Seediscussions,stats,andauthorprofilesforthispublicationat:https://www.researchgate.net/publication/258225906

Comparisonofmethodsfordistinguishingsodiumcarbonateactivatedfromnaturalsodiumbentonites

ARTICLEinAPPLIEDCLAYSCIENCE·DECEMBER2013

ImpactFactor:2.47·DOI:10.1016/j.clay.2013.09.014

CITATIONS

2

READS

165

5AUTHORS,INCLUDING:

StephanKaufhold

BundesanstaltfürGeowissenschaftenund…

101PUBLICATIONS765CITATIONS

SEEPROFILE

KatjaEmmerich

KarlsruheInstituteofTechnology

55PUBLICATIONS484CITATIONS

SEEPROFILE

AnnettSteudel

KarlsruheInstituteofTechnology

16PUBLICATIONS121CITATIONS

SEEPROFILE

Allin-textreferencesunderlinedinbluearelinkedtopublicationsonResearchGate,

lettingyouaccessandreadthemimmediately.

Availablefrom:AnnettSteudel

Retrievedon:04February2016

Applied Clay Science 86 (2013) 23–37

Contents lists available at ScienceDirect

Applied Clay Science

j ourna l homepage: www.e lsev ie r .com/ locate /c lay

Research paper

Comparison of methods for distinguishing sodium carbonate activatedfrom natural sodium bentonites

S. Kaufhold a,⁎, K. Emmerich b,c, R. Dohrmann a,d, A. Steudel c, K. Ufer a

a BGR, Bundesanstalt für Geowissenschaften und Rohstoffe, Stilleweg 2, D-30655 Hannover, Germanyb CMM, Competence Center for Material Moisture, Karlsruhe Institute of Technology, H.-v.-Helmholtz-Platz 1, D-76344 Eggenstein-Leopoldshafen, Germanyc IFG, Institute for Functional Interfaces, Karlsruhe Institute of Technology, H.-v.-Helmholtz-Platz 1, D-76344 Eggenstein-Leopoldshafen, Germanyd LBEG, Landesamt für Bergbau, Energie und Geologie, Stilleweg 2, D-30655 Hannover, Germany

⁎ Corresponding author.E-mail address: s.kaufhold@bgr.de (S. Kaufhold).

0169-1317/$ – see front matter © 2013 Elsevier B.V. All rihttp://dx.doi.org/10.1016/j.clay.2013.09.014

a b s t r a c t

a r t i c l e i n f o

Article history:Received 13 August 2013Received in revised form 26 September 2013Accepted 29 September 2013Available online xxxx

Keywords:Soda activationBentoniteThermal analysisCEC

A lot of natural Ca/Mg-bentonites are turned intoNa-bentonites. By adding sodiumcarbonate to a Ca/Mg-benton-ite in the presence of some water Na+ enters the interlayer and Ca/Mg-carbonates precipitate outside. NaturalNa-bentonites and activated Na-bentonites are rather similar. Therefore, in the present studymethods are testedto distinguish both. This is relevant not only for customs but also for research and development. For activationdifferent amounts of sodium carbonate are added. The dosage ranges from a few % of the CEC to slightly abovethe CEC corresponding to 1–5 mass% Na2CO3. Also the water content may vary from the dried state at whichthe actual activation (cation exchange) does not take place up to the presence of excess water leading to a com-plete reaction. Altogether four cases had to be considered separately (Na2CO3 above CEC+excess water or dryandNa2CO3much below the CEC+excesswater or dry). Ifwaterwas absent (cation exchangewasnot complete)the sodium carbonate phases could be detected by XRD, IR, or with STA–MS measurements. This result wasexpected but surprisingly, STA–MS–CO2 measurements were found to be applicable even in the most difficultcase (sodium carbonate addition below CEC and excess of water=reaction complete). In the case of some sam-ples activated with 2 mass% sodium carbonate only, a weak STA–MS–CO2-peak was observed at about 100 °C.Unprocessed materials are free of any carbonates which decompose around 100 °C. Therefore, this 100 °C peakindicates alkaline activation. This method was applied to five real products with unknown activation and twoofwhichwere found to be activated. The pH of the activatedmaterials was only slightly higher than that of a nat-ural Na-bentonite. The measured difference of 0.3 pH units is not considered to be sufficient to unambiguouslyconclude alkaline activation.

© 2013 Elsevier B.V. All rights reserved.

1. Introduction

A large amount of bentonite is used because of its rheological prop-erties such as thixotropy. Examples are bentonite drilling mud andpaint additive. The rheological properties along with others (such aswater uptake; Kaufhold et al., 2010) strongly depend on the type ofinterlayer cation balancing the permanent charge of the smectites.The common exchangeable cations are Ca2+ and Mg2+. However, afew bentonites (smectites) are dominated by exchangeable Na+ ase.g. the famous Wyoming bentonites (Grim and Güven, 1978). To im-prove the rheological properties of Ca/Mg-bentonites, Hofmann andEndell (1936) suggested the use of sodium carbonate for the exchangeof interlayer Ca2+ or Mg2+ by Na+. Large amounts of NaCl would benecessary because Ca2+ and Mg2+ are selectively adsorbed in contrastto Na+ (e.g. Lagaly, 1993; McBride, 1979). The presence of the carbon-ate or hydrogencarbonate anion, however, leads to the formation of

ghts reserved.

rather insoluble Ca/Mg-carbonates (Lagaly et al., 1981). Precipitationof these carbonates occurs outside the interlayer. Therefore, Ca2+ andMg2+ have to leave the interlayer which in turn indirectly promotesthe cation exchange.

The optimum amount of sodium carbonate to be added to a Ca/Mg-bentonite is commonly determined empirically by investigation of therheology of a set of different bentonites with different ratios of sodiumcarbonate and bentonite. Theoretically, the optimum amount should cor-respond to the CEC and further addition should not change the bentoniteproperties anymore. However, adding slightly more sodium carbonatethan the CEC often yields better results (e.g. rheology), which is probablyrelated to pH effects (Alther, 1986). Assuming a bentonite with a CEC of90meq/100g, considering the molar mass of sodium carbonate account-ing for almost 106 g/mol, and considering the two charges of both Na+

cations of the sodium carbonate leads to 4.8mass% sodium carbonate atwhich the CEC is reached. However, the CEC of most bentonites islower. In the case of 70meq/100g CEC of a bentonite 3.7mass% sodiumcarbonate would be sufficient to saturate all charges.

Technically activated bentonites are produced at production linesall over the world, packed, and shipped to their destination. For

24 S. Kaufhold et al. / Applied Clay Science 86 (2013) 23–37

international shipping different documents are required including acustoms declaration (technically produced or natural materials arecommonly distinguished). For the customs, controlling custom regula-tions is impossible without a suitable method. Distinguishing bothtypes of materials is also relevant in research and development: Whencomparing different products, it is essential to know whether andhow a bentonite was activated or not. Activation is often carried out to-gether with strong shearing (e.g. in extruders) which in turn results inimproved bentonite properties (induced by shearing). Any scientificwork carried out on bentonite, therefore, should be based on materialwith known history.

Different analytical tools were suggested to unambiguously distin-guish natural Na-bentonites from alkaline activated materials. Fahn(1964) suggested to consider alkalinity, pH of the dispersion, CO2 andNa2O3 content, and water soluble salts to detect activation. However,the pH value of natural Na-bentonites can exceed 10 (Kaufhold et al.,2008) and bentonites from some deposits are known to contain naturalsoluble salts. Although time consuming the method proposed by Fahn(1964) is supposed to work for most bentonites but could fail in thecase of natural Na-bentonites with soluble salts (e.g. Morocco).Steudel et al. (2013) proposed using thermal analysis for the detectionof technically activated Na-bentonite. However, Steudel et al. (2013)only used one type of sodium carbonate containing the characteristicimpurities such as thermonatrite and trona. Hence, in order to assessthe proposed methods in the present study different sodium carbonatesamples and a set of well characterized bentonites were used. The tech-nical activation process is commonly run with 30–40 mass% waterwhich may not be sufficient for completing the activation. Hencesome undissolved and unreacted sodiumcarbonatemay still be present.Therefore, in order to be representative of real systems, the water con-tent before and after addition of the sodium carbonate was varied.

The aim of the present study is to compare different methods withrespect to the possibility to unambiguously distinguish technically acti-vated Na-bentonites from natural ones.

Table 1Materials used for the present study.

Sample ID Material Informa

S1 Na2CO3 TechnicS2 Na2CO3 MerckS3 Na2CO3 Merck NaHCOS4 Na2CO3 Trona (P1 Industrial bentonite product CatsanP2 Real bentonite product CatsanP3 Real bentonite product TechnicP4 Real bentonite product UnknowP5 Real bentonite product WaldoBN1 Natural Na bentonite SampleBN2 Natural Na bentonite SampleBN3 Natural Na bentonite SampleLA6-0 Bentonite before activation SampleLA6-2A Bentonite. Lab activated B6 withLA6-2D Bentonite. Lab activated B6 withLA6-5A Bentonite. Lab activated B6 withLA6-5D Bentonite. Lab activated B6 withLA16-0 Bentonite before activation SampleLA16-2A Bentonite. Lab activated B16 wiLA16-2D Bentonite. Lab activated B16 wiLA16-5A Bentonite. Lab activated B16 wiLA16-5D Bentonite. Lab activated B16 wiLA19-0 Bentonite before activation sample B19 (ALA19-2A Bentonite. Lab activated B19 wiLA19-2D Bentonite. Lab activated B19 wiLA19-5A Bentonite. Lab activated B19 wiLA19-5D Bentonite. Lab activated B19 wi

More information about bentonites B6, B8, B11, B16, B19, and B24 is provided by:BGR project publications: Kaufhold and Dohrmann (2008), Kaufhold and Dohrmann (2009), KUfer et al. (2008).

