Electrochemical analysis of water and suds by impedance ... · Electrochemical analysis of water...

J. Sens. Sens. Syst., 3, 133–140, 2014www.j-sens-sens-syst.net/3/133/2014/doi:10.5194/jsss-3-133-2014© Author(s) 2014. CC Attribution 3.0 License.

Electrochemical analysis of water and suds byimpedance spectroscopy and cyclic voltammetry

R. Gruden1, A. Buchholz1, and O. Kanoun2

1Seuffer GmbH & Co. KG, Bärental 26, 75365 Calw-Hirsau, Germany2Technische Universität Chemnitz, Reichenhainer Strasse 70, 09126 Chemnitz, Germany

Correspondence to:R. Gruden ([email protected])

Received: 31 August 2013 – Revised: 10 June 2014 – Accepted: 21 June 2014 – Published: 3 July 2014

Abstract. Optimum detergent dosage during a washing process depends on water quality, degree of pollutionand quantity of laundry. Particularly, water quality is an important factor. Other parameters like carbonate- ornon-carbonate hardness and calcium / magnesium (Ca / Mg) ratio in addition to total hardness of water havean impact on the amount of detergent. This work discusses the possibilities realizing a detergent sensor thatmeasures important parameters for the washing process and assess the ideal necessary amount of detergentduring the washing process. The approach is to combine impedance spectroscopy with cyclic voltammetry inorder to determine both water quality and concentration of detergent in the suds which build up the basis foran optimum detergent dosage. The results of cyclic voltammetry show that it is possible to identify the Ca / Mgratio and the carbonate hardness separately, which is necessary for the optimization of the washing process.Impedance measurements identify total hardness and detergent concentrations.

1 Introduction and motivation

This work is an extension of the conference article (Grudenand Kanoun, 2013a)1.

In 2010 more than 580 000 tons of detergent and more than1 billion liters of water were used in Germany for washingthe laundry. On average, 30 % of the detergent is superfluous,resulting in water wasting during the rinsing process withan impact on environmental burden (Rüdenauer and Gensch,2008).

Due to the higher concentration of detergent in the suds,more and longer rinsing processes are required to get rid ofthe detergent residues in the laundry. This requires not onlyhigher water consumption but also more energy. If the rinsingprocess is adapted to the higher detergent concentration, theremaining residues can cause allergic reactions in sensitiveindividuals.

Furthermore, the washing result, customized to the indi-vidual properties of this load, is only optimal at one detergent

1Gruden, R. and Kanoun, O.: Water quality assessment by com-bining impedance spectroscopy measurement with cyclic voltam-metry, AMA Sensor, Nürnberg, 2013.

quantity (Wagner, 2010; Smulders et al., 2002; Jakobi andLöhr, 1987).

At the present time automatic washing machines use onlya control strategy that does not take the water quality or theactual properties of the laundry into consideration (Wagner,2010; Smulders et al., 2002; Jakobi and Löhr, 1987). Movingto a regulation strategy by maintaining the washing resultneeds key parameters such as water quality, quantity of laun-dry (weight), concentration of dirt and concentration of de-tergent to be able to compute an optimal dosage.

The water parameter currently used for detergent dosage istotal hardness. But it is also important to measure the carbon-ate hardness and the Ca / Mg ratio because these parametersalso influence the washing performance, too. The hydrogencarbonate ion HCO−3 , which is reasonable for the carbonatehardness, reacts to calcium carbonate which attaches to theheating elements in the washing machine and causes dam-ages; calcium and magnesium form insoluble soaps whichdecrease the washing performance. The Ca / Mg ratio influ-ences the washing performance because calcium soaps aremore insoluble than magnesium soaps.

Published by Copernicus Publications on behalf of the AMA Association for Sensor Technology (AMA).

134 R. Gruden et al.: Electrochemical analysis of water and suds by impedance spectroscopy and cyclic voltammetry

These parameters should be measured online during thewashing process with a robust low cost sensor (Gruden et al.,2012), in order to save resources, protect the environmentand avoid allergic reactions (Tröltzsch, 2012).

2 State of the art and new approaches

Optical methods such as turbidity sensors are presently usedfor automatic detergent dosage (Tschulena and Lahrmann,2006; Czyzewski et al., 1999), but they are sensitive to con-tamination and especially to the dirtiness water; furthermore,water parameters cannot be detected with turbidity sensors.

Base–acid titration is able to determine precisely the valueof total hardness and the Ca / Mg ratio (Hütter, 1990) but thismethod is very expensive and works only under laboratoryconditions.

Commercial online methods used for environmental inves-tigation (Huang et al., 2005; Kräuter et al., 2006) and dis-tribution systems (Verberk et al., 2006) focus on other wa-ter components and are too expensive for household low-cost applications. Voltammetric methods combined with ion-selective electrodes are expensive (Cammann and Galster,1996) and not suitable for detergent determination becauseof complex and different detergent composition (Wagner,2010). Low-cost applications need environmental harm-ful mercury electrodes to determine the water parameters(Pungor et al., 1977). Both methods are not feasible insidethe washing machine during the washing process.

Electronic tongues are sensor arrays combined withvoltammetric methods and multi-variable analysis (Winquistet al., 1997; Winquist, 2008) which are tested for large house-hold appliances : applications (Winquist, 2011; Ivarsson,2003; Eriksson et al., 2011). The existent problem is the com-plex data analysis. To have information about total hardness,conductive methods are more suitable, but they are generallynot sufficiently accurate because the conductivity of the elec-trolyte χE depends on the total ion concentration includingeven ions which are not interesting for the washing process.MoreoverχE is frequency dependent and the results of theestablished methods which use only one frequency is incor-rect.

