Solid-contact ion-selective electrode with stable internal electrode

ANALWICA

CHIMICA ACTA Analytica Chimica Acta 320 (1996) 53-61

Solid-contact ion-selective electrode with stable internal electrode Maria Vamvakaki,

Nikolas A. Chaniotakis

*

Department of Chemistry, The University of Crete, 714 09 Iraklion, Crete, Greece

Received 13 June 1995; revised 10 October 1995; accepted 15 October 1995

Abstract A new substrate for the construction of solid contact ion-selective electrodes (SC-ISEs) based on a conductive and porous carbon matrix is presented. The conductive characteristics of the substrate, its controlled porosity in combination with the use of a non-aqueous second kind internal reference element offers considerable advantages over the symmetrical configurations. The SC-ISEs presented here are very easy to construct, can be made very small in size, and can be used in

high pressure environments. At the same time it is shown that the characteristics of the two electrodes constructed (K+ and NO;) exhibit at least the same performance characteristics as the best commercially available symmetric counterparts. The potentiometric response is monitored over a period of one month with excellent results with regard to drift, base line stability, response time, selectivity, Keyw0rd.s: Ion selective electrodes;

and detection limit.

Potentiometry;

Solid contact

1. Introduction The use of ion-selective electrodes (ISEs) in the analysis and monitoring of clinical, environmental and industrial ions and gasses has been continuously expanding [1,2]. For routine analysis, the symmetric configuration is generally preferred in which the sensing membrane comes in direct contact with two aqueous solutions, the internal with fixed ion activity and the external with the ion activity to be measured. Even though the symmetric ISEs have found a wide range of applications, they still have certain inherent limitations [3-61. They are mechanically complicated, and thus difficult to manufacture in small size, the internal reference solution increases the system

* Corresponding

author.

0003.2670/96/$15.00 0 1996 Elsevier Science B.V. All rights reserved SSDI 0003-2670(95)00524-2

impedance, and finally, due to internal compartment, they cannot withstand high pressures. Solid internal contact electrodes (SC-ISEs) refer to a type of ISEs in which the internal reference element is in direct contact with the electroactive membrane matrix and thus contains no aqueous solution [3-121. Due mainly to the elimination of the internal reference aqueous solution, these electrodes will have certain advantages such as their small size, lower cost of production, and ability to operate in high pressure environments where the symmetric ISEs might be damaged. Examples of this sensor design include the coated-wire electrodes (CWE) [13,14], hybrid sensors and the ion-selective field effect transistors [11,12]. Since the introduction of the first graphite rod ISE [S] and the first coated wire electrode [15], a considerable amount of work has been devoted to the construction of SC-ISEs with

M. Vamuakaki, N-4. Chaniotakis/Analytica

54

stable and reproducible potentials [9-121. The key issue in the design of these sensors has been the precise control of the interface between the organic electroactive membrane and the internal reference element. Such control is essential in order to obtain potentials which are stable, are not dependent on redox species, nor are affected by fluctuations of water vapors from the solution. The failure so far to achieve the above specific goals as well as the inability to precisely define the interfacial potential has limited the commercial success of the SC-ISEs. The SC-ISEs presented here are based on a new electroactive membrane support with high conductivity, well controlled and reproducible porosity for organic phases, and it is commercially available at very low cost. These characteristics are combined with the use of a non aqueous second kind reference element, offering a SC-ISE with all the advantages of such a system over the symmetrical configurations. The establishment of a well defined internal electrode potential is achieved with the use of a

& / AgCl 1O.lMKCl ,

,

Chimica Acta 320 (1996) 53-61

Ag/AgX wire in contact with the halide-saturated electroactive membrane phase. The SC-ISEs constructed with the proposed method are very simple to make and easy to maintain, while they can be used in the same way as the symmetric ISEs. Their size can vary considerably (from cm to less than mm) depending on the application requirements while the elimination of the internal compartment would allow their operation under high pressures. For the sake of demonstration SCISEs for one anion, NO;, and one cation, Kf, are presented here. The same configuration can be used for the construction of electrodes for many species as long as the theoretical requirements given in the theoretical section are fulfilled.

