RESEARCH ARTICLE

Aqueous geochemistry of fluoride enriched groundwater in aridpart of Western India

Chander Kumar Singh & Saumitra Mukherjee

Received: 7 March 2014 /Accepted: 21 August 2014# Springer-Verlag Berlin Heidelberg 2014

Abstract Fluoride-enriched water has become a major publichealth issue in India. The present study tries to evaluate thegeochemical mechanism of fluoride enrichment in groundwa-ter of western India. Total 100 groundwater samples werecollected for the study spreading across the entire study area.The results of the analyzed parameters formed the attributedatabase for geographical information system (GIS) analysisand final output maps. A preliminary field survey was con-ducted and fluoride testing was done using Hach make fieldkits. The fluoride concentration ranges from 0.08 to 6.6 mg/L(mean 2.4 mg/L), with 63 % of the samples containing fluo-ride concentrations that exceed theWorld Health Organization(WHO) drinking water guideline value of 1.5 mg/L and 85 %samples exceeding the Bureau of Indian Standards (BIS)guidelines of 1 mg/L. The study also reveals high concentra-tion of nitrate that is found to be above WHO standrads. Thedominant geochemical facies present in water are Na-Cl-HCO3 (26 samples), Na-Ca-Cl-HCO3 (20 samples), Na-Cl(14 samples), and Na-Ca-Mg-Cl-HCO3 (11 samples); howev-er, sodium and bicarbonate being the major component in allthe water types of 100 samples, which in fact has a tendency toincrease fluoride concentration in water by dissolving fluoridefrom fluorite. The thermodynamic considerations between theactivities of calcium, fluoride, and bicarbonate suggest thatfluoride concentration is being governed by activity ofcalcium ion. X-ray diffraction analysis of sediments revealscalcite and fluorite are the main solubility-control minerals

controlling the aqueous geochemistry of high fluoride ground-water. The results indicate that the fluoride concentration ingroundwater is mainly governed by geochemical compositionof rocks, such as metamorphic granites and sedimentaryrocks, alkaline hydrogeological environment, climatic condi-tions, high temperature and lesser rainfall, and geochemicalprocesses such as weathering, evaporation, dissolution, andion exchange.

Keywords Geochemical modeling . Evaporation .

Groundwater .Water facies

Introduction

Fluoride helps in mineralization of bones and formation ofenamel of teeth. A daily dose of 0.5 ppm is required for properformation of enamel and bone mineralization which otherwisemay result in formation of dental fluorosis characterized ini-tially by opaque white patches, staining, mottling and pittingof teeth, lack of enamel formation, and bone fragility(Cao et al. 2000; Rwenyonyi et al. 2000; Vieira et al. 2005;Edmunds and Smedley 2005; Ayenew 2008; Banerjee 2014).Long-term intake of fluoride-enriched water may cause bilat-eral lameness and stiffness of gait (Suttie 1977; Oruc 2008).Common natural fluoride sources in groundwater are thedissolution of some fluoride bearing minerals, such as fluorite(CaF2), muscovite, biotite, hornblende, villianmite, tremolite,fluorapatite, and some micas weathered from silicates, igne-ous, and sedimentary rocks, especially shale (Handa 1975;Pickering 1985; Wenzel and Blum 1992; Datta et al. 1996;Zhang et al. 2003; Fawell et al. 2006; Msonda et al. 2007; Jhaet al. 2010; Singh et al. 2011a, b, c). Unstable minerals such assepiolite and palygorskite may also have a dominant controlon fluoride concentration in groundwater (Kim et al. 2005;Jacks et al. 2005). High fluoride in groundwater is distributed

Responsible editor: Stuart Simpson

C. K. Singh (*)Department of Natural Resources, TERI University, NewDelhi 110070, Indiae-mail: chanderkumarsingh@gmail.com

S. MukherjeeJawaharlal Nehru University, New Delhi, India

Environ Sci Pollut ResDOI 10.1007/s11356-014-3504-5

in all the 31 districts of Rajasthan which is influenced by theregional and local geological settings and hydrological condi-tions (Agrawal et al. 1997). Few studies have been conductedat small administrative units in Rajasthan by several re-searchers (Suthar et al. 2008; Hussain et al. 2012; Singhet al. 2011a) which report fluoride concentration aboveWHO standards of 1.5 mg/L in groundwater of Rajasthan.Handa (1975) observed that high fluoride concentration isaccompanied by high nitrate concentrations in Rajasthan anda solution-evaporation and base-exchange hypothesis wasproposed to explain the genesis of fluoride in the groundwaterin Rajasthan. Fluoride was also reported in the sediments ofAjmer and Jaisalmer district (Madhavan and Subramanian2002; Singh et al. 2012). Geogenic contamination of ground-water due to fluoride, arsenic, and heavy metals has beenperformed by several researchers in different regions ofIndia using graphical methods, statistical, and geochemicalmodeling (Chakrabarti and Bhattacharya 2013; Hussain et al.2013; Singh et al. 2011b, 2014; Rina et al. 2013, 2014).Inverse geochemical modeling in Phreeqc is based on a geo-chemical mole-balance model, which computes the phasemole transfers (the moles of minerals and gases that move inor leave the solution) to comprise the differences in an initialand a final composition of groundwater system along the flowpath (Parkhurst and Appelo 1999; Singh et al. 2011c). Thismass balance approach has been used in recent times toquantify reactions controlling fluoride chemistry alonggroundwater flow paths (Hidalgo and Cruz-Sanjulian 2001;Rafique et al. 2008; Kim et al. 2012; Singh et al. 2011c) andquantify mixing of end-member components in a flow system(Kuells et al. 2000).

With the above background this paper deals with a logicalapproach to evaluate the geochemistry of groundwater in TharDesert of India. The study also examines the spatial distribu-tion of fluoride toxicity and the relation between bedrock andclimatic conditions for fluoride distribution in the groundwa-ter of Thar Desert of Rajasthan. The inter-relationship ofmajor ions and their ionic ratio and different mineral phasesbased on geochemical modeling has also been studied todecipher the actual cause of fluoride enrichment in thegroundwater.