Suds investigations with surface acoustic wave (SAW)sensors have been tested successfully to detect single sufac-tant concentrations (Vivancos et al., 2012). The investigationof water parameters, surfactant mixtures and critical micelleconcentration (CMC) are not shown. Moreover, frequenciesof f > 60 MHz are needed for the SAW application and suchelectronic circuits are too expensive for the application inwhite goods.

Electrochemical impedance spectroscopy (EIS) and cyclicvoltammetry (CV) are well-known analytical methods(Barsoukov and Macdonald, 2005; Brad and Faulkner, 2001)and deliver more specific information. For analysis of sudsduring the washing process they have not yet been applied.

2 Gruden: Electrochemical Analysis of Water and Suds by Impedance Spectroscopy and Cyclic Voltammetry

the washing performace because calcium soaps are moreinsoluble than magnesium soaps.These parameters should be measured online during thewashing process with a robust low cost sensor (Gruden etal., 2012), in order to save resources, protect environmentand avoid allergic reactions (Troltzsch et al., 2012).

2 State of the Art and new Approach

Optical methods such as turbidity sensors are nowadays usedfor automatic detergent dosage (Tschulena and Lahrmann,2006; Czyzewski et al., 1999), but they are sensitive to con-tamination and especially to water dirtiness. Furthermore,water parameters cannot be detected with turbidity sensors.Base-acid titration is able to determine precisely the valueof total hardness and the calcium-magnesium-ratio (Hutter,1990) but this method is very expensive and works only un-der laboratory conditions.Commercial online methods used for environmental investi-gation (Huang et al., 2005; Kraeuter et al., 2006) and distri-bution systems (Verberk, 2006) focus on other water com-ponents and are too expensive for household low-cost appli-cations. Voltammetric methods combined with ion-selectiveelectrodes are expensive (Cammann and Galster, 1996) andnot suitable for detergent determination because of com-plex and different detergent composition (Wagner, 2010).Low-cost applications need environmental harmful mercuryelectrodes to determine the water parameters (Pungor et al.,1977). Both methods are not feasible inside the washing ma-chine during the washing process.Electronic tongues are sensor arrays combined with voltam-metric methods and multivariable analysis (Winquist et al.,1997, 2011) which are tested for white good applications(Winquist et al., 2011; Ivarsson, 2003; Eriksson et al., 2011).The existent problem is the complex data analysis. To haveinformation about total hardness, conductive methods aremore suitable, but they are generally not sufficiently accuratebecause the conductivity of the electrolyte χE depends on thetotal ion concentration including even ions which are not in-teresting for the washing process. Moreover χE is frequency-dependent and the results of the established methods whichuse only one frequency is incorrect.Suds investigations with surface acoustic wave (SAW) sen-sors have been tested successfully to detect single sufactantconcentrations (Vivancos, 2012). The investigation of wa-ter parameters, surfactant mixtures and cmc are not shown.Moreover, frequencies of f > 60MHz are needed for theSAW applicationand and such electronic circuits are too ex-pensive for the application in white goods.Impedance spectroscopy (EIS) and cyclic voltammetry (CV)are well known analytical methods (Barsoukov and Macdon-ald, 2005; Brad and Faulkner, 2001) and deliver more spe-cific information. For analysis of suds during the washing

process they have been not yet applied. The combination ofthese measuring principles provides more data from differ-ent frequency ranges and allows more detailed investigations.The focus of this fluid analysis is the analysis of data and notto electrochemical effects of the sensor element. With thismethod, expensive sensors and environmentally harmful ma-terials can be avoided. Due to the fact that frequencies offmax = 5MHz are needed, the cost of the electronic circuit isadequate for the planned application.The focus of this paper is the investigation of predefined wa-ter samples and one standard detergent and ignore real watersamples and commercial detergents beacuase the feasibilityof this new method should be tested. Tests with real watersamples and commercial detergents can be investigated later.

3 Experimental

The experimental setup for electrochemical impedance spec-troscopy (EIS) and cyclic voltammetry (CV) consisted ofa Zahner Zennium impedance analyser for frequency range0.1 Hz to 4 MHz and an Agilent 9294A impedance anal-yser (used for EIS only) with measurement adapter Agilent16048H for frequency range 40 Hz to 110 MHz. All testswere carried out with a homemade cylindrical sensor elementwith gold electrodes and two-electrode-configuration. Fig.1shows the setup for the Zahner Zennium and fig.2 shows thesetup for the Agilent 9294A.

Figure 1. Measurement setup of Zahner Zennium for EIS and CVmeasurements.

For future application a low-cost electronic is planned anda prototype exists. The sensor element is made of ceramicwith a 500 nm gold surface.

The temperature control was a Julabo LH46 Presto and adouble-walled integral exchanger reactor. High purity waterprepared with a Siemens LaboStar UV7 immediately beforethe experimental procedure was the basis of all applied so-lutions. The synthetic water samples were manufactured ac-cording to IEC 60734:2003 (IEC60734, 2003) and the home-made “Seuffer-Verfahren”.The reference measurement of each water sample was carried

J. Sens. Sens. Syst. www.j-sens-sens-syst.net

Figure 1. Measurement setup of Zahner Zennium for EIS and CVmeasurements.Gruden: Electrochemical Analysis of Water and Suds by Impedance Spectroscopy and Cyclic Voltammetry 3

Figure 2. Measurement setup of Agilent 9294A for EIS measure-ments (Gruden and Kanoun, 2013b).

out with the titrator TA20plus and the software TitriSoft 2.6from SI-Analytics. Reference detergent IEC A / IEC 60456(wfk, IEC60456) of Co. wfk – Testgewebe were used forthe suds. The water samples for the experiments with dif-ferent values of carbonate hardness (see table 1) are equalexcept of the hydrogen carbonate-chloride-ratio (c(HCO–

3 +

Cl–) = 2.56 mmol/L = const.). All these samples have totalion concentration of cion = 4.2mmol/L (Gruden and Kanoun,2013).