2. Theory The symmetrical configuration of an ISE contains the internal and the external interface, which under most circumstances are similar:

1Membrane

1 Sample

t

t

11REF

lnbnal Reference External Interface

In this arrangement, the phase boundary potentials 8, and S, are well defined, and assumed not to influence one another. In order for an ISE to have stable and reversible potential, the following three requirements need to be fulfilled regarding its interfaces [5]: (a) There should be a reversible and stable transition from electronic to ionic conductivity and vice versa at the internal reference electrode interface. (b) The exchange currents taking place at the reference elements which depend on the electrode reactions should be very high. (cl The internal contact must be isolated and well protected from any side reactions. When the reference element of an ISE is replaced

62

by a redox system (Pt, Au) to form a solid internal contact electrode, the following system is obtained:

’

M

I$E 1 Membrani

1 Sample

11REF

It can be seen that the internal boundary potential cannot be precisely controlled due to two basic reasons. 6) The potential of the metal surface depends on the redox species present at its surface. The concentration of these species could easily be altered

M. Vamuakaki, N.A. Chaniotakis /Analytica Chimica Acta 320 (1996) 53-61

by those in the test solution. (ii) Since the charge transfer across the ISE membranes is usually carried by ions, and since there are no redox systems close to the internal reference electrode coupled to the ionic transfer in order to complete the electronic charge conduction, there is no well defined internal reference potential. The replacement of the noble metal with an Ag/AgCl or Cu contact has found considerable use, but still its behavior is not well documented and the interface potentials generated are still not very stable [7,16]. In the case of Ag/AgX, the activity of the anion X- will determine the potential of the M/MX couple. The amount of water that penetrates the membrane becomes the major factor influencing this potential, since it usually drastically alters the activity of the ions present. One approach in solving this problem is to deposit solid electrolyte (i.e., KCl) or an aqueous gel, which contains a certain activity of a primary ion, directly onto the Ag/AgCl or Cu reference elements. Such an approach generates a symmetric ISE, but the osmotic pressure of water inside the membrane cannot be controlled, leading to continuous changes in ion activities. The accumulation of water will result in the “blowing” of the ISE membrane and finally its detachment from the metal support. On the other hand the use of a thicker membrane for the decrease of the water transport will result in a drastic increase in the resistance of the membrane, which introduces noise to the mea-

Internal Reference - Membrane MX + e- __ M++ x-

Membrane Carrier

mem

__

-x-

Interface

surements, and increases the response time of the electrode. The SC-ISEs described in this report overcome these problems, and offer an alternative to the construction of non-symmetrical liquid membrane potentiometric sensors. A schematic diagram of the system presented here is as follows:

I$E Ag / AgCl , KCl, Membrane I Sample 11REF,,, 1

Lipophilic Organic phase

I

The potentiometric half-cell considered here consists of two parts the internal solid reference and the electroactive sensing element (Fig. 1). To be well defined the internal reference system must have the previously described characteristics, that is a stable potential, fast electronic to ionic conductivity transition, and free from side reactions. It is known that if the concentration of a salt MR or RX in the membrane organic phase is high enough, the condition that aM or a, is much larger than is fulfilled [5] where K,, is the solubility \iK,, product of the MX in the membrane phase. Under this condition the organic solution will possess high buffering towards the activities of the Mf and Xions, thus the M/MX internal reference-contact element (such as that of Ag/AgCl) will produce a

Internal Reference Electrode (IWMX)

M MX

- Solution Interface

+X

Fig. 1. Schematic

55

diagram of the SC-ISE. Not shown to scale.

56

M. Vamvakaki, N.A. Chaniotakis/Analytica

stable potential. The additional M+ or X- come from the addition of solid MX on the surface of the internal electrode, as described in the experimental part. The possibility to use different metal reference elements (i.e., Cu, Pb, Hg) as well as salts (RX where X = Cl-, Br-, I-, ClO, or MR where M = K+, Ag+ Cu’> lead to a wide range of potentiometric systems. The present SC-ISE has been designed based on the idea that the amount of a salt in the organic electroactive phase is large in order to stabilize the interface potential as mentioned above, and thus satisfies the internal reference stability requirement. In the case of the K+ -selective SC-ISE the high activity of Cl- in the membrane phase is obtained by the use of a layer of solid KC1 in contact with the organic phase [17,18] (Fig. 1). The same holds in the case of the NO; selective SC-ISE. It should be mentioned that the presence of the ion carriers will help in the dissolution of the salt in the membrane phase. The electronic conductivity requirement is satisfied when the following reversible fast equilibrium exists: MX+e-