Material and methods

Study area

The study area lies within the Jaisalmer district of Rajasthan. Itactually covers two of administrative blocks, Jaisalmer andPokharan. The area lies between 70.81–72.10°E longitudesand 26.29–27.54°W latitudes covering an area of approxi-mately 9,850 km2. The study area along with spatial distribu-tion of fluoride is shown in Fig. 1.

The study area is situated amid Thar Desert where mon-soon is as good as negligible. Numerous saline lakes (playas)are also present in the study area which had formed due toaccumulation of water in natural depressions which later ongot saline due to high evaporation rate (Deotare et al. 2004).The annual precipitation in the region varies from 450 mm atthe eastern margin to 100 mm at the western margin. In theThar, the mean maximum temperature during the summermonths ranges from 40 to 45 °C, and the mean minimumtemperatures during the winter months fluctuate between 3and 10 °C. The evapotranspiration of the region is 3–20 timeshigher than the precipitation, indicating a negative waterbalance.

The Jaisalmer basin is dominated byMesozoic and Tertiaryformations. The western and northwestern parts of the districtare covered by vast blanket of young unconsolidated depositsincluding the blown sand of the Thar Desert of WesternRajasthan. The sedimentary rocks occurring in the area in-clude calcarinite, oolitic calcarenite, calcareous sandstone,marl, shale, and friable sandstone. Pleistocene sand alluvium,blown sand, kankar (calcium nodules), carbonate beds, andevaporite deposits of recent and sub-recent age are found overa large area of West and Eastern Rajasthan. Igneous andmetamorphic rocks of lower Proterozoic age comprising ofslate, quartzite, phyllite, schist, and gneiss are also found inthe area. This igneous suite consisting of basalt and rhyolite isoverlain by the sandstone and limestone (Rai, 1990).Carbonate rocks (limestone, marble and dolomite) are alsofound in this region. The stratigraphy of Jaisalmer district isshown in Table 1. Hydrogeologically, i.e., three types rockformations have been identified depending on characteristi-cally different consolidated, semi-consolidated, and unconsol-idated formations are present in the district, as far as occur-rence and movement of groundwater is concerned. The con-solidated hydrogeological units are the Pre-cambrian, Jaloregranites and Malani rhyolite, and Cambrian rocks of theMarwar supergroup, which are Jodhpur sandstone andChacha limestone. The semi-consolidated hydrogeologicalunits include the Mesozoic and Tertiary formations whichare Lathi, Jaisalmer, Bhadesar, Parewar, sandstone andJaisalmer, Khuiyala and Bandha limestone and Baisakhishale. The unconsolidated formations are the alluvial depositsconsisting of sand, silt, and gravel. The quaternary sedimentscomprising unconsolidated aeolian and alluvium are the im-portant formations due to their wide spread occurrence. Thesediments are composed of sand, silt, and clay gravel, calcar-eous and ferruginous concretions and occur in northern, west-ern, and southern part of the area. However, a major part ofalluvium contains saline ground water. The groundwater oc-curs under unconfined to confined conditions down to depthof 330 m. Depth to water level varies from less than 10 mbglto more than 60 mbgl. Yield of the wells is generally less than10 lps in this formation. Groundwater in granites and rhyolites

Environ Sci Pollut Res

occur under unconfined conditions in joints, fractures, andweathered mantle and is of poor yield.

Water sampling, analysis, and quality assurance

A total of 100 water samples were collected from districtspanning over an area of approx. 10,000 km2. Water sampleswere collected in propylene bottles from shallow hand pumpsand tube wells. The pH, total dissolved solid (TDS), andelectrical conductivity (EC) of the water samples were mea-sured onsite using portable pH, TDS, and EC electrodes(HANNA). The samples were acidified using HNO3

(Ultrapure Merck) for cation analysis. The samples were storedin an icebox, carried to the laboratory, and kept at 4 °C forfurther chemical analysis. Immediately after the water sampleswere transported to the laboratory, the major cations were(Mg2+, Ca2+, Na+, K+) analyzed using an atomic absorptionspectrometer (Thermo Fisher Scientific M series), and themajor anions (F−, Cl−, SO4

2−, NO3−) were analyzed using an

ion chromatograph (Dionex). Bicarbonate (HCO3−) was deter-

mined by titration method using standard procedures as de-scribed in standard methods for the examination of water andwastewater (APHA 2007). Fifty dried surficial sediments werepowdered for X-ray diffractometry (XRD). The XRDmeasure-ments used Cu, Kα radiation on a graphite monochromator ona PANalytical vertical diffractometer with a step size of 0.5 s/0.02°, from 5° to 60° 2θ. Iterative identification of minerals inthe samples was obtained by using X’pert HighScore Plussoftware with search and match of the reference mineral data-base. The analyzed parameters for all the sampling locations

were joined to GPS point locations in Arc GIS 10 whichformed the database to be analyzed GIS platform.

Quality assurance and quality control

All plastics and glassware’s utilized were pre-washed withdetergent water solution, rinsed with tap water, and soakedfor 48 h in 50 % HNO3, then rinsed thoroughly with distilled-deionized water. They were then air-dried in a dust-free envi-ronment. Quality assurance and quality control was main-tained as all reagents were of analytical grade (Merck), andsamples for metal analysis were preserved with 2–3 drops ofconcentrated HNO3 per 250 mL of sample in the field. Theduplicate sample was collected for every 5th sample. Qualitycontrol included reagent and blank analyses, spiked sampledeterminations using standard addition calibration methodusing commercially available standard stock solution. Thestandards were run after every 10 samples analysis. Thereproducibility of the analytical data was found to be within4 %, and the accuracy was estimated to be <10 % based on theresults of standard stock solution of various concentrationsprepared from the purchased standard solution. The analyticalprecision of the ions analyzed was determined by calculatingthe normalized ionic charge balance error, which varied within±5 %.