Table 1. List of water samples with a total hardness of 1.48 mmol/L(1.48 mmol/L) and different carbonat hardness caused by differentHCO–

3−Cl–− ratio.

Carbonate-Hardness (mmol/L)

HCO–3−Cl–

− ratio Conductivity χE

(µ S/cm)

1 0 Cl– only 411.5

2 0.25 1 : 1.35 392.7

3 0.5 1 : 0.63 374.3

4 0.69 1 : 0.25 347.4

5 0.93 HCO–3 only 337.8

4 Results and Discussion

The approach is to combine impedance spectroscopy andcyclic voltammetry in order to have a detailed water anal-ysis as a basis for optimal detergent dosage. By meansof impedance spectroscopy, the exact conductivity of themedium can be determined and changes in the compositionof the medium can be detected. At the same time, the rela-tive concentration αion (see eq. 1 and 2) of the species can bequantified by cyclic voltammetry. Many measurements havebeen carried out and show the feasibility of the combined

0 100 200 300 400−400

−300

−200

−100

0

<(Z)/Ω

=(Z

)/Ω

TH = 0.5 mmol/LTH = 1.0 mmol/LTH = 1.5 mmol/LTH = 2.0 mmol/LTH = 2.5 mmol/L

Figure 3. Dependence if the impedance spectra on total hardnessof water.

measurement principle.

cion = cCa2+ + cMg2+ + cNa+ + cHCO−3 + cCl− + cSO2−4

(1)

1 = αCa2+ +αMg2+ +αNa+ +αHCO−3 +αCl− +αSO2−4

(2)

Fig.3 shows the impedance spectra of water samples withdifferent values of total hardness of water. The impedancespectra are qualitatively equal because of the identical com-position of the water samples. The maximum turning point(see Fig.4) of the impedance spectra equals to the resistanceof the electrolyte RE and is frequency-dependent. The fre-

Figure 4. Qualitative plot of an impedance spectra.

quency values of the MTP of the spectras of fig. 3 are shownin table 2. The advantage of impedance spectroscopy com-pared to the usual single frequency conductance measure-ment is that the maximum turning point, and thus RE, can

www.j-sens-sens-syst.net J. Sens. Sens. Syst.


The combination of these measuring principles providesmore data from different frequency ranges and allows moredetailed investigations. The focus of this fluid analysis is theanalysis of data and not to electrochemical effects of the sen-sor element. With this method, expensive sensors and envi-ronmentally harmful materials can be avoided. Due to thefact that frequencies offmax = 5 MHz are needed, the costof the electronic circuit is adequate for the planned applica-tion.

The focus of this paper is the investigation of predefinedwater samples and one standard detergent, and to ignore realwater samples and commercial detergents because the fea-sibility of this new method should be tested. Tests with realwater samples and commercial detergents can be investigatedlater.

3 Experimental

The experimental setup for electrochemical impedance spec-troscopy (EIS) and cyclic voltammetry (CV) consisted of aZahner Zennium impedance analyzer for frequency range0.1 Hz to 4 MHz and an Agilent 9294A impedance ana-lyzer (used for EIS only) with measurement adapter Agilent

J. Sens. Sens. Syst., 3, 133–140, 2014 www.j-sens-sens-syst.net/3/133/2014/

R. Gruden et al.: Electrochemical analysis of water and suds by impedance spectroscopy and cyclic voltammetry 135

Table 1. List of water samples with a total hardness of1.48 mmol L−1 (1.48 mmol L−1) and different carbonat hardnesscaused by different HCO−3 / Cl− ratio.

Carbonate- HCO−3 / Cl− ConductivityχEhardness (mmol L−1) ratio (µS cm−1)

1 0 Cl− only 411.52 0.25 1 : 1.35 392.73 0.5 1 : 0.63 374.34 0.69 1 : 0.25 347.45 0.93 HCO−3 only 337.8

16048H for frequency range of 40 Hz to 110 MHz. All testswere carried out with a homemade cylindrical sensor ele-ment with gold electrodes and two-electrode configuration.Figure1 shows the setup for the Zahner Zennium and Fig.2shows the setup for the Agilent 9294A.

For future application a low-cost electronic is planned anda prototype exists. The sensor element is made of ceramicwith a 500 nm gold surface.

The temperature control was a Julabo LH46 Presto anda double-walled integral exchanger reactor. High purity wa-ter prepared with a Siemens LaboStar UV7 immediately be-fore the experimental procedure was the basis of all appliedsolutions. The synthetic water samples were manufacturedaccording to IEC 60734:2003 (IEC60734, 2003) and thecustom-made Seuffer procedure.

The reference measurement of each water sample was car-ried out with the titrator TA20plus and the software TitriSoft2.6 from SI-Analytics. Reference detergent IEC A/IEC60456 (wfk, IEC604562) of wfk-Testgewebe GmbH wereused for the suds. The water samples for the experimentswith different values of carbonate hardness (see Table1)are equal except for the hydrogen carbonate / chloride ratio(c (HCO−

3 + Cl−) = 2.56 mmol L−1 = const.). All these sam-ples have a total ion concentration ofcion = 4.2 mmol L−1

(Gruden and Kanoun, 2013a).

4 Results and discussion

The approach is to combine impedance spectroscopy andcyclic voltammetry in order to have a detailed water anal-ysis as a basis for optimal detergent dosage. By meansof impedance spectroscopy, the exact conductivity of themedium can be determined and changes in the compositionof the medium can be detected. At the same time, the relativeconcentrationαion (see Eqs.1 and2) of the species can bequantified by cyclic voltammetry. Many measurements havebeen carried out and show the feasibility of the combinedmeasurement principle.

2wfk: IEC A* Referenzwaschmittel nach IEC 60456 OrderCode: 88010-1, WFK IEC A*, wfk-Testgewebe GmbH, Brüggen-Bracht.