2. Materials and methods

The materials used are listed in Table 1.Alkaline bentonite activation is based on the reaction of a Ca/Mg-

bentonite with Na2CO3 which can be applied as different phases/prod-ucts. For the activation technical Na2CO3 rather than p.a. or food gradeis required. The technical productmay varywith respect to the contentsof different phases,water content, and other chemical impurities. Anhy-drous sodium carbonate is relatively rare because it tends to react withwater to at least form themono-hydrate (thermonatrite). The followingphases may occur in technical sodium carbonate (TSC):

1) Na2CO3=natrite (anhydrous sodium carbonate)2) Na2CO3 ∗H2O=thermonatrite (sodium carbonate monohydrate)3) Na2CO3 ∗ 7H2O = bno mineral nameN (sodium carbonate

heptahydrate)4) Na2CO3 ∗10H2O=natron (sodium carbonate decahydrate)5) Na3(CO3)(HCO3) ∗ 2H2O = trona (sodium carbonate hydrogen

carbonate)6) NaHCO3=nahcolite (sodium hydrogen carbonate).

Because of the variety of phases which may occur in TSC products,different types of sodium carbonate were used (samples beginningwith #S). Samples beginning with #P are actual industrial productswith limited information about a possible alkaline activation. #BN sam-ples are natural Na-rich bentonites which were not in contact with anysodium carbonate. The #LA samples represent Ca/Mg-bentonites (ben-tonites inwhich the interlayer is dominated by Ca2+ and/orMg2+). #LAsamples were treated differently in laboratory. The #LA-0 samples arethe untreated precursor materials of those samples which were modi-fied by sodium carbonate. Both the amount of sodium carbonate andthe type of addition/activation were varied. For technical activation dif-ferentwater contents are used. The actual activation, however, does nottake place in the absence ofwater. The actual activationwhich is the cat-ion exchange of dry activatedmaterials takes place if water is added, e.g.

tion Reference

al Na2CO3 provided by S&B, Landshut S&B, LandshutMerck

3 Merckmineral piece) Owens Lake, CA, USclumbing Mars Inc.Ultra Mars Inc.ally activated with 3% S1 S&B, Landshutn activation S&B, Marl

North (MK11) Klinkenberg (2008)B8 (Wyoming) BGR projects (details below)B11 (India) BGR projects (details below)B24 (Morocco) BGR projects (details below)B6 (Milos), natural state BGR projects (details below)S2 (2mass%, added in dry state) Produced in laboratoryS2 (2mass%, dispersed and dried) Produced in laboratoryS2 (5mass%, added in dry state) Produced in laboratoryS2 (5mass%, dispersed and dried) Produced in laboratoryB16 (Bavaria), natural state BGR projects (details below)th S2 (2mass%, added in dry state) Produced in laboratoryth S2 (2mass%, dispersed and dried) Produced in laboratoryth S2 (5mass%, added in dry state) Produced in laboratoryth S2 (5mass%, dispersed and dried) Produced in laboratorylmeria), natural state BGR projects (details below)th S2 (2mass%, added in dry state) Produced in laboratoryth S2 (2mass%, dispersed and dried) Produced in laboratoryth S2 (5mass%, added in dry state) Produced in laboratoryth S2 (5mass%, dispersed and dried) Produced in laboratory

aufhold and Dohrmann (2010a), Kaufhold and Dohrmann (2010b), Kaufhold et al. (2010),

25S. Kaufhold et al. / Applied Clay Science 86 (2013) 23–37

in the case of producing drillingmud bymixingwater and clay powder.Hence the ratio of reacted sodium carbonate (the Na+ alreadyexchanged for Ca/Mg and the carbonate anion in turn precipitated asCa/Mg-carbonate) to free sodium carbonate (still waiting for somewater molecules facilitating the activation) is not known and immedi-ately changes if water is added (as long as free sodium carbonate andexchangeable Ca/Mg are present). Therefore in the present study bothpossibilities are considered, the air dry state (water content about 10–15%) on one hand and the presence of excess water on the other handto complement the reaction. Furthermore, the amount of sodium car-bonate which is technically added to Ca/Mg-bentonites varies from 2to 5mass%. Commonly 3–4mass% is used for activation which is deter-mined empirically e.g. considering rheological parameters. In some in-stances only partial activation is used. Consequently, in the presentstudy the amount of sodium carbonate added was varied from 2mass%,representing activation below CEC saturation, to 5mass% correspondingto a Na+ concentration slightly larger than the CEC. Reaction time was24 h and samples were shaken end-over-end. Finally all samples weredried at 60 °C.

XRD patterns were recorded using a PANalytical X'Pert PROMPD Θ-Θ diffractometer (Amelo, Netherlands, Cu-Kα radiation generated at40kV and 30mA), equippedwith a variable divergence slit (20mm irra-diated length), primary and secondary soller, Scientific X'Celeratordetector (active length=0.59°), and a sample changer (sample diame-ter=28mm). The samples were investigated from 2° to 85° 2Θ with astep size of 0.0167° 2Θ and a measuring time of 10s per step. For spec-imen preparation the top loading technique was used.

In addition XRD patterns were recorded after stepwise increasingthe relative humidity. The relative humidity (RH) was increased by10% each 6 h and started at 10% RH. Measurements were conductedwith a PANalytical (Amelo, Netherlands) X'Pert PRO MPD Θ-Θ diffrac-tometer (Cu-Kα radiation generated at 40 kV and 40 mA) which wasequipped with an Anton Paar CHC non-ambient chamber flushed withair of different relative humidity at 25 °C. The measurements wereperformed using a proportional counter as detector and a variabledivergence slit (10mm irradiated length). The samples were investigat-ed from 5° to 80° 2Θwith a step size of 0.03° 2Θ. 42min was chosen asmeasuring time per scan to ensure that at least 8 measurements werecollected under identical RH. For specimen preparation again the toploading technique was used.

The total carbon (TC) content was measured with a LECO(Mönchengladbach, Germany) CS-444-Analysator. Samples of 170–180 mg of the dried material were used. The samples were heated inthe device to 1800–2000 °C in an oxygen atmosphere and the CO2 wasdetected by an infrared detector.

For measuring mid infrared (MIR) spectra the KBr pellet technique(1 mg sample/200 mg KBr) was applied. Spectra were collected on aThermo Nicolet (Dreieich, Germany) Nexus FTIR spectrometer (MIRbeam splitter: KBr, detector DTGS TEC; FIR beam splitter: solid sub-strate, detector DTGS PE). The resolution was adjusted to 2 cm−1.

Simultaneous thermal analysis (STA) was performed on a STA449 C Jupiter (NETZSCH-Gerätebau GmbH, Selb, Germany) equippedwith a thermogravimetry/differential scanning calorimetry (TG/DSC) sample holder. The STA is connected by a heated quartz glasscapillary to a quadrupole mass spectrometer 403 C Aëolos (InProcessInstruments (IPI)/NETZSCH-Gerätebau GmbH). Samples were heat-ed to 35–1100 °C with 10 K/min in streaming synthetic air (SA,50mL/min)/nitrogen (N2 20mL/min) atmosphere. Conventional Pt/Rh crucibles with lid (diameter = 5 mm and height = 5 mm) werefilled with 50mg of sample material. An empty Pt/Rh crucible servedas reference. Sodium carbonate samples (S1–S4) were diluted withmetakaolin with a ratio of 10:90 to prevent corrosion of Pt-parts incrucibles, sample holder and thermocouples. 100 mg of thesesamples were used for STA measurements. Prior to STA, sampleswere kept in closed containers at ambient laboratory conditions(T≈ 20–25 °C and 45–55% RH).

The sodium carbonate activation leads to cation exchange. The cat-ion population, in turn, is commonly measured using CEC methods.Because of the importance of the difference of the sum of exchangeablecations (‘sum’) and the amount of index cation adsorbed (CEC) differentCEC methods were applied:

– Cu-trien method (as standard method, Kaufhold and Dohrmann,2003; Meier and Kahr, 1999).

– Cu-trien5x (increasing the Cu-trien concentration (=decreasing theamount of water) reduces the solubility of the partly soluble compo-nents, mainly calcite, Dohrmann and Kaufhold (2009).

– Cu-trien5xcalcite (calcite saturation of the exchange solutionsfurther suppresses dissolution of calcite, Dohrmann and Kaufhold,2009).

– NH4Cl (ammoniumchlorid method carried out with ethanol is sup-posed to reduce carbonate dissolution, Belyayeva, 1967; Dohrmannet al., 2012).