Fluoride testing using field kit

Groundwater sampling and testing was expanded to a fewmore districts lying in the Thar Desert. A Hach pocket

Fig. 1 Study area with fluoridedistribution

Environ Sci Pollut Res

colorimeter (http://www.hach.com/pocket-colorimeter-ii-fluoride-spadns-/product?id=7640445204) was used tomeasure fluoride spectrophotometrically in the field basedon the SPADNS method (APHA, 2008) at 64 well locations.Some samples were diluted in order to remain within theworking range of the colorimeter. Groundwater samplesfrom the same wells were also analyzed in the laboratoryusing a Dionex DX2000 ion chromatograph in gradientmode equipped with an AS-11HC column (Dionex,Sunnyvale, CA). With the exception of a handful of sampleswith >2 mg/L fluoride that should have been diluted in thefield, the two sets of tests were in good agreement. Almost onequarter of the samples did not meet the WHO guideline forfluoride of 1.5 mg/L, within one sample containing as much10.8 mg/L. Relative to the WHO guideline, well classificationaccording to the field kit and laboratory measurements wereconsistent for 82 % of the samples. Well depths reported bythe households indicate that wells which are tapping waterfrom deeper aquifers are more contaminated with fluo-ride than those of shallow wells (Fig. 2).

Geochemical modeling of mineral phases

Inverse modeling is often used for interpreting geochemicalprocesses that account for the hydrochemical evolution

of groundwater (Plummer et al. 1983; Hidalgo andCruz-Sanjulian 2001; Dai et al. 2006). Saturation indicesindicate the thermodynamic tendency of minerals to dissolveor precipitate. A number of assumptions are essential in thesolicitation of inverse geochemical modeling: (1) the twogroundwater analyses from the initial and final water wellsshould represent groundwater that flows along the same path,(2) diffusion and dispersion do not significantly affecthydrochemistry, (3) chemical steady-state prevails in thegroundwater system, and (4) the mineral phases used in theinverse calculation are/were present in the aquifer (Zhu andAnderson 2002). The soundness or validity of the results inthe inverse modeling depends on a valid conceptualization ofthe groundwater system, basic hydrochemical concepts, andaccuracy of input data into the model and level of understand-ing of the geochemical processes in the area. The changes insaturation state are useful to distinguish different stages ofhydrochemical evolution and help to identify which geochem-ical reactions are important in controlling water chemistry(Güler and Thyne 2004; Coetsiers and Walraevens 2006).The saturation index of a mineral can be obtained using thefollowing equation (Garrels and Mackenzie 1971):

SI ¼ log IAP=Ktð Þ ð1Þ

Table 1 General stratigraphy of Jaisalmer District

Age Formation LithologySub-recent to Recent Sand, Rock-waste, Kankars, Gypsite etc.

Subrecent Pebble Spread (Glacial?)

Quarzite, Quartz, Ironstone, Sandstone, Dolomite, Chert, Jasper, Rhyolites, Granites, Basic rocks etc.

--------------------------- Unconformity --------------------------------------------------------------------------------------------------------------------Subrecent

(Pleistocene)Shumar Gritty/ Fragmented Ironstone (subsurface)

--------------------------- Unconformity --------------------------------------------------------------------------------------------------------------------Lower Eocene Khuiala Massive fragmental limestone, marly limestone, foraminiferal limestone and bentonitic clays

(subsurface well section)--------------------------- Unconformity --------------------------------------------------------------------------------------------------------------------

Upper Jurassic (Tithenian)

Badesar Ferrugenised grits and Ironstone

--------------------------- Unconformity --------------------------------------------------------------------------------------------------------------------Upper Jurassic

(Kimmeridgian)Baisakhi A) Not Exposed

B) Shales, siltstone with thin bands of calcareous and ferrugenious sandstoneJaisalmer Friable sandstone ironstones and intercalations of calcareous sandstone

C) Gypseous saline shales with phosphotic nodules bearing ferruginous bandsD) Shales, Oolitic calcarenite and marly limestone

--------------------------- Gradational Contact-------------------------------------------------------------------------------------------------------------------Middle Jurassic

Callovian to Oxfordian

Jaisalmer A) Thick sandstone and shales with interbedded sequence of calcareous sandstone, ferruginous, sandstone calcarenite, calcilutite, marly limestone etc

B) Friable calcareous sandstone with thin bedded fossiliferous limestone, calcilutite, pseudo-oolitic calcarenite on top

C) Calcareous sandstone, friable sandstone, marly limestone with pyrite coated gastropods bearing calcarenite at the top

Environ Sci Pollut Res

where IAP=ion activity product of the dissociated chemi-cal species in solution and Kt=equilibrium solubility at min-eral temperature.

Positive values of saturation index indicate dissolution ofminerals or gas; negative values indicate precipitation oroutgassing. To calculate the saturation index and thus aqueousand mineral phases, inverse parameter modeling with thethermodynamic program Phreeqc v. 3.1.2-8538 (Parkhurstand Appelo 1999) was used. A word of caution hereis that geochemical modeling using Phreeqc is based onassumptions, constraint, and phases so it may be that itrepresents one of the explanation of actual states of min-erals and not only limited to the groundwater composition ofstudy area