Gruden: Electrochemical Analysis of Water and Suds by Impedance Spectroscopy and Cyclic Voltammetry 3



3 +



3−Cl–− ratio.


HCO–3−Cl–


(µ S/cm)

1 0 Cl– only 411.5

2 0.25 1 : 1.35 392.7

3 0.5 1 : 0.63 374.3

4 0.69 1 : 0.25 347.4

5 0.93 HCO–3 only 337.8



0 100 200 300 400−400

−300

−200

−100

0

<(Z)/Ω

=(Z

)/Ω





(1)


(2)





Figure 3. Dependence of the impedance spectra on total hardnessof water.

Table 2. MTP frequencies of the impedance spectras of Fig.3.

TH [mmol L−1] Frequencyf [kHz]

0.5 3.51.0 7.01.5 17.82.0 20.02.5 35.5

cion = cCa2+ + cMg2+ + cNa+ + cHCO−

3+ cCl− + cSO2−

4(1)

1 = αCa2+ + αMg2+ + αNa+ + αHCO−

3+ αCl− + αSO2

4−(2)

Figure3 shows the impedance spectra of water samples withdifferent values of total hardness of water.

The impedance spectra are qualitatively equal because ofthe identical composition of the water samples. The maxi-mum turning point (see Fig.4) of the impedance spectra isequal to the resistance of the electrolyteRE and is frequencydependent.

The frequency values of the MTP of the spectras of Fig.3are shown in Table2.

The advantage of impedance spectroscopy compared tothe usual single frequency conductance measurement is thatthe maximum turning point, and thusRE, can be preciselydetermined by the multi-frequency method. The conductanceGE (reciprocal ofRE) correlates linearly with the total hard-ness of water (see Fig.5) with a correlation coefficient ofr = 0.9997.

Therefore, the resulting sensitivity of STH =

4.85 mS L mmol−1 is very good for the planned application.

www.j-sens-sens-syst.net/3/133/2014/ J. Sens. Sens. Syst., 3, 133–140, 2014





3 +



3−Cl–− ratio.


HCO–3−Cl–


(µ S/cm)

1 0 Cl– only 411.5

2 0.25 1 : 1.35 392.7

3 0.5 1 : 0.63 374.3

4 0.69 1 : 0.25 347.4

5 0.93 HCO–3 only 337.8



0 100 200 300 400−400

−300

−200

−100

0

<(Z)/Ω

=(Z

)/Ω





(1)


(2)






This high correlation is only achievable with synthetic waterby use of the precise impedance spectroscopy method. Astatistic of 124 real water samples of different German areasshows that the correlation coefficient of these water samplesis r = 0.8568. The difference between synthetic and realwater is caused by the different ion concentrations of theions which do not contribute to total hardness (see Table3,Eqs.3 and4).

χ0E = cCa2+ · 30

Ca2+ + cMg2+ · 30Mg2+ + cNa+ · 30

Na+

+ cHCO−

3· 30

HCO−

3+ cCl− · 30

Cl− + cSO2−

4· 30

SO2−

4(3)

χTH = χE − χε → χTH ≈ χE (4)

The relationship between conductanceGE and conductivityχE is given by the cell constantK of the sensor element.

χE = GE · K (5)

The conductivity,χE, of the water sample is composed ofthe concentration of each species,cSP, multiplied by its mo-lar limit conductivity,30. The portion of conductivity whichrepresents the total hardness of water,χTH, cannot be clearlydefined because of the different combination possibilities ofthe hardness forming cations Ca2+ and Mg2+ with its pos-sible corresponding anions HCO−

3 , Cl− and SO2−

4 . All com-binations between the cations and anions are possible andlead to different conductivities which do not correlate to thechemical effects. The arithmetic mean ofχTH equalsχE ifthe faultχε is low or the requirements to the results are notstrict. This results in the need for a differentiated analysis ofions in case of precise measurement requirements.



TH/mmol/L Frequency f kHz

0.5 3.5

1.0 7.0

1.5 17.8

2.0 20.0

2.5 35.5

0 0.5 1 1.5 2 2.5 32

4

6

8

10

12

14

Total Hardness /mmol/L

Con

duct

ance

/mS

Figure 5. Dependence of the conductance on total hardness of wa-ter.

be precisely determined by the multi-frequency method. Theconductance GE (reciprocal of RE) correlates linearly withthe total hardness of water (see Fig.5) with a correlation co-efficient of r = 0.9997. Thereof the resulting sensitivity ofS TH = 4.85mS ·L/mmol is very good for the planned appli-cation. This high correlation is only achievable with syntheticwater by use of the precise impedance spectroscopy method.A statistic of 124 real water samples of different German ar-eas shows that the correlation coefficient of these water sam-ples is r = 0.8568. The difference between synthetic and realwater is caused by the different ion concentrations of the ionswhich do not contribute to total hardness (see table 3, eq. 3and eq. 4).

χ0E = cCa2+ ·Λ0

Ca2+ + cMg2+ ·Λ0Mg2+ + cNa+ ·Λ0

Na+

+cHCO−3 ·Λ0HCO−3

+ cCl− ·Λ0Cl− + cSO2−

4·Λ0

SO2−4

(3)

χTH = χE − χε→ χTH ≈ χE (4)

Table 3. List of the ions which contribute mainly to the total ionconcentration of water.

Ion Contribution to Remarks

Ca2+ Total hardness (TH) Together with Mg2+ mainion for TH

Mg2+ Total hardness (TH) Together with Ca2+ main ionfor TH

Na+ Only total ion concentration(TIC)

-

K+ Only total ion concentration(TIC)

-

HCO–3 Carbonate hardness (KH)

Cl– Non carbonate hardness(NKH)

All anions contribute to theconductivity of the totalhardness, depending on thecomposition.

SO2–4 Non carbonate hardness

(NKH)

The relationship between conductance GE and conductivityχE is given by the cell constant K of the sensor element.