The pHwasmeasured using aWTW Inolab pH-meter (WTWGmbH,Weilheim, Germany) equipped with a WTW pH electrode SenTix 81.The electrode was calibrated using buffer solutions of pH 4, 7, and 10.The pH of the dispersion was measured with continuous stirring. ThepH values were recorded after 5, 10, and 15 min in order to ensurethat equilibrium was attained.

The electrical conductivity was measured with the conductometerHI 9033 HANNA.

To determine dissolved carbonate, a CaCl2 solution was added andthe solution was inspected with respect to the possible formation ofany turbidity which in turn would indicate free (dissolved) carbonate.

3. Results and discussion

3.1. XRD

The sodium carbonate samples (S-samples) were analyzed byXRD few days before addition to the bentonite samples withoutany pretreatment. In this state, S1 and S2 were dominated by natritefollowed by thermonatrite and traces of trona. XRD pattern werenearly identical. S3 was nearly pure nahcolite with traces of trona;S4 was dominated by trona with minor amounts of burkeite, halite,and possibly nahcolite. None of the samples was pure natritewhich indicates that anhydrous sodium carbonate (natrite) readilyhydrates and that the occurrence of the hydrates is a characteristicfeature of sodium carbonate. To identify the conditions at whichpurely anhydrous sodium carbonate exists, the Merck sample S2was dried at 105 °C for four days and then investigated by XRDwith stepwise increasing the relative humidity. Every 6h the relativehumidity was increased by 10%. At the beginning, the previouslydried Merck sample consisted of natrite only. The formation of hy-dration products (thermonatrite) was observed above 70% RHwhich would indicate that the hydrous sodium carbonate phaseswere not typical minor components, i.e. that storage of anhydroussodium carbonate may be possible on industrial scale if the RH re-mains below 70%. The formation of hydrous sodium carbonate oc-curred at unexpectedly large relative humidity (above 70%). To testif kinetic aspects were important, anhydrous sodium carbonatewas stored at 30% relative humidity in a climate oven for 4 days(Fig. 2). Afterwards thermonatrite was found by XRD which at firstglance contradicts the climate chamber XRDmeasurements. Howev-er, the 4day storage test indicates that most of the technical sodiumcarbonate used for activation may contain some hydrous sodiumcarbonates. The XRD measurements prove that i) TSC if not storedunder dry conditions will contain some hydrous sodium carbonatephases and ii) the ratio of hydrous phase to anhydrous phase mayvary depending on the ambient conditions. Notably, the hydrationand formation of even more trona in S1 & S2 investigated by thermalanalysis was more advanced compared to Fig. 1. This may be

Fig. 1. XRD patterns of the four sodium carbonate samples (S1 bottom (black), S2 (blue), S3 (green), S4 top (red)). [Å].

26 S. Kaufhold et al. / Applied Clay Science 86 (2013) 23–37

explained by hydration throughout storage. This was proven bycomparing XRD traces measured in both labs. The XRD traces of thedifferently treated samples #B06, #B16, and #B19 are given in Fig. 3.

Samples activated with excess water showed formation of calcitewhichwas detected by XRD. The dry added samples were not contacted

Fig. 2. XRD patterns of Merck sodium carbonate predried at 105 °C before (black line) and aftpositions of thermonatrite peaks.

with water prior to XRD, hence the natrite reflections were found. Themain reflection could be resolved even when 2mass% only was added.In the case of sample #LA16 dissolution of gypsum throughout the dis-persion step (“D” samples) was observed. The XRD traces of the actualproducts are given in Fig. 4.

er storage at 30% relative humidity for four days (red line). The black arrows indicate the

Fig. 3.XRDpatterns of the differently activated samples #B06 (left) and#B16 (right); bottom, red: precursormaterial (#LA6-0 and#LA16-0, respectively), blue: -2A, green: -2D, pink: -5A,and black: -5D. Only the peaks relevant for the study were assigned.

27S. Kaufhold et al. / Applied Clay Science 86 (2013) 23–37

In the products neither natrite nor thermonatrite could be detectedwhich indicates that – if activated at all – activation was performedwith elevated water contents. The sodium carbonate component could

Fig. 4. XRD patterns of the actual industrial products (from bottom to top: #P1–#P5, natrite XR

not be detected in the solid state. In these XRD traces, of course, nodistinction can be made between natural calcite and activation derivedcalcite.

D pattern (blue) given as reference). Only the peaks relevant for the study were assigned.

Table 2CEC equivalents obtained by investigation of pure sodium carbonate samples withdifferent CEC methods. n.a.= not analyzed.

Mass Na K Mg Ca Sum CEC Sum–CEC

Soda 1 Technical Cu–trien 0.080 1800 0.0 0.0 0.0 1800 6 1794

Na2CO3 Cu–trien 0.120 1800 0.0 0.3 0.0 1800 3 1797

provided by Cu–trien5x 0.400 1950 0.0 0.2 0.3 1950 10 1940

S&B, Cu–trien5x 0.600 1950 0.0 0.2 0.3 1950 8 1942

Landshut Cu–trien5xcalcite 0.400 n.a. n.a. 0.1 0.0 n.a. 7 n.a.

Cu–trien5xcalcite 0.600 n.a. n.a. 0.1 0.0 n.a. 6 n.a.

NH4Cl–ethanolic 0.080 768 0.0 0.0 0.0 768 n.a. n.a.

NH4Cl–ethanolic 0.120 511 0.2 0.0 0.0 511 n.a. n.a.

Soda 2 Merck Cu–trien 0.080 1850 1.5 0.6 0.5 1850 4 1846

Cu–trien 0.120 1800 0.0 0.0 0.0 1800 4 1796

Cu–trien5x 0.400 2000 0.5 0.4 0.3 2000 10 1990

Cu–trien5x 0.600 1950 0.0 0.1 0.2 1950 8 1942

Cu–trien5xcalcite 0.400 n.a. n.a. 0.5 0.0 n.a. 9 n.a.

Cu–trien5xcalcite 0.600 n.a. n.a. 0.0 0.0 n.a. 8 n.a.

NH4Cl–ethanolic 0.080 773 1.1 0.0 0.4 775 n.a. n.a.

NH4Cl–ethanolic 0.120 525 0.0 0.0 0.0 525 n.a. n.a.

Soda 3 Merck Cu–trien 0.080 1740 0.0 0.0 0.0 1740 2 1738

NaHCO3 Cu–trien 0.120 1700 0.0 0.0 0.0 1700 2 1698

Cu–trien5x 0.400 1950 0.6 0.1 0.1 1950 8 1942

Cu–trien5x 0.600 1900 0.0 0.1 0.1 1900 6 1894

Cu–trien5xcalcite 0.400 n.a. n.a. 0.0 0.0 n.a. 6 n.a.

Cu–trien5xcalcite 0.600 n.a. n.a. 0.1 0.0 n.a. 6 n.a.

NH4Cl–ethanolic 0.081 899 0.1 0.0 0.0 899 n.a. n.a.

NH4Cl–ethanolic 0.121 590 0.0 0.0 0.0 590 n.a. n.a.

Soda 4 Trona Cu–trien 0.080 1600 10.5 0.0 0.0 1600 3 1597

mineral Cu–trien 0.120 1600 10.6 0.0 0.0 1600 3 1597

piece, Cu–trien5x 0.400 1800 10.7 0.0 0.2 1800 8 1792

Owens Cu–trien5x 0.600 1700 10.7 0.0 0.2 1700 6 1694

Lake, Cu–trien5xcalcite 0.400 n.a. n.a. n.a. n.a. n.a. n.a. n.a.

CA, US Cu–trien5xcalcite 0.600 n.a. n.a. n.a. n.a. n.a. n.a. n.a.

NH4Cl–ethanolic 0.082 914 1.3 0.0 0.0 915 n.a. n.a.

NH4Cl–ethanolic 0.120 725 1.1 0.0 0.0 726 n.a. n.a.

(meq/100 g)(g)

Table 3CEC results of the three natural Na bentonites (four different methods).

Mass Na K Mg Ca Sum CEC Sum–CEC

BN1 Sample B08 Cu–trien 0.080 66.3 0.0 0.6 18.4 86 62 23

Wyoming Cu–trien 0.120 65.5 0.0 0.4 16.2 82 63 19

Natural state Cu–trien5x 0.400 68.2 0.3 0.3 11.6 81 65 15

Cu–trien5x 0.600 69.0 0.3 0.2 9.3 79 65 14

LECO 0.3 mass% C Cu–trien5xcalcite 0.400 69.4 0.8 0.5 4.5 75 66 9

Nominally 2.5 mass % Cc Cu–trien5xcalcite 0.600 68.3 0.2 0.1 4.3 73 65 8

NH4Cl–ethanolic 0.080 68.0 0.7 0.9 3.1 73 n.a. n.a.

NH4Cl–ethanolic 0.120 66.6 0.6 0.7 1.3 69 n.a. n.a.