Hierarchical cluster analysis

Cluster analysis (CA) is a technique of grouping samples/objects into unknown groups. This is a classification method(Sneath and Sokal 1973) which enables the grouping of sim-ilar samples/objects on basis of distance criteria and of aspecific aggregative algorithm in order to create a typologywhich will characterize the elements to be classified. It groupsthe objects into the classes or clusters on the basis of similar-ities within a class and dissimilarities between different clas-ses. The results of CA help in interpreting the data and indicatepatterns. Many authors have utilized cluster analysis to studygroundwater samples (Woocay and Walton 2008; Singh et al.2011b) In hierarchical clustering, clusters are formed sequen-tially by starting with the most similar pair of objects andforming higher clusters step by step. The samples weregrouped according to their “similarity” to each other.Hierarchical agglomerative CAwas performed on the normal-ized data set (mean observations over the whole period) byusing Euclidean distance (straight line distance between twopoints in c-dimensional space defined by c variables) forsimilarity measurement, together with Ward’s methodfor linkage. Cluster analysis was applied to the ground-water quality data set with a view to group the similarsampling sites (spatial variability) spread in the region.The method is a hierarchical one, providing a branchingdiagram with junction levels proportionate to the rate of

inertia associated with each junction. The results indicate thatCA techniques are useful in offering reliable classification ofwater resources in the study area and will make it possible todesign future spatial sampling strategy. We also utilized thedata of the four clusters obtained from CA in geochemicalmodeling to calculate phase mole transfer between theseclusters.

Results and discussion

Hydrochemistry

Groundwater samples of the study area are alkaline innature, pH in the area ranged from 7.4 to 8.6 withaverage value of 7.9 depicting weak alkaline conditionsdominant within groundwater system in the JaisalmerDistrict. Electrical conductivity in the area ranged from890 to 12,000 μS/cm with average value of 2,886 μS/cmwhereas TDS ranged from 450 to 5,300 μS/cm. Very highvalues of TDS and EC in the region are due to salt encrusta-tions because of the presence of several playas present in

Fig. 2 Results for fluoride in wellwater using (a) a field kit and (b)ion chromatography with error ofstandard deviation

Table 2 Statistical summary of the physicochemical parameters ofgroundwater [all units in mg/L except pH, EC (μS/cm)]

Avg Min Max Std dev

pH 7.9 7.4 8.6 0.3

EC 2,886 890 12,000 1,970

TDS 1,790 449 5,300 1,020

Na 433 67 1,530 316

K 18.5 2.7 117 17

Ca 84 20 228 33

Mg 51 12 170 32

Cl 540 70 2,230 395

SO4 183 19 852 154

HCO3 414 158 1,220 175

NO3 149 5 1,050 181

F 2.2 0.1 6.6 1.4

HCO3/Ca 5.5 2.0 28 3.6

Na/Ca 5.8 1.1 24.5 4.4

Environ Sci Pollut Res

this region. The values of EC and TDS also indicatehighly saline water in the area. Na+, Ca2+, and Mg2+ aredominant cations contributing 74.3, 14.1, and 8.43 %,respectively. Among the anions Cl−, HCO3

−, SO42−, and

NO3− are major contributors. Cl− contributes 41.9 % whereas

HCO3−, SO4

2−, and NO3− contribute 32.1, 14.1, and 11.5 %,

respectively. The statistical variables for all the analyzed pa-rameter is given in Table 2.

Fluoride distribution and contamination

The fluoride concentration in the groundwater samples rangesfrom 0.08 to 6.6 mg/L (mean 2.4 mg/L), with 63 % of thesamples containing fluoride concentrations that exceed theWHO drinking water guideline value of 1.5 mg/L and 85 %samples exceeding the BIS guidelines of 1 mg/L. Almost,whole of the study area exhibits fluoride concentration morethan the permissible values as per BIS guidelines, with fewexceptions in the central and western region of the study area.The current study shows that the nitrogen accumulation in thestudy area is a long-term accumulation of nitrate as nitrateconcentration ranges from 5 mg/L to 1,050 mg/L with anaverage value of 136 mg/L. The ionic ratio of NO3

− and Cl−

suggests that the nitrate input from the manures is not takingplace in the region. The build up nitrate has taken place as aresult of long-term leaching of nitrate into the groundwater.

The correlation matrix for the water quality parameterswere examined to understand the hydro-geochemical process-es leading to enrichment of fluoride in groundwater (Table 3).A positive correlation between fluoride concentration and pHis commonly observed in different parts of the world. Fluoridein groundwater is associated with a pH >7.0 (Genxu andGuodong 2001). In the present study, moderate positive cor-relation of fluoride and pH (r=0.5, r value at significance levelof p<0.05) showed that fluoride in groundwater has resultedfrom leaching of fluoride-containing minerals (Fig. 3a).

Fluoride exhibits positive correlation (r=0.37) with bicarbon-ate ion. At acidic pH, the fluoride is adsorbed on the surface ofthe clay. A higher value of pH favors the enrichment offluoride in groundwater. The OH− in groundwater with highvalue of pH can replace the exchangeable fluoride of clayminerals (biotite/muscovite) thus can increase the concentra-tion of fluoride in groundwater (Guo et al. 2007; Singh et al.2011a; Gupta et al. 2012). The hydroxyl ions replace fluoridefrom the muscovite as shown below:

MuscoviteKAl3Si3O10 OH; Fð Þ2 þ CO2

þ 2:5H2O→1:5Al2Si2O5 OHð Þ4 þ Kþ þ HCO3‐ þ 2F−

ð2ÞIn granitic or sandstone dominant aquifers, dissolution of

fluorite can be a possible reason for presence of fluoride ingroundwater. The hydrolysis of alumino-silicate minerals inthe hard rock aquifers produces bicarbonate ion, which canenhance fluorite dissolution

CaF2 þ 2HCO‐3→CaCO3 þ 2F‐ þ H2Oþ CO2 ð3Þ

Water with high F- concentration can form in the areaswhere alkaline (carbonate rocks) waters are in contact withfluoride-bearing minerals. Fluoride concentration generallydepends on few water-soluble components, but a good corre-lation exists between fluoride and pH. In the present study, anoticeable moderate correlation (R=0.5) exists between fluo-ride and pH (Fig. 3a). The fluoride solubility is lowest in thepH range of 5–6.5 (Adriano 1986). At higher pH, ionicexchange occurs between F and OH ions (illite, mica, andamphiboles) resulting in increase of F ion concentration ingroundwater (Datta et al. 1996). The presence of highHCO3, Na, and pH favors the release of F from aquifermatrix into groundwater. It was observed that the