χE = GE ·K (5)

The conductivity χE of the water sample is composed ofthe concentration of each species cSP multiplied by its mo-lar limit conductivity Λ0. The portion of conductivity whichrepresents the total hardness of water χTH cannot be clearlydefined because of the different combination possibilities ofthe hardness forming cations Ca+

2 and Mg+2 with its possible

corresponding anions HCO–3, Cl– and SO2–

4 . All combinationsbetween the cations and anions are possible and lead to dif-ferent conductivities which do not correlate to the chemicaleffects. The arithmetic mean of χTH equals to χE if the faultχε is low or the requirements to the results are not strict. Thisresults in the need for a differentiated analysis of ions in caseof precise measurement requirements.Fig. 6 shows the impedance spectra of five water samples, allof them has a total hardness of 1.48 mmol/L (8.31 °dH) butdifferent carbonate hardness values (see table 1). All spec-tra are qualitatively equal. The frequency of the MTP isf = 7kHz at all spectras and the highest frequency point isf = 4MHz. To distinguish them from the spectra of differenttotal hardness values, further information by an additionalmeasurement principle is necessary.Fig. 7 shows a schematic diagram of two cyclic voltamme-try measurements. At tI the different current responses (I1and I2) are used to distinguish carbonate hardness. At It thedifferent relaxation times (t1 and t2) are used to distinguishcalcium-magnesium-ratios.


Figure 5. Dependence of the conductance on total hardness of wa-ter.Gruden: Electrochemical Analysis of Water and Suds by Impedance Spectroscopy and Cyclic Voltammetry 5

150 200 250−150

−100

−50

0

<(Z)/Ω

=(Z

)/Ω

KH: 0 mmol/LKH: 0.5 mmol/LKH: 1 mmol/LKH: 1.4 mmol/LKH: 1.85 mmol/L

Figure 6. Impedance spectra of water samples with the same totalhardness and different carbonate hardness.

Figure 7. Two typical cyclic voltammetry measurements of differ-ent water samples.

Fig. 8 shows the results of the water investigation bycyclic voltammetry. A sensitivity of S i = −6.7µA ·L/mmolwas reached, which is sufficient for the target application.The absolute value of current I decreases linearly with theincreasing carbonate hardness. The reason is that the ratioof the anions changes. The concentration of HCO–

3 increasesand the concentration of Cl– ions decrease. Cl– ions have ahigher contribution to the conductivity than HCO–

3.Fig. 9 shows the impedance spectra of water samples withidentical total and carbonate hardness but with a differentcalcium-magnesium-ratio. The frequency of the MTP is f =

4kHz at all spectras and the maximum frequency is f =

4MHz. The impedance spectra are qualitatively and quantita-tively identical and thus it is not possible to distinguish differ-ent calcium-magnesium-ratios by impedance spectroscopy.The water samples are nearly identical from electrochemical

0.6 0.8 1 1.2 1.4−92

−90

−88

−86

Carbonate Hardness /mmol/L

Cur

rent

/µA

Figure 8. Current response as a function of carbonate hardness.

200 220 240 260 280

−150

−100

−50

0

<(Z)/Ω

=(Z

)/Ω

Ca/Mg = 3:1Ca/Mg = 1:1Ca/Mg = 1:3

Figure 9. Impedance spectra of water samples with the same to-tal and carbonate hardness but different calcium-magnesium-ratio(mmol/L).



Figure6 shows the impedance spectra of five water sam-ples, all of them has a total hardness of 1.48 mmol L−1

(8.31 dH) but different carbonate hardness values (see Ta-ble1).

All spectra are qualitatively equal. The frequency of theMTP is f = 7 kHz at all spectras and the highest frequencypoint is f = 4 MHz. To distinguish them from the spectraof different total hardness values, further information by anadditional measurement principle is necessary.

Figure7 shows a schematic diagram of two cyclic voltam-metry measurements.

At tI the different current responses (I1 andI2) are used todistinguish carbonate hardness. AtIt the different relaxation



Table 3. List of the ions which contribute mainly to the total ion concentration of water.

Ion Contribution to Remarks

Ca2+ Total hardness (TH) Together with Mg2+ main ion for THMg2+ Total hardness (TH) Together with Ca2+ main ion for TH

Na+ Only total ion concentration (TIC) –K+ Only total ion concentration (TIC) –

HCO−

3 Carbonate hardness (KH)Cl− Non carbonate hardness (NKH) All anions contribute to the conductivity of the total hardness, depending on the composition.Cl− Non carbonate hardness (NKH) All anions contribute to the conductivity of the total hardness, depending on the composition.SO2−

4 Non carbonate hardness (NKH)


150 200 250−150

−100

−50

0

<(Z)/Ω

=(Z

)/Ω









0.6 0.8 1 1.2 1.4−92

−90

−88

−86


Cur

rent

/µA


200 220 240 260 280

−150

−100

−50

0

<(Z)/Ω

=(Z

)/Ω

Ca/Mg = 3:1Ca/Mg = 1:1Ca/Mg = 1:3




times (t1 andt2) are used to distinguish calcium-magnesium-ratios.

Figure 8 shows the results of the water investigation bycyclic voltammetry.

A sensitivity of Si = −6.7 µA L mmol−1 was reached,which is sufficient for the target application. The absolutevalue of currentI decreases linearly with the increasing car-bonate hardness; the reason is that the ratio of the anionschanges. The concentration of HCO−

3 increases and the con-centration of Cl− ions decrease. Cl− ions have a higher con-tribution to the conductivity than HCO−3 .

Figure 9 shows the impedance spectra of water sampleswith identical total and carbonate hardness but with a differ-ent calcium-magnesium-ratio.