BN2 Sample B11 Cu–trien 0.080 57.3 0.0 16.3 30.3 104 88 16

India Cu–trien 0.120 56.9 0.6 15.7 29.8 103 88 15

Natural state Cu–trien5x 0.400 59.3 0.7 15.8 27.1 103 92 11

Cu–trien5x 0.600 58.9 0.7 15.6 25.4 101 91 10

LECO 0.2 mass% C Cu–trien5xcalcite 0.400 59.2 0.6 15.2 19.4 95 92 2

Nominally 1.7 mass % Cc Cu–trien5xcalcite 0.600 58.9 0.6 15.2 19.8 95 91 4

NH4Cl–ethanolic 0.082 53.6 0.7 17.7 22.4 94 n.a. n.a.

NH4Cl–ethanolic 0.120 53.9 0.5 16.6 21.2 92 n.a. n.a.

BN3 Sample B24 Cu–trien 0.080 44.9 3.7 26.5 33.1 108 91 18

Morocco Cu–trien 0.120 44.7 1.8 26.1 32.7 105 90 15

Natural state Cu–trien5x 0.400 47.1 2.5 26.5 31.1 107 93 15

Cu–trien5x 0.600 46.5 2.0 26.0 30.0 105 92 12

LECO 0.1 mass% C Cu–trien5xcalcite 0.400 46.4 1.8 25.8 27.0 101 94 7

Nominally 0.8 mass % Cc Cu–trien5xcalcite 0.600 46.6 1.9 25.5 27.2 101 93 9

NH4Cl–ethanolic 0.081 47.8 2.3 26.4 28.6 105 n.a. n.a.

NH4Cl–ethanolic 0.120 46.9 1.3 27.4 29.1 105 n.a. n.a.

(meq/100 g)(g)

28 S. Kaufhold et al. / Applied Clay Science 86 (2013) 23–37

3.2. CEC

Sodium carbonate activation is performed to exchange the interlayerCa2+ andMg2+ by Na+. Therefore, CECmethodswere supposed to pro-vide some information about activated and non-activated materials. Ofparticular interest is the comparison of the sum of exchangeable cations(‘sum’) and the amount of the index cation adsorbed (CEC). If solublecomponents are absent both values match (‘sum’ = CEC) but in caseof the presence of some soluble or at least partly soluble phases suchas calcite or gypsum, the ‘sum’ value exceeds the CEC value. If partly sol-uble phases are present the actual difference (‘sum’−CEC) depends onthe solid liquid ratio and hence on themass used in the exchange exper-iment. This is discussed in detail by Dohrmann and Kaufhold (2009).Compared to calcite and gypsum, the solubility of the sodium carbonatephases is much better. Therefore, the addition of sodium carbonate isexpected to increase the parameter ‘sum’ − CEC. Accordingly, the‘sum’− CEC value determined with different methods of all materialsis discussed in the following. Moreover, to be able to draw further con-clusions, different CEC-methods were applied. Both the “Cu-trien5x”and the “Cu-trien5xcalcite”methods were used to reduce the solubilityof partly soluble phases. This is achieved by increasing the solid/liquidratio (by the factor of 5=5x) and by saturation of the exchange solutionwith calcite before the experiment. The Cu-trien5xcalcite-method ismost effective in eliminating the effect of calcite. Values obtained withthis method are not affected by calcite solubility. Hence, if the ‘sum’ de-terminedwith thesemethods exceeds the CEC, phases other than calcitewere dissolved. Also the ammonium-chloride method performed withethanol reduces the calcite dissolution. The comparison of the resultsof different CEC methods was thought to provide a deeper insight intothe interlayer composition. First, the pure sodium carbonate sampleswere investigated with respect to the CEC and exchangeable cations asif they were bentonite samples (Table 2).

Owing to the large solubility of the sodium carbonate phases(N200 g/L at 25 °C) a large Na+ concentration was found after the ex-change experiment which can be simply explained by dissolution. Thisdata is marked in blue in Table 2. The data prove the much lower solu-bility of the carbonates even of sodium carbonate in ethanolic solution.The natural trona contains someK+which could be detected alongwiththe Na+. The red marked K-values represent analytical errors.

The CEC data of the natural Na-bentonites is given in Table 3.All natural Na-bentonites considered in the present study contain

some Ca-carbonates which is evident from the larger ‘sum’ − CECvalue determined with the standard methods and by the larger Ca-values (marked in blue). This was confirmed by XRD and C–S-analysis.Results shown in Table 3 also prove that the Cu-trien5xcalcite methodis much more effective in suppressing calcite dissolution than the Cu-trien5x method. Interestingly, the ‘sum’−CEC value varies from about3 to 9 meq/100 g and hence indicates the presence of further solublecontents. Considering the Cl− content of aqueous extracts of the sample#B24 from Morocco Kaufhold and Dohrmann (2008) suggested thepresence of some halite in this material. However, no chloride andhence no halite was found in sample #B08. This material containssome gypsum which can explain the positive ‘sum’− CEC value. Thismay also explain the fact that the ethanolic NH4Cl exchangeable Ca2+

value of sample #B08 is even lower than the Cu-trien5xcalcite ex-changeable Ca2+ value. These results show the complexity of determin-ing the correct amount of exchangeable cations versus dissolved cationswhich of course is relevant for sodium carbonate activated bentonites aswell. The CEC results of the samples activated in laboratory are shown inTable 4 and the most important values which are discussed below aremarked.

According to C–S-analysis sample #B06 contains b0.1 mass% Cwhich does not affect the ‘sum’ − CEC value. In contrast, the‘sum’ − CEC values of the other materials (with more carbonate)are significantly above 0 when using the standard method and al-most 0 if the Cu-trien5xcalcite method is used to suppress calcitedissolution. This proves that the Ca-values of these samples, atleast when using standard CEC methods, are affected by calcite.

After dry addition of 2mass% sodium carbonate the Na+ content in-creased by about 35meq/100g. The activation, i.e. the actual cation ex-change, is supposed to occur throughout the CEC experiment and starts

Table 4CEC data of all samples activated in laboratory.

sample B06 <0.1 mass% C sample B16 0.3 mass% C sample B19 0.5 mass% C

mass Na K Mg Ca S CEC sum– Na K Mg Ca S CEC sum– Na K Mg Ca S CEC sum–

[g] [meq/100g] CEC [meq/100g] CEC [meq/100g] CEC

natural state natural state natural state

Cu–trien 0.08 14 1 33 44 93 94 –1 1 4 21 45 71 59 12 18 2 30 45 96 78 18

Cu–trien 0.12 14 1 33 44 93 93 0 0 2 20 45 67 59 8 18 2 30 39 90 77 12

Cu–trien5x 0.40 15 2 33 45 95 96 –1 1 3 20 46 71 64 6 19 2 30 31 83 81 2

Cu–trien5x 0.60 15 2 33 44 94 96 –2 0 2 20 46 68 63 5 19 2 30 30 81 80 1

Cu–trien5xcal. 0.40 15 1 32 43 91 97 –7 0 2 19 44 66 65 0 19 3 30 27 80 81 –1

Cu–trien5xcal. 0.60 15 1 32 43 91 97 –5 0 2 19 44 66 64 2 19 2 30 27 78 81 –3

NH4Cl–ethan. 0.08 16 1 33 46 96 n.a. n.a. 1 3 21 40 65 n.a. n.a. 17 1 33 40 91 n.a. n.a.

NH4Cl–ethan. 0.12 14 1 32 46 93 n.a. n.a. 1 2 19 38 60 n.a. n.a. 16 1 31 37 86 n.a. n.a.

2 mass% S2 added dry 2 mass% S2 added dry 2 mass% S2 added dry

Cu–trien 0.08 33 3 10 40 87 55 32 53 2 18 30 103 73 30

Cu–trien 0.12 49 1 20 37 107 88 19 33 2 10 39 83 56 28 54 3 17 20 94 72 22

Cu–trien5x 0.40 51 2 20 25 98 91 7 33 2 10 28 73 59 14 54 2 17 11 84 76 8

Cu–trien5x 0.60 51 1 20 22 94 90 4 33 2 10 25 71 58 12 54 2 16 8 81 75 5

Cu–trien5xcal. 0.40 50 1 20 17 89 91 –3 33 2 10 21 65 59 6 56 2 17 5 79 76 3

Cu–trien5xcal. 0.60 51 1 20 16 88 90 –2 33 2 10 20 65 58 7 55 2 16 3 77 76 1

NH4Cl–ethan. 0.08 50 2 23 44 119 n.a. n.a. 33 2 13 27 75 n.a. n.a. 55 2 22 31 109 n.a. n.a.

NH4Cl–ethan. 0.12 51 1 22 40 114 n.a. n.a. 32 2 11 21 67 n.a. n.a. 54 2 21 24 101 n.a. n.a.

2 mass% S2, wet activated 2 mass% S2 wet activated 2 mass% S2, wet activated

Cu–trien 0.08 45 3 25 43 116 91 25 29 3 15 43 90 60 30 80 4 5 27 116 77 39

Cu–trien 0.12 44 1 24 37 106 90 16 28 2 14 38 82 60 21 79 3 3 18 104 77 27

Cu–trien5x 0.40 46 2 24 27 98 95 4 29 2 14 28 74 64 10 81 3 2 7 93 81 12

Cu–trien5x 0.60 45 1 23 25 94 93 1 29 2 14 26 71 63 8 80 2 1 4 88 81 7

Cu–trien5xcal. 0.40 46 2 23 20 90 95 –5 29 2 13 21 65 65 0 82 2 2 0 87 82 5

Cu–trien5xcal. 0.60 46 1 23 20 90 93 –4 29 1 13 21 65 63 2 81 2 1 0 85 81 3

NH4Cl–ethan. 0.08 42 2 27 40 111 n.a. n.a. 26 3 17 25 71 n.a. n.a. 78 3 7 17 104 n.a. n.a.

