Miocene and location of the sampling sites

groundwater—which is under subartesian condi-tions within the upland areas, and under the arte-sian conditions within the Warta River Valley.The confining bed is composed of an assembly ofclay and muddy-clayey sediments of the PoznanClay, Upper Miocene in age and variable in termsof thickness. In the study period (May–June,2006), the water level in the study area was mea-sured at the altitude of 54 to 75 m a.s.l. and showeda distinctly shaped cone of depression within theright-bank side of Poznan. The cone developeddue to the extensive extraction of groundwater bythe industrial plants.

Analytical techniques in determination of Al

Four atomic absorption spectrometers with differ-ent types of atomization were used in the deter-mination of aluminium in groundwater samples:GF-AAS produced by Perkin Elmer AAnalyst300 (Norwalk, Connecticut, USA), GF-AAS pro-duced by Varian (Mulgrave, Victoria, Australia),

ICP-MS spectrometer Agilent 7500 CE producedby Agilent (Agilent, Japan) and ICP-OES spec-trometer Jobin Yvon 170Ultrace (Jobin Yvon,France). The Spectra-20 Plus spectrometer isequipped with a GTA-96 electrothermal atom-izer, which enables the temperature programmingwithin the range from 40 to 3,000◦C, ±1◦C, andthe speed of 2,000◦C/s. In order to inject thesamples to the graphite furnace, the PSC-56 auto-matic sample injector was used. The AAnalyst 300spectrometer is equipped with an electrothermalatomizer. The atomizer has the function of pro-gramming the temperature within the range from20 to 3,000◦C, ±1◦C. In order to inject the samplesto the graphite furnace, the AS-72 automatic sam-ple injector produced by Perkin Elmer was used.During the analysis, graphite furnaces with a Lvovplatform, covered with pyrolytic graphite andHCL lamps were used. Particular phases of thetemperature scheme of both AA spectrometerswere subjected to optimization. The performeddeterminations in groundwater samples did not

Environ Monit Assess (2011) 182:71–84 75

Table 1 Conditions of Aldeterminations byGF-AAS technique

aSpectra-20 PlusspectrometerbAAnalyst 300spectrometer(explanation of 15 + (30)mean that 15 [s] it is ramptime and 30 [s] it is holdtime)

Parameter GF-AASa GF-AASb

Wavelength [nm] 309.3Slit [nm] 0.5Lamp current [mA] 6 25Argon flow [L min−1] 3.0 0.25Drying [0C] 120 300Drying time [s] 40 15 + (30)Ashing [0C] 1000 1350Ashing time [s] 2 1 + (30)Atomization [0C] 2600 2700Atomization time [s] 2 0 + (5)Cleaning [0C] 2700 2700Clearing time [s] 2 5Number of replications – 3

require the use of modifiers. The appropriatelyselected parameters of the GF-AAS spectrometerfor the aluminium determinations have been listedin Table 1.

Quadrupole ICP-MS spectrometer Agilent7500 CE (Agilent, Japan) is equipped by octopolereaction cell for removing of isobaric interfer-ences. The solution was nebulized into double-pass Scott chamber using Babington nebulizer.Spray chamber was cooled to 2◦C. Sample flowrate was 0.4 ml L−1. ICP-MS conditions (summa-rized in Table 2) were optimized with respect tomaximum S/N ratio and minimum oxide forma-tion. Integration time for isotope 27 measurementwas set to 0.1 s. ICP-AES spectrometer JobinYvon, 170Ultrace (Jobin Yvon, France) with lat-erally viewed plasma was used. Water sampleswere nebulized using Meinhard concentric nebu-lizer. The basic parameters of the ICP-AES and

ICP-MS spectrometers have been presented inTable 2.