Table 3 Correlation matrix ofanalyzed physicochemicalparameters

Significant correlation has beenitalicized

pH EC TDS Na K Ca Mg Cl SO4 HCO3 NO3 F

pH 1.00

EC −0.01 1.00

TDS 0.12 0.78 1.00

Na 0.01 0.84 0.77 1.00

K 0.15 0.17 0.35 0.09 1.00

Ca 0.11 0.41 0.64 0.27 0.63 1.00

Mg 0.02 0.55 0.71 0.46 0.31 0.63 1.00

Cl 0.01 0.89 0.84 0.85 0.13 0.43 0.63 1.00

SO4 −0.04 0.76 0.88 0.78 0.21 0.50 0.65 0.80 1.00

HCO3 0.13 0.50 0.58 0.54 0.08 0.28 0.26 0.40 0.51 1.00

NO3 0.17 0.36 0.54 0.42 0.46 0.46 0.35 0.25 0.48 0.35 1.00

F 0.50 0.15 0.25 0.08 0.21 0.16 −0.05 0.11 0.15 0.37 0.17 1.00

Environ Sci Pollut Res

groundwater in the study area was alkaline as the pHvalue of groundwater in the study area varied from 7.9–8.5. The solubility of fluoride-bearing minerals increasesdue to alkaline nature of the groundwater. In an alkalineenvironment, the fluoride ions are desorbed thussupplementing the dissolution of fluoride-bearing minerals(Singh et al. 2011a, 2012; Gupta et al. 2012).

Hydrochemical facies and fluoride concentration

The dominant hydrochemical facies of the groundwater areNa-Cl-HCO3 and Na-Ca-Cl-HCO3 followed by Na-Cl andNa-Ca-Mg-Cl-HCO3 type. Na

+ and HCO3− are predominant

ions in almost all the samples of groundwater.It has been observed that sodium bicarbonate type waters

are capable of releasing fluoride from fluorite mineral (Handa1975; Chae et al. 2007; Guo et al. 2007). The concentration offluoride can reach as high as 20 mg/l due to fluorite dissolu-tion in water with sodium bicarbonate facies (Rao et al. 1993);nevertheless, ionic constituents of water and mineralsmight influence fluorite dissolution in a regional setting.The dissolution of fluorite takes place as (Jacks et al. 2005;Guo et al. 2007):

CaF2 þ Na2CO3→CaCO3 þ 2Naþ 2F ð4Þ

CaF2 þ 2NaHCO3→CaCO3 þ 2Naþ 2Fþ H2Oþ CO2 ð5Þ

XRD analysis

The X-ray diffraction was carried out for 50 sediment sam-ples. The XRD analysis of samples revealed that quartz,calcite, feldspar, and micas forms the major mineralogy ofthe samples with quartz most high prevalent in the region.Along with these minerals, biotite, muscovite, albite, micas,and clay minerals such as kaolinite, montmorillonite, andchlorite are also present in the samples.

The mineral composition points towards the granitic natureof rocks, and it has been observed that these rocks serve asgenesis of fluoride-rich groundwater (Veksler et al. 2005;Fordyce et al. 2007; Singh et al. 2012).

The plot of Na+/Na++Ca2+ vs log TDS suggest rock-dominant weathering, and the salinization due to evaporationand crystallization dominance (Fig. 3b). The Na/Na+Cl ratiois supportive to differentiate water of different origin(Hounslow 1995). The average Na/Na+Cl ratio of groundwa-ter is 0.45 indicating presence of albite (Naseem et al. 2010).The normative mineral composition indicates high proportionof albite in the granitic rocks. Ozsvath (2006) identified goodrelationship between albite and high F− bearing groundwater.In general, concentrations of F− in granitic rocks range be-tween 500 and 1,400 mg kg−1 (Krauskopf and Bird 1995;Naseem et al. 2010) which is consideredmuch higher than anyother rock type. Sedimentary rocks have a fluorine concentra-tion from 200 ppm in CaCO3 up to 1,000 ppm in shales(Frencken 1992). Fluoride is present as fluorite in carbonatesedimentary rocks. Clastic sediments have higher fluoride

Fig. 3 a Scatter plot between pH and Fluoride. b Scatter plot between TDS and Na/Na+Ca. c Samples with under or over saturation with respect tofluorite and calcite

Environ Sci Pollut Res

concentrations as the fluoride is concentrated in micas andillites in the clay fractions however high fluoride may also befound in phosphate beds of sedimentary rocks (Frencken1992). The F− distribution in the rocks of study area showswide range because of diversity in the composition and type ofrocks. The high sodium concentration imparts bicarbonatetype character to groundwater, which permits higherfluoride concentration as and when equilibrium is reachedwith fluorite.

Geochemical modeling of mineral phases

The values for saturation index were indicated that nearly allthe samples (except three samples) are oversaturated withrespect to calcite. However, leaving a handful of samples(approx. 15 samples) undersaturated with respect to fluorite(Fig. 3c). Undersaturation reflects the character of water froma formation with insufficient amount of the mineral for solu-tion or short residence time. The high rate of evaporation in

semiarid condition assisted by higher temperature andlesser rainfall might enhance the calcite precipitation inan alkaline environment of groundwater and thus creat-ing deficiency of calcium ions and creating a harmoni-ous environment for dissolution of fluorite. Oversaturationreflects that the groundwater discharging from an aquifercontaining ample amount of the mineral with sufficient resi-dence time to reach equilibrium. After oversaturation, theconcentration of Ca2+ ions overrides the solubility limit offluorite as fluorite dissolution is suppressed by common ioneffect, which might lead to a negative correlation between thetwo ions (Ca2+ and F−) (Handa 1975) which in this case is alsoshown by poor correlation between the two ions (r=0.16).The fluoride concentration in groundwater is mainly governedby oversaturation of calcite due to evaporation and calciumion exchange process.