NH4Cl–ethan. 0.12 44 1 28 39 112 n.a. n.a. 26 2 15 22 65 n.a. n.a. 75 2 5 11 92 n.a. n.a.

5 mass% S2 added dry 5 mass% S2 added dry 5 mass% S2, added dry

Cu–trien 0.08 99 1 10 26 136 87 49 84 2 4 28 118 59 59 105 2 4 19 131 73 58

Cu–trien 0.12 99 1 9 16 125 87 38 84 2 3 17 106 59 47 105 2 3 12 122 72 50

Cu–trien5x 0.40 101 1 7 3 113 92 21 85 2 2 4 93 63 30 106 2 1 2 111 77 34

Cu–trien5x 0.60 100 1 6 2 110 91 19 85 2 1 2 91 62 28 106 2 1 1 110 76 34

Cu–trien5xcal. 0.40 103 2 6 –1 110 93 17 86 2 2 –3 88 64 24 109 3 2 –5 109 77 32

Cu–trien5xcal. 0.60 103 1 5 –3 107 91 16 86 2 1 –2 86 63 24 106 2 1 –4 106 76 29

NH4Cl–ethan. 0.08 99 1 9 28 137 n.a. n.a. 85 2 6 13 107 n.a. n.a. 108 2 6 23 139 n.a. n.a.

NH4Cl–ethan. 0.12 104 1 8 19 132 n.a. n.a. 85 2 5 6 97 n.a. n.a. 106 2 4 15 126 n.a. n.a.

5 mass% S2, wet activated 5 mass% S2, wet activated 5 mass% S2, wet activated

Cu–trien 0.08 93 1 6 25 126 92 34 83 1 3 21 108 60 48 103 4 2 20 130 78 52

Cu–trien 0.12 93 1 4 16 116 91 25 83 1 2 13 99 62 37 102 3 1 15 122 80 42

Cu–trien5x 0.40 96 1 3 6 106 96 10 84 1 1 3 89 67 23 102 3 0 2 108 83 25

Cu–trien5x 0.60 94 1 2 4 101 95 6 83 1 0 2 87 66 21 102 2 0 1 106 82 24

Cu–trien5xcal. 0.40 96 1 2 1 100 96 4 83 1 1 –2 83 67 16 104 2 0 –4 102 82 20

Cu–trien5xcal. 0.60 95 1 2 0 98 95 3 84 1 0 –2 83 67 17 104 2 0 –3 103 82 20

NH4Cl–ethan. 0.08 94 1 8 20 123 n.a. n.a. 80 1 4 1 86 n.a. n.a. 98 3 4 16 120 n.a. n.a.

NH4Cl–ethan. 0.12 91 1 7 11 109 n.a. n.a. 80 1 3 0 85 n.a. n.a. 98 2 3 10 113 n.a. n.a.

S2 = soda sample 2; added dry = S2 was added to sample in dry state (activation throughout CEC experiment)

wet activated = S2 was added, sample was mixed with water with sufficient reaction time, centrifuged, washed, dried.

49 3 21 41 114 89 25

29S. Kaufhold et al. / Applied Clay Science 86 (2013) 23–37

when the powder (mixture of bentonite and sodium carbonate) isadded to the exchange solution. Interestingly, the Na+ content of sam-ple #B16 was lower compared to the others despite the same amountof sodium carbonate was added. This finding, however, may be relatedto different exchange selectivities but needs no further discussion inthe present study.

The lower amount of exchangeable Na+ of sample #B16 results fromthe lower amount of initially present exchangeable Na+. For samples#B06 and #B16 no significant differences of dry and wet activationwere found. Sample #B19, however, showed much higher Na+ andmuch higher Mg2+ contents after wet activation using 2% sodium

carbonate. Interestingly, the Na+ content of the other samples wasslightly lower compared to the dry activation.

In the case of the 5% activated materials also less Na+ of the wet ac-tivated sample but also less Mg2+ was found. This trend was also evi-dent for sample #B19 which showed a different trend when activatedwith 2% sodium carbonate. The lower Na+-content results in a lower‘sum’− CEC value of the material activated with 5% sodium carbonateby dispersion. Comparison of the ‘sum’−CEC values of the other sam-ples activated with different amounts of sodium carbonate suggeststhat this is a generally valid trend. To simulate the activation thewet ac-tivated dispersions were not washed with water but simply allowed to

Table 5CEC data of the commercial products.

Mass Na K Mg Ca Sum CEC Sum− CEC

(g) (meq/100 g)

P1 Catsan clumbing Cu-trien 0.080 11.0 1.6 7.4 53.8 74 53 21Cu-trien 0.120 11.0 1.5 7.2 47.3 67 54 13Cu-trien5x 0.400 11.5 1.6 7.0 38.3 59 56 3Cu-trien5x 0.600 11.4 1.4 7.1 36.9 57 56 1Cu-trien5xcalcite 0.400 12.1 2.3 7.0 31.7 53 56 −3Cu-trien5xcalcite 0.600 11.6 1.5 6.7 32.4 52 56 −4NH4Cl–ethanolic 0.082 12.3 2.9 8.5 49.5 73 n.a. n.a.NH4Cl–ethanolic 0.121 13.7 2.9 8.4 48.7 74 n.a. n.a.

P2 Real bentonite product, Catsan Ultra Mars Inc. Cu-trien 0.080 7.5 1.8 14.8 45.9 70 65 5Cu-trien 0.120 7.6 1.8 14.4 45.3 69 66 4Cu-trien5x 0.400 8.1 1.8 14.3 45.5 70 69 1Cu-trien5x 0.600 8.5 1.8 14.2 44.6 69 69 0Cu-trien5xcalcite 0.400 8.1 2.0 13.7 41.9 66 71 −5Cu-trien5xcalcite 0.600 7.9 1.8 13.6 41.8 65 70 −4NH4Cl–ethanolic 0.081 8.1 1.4 16.1 48.5 74 n.a. n.a.NH4Cl–ethanolic 0.120 8.1 1.3 15.9 47.9 73 n.a. n.a.

P3 Real bentonite product, technically activated with 3% S1, S&B, Landshut Cu-trien 0.080 58.0 2.5 19.8 36.7 117 73 44Cu-trien 0.120 56.3 0.8 17.7 27.1 102 73 29Cu-trien5x 0.400 59.5 1.5 15.8 12.6 89 77 13Cu-trien5x 0.600 58.5 1.3 14.4 9.3 83 76 7Cu-trien5xcalcite 0.400 58.0 1.0 15.1 5.2 79 77 2Cu-trien5xcalcite 0.600 58.3 1.0 14.3 4.0 78 76 1NH4Cl–ethanolic 0.080 59.0 1.8 18.5 25.1 104 n.a. n.a.NH4Cl–ethanolic 0.120 57.5 1.2 17.3 19.0 95 n.a. n.a.

P4 Real bentonite product, unknown activation, S&B, Marl Cu-trien 0.080 81.3 2.3 14.0 28.0 126 89 37Cu-trien 0.120 80.5 2.5 12.7 24.2 120 88 33Cu-trien5x 0.400 83.0 2.6 10.7 11.5 108 92 16Cu-trien5x 0.600 83.2 2.7 9.5 8.8 104 91 13Cu-trien5xcalcite 0.400 83.4 3.1 10.5 4.5 102 91 10Cu-trien5xcalcite 0.600 83.7 2.5 9.6 3.4 99 91 9NH4Cl–ethanolic 0.080 81.7 1.8 15.0 20.9 119 n.a. n.a.NH4Cl–ethanolic 0.120 84.2 1.8 14.3 17.4 118 n.a. n.a.

P5 Real bentonite product, Waldo North (MK11), Klinkenberg (2008) Cu-trien 0.080 28.9 2.0 9.0 42.6 83 59 23Cu-trien 0.120 27.7 0.4 8.3 40.5 77 59 18Cu-trien5x 0.400 29.4 1.0 8.3 36.9 76 61 15Cu-trien5x 0.600 28.8 0.5 8.1 35.6 73 61 12Cu-trien5xcalcite 0.400 28.9 0.4 7.8 29.1 67 62 4Cu-trien5xcalcite 0.600 29.0 0.4 7.9 29.5 67 61 6NH4Cl–ethanolic 0.081 28.4 1.4 8.3 22.2 60 n.a. n.a.NH4Cl–ethanolic 0.120 27.5 0.6 7.5 20.9 57 n.a. n.a.

Fig. 5. Thermal analysis of the sodium carbonate samples, left side DSC, center: evolved gas analysis with mass spectrometry of water, and right: evolved gas analysis with massspectrometry of CO2.