Determinations of F− and SO2−4

Concentrations of F− and SO2−4 were determined

by Dionex DX-120 Ion Chromatograph equippedwith a CDM-3 conductivity detector, 25 μl injec-tion loop, and an AS14 analytical column (4 ×250 mm), AG14 guard column (4 × 50 mm), andASRS-II 4-mm suppressor. The applied proce-dures made it possible to obtain reliable resultswith a good precision (RSD < 4%) and accuracy(RSD < 6%), which was confirmed by analysesof the certified reference material (EnviroMAT™Ground Water, ES H 1; SCP SCIENCE, Canada)with a matrix corresponding to the samples un-der study. The recoveries were 102% for fluorideand 105% for sulphate, whereas method detection

Table 2 Conditions of Aldeterminations by theICP-AES spectrometer(Jobin Yvon) andICP-MS spectrometer(Agilent)

Parameter ICP-AES ICP-MS

Wavelength [nm] 308.215 –396.152

Isotope – 27RF power [W] 1200 1500Argon flow rates

–Plasma gas [l min−1] 12 15–Sheath gas 0.4 0.21–Carrier gas 0.6 0.84

Sample flow rate [ml min−1] 1.0 0.4Rinsing time [s] 30 30Replicates 3 5

76 Environ Monit Assess (2011) 182:71–84

Fig. 2 Results obtainedfor determination ofaluminium ingroundwater samples byGF-AAS (Varian, PerkinElmer), ICP-AES andICP-MS techniques

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39sample

0

100

200

300

400

500

600

700

800

Al [

ug

L-1

]

GF-AAS VARIAN GF-AAS PERKIN ICP-AES ICP-MS

limits were 0.01 mg L−1 for F− and 0.01 mg L−1 forSO2−

4 .

Reagents

The reagents used were analytically pure and wa-ter was deionized to a resistivity of >18 M� in aMilli Q-RG apparatus (Millipore, France). Duringthe determinations by the IC, GF-AAS, ICP-AESand ICP-MS techniques, standard solutions pro-duced by Merck (Merck, Darmstadt, Germany),Fluka (Sigma-Aldrich, Steinheim, Switzerland),Astasol (Analytika Praha, Czech Republic) andthe Czech Metrology Institute (CMI, Brno, CzechRepublic) were used. The mobile phase for thedetermination of anions by the IC technique wasprepared using the solutions produced by Merck(Merck, Darmstadt, Germany).

Results and discussion

Statistical analysis—comparison of analyticaltechniques

In order to compare the analytical techniques interms of the results obtained for aluminium in thegroundwater samples from the Miocene aquifer,statistical tests were used. For the determinationsof aluminium (n = 39) performed using the threeanalytical techniques (GF-AAS—a Varian appa-ratus, GF-AAS—a Perkin Elmer apparatus, ICP-MS and ICP-AES), the W Shapiro–Wilk test wasused. The test p values constituted p < p of thesignificance level p > p = 0.05, which indicates

that the samples do not originate from the pop-ulation with a normal distribution. The nonpara-metric Kruskal–Wallis test was then performed.The test H value amounted to 5.505 at the sig-nificance level p = 0.138. p > p = 0.05, whichindicates the equal distribution of the comparedanalytical techniques. The results obtained in theKruskal–Wallis test showed the lack of differencesbetween the analytical techniques for the set of39 aluminium determinations. The nonparamet-ric Kolmogorov–Smirnov test for two variablesin all possible combinations between the tech-niques was performed. The results obtained forparticular groups (for n = 39 in three analyticaltechniques) amounted to p > 0.1 for the groups ofGF-AAS techniques and the ICP-AES technique.At the significance level α = 0.05, there are no

GFAAS* GFAAS** ICP-AES ICP-MS0

100

200

300

400

500

600

700

800

Al [

ug L

-1]

Median 25%-75% Min-Max

Fig. 3 The values of median, upper and lower quartile andconcentration range of aluminium obtained in ground wa-ter samples of Miocene by analytical techniques (*VarianSpectraAA, **Perkin Elmer AAnalyst 300)

Environ Monit Assess (2011) 182:71–84 77

grounds to reject the mean equality hypothesisin the investigated populations. For the ICP-MSand ICP-AES techniques, the value of p > 0.1in the Kolmogorov–Smirnov test was obtained. Itshould be underlined that the ICP and GF-AAStechniques had different injection systems whichmay influence the differences in results betweenthese techniques. From the point of view of theenvironmental analytics, the results obtained foraluminium using different analytical techniquesshow too wide divergence, especially the resultsobtained using the ICP-MS and ICP-AES tech-niques. Despite the similar median values for allthe analytical techniques for a certain number ofsamples, the minimum and maximum values inthe ICP-MS technique significantly differ fromthe values obtained by the GF-AAS technique.Figure 2 presents the results of aluminium con-centration analysis (GF-AAS, ICP) for the watercollected from the wells at the Miocene aquifer.Figure 3 presents the basic statistical parametersfor the studied group of samples (n = 39).