The SI values for calcite and aragonite shows that they aresupersaturated at most of the locations. An empirical relationbetween calcite precipitation and aragonite dissolution shows

P28

J23

P3

P22

P18

J18

P12

P21

J13

P9

P26

P50

P55

J31

P59

P62

J14

J28

J11

J21

P39

J19

J8

P8

J22

P33

J27

J29

P40

P53

P45

P61

P56

P63

P65

P54

P64

P47

P49

P60

P66

P38

P48

P37

P46

P51

P42

P41

P43

P57

P58

P44

P52

P2

P11

J33

P31

P24

J24

P14

P36

P19

J2

J30

J15

J20

J9

J3

J7

P16

P27

J1

P5

J6

J10

P4

J34

P15

P23

P35

P7

P30

P6

P32

J16

P34

J26

P20

J17

J25

P17

J12

P25

P29

P10

J32

J4

J5

P1

P13

0

1000000

2000000

3000000

4000000

5000000

6000000

7000000

8000000

9000000

10000000

Dissim

ilarity

Dendrogram

C4

C3

C2

C1

24043546

74043546

124043546

174043546

224043546

274043546

324043546

374043546

Dissim

ilarity

Fig. 4 Dendrogram showingclusters of samples

Table 4 Mass transfer betweenthe four clusters based on geo-chemical parameters

Mineral phases Phase state Phase mole transferfrom cluster 1–2

Phase mole transferfrom cluster 2–3

Phase mole transferfrom cluster 3–4

Halite Precipitating −2.84E-02 −3.73E-02 −1.72E-01Gypsum Dissolving 4.55E-02 4.43E-01 1.05E-01

Kaolinite Precipitating −2.05E-03 − −CO2 Precipitating −4.96E-02 −1.70E+00 −1.18E-01Calcite Precipitating −6.87E-02 −1.31E+00 −1.50E-01Plagioclase Dissolving 2.97E-03 2.48E+00 −Fluorite Dissolving 2.91E-04 2.24E-01 −Ca-Montmorillonite Precipitating − −1.47E+00 −Biotite Dissolving − 1.60E-02 −Calcedony Precipitating − −1.13E+00 −9.77E-03

Environ Sci Pollut Res

that these two processes are most likely sources of Ca2+ andHCO3

− in the hydro-geochemistry. The hydro-chemical datawere classified by HCA into 13-dimensional spaces (pH, EC,TDS, Na, K, Ca, Mg, Cl, SO4, HCO3, NO3, F, SiO2), and theresult is presented as a dendrogram (Fig. 4). Four cluster wereidentified based on the dendrogram each representing ahydrochemical facies. Cluster 1 encompasses a total of 26samples; cluster 2 consisting of 63 samples and cluster 3 and4 consisting of 10 and 1 samples, respectively. Electricalconductivity, Cl and NO3 seems to be a major distinguishingfactor among these clusters. The dissolution of gypsum andfluorite provides inputs for SO4 and F to the groundwater andtherefore were considered as phases in the model and elementsas S and F as corresponding constraints. The rock waterinteraction is majorly due to weathering of silicate mineralssuch as biotite, kaolinite, montmorillonite, and plagioclase;thus, these were also included as phases and elements Si, Ca,Mg, Na, and K as constraints. The phase mole transfer wascalculated between these four clusters. Potential phases in theinverse modeling were constrained (precipitation/dissolution)using compiled data of saturation indices derived fromPhreeqc and a conceptual model inferred from general trendsin chemical analyses data of groundwater. The inverse modelwas constrained so that primary mineral phases includinghalite, kaolinite, calcite, biotite, chalcedony, and CO2 wereset to dissolve until they reached saturation, and gypsum,plagioclase, fluorite, and biotite were set to precipitate oncethey reached saturation. Carbon dioxide was assumed to beavailable throughout the flow path by incorporating reactionsthat result in oxidation. The calculated phase mole transfergiven in Table 4 depicts that calcite is precipitating and thuscreating a favorable environment for dissolution of fluorite.Since the inverse modeling is based on assumptions andconstraints so it may be worth mentioning that the explanationmay not be the only solution in the study area and thus other aholistic approach must be adopted for inverse modeling.

Conclusion

High levels of fluoride in groundwater lead to health hazardssuch as dental fluorosis and skeletal fluorosis, leading tomolting and pitting of teeth, stiffness and rigidity of joints,and bending of spinal cord. The concentration of fluoride(almost 85 % of the samples) was well above the maximumpermissible limit set by Bureau of Indian Standards. The studyalso finds that field kits can be a reliable method for onsitetesting of fluoride and the households could be conveyed theresults so that they avoid drinking and cooking with fluoride-enriched groundwater and thus lower exposure to fluoride andits health impacts. The results indicate that the high fluoride ingroundwater is basically geogenic in nature. Hydro-geologicalsettings coupled with high evaporation rate and high

temperature are controlling factors for fluoride enrichment.The interaction of water with fluoride-rich minerals enforcesthe geochemical facies of water towards Na-HCO3 type,which in turn favors dissolution of these minerals. The pres-ence of high HCO3, sodium, and pH favors the release offluoride from aquifer matrix into groundwater. The granite inthe area contains abundant fluoride and during weathering,fluoride can leach and dissolve the aquifers. The oversatura-tion of samples with respect to calcite and undersaturationwith respect to fluorite makes it feasible for fluoride to getreleased in groundwater. The climate coupled with geochem-ical processes is found to be main controlling factors forhigher concentration of fluoride.