The frequency of the MTP isf = 4 kHz at all spectras andthe maximum frequency isf = 4 MHz. The impedance spec-tra are qualitatively and quantitatively identical and thus itis not possible to distinguish different calcium-magnesium-ratios by impedance spectroscopy. The water samples arenearly identical from electrochemical point of view becausethe earth alkaline metals calcium and magnesium behavevery similar.

Figure10shows the result of an investigation of the watersamples by cyclic voltammetry.


150 200 250−150

−100

−50

0

<(Z)/Ω

=(Z

)/Ω









0.6 0.8 1 1.2 1.4−92

−90

−88

−86


Cur

rent

/µA


200 220 240 260 280

−150

−100

−50

0

<(Z)/Ω

=(Z

)/Ω

Ca/Mg = 3:1Ca/Mg = 1:1Ca/Mg = 1:3





150 200 250−150

−100

−50

0

<(Z)/Ω

=(Z

)/Ω









0.6 0.8 1 1.2 1.4−92

−90

−88

−86


Cur

rent

/µA


200 220 240 260 280

−150

−100

−50

0

<(Z)/Ω

=(Z

)/Ω

Ca/Mg = 3:1Ca/Mg = 1:1Ca/Mg = 1:3



Figure 9. Impedance spectra of water samples with the same totaland carbonate hardness but different Ca / Mg ratio (mmol L−1).




point of view because the earth alkaline metals calcium andmagnesium behave very similar.Fig. 10 shows the result of an investigation of the watersamples by cyclic voltammetry. A correlation of the relax-ation time and the calcium-magnesium-ratio can be clearlyseen. The relaxation time depends on the ionic radii of thespecies. Different radii mean different speeds of the ions.The resolution of the sensor signal to distiguish the differ-ent calcium-magnesium-ratios is an important factor for theplanned application. For this experiment solutions with dif-ferent calcium-magnesium-ratios were used, the total con-centration was constant.

ctot = c(Ca) + c(Mg) = 1.5mmol/L = const. (6)

Fig. 11 shows the results of the relative current responses ofdifferent calcium-magnesium-ratios. The curves are scaledthat if the calcium-magnesium-ratio is 1:1 (c(Mg) = c(Cl))the current response is zero (I = 0). Differences from a bal-anced calcium-magnesium-ratio can be seen easily. Devia-tions leads to a current response I , 0 and the sensitivity ofthe sensor is S ≈ 0.9s ·L/mmol of Ca which is adequate forthe planned application.

The results of calcium-magnesium-ratio measurementcould be confirmed only with synthetic water samples. Testswith real water samples could not confirm the laboratory re-sults in the needed precision. The reasons for these differ-ent results could be the cross-sensitivity of the sensor systemto other ions. A new approch to enlarge the measurementmethod by further tests to suspend the influence of other ionsis in progess.These three characteristics of water (total hardness of wa-ter, carbonate hardness and calcium-magnesium-ratio) areimportant for detergent dosage and have directly influenceon the washing result (Detergent Manufacturer, 2012). Thewashing performance depends on the carbonate hardness andthe calcium-magnesium-ratio because hydrogen carbonateions react to insoluble calcium carbonate and attaches on thelaundry and on the heating elements of the washing machineand calcium and magnesium form salts with different solu-bility. All other ions, which have also impact on the washingprocess are not respected in this investigations. These ionswill investigated later.The new unit AQrate will be introduced to characterize water.

AQrate = f (T H,KH,Ca/Mg) (7)T H = c(Ca+

2 + Mg+2 ) (8)

KH = c(HCO−3 ) (9)

Ca/Mg =c(Ca+

2 )c(Mg+

2 )(10)

For detergent dosage recommendation AQrate will replacetotal hardness of water in a first step because it is much moreprecise and meaningful.

Fig. 12 shows the impedance spectra of suds with differentconcentrations cdet and table 4 shows the frequencies of the

0.2 0.4 0.6 0.8 1 1.2 1.4

9.6

9.8

10

10.2

10.4

10.6

Share in Ca /mmol/L

Rel

axat

ion

Tim

e/s

Ca/Mg=1:3

Ca/Mg=1:1

Ca/Mg=3:1

Figure 10. Relaxation time depending on calcium-magnesium-ratio.


Figure 10. Relaxation time depending on calcium-magnesium-ratio.

A correlation of the relaxation time and the Ca / Mg ratiocan be clearly seen. The relaxation time depends on the ionicradii of the species. Different radii mean different speeds ofthe ions. The resolution of the sensor signal to distinguishthe different calcium-magnesium-ratios is an important fac-tor for the planned application. For this experiment, solutionswith different calcium-magnesium-ratios were used, the totalconcentration was constant.

ctot = c(Ca) + c(Mg) = 1.5 mmolL−1= const. (6)

Figure11 shows the results of the relative current responsesof different calcium-magnesium-ratios. The curves are scaledthat if the Ca / Mg ratio is 1: 1 (c(Mg) = c(Cl)) the cur-rent response is zero (I = 0). Differences from a balancedCa / Mg ratio can be easily seen. Deviations lead to a cur-rent responseI 6= 0 and the sensitivity of the sensor isS ≈

0.9 s L mmol−1 of Ca which is adequate for the planned ap-plication.

The results of Ca / Mg ratio measurement could be con-firmed only with synthetic water samples. Tests with real wa-ter samples could not confirm the laboratory results withinthe needed precision. The reasons for these different resultscould be the cross-sensitivity of the sensor system to otherions. A new approach to enlarge the measurement methodby further tests to suspend the influence of other ions is inprogress.

These three characteristics of water (total hardness of wa-ter, carbonate hardness and calcium-magnesium-ratio) areimportant for detergent dosage and have direct influenceon the washing result (personal communication, 2012). Thewashing performance depends on the carbonate hardness andthe Ca / Mg ratio because hydrogen carbonate ions react toinsoluble calcium carbonate and attach to the laundry and onthe heating elements of the washing machine, and calcium


0 5 10 15 20 25 30 35−2

−1

0

1

2·10−5

Time /s

Cur

rent

/A

Mg=0Ca/Mg=2:1Ca=MgCa/Mg=1:2Ca/Mg=1:5Ca=0

Figure 11. Relative current response of different Ca/Mg-ratiosbased on a balanced Ca/Mg-ratio.