Based on the obtained study results, it maybe stated that the trend of aluminium occur-rence in particular samples is preserved (Fig. 2).The similar values between the two spectrometerswere stated for GF-AAS. In order to graphicallypresent the results of the analysis of samples col-lected at the Miocene aquifer, the map of concen-tration with the distribution network dependingon the concentration were made using the Surfer8 program (Golden Software Inc. USA) for GF-AAS (Perkin Elmer) analytical techniques as anexample (Fig. 4)

Based on the results of the analysis presentedgraphically on the map (Fig. 4), it may be statedthat the results obtained for aluminium by meansof different techniques are marked by the highestconcentrations at points No. 7–10, 17, 23. At thesepoints, the highest concentrations of chlorides inthe Miocene aquifer groundwater were also stated(Siepak et al. 2006). This proves that, along withthe saline water in the area of the influence ofthe tectonic structure, aluminium occurs in higher

Fig. 4 Concentrationmap of aluminium[μg·L−1] determined byGF-AAS (Perkin Elmer)

78 Environ Monit Assess (2011) 182:71–84

concentrations than in the other parts of the cityof Poznan. The higher concentrations in the areaof the tectonic fault graben may result from theascent of water from the Mesozoic aquifer inthe zone of hydraulically active tectonic faultscaused by the pumping of water from the Mioceneaquifer. At the same time, considering the factthat the water of this aquifer is isolated fromthe surface by a layer of boulder clays and siltsof the Upper Miocene of the Poznan series, the

pollution of this water with anthropogenic factorsmay be excluded. A greater threat is posed to thegroundwater of the Miocene aquifer by extensivepumping of water taking place in the eastern partof the city, which has been graphically presentedon the hydroizohips map (Fig. 1). Moreover, theintensive exploitation may cause the dislocationof the front of coloured water and the ascent ofwater marked by higher mineralization from theMesozoic aquifer.

Table 3 The values ofparameters used formodeling of aluminiumforms in ground watersamples (*A—upper,B—middle, C—downlevel of Miocene)

Sample Water-bearing Temperature pH Altot F− SO2−4

horizon of Miocene* [◦C] [μg L−1] [mg L−1] [mg L−1]

1 C 13.4 6.60 37.79 0.46 10.992 B 12.3 7.23 29.73 0.48 7.353 A 12.1 6.90 39.76 0.53 5.764 B 12.3 6.94 60.68 0.57 4.875 C 11.9 6.61 67.14 0.67 0.096 C 13.9 6.88 106.6 0.76 0.187 C 13.2 7.16 97.47 0.23 112.18 C 11.1 8.08 91.66 0.82 0.289 C 13.1 6.85 123.5 0.68 1.2110 C 13.2 6.92 196.9 0.73 4.011 B 12.2 7.18 15.77 0.39 0.4912 C 13.1 7.08 25.42 0.57 0.8113 A 13.2 7.00 10.9 0.31 7.0714 A 11.8 6.8 70.27 0.24 36.115 B 11.5 6.96 32.93 0.30 27.416 B 10.9 6.68 43.08 0.51 2.3317 C 15.7 6.66 154.7 0.56 5.518 C 13.7 7.08 64.51 0.85 0.3319 A 13.5 6.6 65.24 0.54 0.3320 B 13.1 7.07 23.29 0.59 1.7221 A 12.1 6.86 31.38 0.66 0.0322 A 12.4 6.76 5.876 0.62 0.1123 A 12.0 6.75 195.8 0.66 0.0724 A 13.3 6.62 21.8 0.67 0.125 A 12.3 6.77 22.71 0.58 0.2326 A 12.9 6.73 48.95 0.13 32827 A 13.1 6.54 10.43 0.63 0.0828 A 13.2 6.70 25.1 0.66 0.3929 A 10.8 6.61 3.71 0.69 0.0430 B 14.3 6.79 31.23 0.59 1.3631 B 13.9 6.67 15.12 0.58 0.5232 B 14.35 6.84 9.376 0.58 2.1133 B 14.4 6.81 12.37 0.57 3.6734 B 12.3 6.82 15.12 0.68 1.9735 B 13.0 6.88 15.50 0.69 2.7236 B 12.4 6.28 8.691 0.67 2.6837 B 12.4 6.20 15.31 0.62 3.0938 A 12.3 6.50 21.85 0.49 0.8239 A 12.5 6.71 21.78 0.52 0.04