Acknowledgments The authors are most grateful to the BrianMailloux and Alexander van Geen of Columbia University for carry-ing out Ion chromatography of water samples which were tested onsiteusing Hach field kit for fluoride. The authors also wish to thank all ofthe research team members that participated in the sampling andsample analysis

References

Adriano DC (1986) Trace elements in the terrestrial environment.Springer, New York

Agrawal V, Vaish AK, Vaish P (1997) Groundwater quality: focus onfluoride and fluorosis in Rajasthan. Curr Sci 73(9):743–746

American Public Health Association (APHA) (2007) Standard methodsfor the examination of water and waste water, 21st edn. AmericanPublic Health Association, Washington DC

Ayenew T (2008) The distribution and hydrogeological controls of fluo-ride in the groundwater of central Ethiopian rift and adjacent high-lands. Environ Geol 54:1313–1324

Banerjee A (2014) Groundwater fluoride contamination: a reappraisal.Geosci Front. doi:10.1016/j.gsf.2014.03.003

Cao J, Zhao Y, Lin JW, Xirao RD, Danzeng SB (2000) Environmentalfluoride in Tibet. Environ Res 83:333–337

Chae GT, Yun ST, Bernhard M, Kim KH, Kim SY, Kwon JS, Kwon K,Koh YK (2007) Fluorine geochemistry in bedrock groundwater ofSouth Korea. Sci Total Environ 385:272–283

Chakrabarti S, Bhattacharya HN (2013) Inferring the hydro-geochemistryof fluoride contamination in Bankura district, West Bengal: a casestudy. J Geol Soc India 82(4):379–391

Coetsiers M, Walraevens K (2006) Chemical characterization of theneogene aquifer, Belgium. Hydrogeol 14(8):1556–1568

Dai Z, Samper J, Ritzi R (2006) Identifying geochemical processes byinverse modeling of multicomponent reactive transport in the Aquiaaquifer. Geosph 2(4):210–219

Datta PS, Deb DL, Tyagi SK (1996) Stable isotope (O18) investigationson the processes controlling fluoride contamination of groundwater.J Contam Hydrol 24:85–96

Deotare BC, Kajale MD, Rajaguru SN, Basavaiah N (2004) Late quater-nary geomorphology, palynology and magnetic susceptibility ofplayas in western margin of the Indian Thar Desert. J IndianGeophys Union 8(1):15–25

Edmunds M, Smedley P (2005) Fluoride in natural waters. In: Selnius O,Alloway B, Centeno JA, Finkleman RB, Fuge R, Lindh U, Smedley

Environ Sci Pollut Res

P (eds) Essentials of medical geology-impacts of the natural envi-ronment on public health. Academic, Amsterdam

Fawell J, Bailey K, Chilton J, Dahi E, Fewtrell L, Magara Y (2006)Fluoride in drinking water. IWA, London

Fordyce FM, Vrana K, Zhovinsky E, Povoroznuk V, Toth G, Hope BC,Iljinsky U, Baker JA (2007) Health risk assessment for fluoride inCentral Europe. Environ Geochem Health 29:83–102

Frencken JE (1992) Endemic Fluorosis in developing countries, causes,effects and possible solutions. Publication number 91.082, NIPG-TNO, Leiden, The Netherlands

Garrels RM, Mackenzie FT (1971) Evolution of sedimentary rocks.Norton, New York

Genxu W, Guodong C (2001) Fluoride distribution in water and thegoverning factors of environment in the arid and north-westChina. J Arid Environ 49:601–614

Güler C, Thyne GD (2004) Hydrologic and geologic factors controllingsurface and groundwater chemistry in Indian Wells-Owens Valleyarea, southeastern California, USA. J Hydrol 285(1):177–198

Guo QH, Wang YX, Ma T, Ma R (2007) Geochemical processes con-trolling the elevated fluoride concentrations in groundwaters of theTaiyuan basin, Northern China. J Geochem Explor 93:1–12

Gupta S, Mondal D, Bardhan A (2012) Geochemical provenanceand spatial distribution of fluoride in groundwater in parts ofRaniganj coal field, West Bengal, India. Arch Appl Sci Res 4(1):292–306

Handa BK (1975) Geochemistry and genesis of fluoride containinggroundwater in India. Ground Water 13:275–281

Hidalgo MC, Cruz-Sanjulian J (2001) Groundwater composition,hydrochemical evolution and mass transfer in a regionaldetrital aquifer (Baza basin, southern Spain). Appl Geochem 16(7):745–758

Hounslow A (1995) Water quality data: analysis and interpretation, 1stedn. CRC Lewis, Boca Raton

Hussain I, Arif M, Hussain J (2012) Fluoride contamination in drinkingwater in rural habitations of Central Rajasthan, India. EnvironMonitAssess 184(8):5151–5158

Hussain J, Husain I, Arif M (2013) Fluoride contamination in groundwa-ter of central Rajasthan, India and its toxicity in rural habitants.Toxicol Environ Chem 95(6):1048–1055

Jacks G, Bhattacharya P, Chaudhary V, Singh KP (2005) Controls on thegenesis of some high-fluoride groundwater in India. Appl Geochem20:221–228

Jha SK, Nayak AK, Sharma YK (2010) Potential fluoride contaminationin the drinking water of Marks Nagar, Unnao district, Uttar Pradesh,India. Environ Geochem Health 32:217–226

Kim K, Jeong G (2005) Factors influencing natural occurrence offluoride-rich groundwaters: a case study in the southeastern part ofthe Korean Peninsula. Chemosphere 58:1399–1408

Kim SH, Kim K, Ko KS, Kim Y, Lee KS (2012) Co-contamination ofarsenic and fluoride in the groundwater of unconsolidated aquifersunder reducing environments. Chemosphere 87:851–856

Krauskopf KB, Bird DK (1995) Introduction to geochemistry. McGraw-Hill, New York

Kuells C, Adar EM, Udluft P (2000) Resolving patterns of ground waterflow by inverse hydrochemical modeling in a semiarid Kalaharibasin. Tracers Model Hydrogeol IASH Publ 262:447–451