30 S. Kaufhold et al. / Applied Clay Science 86 (2013) 23–37

Fig. 6. Thermal analysis of the samples activated in laboratory (#LA samples). (1) natrite, (2) initial calcite, and (3) precipitated calcite (OM organic matter, DHX temperature range ofdehydroxylation of the smectites).

31S. Kaufhold et al. / Applied Clay Science 86 (2013) 23–37

settle and dry afterwards. Throughout this slow drying some of thedissolved salts could have been precipitated on the wall of the beakerwhereas the smectites settled. This could explain the differences of the‘sum’ − CEC values of the wet and dry activated materials. However,the 2% activated #B19 does not follow this trend (marked in red inTable 4) which cannot be explained, yet.

Samples #B16 and #B19 contain more soluble Ca-phases (calcite +gypsum) than #B06which hence is evident from the difference of the ex-changeable Ca2+ values determined by the standard method and by Cu-trien5xcalcite. Compared to the 2% activated materials all 5% activatedones show significant ‘sum’− CEC values which indicates the presenceof excess sodium carbonate because not all sodium carbonate was con-sumed for the activation. Hence the “exchangeable Na+ values” are sup-posed to represent the sum of true exchangeable Na+ plus Na+ fromsoluble Na-phases (Na+exc-inflated, Dohrmann and Kaufhold, 2010).

The results of the CEC measurements of the five unknown productsare given in Table 5. The content of exchangeable Na of #P1, #P2, andprobably #P5 is too low to be activated. These samples were not

activated. For sample #P3 almost 60meq/100 g Na+ was found, whichis approximately 75% of the CEC. The ‘sum’ − CEC value of the Cu-trien5xcalcite method is low, indicating that calcite is the only solublephase. This sample is known to be activated with approximately 3% so-dium carbonatewhich obviously does not lead to the presence of excessNa+ (Na+exc-inflated) and hence cannot be detected by CEC measure-ments. For sample #P4 amuch larger ‘sum’−CEC value could be detect-ed which indicates that phases other than calcite are presentcontributing to the sum of exchangeable cations. Hence it is possibleto conclude that this material was activated with sodium carbonate ina way that excess sodium carbonate is present.

3.3. STA

The DSC curves of the sodium carbonate samples #S1–#S4 show astrong endothermal reaction in a low temperature range below 200 °C(Fig. 5, left, DSC). The maximum peak temperature for #S3 was 146 °Cwhich was 30–40 °C higher compared to the others. Only the #S1 and

Fig. 7. Thermal analysis of the products.Multiplication factors for amplifying the intensities are indicated. In the right part of the figure amagnification of the low-T range of two samples isshown.

32 S. Kaufhold et al. / Applied Clay Science 86 (2013) 23–37

#S2 DSC curves showed an endothermal reaction above 800°C. The peakmaximum in theMS curves of evolvedwater (m/e=18; Fig. 5, center) for#S1, #S2, and #S4 at 105 to 115 °C corresponded to the maximum peaktemperature in DSC curves according to the release of constitutionalwater of thermonatrite and trona (Na2CO3∗H2O→Na2CO3+H2O↑; Na3(-CO3)(HCO3)∗2H2O→Na3(CO3)(HCO3)+2H2O↑). The peakmaximum inthe MS curve of evolved water and CO2 of #S3 at 146 °C was related tothermal decomposition of nahcolite (2NaHCO3 → Na2CO3 + CO2

↑ + H2O↑). The peak maximum in the MS curves of evolved CO2 (m/e=44) for #S1, #S2, and#S4 at 122–125°C corresponded to the shoulderin the DSC curves and in the MS curves of evolved water at higher tem-peratures due to the first thermal decomposition of trona (Na3(CO3)(-HCO3)→Na2CO3+CO2↑+H2O↑; Bain and Morgan, 1969; Beck, 1950).

Table 6Conductivity, pH, and turbidity after CaCl2 addition of all samples. The qualitative CaCl2 turbid

Sample 50mg in 50mL conductivity [μS/m] pH + CaCl2 turbidity

S1 1372 10.4 +S2 1331 10.4 +S3 1027 9.0 –

S4 1257 10.3 +P1 36 9.4 –

P2 11 8.1 ±P3 61 10.0 ±P4 61 10.0 ±P5 27 9.3 ±BN1 41 9.7 –

BN2 22 8.9 +BN3 20 8.0 –

LA6-0 7 7.4 –

LA6-2A 32 9.5 ±LA6-2D 26 9.4 ±LA6-4A 75 10.0 ±LA6-4D 112 10.1 +LA16-0 13 6.8 –

LA16-2A 32 9.1 ±LA16-2D 27 9.2 ±LA16-4A 75 9.8 ±LA16-4D 74 9.9 +LA19-0 35 9.5 –

LA19-2A 59 9.9 ±LA19-2D 80 10.0 +LA19-4A 104 10.1 ±LA19-4D 97 10.1 +

Dehydration and thermal decomposition of sodium carbonate hydratesresulted in anhydrous sodium carbonate (ASC). Melting and decomposi-tion occurred simultaneously for the initial natrite in the TSC samples andthe natural trona (#S4) and for the ASC. Thereby, decomposition of theASC resulted in a broad peak in theMS curve of the evolved CO2 between400 and 800 °C followed by a sharp peak resulting from the decomposi-tion of natrite above 800 °C. No difference was observed between #S1,#S2, and#S3. In the natural sample of trona (#S4)melting anddecompo-sition of ASC occurred atmuch lower temperatures (b700°C). Burkeite in#S4 resulted in additional evolved SO2 with a peak maximum in the MScurve (m/e= 64) at 966 °C (curve not shown). At this temperature noCO2 was liberated. Thus decomposition of burkeite may be a two-stepreaction. Using thermal analysis, trona and thermonatrite can be best

ity tests were performed to probe dissolved carbonate.

? 200mg in 50mL conductivity [μS/m] pH + CaCl2 turbidity?

3830 10.3 +3750 10.3 +3040 9.0 +3880 10.2 +

50 9.6 ±21 9.2 ±

172 10.4 +198 10.4 +93 9.6 +

140 10.0 +76 9.7 +74 9.5 +17 9.4 +

100 10.1 +83 9.9 +

240 10.4 +200 10.4 +40 7.0 –

110 9.8 ±91 9.7 ±

257 10.3 +245 10.4 +71 10.0 –

158 10.2 ±219 10.4 +288 10.5 +266 10.5 +

Fig. 8. Time depending pH (time aftermixing the powder and thewater) of a samplewithsodium carbonate added in the dry state and the precursor material.

33S. Kaufhold et al. / Applied Clay Science 86 (2013) 23–37

identified bymonitoring evolved H2O and CO2 below 200°C and evolvedCO2 at higher temperatures by mass spectrometer or other gas analysislinked to TG or DSC.

Before the addition of sodium carbonate and any treatment thedehydration shoulder towards higher T being characteristic of bivalentinterlayer cations was observed for all samples (Fig. 6). The peaks ob-served at higher temperatures either in the DSC or TG curves (notshown) along with data presented in Fig. 6 correspond to XRD results(Fig. 3). Some additional information can be obtained from the investi-gation of the shape of the carbonate peaks. Generally, the position andshape of decomposition peaks depend on the concentration of thephase, particle size, and even crystallinity. Considering the MS curve ofCO2 of sample #LA19-0, as an example, suggests the presence of a sec-ond (Ca,Mg-)carbonate along with calcite, which was not detected byXRD. This could be further investigated for example by SEM–EDX butis not important for the present study. After dry activation and wet ac-tivation the water content of all samples decreased (data not shown).

Fig. 9. IR spectra of the four differe

After dry activation with 2 and 5mass% of #S2 all samples still showeddehydration typical of divalent cations in the interlayer of smectites, i.e.a shoulder towards higher temperature (Fig. 6). This proves that the ac-tivation (the actual cation exchange) did not take place in the dry state.In contrast, after wet activation with 2mass% this shoulder significantlydecreased and after wet activation with 5mass% all samples showed adehydration typical of monovalent cations in the interlayer of the smec-tite (no shoulder observable anymore) as observed for samples #BN1–#BN3 (data not shown). The MS CO2 curves appear to be more compli-cated than the water curves. After dry activation (regardless of theamount, 2 or 5 mass%), all samples showed a peak doublet below200 °C in the MS curve of evolved CO2 at 97/137 °C (#LA6-2A) and106/138 °C (#LA6-5A); at 119/143 °C (#LA16-2A) and 122/144 °C(#LA16-5A) and 125/144 °C (#LA19-2A) and 125/143 °C (#LA19-5A)(Fig. 6). This peak doublet was thought to be related to the decomposi-tion of remaining trona but the shift to higher peak temperatures sug-gests a transformation of trona into nahcolite (Fig. 5) with the waterevolved from the smectites liberated by increasing temperaturethroughout the STA measurements (Wegscheider and Mehl, 1928)hence being a kind of water redistribution from clay to salt as studiede.g. by Kaufhold et al. (2009). Samples activated with excess waterusing 2mass% sodium carbonate only showed a very weak peak in theMS curve of evolved CO2 centered at 130 °C (#LA6-2D and #LA19-2D)and a shoulder at lower temperatures (at about 100°C). This peak is as-sumed to result from traces of a Na2CO3 phasewhichwas surprising be-cause at least in the case of the ‘-2D’ samples a complete reaction of thesoda without any remaining traces of Na2CO3 was expected. However,the really low intensity of this peak proves that only traces (≪1%) ofthe Na2CO3 phase are still present.