Environ Monit Assess (2011) 182:71–84 79

Aluminium determination and the comparisonof analytical techniques

For all the analytical techniques applied inaluminium determinations, the analysis of thecertified reference material Enviromat GroundWater ES-H-1 was performed and the recoveryof 98–105% for all the techniques was obtained.The aluminium concentration in the Miocene

aquifer occurred on the level from <0.0001 to752.7 μg·L−1 (Fig. 3).

Modelling of aluminium complexes

The pH reaction of water at the Miocene aquiferamounted to 6.60–8.08, which may be the evidenceof the occurrence of mainly hydroxy aluminiumcomplexes. The modelling was applied to the

Table 4 The results of calculated forms of aluminium complexes in [%] from Mineql+ program (value 0 means complexesin molar concentration of <E-10)

Sample Al3+ Al(OH)+2 bf Al(OH)3 Al(OH)−

4 AlOH2+ AlF+2 AlF3 AlF+2 AlF−

4 AlSO+4 Al(SO4)−

2

1 0 24.8 19.8 5.9 0 34.9 9.9 4.1 0 0 02 0 12.9 43.2 42.2 0 1.2 0 0 0 0 03 0 25.2 39.7 17.2 0 12.3 4.1 1.3 0 0 04 0 21.9 37.9 18.9 0 14 5.9 1.2 0 0 05 0 17.4 14.1 3 0 44 17.2 3.8 0 0 06 0 19.6 29.7 19.3 0 20.3 9.3 1.5 0 0 07 0 14.7 40.6 44 0 0 0 0 0 0 08 0 0 16.5 82.8 0 0 0 0 0 0 09 0 22.4 31.6 15.6 0 20.2 8.2 1.7 0 0 010 0 21 34.8 20.8 0 15.4 6.5 1.2 0 0 011 0 15 45.3 38 0 1.2 0 0 0 0 012 0 17.2 41.3 34.7 0 4.5 1.7 0 0 0 013 0 21.9 43.3 31.3 0 2.4 0 0 0 0 014 0 35.1 43.2 14 0 5.3 0 1.3 0 0 015 0 26.9 47.9 20.7 0 3.2 0 0 0 0 016 0 26.4 25.1 4.8 0 30.4 9.5 3.4 0 0 017 0 23.6 21.4 13.3 0 29.2 8.7 3.3 0 0 018 0 15 35.9 35.1 0 8.6 4.7 0 0 0 019 0 21.9 17.4 5.4 0 38.6 12.2 4.1 0 0 020 0 17.5 41.1 33.7 0 5.1 2 0 0 0 021 0 22.7 32.8 12.8 0 20.9 8.8 1.7 0 0 022 0 22 25.3 8.5 0 29.5 11.8 2.5 0 0 023 0 24.7 27.8 8.2 0 26.8 9.5 2.6 0 0 024 0 15.6 13 4 0 44.5 18.8 3.6 0 0 025 0 24 28.3 9.5 0 26 9.6 2.4 0 0 026 0 40.3 39.8 15.4 0 2.3 0 1.3 0 0 027 0 13.4 9.3 2.2 0 50.1 20.2 4.2 0 0 028 0 19.1 19.1 6.8 0 36.4 15.1 3 0 0 029 0 14.4 11.8 1.9 0 46.9 20.9 3.6 0 0 030 0 22.4 27.5 16.1 0 23 8.6 2.1 0 0 031 0 20.5 19.1 7.6 0 35.8 13.2 3.3 0 0 032 0 22.1 30.4 20.5 0 18.2 6.8 1.7 0 0 033 0 22.6 29 18.3 0 20.4 7.4 1.9 0 0 034 0 21.6 28.4 10.7 0 25.8 11.2 2 0 0 035 0 21.2 31.8 16.5 0 19.9 8.8 1.5 0 0 036 0 4.5 1.7 0 0 61.8 26.2 5 0 0 037 0 3.9 1.2 0 0 63.9 24.6 5.7 0 0 038 0 18.6 11.7 2.1 0 47.4 14.4 5.3 0 0 039 0 25 25.7 7.9 0 28.6 9.4 2.9 0 0 0