Madhavan N, Subramanian V (2002) The natural abundance offluoride in soils of the Ajmer district, Rajasthan. J Environ Monit4:821–822

Maithani PB, Gurjar R, Banerjee R, Balaji BK, Ramachandran S, SinghR(1998) Anomalous fluoride in groundwater from western part ofSirohi district, Rajasthan and its crippling effects on human health.Curr Sci 74(9):773–777

Msonda KWM, Masamba WRL, Fabiano E (2007) A study of fluoridegroundwater occurrence in Nathenje, Lilongwe, Malawi. PhysChem Earth 32:1178–1184

Naseem S, Rafique T, Bashir E, Bhanger MI, Laghari A, Usmani TH(2010) Lithological influences on occurrence of high-fluoridegroundwater in Nagar Parkar area, Thar Desert, Pakistan.Chemosphere 78(11):1313–1321

Oruc N (2008) Occurrence and problems of high fluoride water inTurkey: an overview. Environ Geochem Health 30:315–323

Ozsvath DL (2006) Fluoride concentrations in a crystalline bed-rock aquifer Marathon County, Wisconsin. Environ Geol 50(1):132–138

Parkhurst DL, Appelo CAJ (1999) User’s guide to PHREEQC (Version2)-A Computer Program for Speciation, Batch-Reaction, One-Dimensional Transport, and Inverse Geochemical Calculations.United States Geological Survey, Water Resources InvestigationsReport 99–4259, Washington, DC, 326

Pickering WF (1985) The mobility of soluble fluoride in soil. EnvironPollut 9:281–308

Plummer LN, Parkhurst DL, Thorstenson DC (1983) Development ofreaction models for ground-water systems. Geochim CosmochimActa 47:665–686

Rafique T, Naseem S, Bhanger MI, Usmani TH (2008) Fluoride ioncontamination in the groundwater of Mithi sub-district, the TharDesert, Pakistan. Environ Geol 56:317–332

Rai V, Sinha AK (1990) Geological evolution of Kuchaman Lake, districtNagaur, Rajasthan. J Palaeontological Soc India 35:137–142

Rao NVR, Rao N, Rao KSP, Schuiling RD (1993) Fluorine distribution inwaters of Nalgonda District, AP, India. Environ Geol 21:89

Rina K, Datta PS, Singh CK, Mukherjee S (2013) Characterization andevaluation of processes governing the groundwater quality in partsof the Sabarmati basin, Gujarat using hydrochemistry integratedwith GIS. Hydrol Process 26(10):1538–1551

Rina K, Datta PS, Singh CK, Mukherjee S (2014) Determining thegenetic origin of nitrate contamination in aquifers of NorthernGujarat, India. Environ Earth Sci 71(4):1711–1719

Rwenyonyi CM, Birkeland JM, Haugejorden O, Bjorvatn K (2000) Ageas a determinant of severity of dental fluorosis in children residing inareas with 0.5 and 2.5 mg fluoride per liter in drinking water. ClinOral Investig 4:157–161

Singh CK, Kumari R, Singh RP, Shashtri S, Kamal V, Mukherjee S(2011a) Geochemical modeling of high fluoride concentration ingroundwater of Pokhran area of Rajasthan, India. Bull EnvironContam Toxicol 86(2):152–158

Singh CK, Shashtri S, Mukherjee S (2011b) Integrating multivariatestatistical analysis with GIS for geochemical assessment of ground-water quality in Shiwaliks of Punjab, India. Environ Earth Sci 62(7):1387–1405

Singh CK, Shashtri S, Mukherjee S, Kumari R, Avatar R, Singh A, SinghRP (2011c) Application of GWQI to assess effect of land use changeon groundwater quality in lower Shiwaliks of Punjab: remote sens-ing and GIS based approach. Water Resour Manag 25(7):1881–1898

Singh CK, Kumari R, Singh N, Mallick J, Mukherjee S (2012)Fluoride enrichment in aquifers of the Thar Desert: controllingfactors and its geochemical modeling. Hydrol Process. doi:10.1002/hyp.9247

Singh CK, Rina K, Singh RP, Mukherjee S (2014) Geochemical charac-terization and heavy metal contamination of groundwater in SatlujRiver Basin. Environ Earth Sci 71(1):201–216

Suthar S, Garg VK, Jangir S, Kaur S, Goswami N, Singh S (2008)Fluoride contamination in drinking water in rural habitations ofNorthern Rajasthan, India. Environ Monit Assess 145(1–3):1–6

Suttie JW (1977) Effects of fluoride on livestock. J Occup Med 19:40–48Veksler IV, DorfmanAM, KamenetskyM, Dulski P, Dingwell DB (2005)

Partitioning of lanthanides and Y between immiscible silicate andfluoride melts, fluorite and cryolite and the origin of the lanthanidetetrad effect in igneous rocks. Geochim Cosmochim Acta 69(11):2847–2860

Environ Sci Pollut Res

Vieira APGF, Hanocock R, Eggertsson H, Everett ET, Grynpas MD(2005) Tooth quality in dental fluorosis: genetic and environmentalfactors. Calcif Tissue Int 76:17–25

Wenzel W, Blum WEH (1992) Fluoride speciation and mobility influoride contaminated soil and minerals. Soil Sci 153:357–364

Wu Y, Wang Y (2014) Geochemical evolution of groundwatersalinity at basin scale: a case study from Datong basin, northern

China. Environ Sci Process Impacts. doi:10.1039/C4EM00019F

Zhang B, Zhao HM, Yongsheng XL, Xuelin Z, Jun D (2003) Distributionand risk assessment of fluoride in drinking water in the west plainregion of Jilin province, China. Environ Geochem Health 25:421–431

Zhu C, Anderson G (2002) Environmental application of geochemicalmodeling. Cambridge University Press, Cambridge

Environ Sci Pollut Res