MTP. The maximum turning points correlate strongly with

0 50 100 150 200−200

−150

−100

−50

0

<(Z)/Ω

=(Z

)/Ω

4.06 g/L2.03 g/L1.02 g/L0.51 g/L0.25 g/L

Figure 12. Dependence of the IS on detergent concentration.

the concentration of the suds (r ≈ 1) as long as they do notexceed a certain value (see fig. 13).

Above this concentration ccmc, the conductivity increasesagain linearly with the concentration, but with a lesser gra-dient. Thus there are two lines and their intersection pointmarks the critical micelle concentration (CMC) (Nakamuraet al., 1998; Neto et al., 2006; Chang et al., 1998). There is astrong presumption that the optimum detergent concentration


c/g/L Frequency f kHz

0.25 48.4

0.51 71.4

1.02 153

2.03 172

4.06 191

Figure 13. Dependence of the conductance on detergent concentra-tion

is close to the critical micelle concentration. On this issue in-tensive tests are carried out currently (Gruden and Kanoun,2013).

5 Conclusion and Outlook

The results indicate that it is possible to determine the deter-gent concentration of suds and the parameters of water whichhas impact on the washing process by a combination ofimpedance spectroscopy (EIS) and cyclic voltammetry (CV).The results of calcium-magnesium-ratio measurement couldbe confirmed only with synthetic water samples. Tests withreal water samples could not confirm the laboratory resultsof calcium-magnesium-ratio in the needed precision. A bigadvantage of impedance spectroscopy as a multi-frequencymethod in comparison to conventional conductance measur-ing methods is the possibility to determine RE precisely andthus form a solid basis for all further investigations. More-over, further investigations are in progress to get more infor-mation from the impedance spectra to improve the plannedapplication.


Figure 11. Relative current response of different Ca / Mg-ratiosbased on a balanced Ca / Mg-ratio.

and magnesium form salts with different solubility. All otherions, which also have impact on the washing process are notrepresented in this investigation. These ions will investigatedlater.

The new unit AQrate will be introduced to characterizewater.

AQrate= f (TH,KH,Ca/Mg) (7)

TH = c(Ca2++ Mg2+) (8)

KH = c(HCO−

3 ) (9)

Ca/Mg =c(Ca2+)

c(Mg2+)(10)

For detergent dosage recommendation AQrate will replacetotal hardness of water in a first step because it is much moreprecise and meaningful.

Figure12 shows the impedance spectra of suds with dif-ferent concentrationscdet and Table4 shows the frequenciesof the MTP.

The maximum turning points (MTP) correlate stronglywith the concentration of the suds (r ≈ 1) as long as theydo not exceed a certain value (see Fig.13).

Above this concentrationcCMC, the conductivity increasesagain linearly with the concentration, but with a lesser gra-dient. Thus there are two lines and their intersection pointmarks the critical micelle concentration (CMC) (Nakamuraet al., 1998; Neto et al., 2006; Chang et al., 1998). Thereis a strong presumption that the optimum detergent concen-tration is close to the critical micelle concentration. On thisissue intensive tests are currently being carried out (Grudenand Kanoun, 2013a).




0 5 10 15 20 25 30 35−2

−1

0

1

2·10−5

Time /s

Cur

rent

/A

Mg=0Ca/Mg=2:1Ca=MgCa/Mg=1:2Ca/Mg=1:5Ca=0



0 50 100 150 200−200

−150

−100

−50

0

<(Z)/Ω

=(Z

)/Ω

4.06 g/L2.03 g/L1.02 g/L0.51 g/L0.25 g/L






0.25 48.4

0.51 71.4

1.02 153

2.03 172

4.06 191








c [g L−1] Frequencyf [kHz]

0.25 48.40.51 71.41.02 1532.03 1724.06 191

5 Conclusion and outlook

The results indicate that it is possible to determine the deter-gent concentration of suds and the parameters of water whichhave impact on the washing process by a combination ofimpedance spectroscopy (EIS) and cyclic voltammetry (CV).The results of Ca / Mg ratio measurement could be confirmedonly with synthetic water samples. Tests with real water sam-ples could not confirm the laboratory results of Ca / Mg ratiowithin the needed precision. A big advantage of impedancespectroscopy as a multi-frequency method in comparison toconventional conductance measuring methods is the possi-bility to determineRE precisely and thus form a solid basisfor all further investigations. Moreover, further investigationsare in progress to get more information from the impedancespectra to improve the planned application.

The detergent concentration of the suds can be measuredprecisely by impedance spectroscopy as well. The implemen-tation of this method for the CMC detection and thus the op-timization of the detergent amount and the washing processis in progress.

To determine the carbonate hardness and the calcium-magnesium-ratio, a combination of the results of impedancespectroscopy and cyclic voltammetry is necessary and the re-sults show the feasibility.


0 5 10 15 20 25 30 35−2

−1

0

1

2·10−5

Time /s

Cur

rent

/AMg=0Ca/Mg=2:1Ca=MgCa/Mg=1:2Ca/Mg=1:5Ca=0



0 50 100 150 200−200

−150

−100

−50

0

<(Z)/Ω

=(Z

)/Ω

4.06 g/L2.03 g/L1.02 g/L0.51 g/L0.25 g/L






0.25 48.4

0.51 71.4

1.02 153

2.03 172

4.06 191






Figure 13. Dependence of the conductance on detergent concentra-tion.