80 Environ Monit Assess (2011) 182:71–84

aluminium complexes combined with fluoridesand sulphates. The concentration of phosphateswas so low and uncompetitive in relation to F−and SO2−

4 that these ligands could be omitted inthe modelling. The determination results obtainedusing the Perkin Elmer spectrometer (GF-AAS)were used. A simulation of aluminium formsbased on the parameters presented in Table 3was performed for all the groundwater samples.The ionic strength value was calculated basedon the introduced concentration values of alu-minium, fluorides and sulphates. The real tem-perature conditions of groundwater samples wereadopted for the modelling.

For the groundwater samples of the Mioceneaquifer, the modelling of aluminium forms whichmay occur in this environment was performed.In terms of aluminium toxicity, the values of thiselement obtained for groundwater are markedby low concentration. Hence, it may be assumedthat they constitute the geochemical background.The fluorides, sulphates or phosphates present inthe water can form complexes with aluminiummainly in the conditions of low pH reaction. How-ever, taking into consideration the results ob-tained for the fluorides which prevail over thealuminium, aluminium complexes with fluoridesmay be highly competitive for the hydroxyalu-minium forms. The results obtained based onthe modelling have been presented in Table 4.The analysis of data obtained from the modellingconfirms the lack of aluminium complexes with

sulphates in the whole spectrum of samples. Evenin sample No. 26, where the sulphate concentra-tion amounted to 328 mg L−1, there are no condi-tions to form aluminium complexes (pH = 6.73).It should be underlined that the pH conditionsin the analysed water are not advantageous forthe formation of complexes with inorganic andorganic ligands. In the case of aluminium com-plexes with humuses, the samples with the highcontent of total organic carbon were subjectedto modelling. No significant aluminium complexeswith humuses were stated (Al-humuses valuesat E−10 M). The complexes of aluminium withfluorides are more competitive. The formation ofaluminium complexes with AlF+

2 , AlF3, AlF2+wasstated. For the majority of samples, the hydroxya-luminium complexes dominated.

The graph of the pH reaction and the mo-lar proportion Al:F (based on aluminium resultsobtained by GF-AAS Perkin Elmer analyticaltechnique) drawn to illustrate the dependencesresulting from the formation of aluminium com-plexes (Fig. 5).

Based on the obtained results, it may be statedthat for the groundwater samples, the dominat-ing complexes are hydroxyaluminium ones, which,due to the pH value, occur in each of the analysedsamples. Along with the increase of pH reac-tion, the number of Al(OH)−

4 complexes increases(sample No. 5). On the other hand, along withthe decrease of pH reaction, which is advanta-geous for the formation of aluminium complexes

Fig. 5 Variability ofconcentration of [H+]and molar ratio ofaluminium to fluorides

1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

SAMPLE

6,0

6,2

6,4

6,6

6,8

7,0

7,2

7,4

7,6

7,8

8,0

8,2

pH

0,00

0,02

0,04

0,06

0,08

0,10

0,12

0,14

0,16

0,18

0,20

MO

LA

R R

AT

IO A

l:F

pH

MOLAR RATIO Al:F

Environ Monit Assess (2011) 182:71–84 81

Fig. 6 Modelling ofaluminium forms forsample No 8 in the rangeof reaction 5.5–8.0

with fluorides, the Al(OH)−4 form does not occur

and the whole available aluminium is bound byfluorides, mainly by AlF+

2 (samples 36, 37). Thus,the factor which conditions the occurrence of agiven form of aluminium is the pH value. On theother hand, an equally important factor is the con-centration of fluoride, sulphate, phosphate andorganic ligands which are competitive in terms offorming complexes. The formation of complexesoccurs in the conditions of the dominance offluoride concentration over aluminium concentra-tion. In the case of sulphates, the concentrationmust be very high (Frankowski et al. 2009). Itshould be underlined that the complexes of alu-minium with fluorides are the most competitiveand, as it results from the Mineql+ program mod-