Activation under wet conditions with 5mass% resulted in a well re-solved CO2 peak doublet in the same temperature range (Fig. 6) with amaximum peak temperature equal to the dehydration peak (#LA6-5D)or 10K higher than dehydration (#LA16-5D and #LA19-5D). This peakdoublet in the low temperature range of the CO2 curve possibly corre-sponds to some trona/nahcolitewhichwas not expected because partic-ularly in the case of wet activation with sodium carbonate below theCEC (in the present study = 2 mass%) the complete consumption of

nt sodium carbonate samples.

Fig. 10. IR spectra of the differently activated bentonites #B06, #B16, and #B19.

34 S. Kaufhold et al. / Applied Clay Science 86 (2013) 23–37

the sodium carbonate was thought to be likely. Obviously, thanks to therather low detection limit of the minor components trona andthermonatrite, even rather small amounts can be detected.

Additional informationwas expected in the high temperature range,because the shape of the carbonate decomposition peaks may corre-spond to their crystallinity which is supposed to be low in the case ofthe freshly precipitated Ca/Mg-carbonates. Accordingly, the position ofthe decomposition peak of low crystallinity phases is thought to belower. After dry activation with 2 mass% and 5 mass% all samplesshowed four or more superimposing CO2 releases between 450 and800°C. The shoulder at about 540°C may correspond to the decomposi-tion of a double carbonate that formed by the reaction of sodium car-bonate with calcium carbonate (Wilburn et al., 1965). A broad peak

Fig. 11. IR spectra of the five a

with a maximum at 623 °C (#LA6-2A) and 646 °C (#LA6-5A) in theMS curves (m/e = 44) results from precipitated calcite together withthe initially present calcite (calcitewhichwas present before the activa-tion), which is fully masked, indicating the beginning of activation. Thesomewhat sharper peaks with a maximum at 655 °C (#LA6-2A) and670 °C (#LA6-5A) may correspond to the decomposition of remainingnatrite from #S2. Further CO2 releases can possibly be attributed toun-reacted #S2. In contrast, wet activation resulted in dissolution of#S2 and cation exchange. After drying only precipitated calcite was de-tected in the MS curves of evolved CO2 with peak maxima at 674 °C(#LA6-2D) and 700°C (#LA6-5D) (Fig. 6).

Dry activated samples of #LA16 showed an additional CO2 release at227°C (sharp peak). The CO2 is released together with somewater in an

ctual products #P1–#P5.

Fig. 12. STA–MS–CO2 signals of the partially activated and dispersed bentonites (left) compared to the unactivated materials (right).

35S. Kaufhold et al. / Applied Clay Science 86 (2013) 23–37

exothermal reaction which cannot be explained, yet. #LA16-2A and#LA16-5A showed a broad peak in the MS curves (m/e = 44) with amaximum slightly above 600 °C which could correspond to neoformeddouble carbonate and precipitated calcite. The sharp peaks at 653 °C(#LA16-2A) and 673°C (#LA16-5A) probably correspond to the decom-position of remaining natrite from#S2.Wet activation resulted in disso-lution of #S2. After drying only precipitated calcite was detected in theMS curves of evolved CO2 with peakmaxima at 675°C (#LA16-2D) and696 °C (#LA16-5D) (Fig. 6).

Dry activated samples of #LA19 again show CO2 release fromneoformed double carbonate followed by a small peakwith amaximumat about 630°C (#LA19-2A and #LA19-5A). In analogy to the above ex-planations this peak is attributed to the decomposition of remainingnatrite from #S2. CO2 release from precipitated carbonates (mainly cal-cite) with a peak maximum at 700 to 710 °C is superimposing the CO2

release at 690 °C from the initially present (natural) calcite in both dryand wet activated #LA19 samples (Fig. 6).

Finally, thermal analysis was performed to investigate the actualbentonite products. Dehydration of #P1, #P2, and #P5 is typical ofsmectites with bivalent cations (shoulder to higher T in Fig. 7, leftside) as described for #LA6-0, #LA16-0 and #LA19-0. These sampleswere definitely not activatedwith sodium carbonate. Activation of sam-ples #P3 and #P4, on the other hand, cannot be excluded based on theshape of the dehydration peak. Na+ seems to be the dominating ex-changeable cation of #P3 and #P4 being in accordance with CECmeasurements. Further information may be available from the investi-gation of theMS CO2 curve which showed calcite with a decompositionpeak temperature at 712 °C (#P1), 629 °C (#P2), and 671 °C (#P5). Thecharacteristic peak doublet b200°C, however, could not be observed forthese samples thus confirming the conclusion that the samples werenot activated. In contrast, the MS-curves of evolved CO2 of samples#P3 and #P4 show a weak peak doublet below 200 °C with the peakmaximum of the dominating peak at 120°C (#P3) and 124°C (#P4), re-spectively. Based on the above interpretation of Fig. 6 it can be conclud-ed that samples #P3 and #P4 were activated with sodium carbonateand that water was present. The relative intensities of the peaks suggestthat #P3 was possibly activated with less sodium carbonate comparedto the CEC while #P4 was possibly activated with excess sodiumcarbonate.

At high temperatures the decomposition of initially present (natu-ral) calcite in #P3 at 728 °C was observed, while sample #P4 containedvery low amounts of carbonates. Nevertheless, two CO2 releases at 501and 639°Cwere observed which could not be assigned to a special typeof carbonate. The CO2 release at 639°Cmay be from non-reacted TSC asobserved for #LA16-2A.

3.4. pH, conductivity and turbidity after CaCl2 addition

Sodium carbonate dissolution in water causes the pH to increase.This large pH value is occasionally used to try to distinguish activatedfrom non-activated materials. As can be seen from Table 6 the pH can-not be used to identify sodium carbonate activated materials becausethe pH of natural Na-bentonites may be as high as 10. As an example,the pHof #BN1 is almost 10. The difference of 0.3 pH units is not consid-ered to be significant. The pH value, however, is suitable to distinguishCa/Mg- from Na-bentonites (Kaufhold et al., 2008), but not to distin-guish natural Na-bentonites from activated bentonites. The electricalconductivity of dispersions or extracts of the activated materials is gen-erally larger but the natural Na-bentonite #BN1 also showed a largeconductivity of the water. Hence this parameter is also not suitable forthe detection of sodium carbonate activated materials.

To at least be able to identify the dry addedmaterials one ideawas tomeasure the time dependent pH, i.e. the pH value depending on thetime after mixing the powder and the water. This was supposed to beinteresting because in the case of these samples the activation, i.e. theactual cation exchange, could possibly be followed indirectly. Resultsare shown in Fig. 8 and prove that the time dependence of the pH ofboth the precursor and the material, which was mixed with 5% sodiumcarbonate is similar and hence is not considered to be suitable to iden-tify activated materials.

3.5. IR

Infrared spectroscopy was used as complementary tool for phaseanalysis. First the sodium carbonate samples were investigated (Fig. 9)and then compared with the differently treated samples (Fig. 10).Three of the four sodium carbonate samples showed the characteristicmain vibration of carbonates, in this case at 1446 cm−1. In the case of

36 S. Kaufhold et al. / Applied Clay Science 86 (2013) 23–37

the hydrogencarbonate this band was also observed but it is much lesspronounced.

The IR spectra shown in Fig. 10 confirm the XRD results in that thecarbonates of the dry added samples mainly correspond to natrite(1446cm−1). In contrast the main carbonate band of the dispersedma-terials is between 1420 and 1430 cm−1 corresponding to calcite. Com-pared to XRD no further information could be gained from IR. Thedetection limit of both methods (IR+XRD)with respect to free sodiumcarbonate (natrite) seems to be similar. For the 2% dry added samples areally close look on the XRD traces is needed to identify the natritemainreflection. The characteristic IR band of natrite of the same sample at1446 cm−1 is well resolved which points towards a somewhat lowerIR detection limit of natrite compared to XRD. However, the presenceof dolomite or magnesite has to be excluded because the main band ofthese minerals interferes with the others. Hence the IR detection limitof dolomite or magnesite in such bentonite materials may be evenworse than that of XRD.

The IR spectra of the actual products are shown in Fig. 11. The differ-ent bands in the 1400–1500cm−1 region correspond to the carbonates.#P3 contains the largest amount of carbonates but a range of the maxi-mum of the carbonate main vibration was found ranging from 1430 to1480cm−1 and hence reflects the case discussed above — the intensityat about 1440cm−1 cannot be unambiguously assigned to either natriteor dolomite.