Both methods can be realized with the same sensor ele-ment and the same hardware. The planned application canbe realized as a low cost solution for an online measurementduring the washing process without environmentally harmfulsubstances. The realized measurements indicate the feasibil-ity of the planned application. For a final version some pointslike the online detection of the critical micelle concentrationand the precise correlation of the measurement data to thewater parameters still has to be clarified. Such tests are inprogress.

The new unit AQrate has to be specified and will be ex-tended by additional ion effects. With such adaptations, thenew unit AQrate can consider region-specific characteristicsof water.

Edited by: A. L. SpetzReviewed by: two anonymous referees

References

Barsoukov, E. and Macdonald, J. R.: Impedance Spectroscopy,Wiley-Interscience, 2005.

Brad, A. J. and Faulkner, L. R.: Electrochemical Methods, John Wi-ley & Sons, 2001.

Cammann, K. and Galster, H.: Das Arbeiten mit ionenselektivenElektroden, Springer Verlag, 1996.

Chang, H.-C., Hwang, B.-J., Lin, Y.-Y., Chen, L.-J., and Lin, S.-Y.:Measurement of critical micelle concentration of nonionic sur-factant solutions using impedance spectroscopy technique, Rev.Sci. Instrum., 69, 2514–2520, 1998.

Czyzewski, G., Schulze, I., and Engel, C.: Europäische Patentan-meldung EP 0 992 622 A2, 1999.

Eriksson, M., Lindgren, D., Bjorklund, R., Winquist, F., Sundgren,H., and Lundström, I.: Drinking Water monitoring with voltam-metric sensors, Procedia Engineering, 25, 1165–1168, 2011.

Gruden, R., Kanoun, O., and Tröltzsch, U.: Influence of surfaceeffects on the characterisic curves of detergent sensors, 9th In-



ternational Multi-Conference on Signals, Sensors and Devices,Chemnitz: 20–23 März 2012.

Gruden, R. and Kanoun, O.: Water quality assessment by combiningimpedance spectroscopy measurement with cyclic voltammetry,AMA Sensor, Nürnberg, 2013a.

Gruden, R. and Kanoun, O.: Fast and Low-Cost Online Detec-tion of Critical Micelle Concentration based on ImpedanceSpecroscopy, The 7th International Conference on Sensing Tech-nology, 3–5 December, Wellington, New Zealand, 2013b.

Huang, Z., Li, K., and Xu, H.: Research of an online measurementand control system of water-quality on FIA, Chinese Journal ofScientific Instrument 36, 343–346 + 385, 2005.

Hütter, L. A.: Wasser und Wasseruntersuchung, Otto Salle Verlagund Verlag Sauerländer, 1990.

IEC60734: Household electrical applicances Performance – Hardwater for testing (IEC 60734:2003), CENELEC (European Com-mittee for Electrotechnical Standardization), Brussels, 2003.

Ivarsson, P.: Electronic Tongues – New Sensor Technology inHousehold Appliances, Linköping, Sweden, 2003.

Jakobi, G. and Löhr, A.: Detergents and Textile Washing, VCH,1987.

Kräuter, S., Lange, R., and Seifert, F.: Surface water quality mea-suring station adaptable to changing requirements, Water andWastewater International, 21, p. 41, 2006.

Nakamura, H., Sano, A., and Matsuura, K.: Determination of Crit-ical Micelle Concentration of Anionic Surfactants by CapillaryElectrophoresis Using 2-Naphthalenemethanol as a Marker forMicelle Formation, Journal of Analytical Sciences, 14, 379–382,1998.

Neto, J. M. G., da Cunha, H. N., Neto, J. M. M., and Ferreira,G. F. L.: Impedance spectroscopy analysis in a complex sys-tem: Sodium dodecyl sulfate solutions, J. Sol-Gel Sci. Techn.,38, 191–195, 2006.

Pungor, E., Nagy, G., and Feher, Z.: The flat surfaced membranecoated mercury electrode as analytical tool in the continuousvoltammetric analysis, J. Electroanal. Chem., 75, 241–254, 1977.

Rüdenauer, I. and Gensch, C.-O.: Einsparpotentiale durch automa-tische Dosierung bei Waschmaschinen, Öko-Institut e.V., Studieim Auftrag der Miele & Cie. KG, 2008.

Smulders, E., Rähse, W., von Rybinski, W., Steber, J., Sung, E., andWiebel, F.: Laundry Detergents, Wil, 2002.

Tröltzsch, U.: Anwendungspotential der Impedanzspektroskopiefür die Waschlaugensensorik, 16. GMA/ITG-Fachtagung Sen-soren und Messsysteme, 650–661, 2012.

Tschulena, G. and Lahrmann, A.: Sensors in Household Appli-ances, Wiley-VCH, 2006.

Verberk, J. Q. J. C., Hamilton, L. A., O’Halloran, K. J., Van DerHorst, W., and Vreeburg, J.: Analysis of particle numbers, sizeand composition in drinking water transportation pipelines: Re-sults of online measurements, Wa. Sci. Technol., 6, 35–43, 2006.

Vivancos, J.-L., Rácz, Z., Cole, M., and Gardner, J. W.: Surfaceacoustic wave based analytical system for the detection of liquiddetergents, Sensor. Actuat. B-Chem., 171, 469–477, 2012.

Wagner, G.: Waschmittel, Viley VCH, 2010.Winquist, F.: Voltammetric electronic tongues – basic principles and

applications, Microchim. Acta, 163, 3–10, 2008.Winquist, F., Wide, P., and Lundström, I.: An electronic tongue

based on voltammetry, Anal. Chim. Acta, 357, 21–31, 1997.Winquist, F.: Multicomponent analysis of drinking water by a

voltammetric electronic tongue, Anal. Chim. Acta, 683, 192–197, 2011.


Electrochemical analysis of water and suds by impedance ... · Electrochemical analysis of water...

Documents

Transcript of Electrochemical analysis of water and suds by impedance ... · Electrochemical analysis of water...