elling, they start forming at the pH of about 7. TheAlF+

2 > AlF3 > AlF2+ dominate, and at very highconcentrations of aluminium the AlF−

4 complexesbegin to form. The Mineql+ program also enablesthe simulation of variable conditions, i.e., of pH,the concentration of aluminium, fluorides and sul-phates in order to estimate the aluminium fractionin given conditions. And so, for sample No. 8,where the highest concentration of fluorides wasdetermined (0.82 mg·L−1), the modelling was per-formed within the range of pH = 5.5–8.0 (Fig. 6).

Based on Fig. 6 it may be observed that thebiggest changes in the formation of complexesoccur in the range of pH = 6.5–7.0. At pH <

6.5, the aluminium forms with fluorides dominate(AlF+

2 and AlF3). Then the range of pH reaction

Fig. 7 Modelling ofaluminium forms forsample No 29

82 Environ Monit Assess (2011) 182:71–84

Fig. 8 Modelling ofaluminium forms (Al-SO4complexes) in the rangeof reaction 5.5–8.0 forconcentration offluorides = 0

(5.5–8.0) from the sample collected at well No.29 was subjected to simulation, where the highestsulphate concentration was determined (328 mgL−1). Figure 7 presents the graph of these simu-lation conditions.

The bonds of aluminium with sulphates startforming at pH < 6.5 (AlSO+

4 ) but, despite sucha high concentration of sulphates (328 mg·L−1),only 10% of aluminium creates bonds withsulphates. Similarly to Fig. 6, the aluminiumcomplexes with fluorides dominate. It is note-worthy that a lower concentration of fluorides(0.13 mg·L−1) most probably causes the fact thataluminium complexes with fluorides only start todominate at pH < 6.0. At pH = 6.0–7.0, the dom-inating form is Al(OH)+

2 . Assuming that pH =5.5 causes the formation of aluminium complexeswith sulphates, it should be stated that the lack offluorides in the sample will cause the binding ofaluminium by sulphates (Fig. 8)

In the conditions of fluoride concentration = 0,at pH < 5.5, aluminium complexes with sulphatesstart to dominate and in fact only the AlSO+

4 ,Al(SO4)

−2 occurs at the low level <5%. However,

it should be especially underlined that the form ofAl3+ion at the level of total concentration of about18% still occurs, which in the case of this form isnot advantageous as it is marked by the highesttoxicity. The Mineql+ program provides a lot ofvaluable information, but it is only a tool helpfulin the determination of aluminium forms in bondswith complex-forming ligands. The previous stud-ies using computer programs are often used atpresent due to their usefulness in the interpreta-

tion of the results. However, in the environmentalanalysis, due to the rich sample matrix, it is notpossible to fully render the bonds of a complexoccurring in a sample.

Conclusions

Based on the obtained results, it may be statedthat the groundwater at the Miocene aquifer,which is isolated from the surface with a layer ofboulder clays and the upper Miocene silts of thePoznan series, is to a small extent exposed to an-thropogenic pollution. Definitely, a more seriousthreat to the groundwater of the Miocene aquifer,and especially to the lower Miocene aquifer, isposed by the ascent of water marked by highermineralization from the Mesozoic aquifer in thearea of the fault graben as well as the possibleascent of water caused by the intensified and con-centrated exploitation of well fields in the easternpart of the Warta River, were a deep cone ofdepression has been formed.

The groundwater of the investigated water-bearing level indicates the low concentration oftotal aluminium and the occurrence of aluminiumcomplexes with fluorides. Above pH = 7.0, thealuminium solubility decreases, which in conse-quence causes the decrease of concentration ofthe toxic Al3+form. Below pH < 7.0, the Al3+formstarts to dominate, but the mainly inorganic lig-ands present in the sample form fluoride com-plexes, which is advantageous for decreasing thetoxicity of this element. Fluoride ligands actually

Environ Monit Assess (2011) 182:71–84 83

form complexes with the whole aluminium belowpH = 6.5 present in the sample.