4. Summary and conclusions

The main problem with respect to the identification of sodium car-bonate activated bentonites is the number of parameters being relevantbefore, throughout, and probably even after activation. In an unknownactivation process, sodium carbonate could have been added in excessormuch below excess and activationmayhave been completed becauseof the presence of water or the dry sodium carbonate that may still bepresent ‘waiting’ for the activationwhich finally takes place in the appli-cation where water is added (example drilling mud). Hence four caseshave to be distinguished:

1 Sodium carbonate added up to or above CEC, water content large⇒ excess+completed,

2 Sodium carbonate added below CEC, water content large⇒ partial activation complete,

3 Sodium carbonate added up to or above CEC, no water⇒ excess+waiting for water,

4 Sodium carbonate added below CEC, no water⇒ partial activation waiting for water.

In cases 3 and 4 (addition of sodium carbonate in the dry or moder-ately dried state) sodium carbonate (natrite+hydrous sodium carbon-ates) can be detected by typical phase analytical methods as XRD, STA–MS, and IR. By far the lowest detection limit was found for STA–MS. Toour surprise, STA–MS–CO2 measurements were even able to detecttraces of a Na2CO3 phase in case 2. The MS–CO2 peak indicating thepresence of traces of sodium carbonate is around 100 °C and rathersmall and hence needs significantmagnification for a reasonable assess-ment (Fig. 12, left). Nevertheless, none of the un-activated materialsshowed any CO2 peak b200 °C (Fig. 12, right) and hence the most im-portant conclusion of the present study is that the presence of anypeak or intensity b200 °C of the MS–CO2 curve indicates technical acti-vation of bentonite with sodium carbonate — particularly if the charac-teristic peak doublet is observed.

A further possibility to identify activation is to consider the cation ex-change data, particularly the ‘sum’−CEC value determinedwith differentmethods. Generally, a large ‘sum’−CECvalue indicates thepresenceof sol-uble Na-, K-, Ca-, or Mg-phases, the most common being calcite which, asdiscussed above, may be natural or a product of sodium carbonate activa-tion. For a comprehensive understanding of the cation population and thesoluble phases the abundance of gypsum, calcite, and halite has to be

assessed. Gypsum and calcite can be detected with STA–MS (and quanti-fied by C–S-analysis). For the determination of halite abundance an aque-ous extract can be investigated with respect to the Cl− content. Thisinformation has to be available for the assessment of the CEC data.

A significant (N5meq/100 g) ‘sum’− CEC value with Na+ as domi-nant exchangeable cation and a low ‘sum’ − CEC value determinedwith the Cu-trien5xcalcite method indicate the presence of either natu-ral Na-bentonite with calcite as minor component or partially activatedbentonite. The absence of calcite, or better carbonate in general, in thiscase indicates the presencze of a natural Na-bentonite because any ad-dition of sodium carbonate to Ca/Mg leads to the abundance of inorgan-ic C. As an example, sodium carbonate activation with 3mass% wouldincrease the C content by 0.4 mass%. On the other hand, a significant‘sum’ − CEC value determined with the Cu-trien5xcalcite method incombinationwith the absenceof halite andgypsumclearly proves sodiumcarbonate activated bentonites. The calculation of exchangeable Ca2+

values is explained in detail by Dohrmann and Kaufhold (2010). Calcu-lating exchangeable Ca2+ values, however, with “negative differences”results in “negative exchangeable Ca2+values”.Measuring such “negativeexchangeable Ca2+ values” is a clear indication for soda activation (freesoda). The data presented was used to assess the real products. #P1,#P2, and #P5 were not activated with sodium carbonate (the content ofexchangeable Na+ is too low). #P3 and #P4 could be activated materials.For #P3, a product which is known to be produced with 3mass% sodiumcarbonate, a large ‘sum’− CEC value determined with the standardCu-trien method was found. The low values found with the Cu-trien5xcalcite, on the other hand, indicate that the bentonite was notactivated with excess sodium carbonate. The CEC method was notfound to be suitable to detect technical activation much below the CEC.

Occasionally the pH of dispersions is measured to detect technical so-dium carbonate activation. In the present study the pH values of the tech-nically activated materials were slightly above the natural Na-bentonitesbut the difference accounting for 0.3 pH units is considered to be insignif-icant. Therefore, at least based on the data discussed in the present paper,measuring the STA–MS–CO2 curve is suggested as alternative.

References

Alther, G.R., 1986. The effect of the exchangeable cations on the physico-chemical proper-ties of Wyoming bentonites. Appl. Clay Sci. 1, 273–284.

Bain, J.A., Morgan, D.J., 1969. The role of thermal analysis in the evaluation of impure claydeposits as mineral raw materials. Clay Miner. 8, 171–192.

Beck, C.W., 1950. Differential thermal analysis curves of carbonate minerals. Am. Mineral.35, 985–1013.

Belyayeva, N.I., 1967. Rapid method for the simultaneous determination of the exchangecapacity and content of exchangeable cations in solonetzic soils. Soviet Soil Sci.1409–1413.

Dohrmann, R., Kaufhold, S., 2009. Three new, quick CEC methods for determining theamounts of exchangeable cations in calcareous clays. Clay Clay Miner. 57, 338–352.

Dohrmann, R., Kaufhold, S., 2010. Determination of exchangeable calcium of calcareousand gypsiferous bentonites. Clay Clay Miner. 58, 79–88.

Dohrmann, R., Genske, D., Karnland, O., Kaufhold, S., Kiviranta, L., Olsson, S., Plötze, M.,Sandén, T., Sellin, P., Svensson, D., Valter, M., 2012. Interlaboratory CEC and ex-changeable cation study of bentonite buffer materials: II. Alternative methods. ClayClay Miner. 60, 176–185.

Fahn, R., 1964. Möglichkeiten und Methoden zur Unterscheidung industriell erzaugterAktivbentonite von natürlichen Natriumbentoniten. Ber. Dtsch. Keram. Ges. 41,546–550.

Grim, R.E., Güven, N., 1978. Bentonites, Geology, Mineralogy, Properties and Uses. DevelSedimentol, 14. Elsevier, Amsterdam – Oxford – New York, p. 256.

Hofmann, U., and Endell, K., 1936. British Patent 458240.Kaufhold, S., Dohrmann, R., 2003. Beyond the methylene blue method: determination of

the smectite content using the Cu-trien method. Z. Angew. Geol. 13–18 (ISSN0044–2259, 2/2003).

Kaufhold, S., Dohrmann, R., 2008. Detachment of colloidal particles from bentonites inwater. Appl. Clay Sci. 39, 50–59.

Kaufhold, S., Dohrmann, R., 2009. Stability of bentonites in salt solutions I sodium chlo-ride. Appl. Clay Sci. 45 (3), 171–177.

Kaufhold, S., Dohrmann, R., 2010a. Effect of extensive drying on the cation exchange ca-pacity of bentonites. Clay Miner. 45, 441–448.

Kaufhold, S., Dohrmann, R., 2010b. Stability of bentonites in salt solutions II. Potassiumchloride solution — initial step of illitization? Appl. Clay Sci. 49, 98–107.

Kaufhold, S., Dohrmann, R., Koch, D., Houben, G., 2008. The pH of aqueous bentonite sus-pensions. Clay Clay Miner. 56, 338–343.

37S. Kaufhold et al. / Applied Clay Science 86 (2013) 23–37

Kaufhold, S., Stührenberg, D., Dohrmann, R., 2009. Water redistribution between benton-ite and halite at elevated temperature. Appl. Clay Sci. 46, 245–250.

Kaufhold, S., Dohrmann, R., Klinkenberg, M., 2010. Water uptake capacity of bentonites.Clay Clay Miner. 58, 37–43.

Klinkenberg, M., 2008. Einfluss des Mikrogefüges auf ausgewählte petrophysikalischeEigenschaften von Tongesteinen und Bentoniten. - PhD thesis of the Georg-August-Universität Göttingen, 144 S., online available at: http://webdoc.sub.gwdg.de/diss/2008/klinkenberg/klinkenberg.pdf.

Lagaly, G., 1993. Reaktionen der Tonminerale. In: Jasmund, K., Lagaly, G. (Eds.),Tonminerale und Tone – Struktur, Eigenschaften, Anwendung und Einsatz inIndustrie und Umwelt. Steinkopff Verlag Damrstadt (490 pp.).

Lagaly, G., Müller-Vonmoos, M., Kahr, G., Fahn, R., 1981. Vorgänge bei derSodaaktivierung von Bentoniten am Beispiel eines Bentonits von Neuseeland.Keram. Z. 33, 278–283.

McBride, M.B., 1979. An interpretation of cation selectivity variations in M+–M+

exchange on clays. Clay Clay Miner. 27, 417–422.Meier, L.P., Kahr, G., 1999. Determination of the cation exchange capacity (CEC) of clay

minerals using the complexes of copper(II) ion with triethylenetetramine andtetraethylenepentamine. Clay Clay Miner. 47, 386–388.

Steudel, A., Mehl, D., Emmerich, K., 2013. Simultaneous thermal analysis of different ben-tonite–sodium carbonate systems: an attempt to distinguish alkali-activated benton-ites from raw materials. Clay Miner. 48, 117–128.

Wegscheider, R., Mehl, J., 1928. Über Systeme Sodium carbonate-NaHCO3-H2O und dasExistenzgebiet der Trona. Monatsh. Chem. 49, 283–315.

Wilburn, F.W., Metcalfe, S.A., Warburton, R.S., 1965. Differential thermal analysis,differential thermogravimetric analysis, and high temperature microscopy of re-actions between the major components of a sheet glass batch. Glass Technol. 6,107–114.