4M”+X-

(1)

where M is the metal conductor (i.e., Ag), and MX is the metal salt such as AgCl. The ionic to electronic conductivity will be satisfied only in the case that one of the following equilibria applies: M”+ + Carrier(,,,)

$ M”f-Carrier(,emJ

(2)

X”- + Carrier(,,,)

% X”--Carrier(,,,)

(3)

These equilibria will offer the required buffering capacity in the system to support the equilibrium shown in Eq. (1). It is clear that when either a complexing carrier such as valinomycin or Niphenanthroline, or an ion exchanger, such as tridodecyl methyl ammonium chloride (TDMACL) or potassium tetraphenyl borate (TPhB), which has selectivity for either the anion of the metal salt that takes part in the reference couple (Cl- in the case of AgCl), or a counterion of Cl- (such as K+ in the case of AgCl) is employed for the development of anionic or cationic sensitive membrane, equilibrium (1) is satisfied. In the case of the K’-selective

Chimica Acta 320 (1996) 53-61

SC-ISE valinomycin will complex Eq. (2) yielding Eq. (4):

K+ according

to

KCl, + Valinomycin(,,,, % K+-Valinomycin(,,,,

+ Cl-

(4)

Similarly the NO, selective SC-ISE, the Niphenanthroline exchanger possesses affinity for Cl(see NO, selectivity Fig. 4), thus Eq. (3) is satisfied yielding Eq. (5): KCl, + Ni-Phen(,,,,

% Cl--Ni-Phen,,,,,

+ K+ (5)

The third requirement, protection of the internal reference element from any side reactions, can be achieved by preventing water from the aqueous solution to reach the internal reference element, and thus change the activities of the electroactive species. Since the presence of just a polymeric membrane is not sufficient to prevent the transport of water vapor to the reference contact, it is thought that the introduction of a very lipophilic and conductive spacer between these two phases will help, at least for a certain period of time. The spacer used for the construction of the electrodes presented here has very high conductivity (25 s1 cm-‘) allowing a more remote position of the internal reference element, while its impregnation with the organic cocktail (plasticizer, carrier, and additives) provides the required high lipophilicity. The assessment of the potential stability of these SC-ISEs is the purpose of this study.

3. Experimental

3. I. Reagents Valinomycin used in potassium electrodes was purchased from Sigma (St. Louis, MO). Niphenanthroline in o&o-nitrophenyl octyl ether (oNPOE) used as a NO, carrier-plasticizer cocktail was obtained from AT1 Orion Research (Boston, MA). Bis(2-ethylkexyl) sebacate (DOS) and highmolecular-weight PVC used in the membrane cocktail were obtained from Fluka. The carbon rods were No. 192/l. Masterflex@ 6409-16 Tygon@ tube, manufactured by Norton was used in the electrode

M. Vamuakaki, N.A. Chaniotakis/Analytica

construction. All other chemicals were of reagent grade. Tetrahydrofuran (THF) was freshly distilled before use. Symmetric commercial electrodes, for comparison, were obtained from ATI Orion: Model 93-07 Nitrate and 93-19 Potassium. Deionized, 18.3 MR water was used to prepare all aqueous solutions. 3.2. Electrode

51

Chimica Acta 320 (1996) 53-61

tubing, and subsequently dipped into the liquid carrier cocktail. Capillary action enables the carbon rod to absorb the carrier cocktail, and to completely saturate it in a few hours. Finally the exposed side of the electrode and the side of the Tygon tubing is covered with the PVC-based membrane cocktail, which is allowed to evaporate overnight, forming a 0.01 cm thick membrane.

construction 3.2.1. EMF measurements All EMF measurements were performed at 25 ? 1°C. The temperature was controlled with an Edmund Biiller, D-7400 water heater-circulator in double wall beakers. Millivolt measurements were obtained using a Xenon CI-317, &channel Electrochemical Measuring System (Xenon Halanthri, Athens) with input impedance > 10’” 0, versus an AT1 Orion Research 900200 double junction reference electrode. All other measurement procedures were obtained from the Orion Model 93-07 Nitrate