In the situation of condition simulation atfluoride concentration = 0, the aluminium com-plexes with fluorides present in the sample domi-nate (AlSO+

4 ). However, about 18% of aluminiumremains in the Al3+form.

Statistical analysis performed based on theobtained study results shows that the resultsobtained using the GF-AAS and ICP-AES tech-niques differ statistically from the results obtainedfrom the ICP-MS spectrometer. It should be un-derlined that in the case of the study of that type ofthe environment, the dependences may be causedby the injection systems for ICP and GF-AASanalytical techniques.

Acknowledgement The research was supported by thePolish Ministry of Science and Higher Education throughresearch grant No. 304 374 338 (2010–2011).

Open Access This article is distributed under the termsof the Creative Commons Attribution Noncommercial Li-cense which permits any noncommercial use, distribution,and reproduction in any medium, provided the originalauthor(s) and source are credited.

References

Bérubé, D., & Brûlé, D. G. (1999). A field aluminum speci-ation method to study the aluminium hazard in water.Fresenius Journal of Analytical Chemistry, 363, 566–570.

Bloom, P. R., & Erich, S. M. (1996). The quantitation ofaqueous aluminium. In: G. Sposito (Ed.), The envi-ronmental chemistry of aluminium (pp. 363–419). BocaRaton: CRC.

Boudot, J. P., Becquer, T., Merlet, D., & Rouiller, J. (1994).Aluminium toxicity in declining forests: A generaloverview with a seasonal assessment in a silver fir’forest in the Vosges mountains (France). Annales desSciences Forestières, 51, 27–51.

Cory, N., Buffam, I., Laudon, H., Björkvald, L., Mörth,C. M., Köhler, S., et al. (2009). Particulate aluminiumin boreal streams: Towards a better understanding ofits sources and influence on dissolved aluminium spe-ciation. Applied Geochemistry, 24, 1677–1685.

Dabrowski, S, Górski, J., & Przybyłek, J., (2007). The low-land sub-region of the Warta River; Regional hydro-geology of Poland. Fresh water (Vol. I, pp. 369–387).Warsaw: Polish Geological Institute, (in Polish).

Danielsson, L. G., & Sparén, A. (1995). A mechanizedsystem for the determination of low levels of quickly

reacting aluminium in natural waters. Analytica Chim-ica Acta, 306, 173–181.

Driscoll, C. T., & Schecher, W. D. (1990). The chemistryof aluminum in the environment. Environmental Geo-chemistry and Health, 12, 28–49.

Fairman, B., Sanz-Medel, A., Jones, P., & Evans, E. H.(1998). Comparison of fluorometric and inductivelycoupled plasma mass spectrometry detection systemsfor the determination of aluminium species in watersby high-performance liquid chromatography. Analyst,123, 699–703.

Frankowski, M., & Zioła-Frankowska, A. (2010). Spe-ciation analysis of aluminium and aluminium flu-oride complexes by HPIC-UVVIS. Talanta, 82,1763–1769.

Frankowski, M., Zioła-Frankowska, A., & Siepak, J.(2009). Study of aluminium sulphate complexes of sur-face water and fractionation of aluminium from bot-tom sediment. Archives of Environmental Protection,35, 55–67.

Frankowski, M., Zioła-Frankowska, A., & Siepak, J.(2010a). Speciation of aluminium fluoride complexesand Al3+ in soils from the vicinity of an aluminiumsmelter plant by hyphenated High Performance IonChromatography Flame Atomic Absorption Spec-trometry technique. Microchemical Journal, 95, 366–372.

Frankowski, M., Zioła-Frankowska, A., & Siepak, J.(2010b). New method for speciation analysis of alu-minium fluoride complexes by HPLC-FAAS hyphen-ated technique. Talanta, 80, 2120–2126.

Górski, J. (1989). The main hydrochemical problems ofthe Cainozonic aquifer located in central Wielkopolska(Great Poland). Scientific Bulletins of StanisławStaszic Academy of Mining and Metallurgy, no.1308,Geology bulletin 45, Cracow (in Polish).

Górski, J., & Przybyłek, J. (1996). Geological structure andground water (pp. 23–43). The natural environment ofthe city of Poznan. Poznan (in Polish).

Hills, A., Grote, M., Janßen, E., & Eichhorn (1999). Spe-ciation of trace amounts of aluminum in percolatingwater of forest soil by online coupling HPLC-ICP-MS.Fresenius Journal of Analytical Chemistry, 364, 457–461.

Kabata-Pendias, A., & Pendias, H. (1999). Biogeochem-istry of trace elements (pp. 192–198). Warszawa:Wydawnictwo Naukowe PWN.

Macioszczyk, A., & Dobrzynski, D. (2002). Hydrogeo-chemistry of short turn-over time zone (pp. 322–324).Warszawa: Wydawnictwo Naukowe PWN (in Polish).

Mitrovic, B., Milacic, R., Pihlar B., & Simoncic, P.(1998). Specification of trace amounts of aluminiumin environmental samples by cation-exchange FPLC—ETAAS. Analusis, 26, 381–388.

Narin, I., Tuzen, M., & Soylak, M. (2004). Aluminum de-termination in environmental samples by graphite fur-nace atomic absorption spectrometry after solid phaseextraction on amberlite XAD-1180/pyrocathechol vi-olet chelating resins. Talanta, 63, 411–418.

Préndez, M., & Carrasco, M. A. (2003). Elemental com-position of surfacewaters in the antarctic peninsula

84 Environ Monit Assess (2011) 182:71–84

and interactions with the environment. EnvironmentalGeochemistry and Health, 25, 347–363.

Przybyłek, J. (1986). Groundwater in the vicinity of thePoznan-Gostyn fault graben. Papers of Wrocław Uni-versity of Technology. Wrocław. no. 49 (in Polish).

Rezaee, M., Yamini, Y., Khanchi, A., Faraji, M., & Saleh,A. (2010). A simple and rapid new dispersive liquid-liquid microextraction based on solidification of float-ing organic drop combined with inductively coupledplasma-optical emission spectrometry for preconcen-tration and determination of aluminium in water sam-ples. Journal of Hazardous Materials, 178, 766–770.

Rodushkin, I., & Ruth, T. (1997). Determination of tracemetals in estuarine and sea-water reference materialsby high resolution inductively coupled plasma massspectrometry. Journal of Analytical Atomic Spectrom-etry, 12, 1181–1185.

Rossiter, H. M. A., Owusu, P. A., Awuah, E.,MacDonald, A. M., & Schäfer, A. I. (2010). Chemicaldrinking water quality in Ghana: Water costs andscope for advanced treatment. Science of the TotalEnvironment, 408, 2378–2386.

Salomon, S., Giamarchi, P., Le Bihan, A., Becker-Roß, H.,& Heitmann, U. (2000). Improvements in the deter-

mination of nanomolar concentrations of aluminiumin seawater by electrothermal atomic absorption spec-trometry. Spectrochimica Acta B, 55, 1337–1350.

Šcancar, J., & Milacic, R. (2006). Aluminium speciationin environmental samples: A review. Analytical andBioanalytical Chemistry , 386, 999–1012.

Siepak, M., Frankowski, M., Zioła, A., Kurzyca, I., &Przybyłek, J. (2006). Chemistry in sustainable develop-ment (pp. 453–460) (in Polish).

Sjöstedt, C., Wällstedt, T., Gustafsson, J. P., & Borg, H.(2009). Speciation of aluminium, arsenic and molyb-denum in excessively limed lakes. Science of the TotalEnvironment, 407, 5119–5127.

Xia, L., Hu, B., Jiang, Z., Wu, Y., Li, L., & Chen, R. (2005).8-Hydroxyquinoline-chloroform single drop microex-traction and electrothermal vaporization ICP-MS forthe fractionation of aluminium in natural waters anddrinks. Journal of Analytical Atomic Spectrometry, 20,441–446.

Zioła-Frankowska, A., Frankowski, M., & Siepak, J.(2009). Development of a new analytical method foronline simultaneous qualitative determination of alu-minium (free aluminium ion, aluminium-fluoride com-plexes) by HPLC-FAAS. Talanta, 78, 623–630.