Round carbon rods with 0.55 cm diameter and 1.0 cm length were used for the construction of the electrode sensor element. A hole of the same diameter as that of the reference silver wire (0.05 cm diameter and 0.2 cm depth) is drilled in one end of the carbon rod. The anodized silver wire is then placed tightly into the hole of the carbon rod, and soldered to a shielded wire. This assembly is placed into a Masterflex 6409-16 Tygon tubing, 15 cm long, with the carbon rod extending 0.05 cm outside the

t

lcoj

loo

40

40

20

20

0

0 0

100

200

300

400

500

600

Time (hrs) Fig. 2. Potentiometric

stability of NO;

electrodes

over 30 days.

0 = SC-ISEs; 0 = commercial electrodes

58

M. Vamvakaki, NA. Chaniotakis/Analytica

Electrode and Model 93-19 Potassium Electrode instruction manuals [19,20]. The long term stability measurements were performed in a solution of 0.01 M sodium nitrate or potassium chloride, respectively, also containing the ionic strength adjustment buffers suggested in the Orion manuals. The electrodes were tested and stored in the appropriate solution until the next measurement at which time the test solution was replaced with a fresh one. The selectivity measurements were performed by the addition of 2.50 ml of 0.10 M of the appropriate salt solution into 25.00 ml of the ionic strength adjustment buffer, bringing the final concentration of the salt to 0.0153 M. The calibration curves were obtained by serial addition of concentrated sample. The data was collected in a Dale 486 DX personal computer via a ADC-16, 16-bit Keithley A/D converter board, and displayed using a program written in BASIC. The measurements were not corrected for junction potentials or activity

0

200

400

Chimica Acta 320 (1996) 53-61

coefficients. No pretreatment was performed on any of the electrodes before use.

4. Results and discussion The experiments performed with this type of electrode were designed in order to compare directly the potentiometric behavior of the SC-ISEs with that of the commercially available symmetric ISEs, which are the state of the art. In particular, the potentiometric stability over time, response time, baseline return, selectivity, and detection limit were examined. 4.1. Potential stability of the SC-ISEs over time The major drawback of SC-ISEs has been their large potential drift. In order to determine the stability of the SC-ISE systems presented here, their po-

600

800

1000

Samples (lsamplekc)

Fig. 3. Selectivity

of K+ electrodes.

0 = SC-ISEs; 0 = commercial

electrodes (the n-axis of the latter was offset by 30 s for clarity).

M. Vamuakaki, N.A. Chaniotakis/Analytica

Chimica Acta 320 (1996) 53-61

tential was monitored over a one-month period. Fig. 2 shows the EMF stability of the NO; electrodes. It is clear that the potential drift of the SC-ISE electrode is comparable to the symmetric commercial ISE. The small variations of the potentials between measurements observed with both electrodes could be due to drifts of the measuring systems, such as that of the reference potential and small temperature fluctuations. This high potential stability displayed by the SC-ISEs indicates that the internal reference element remains free from any water transport over at least one month of continuous examination, while the electroactive membrane remains intact, and fully operational (see also calibration results). The same results were obtained in the case of the K+ electrode (results not shown). It should also be mentioned that initially the impregnated carbon rod was examined for response

54

to oxygen or carbon dioxide. When the blank electrode with or without the polymeric membrane was immersed in solutions of different partial pressures of the two gases, no detectable potential change or drift was observed for a period of 1 h. 4.2. Selectivity The selectivities of the electrodes were examined in order to determine the chemical stability of the SC-BE systems, and the possible influence of the carbon rod on the electroactive membrane. Even though the SC-BE K+ selective membrane did not contain any lipophilic additives, the obtained selectivities are comparable. This is probably due to the fact that the carbon matrix contains enough charged sites [21]. Figs. 3 and 4 show the selectivities of the SC-ISEs and the commercial ISEs. The results pre-

SCN

0

500

1500

loo0

2G00

Samples (1 sample/set) Fig. 4. Potentiometric for clarity).

selectivity

of NO;

electrodes.

0 = SC-ISEs;

0 = commercial

electrodes

(the x-axis of the latter was offset by 30 s

60

M. Vamuakaki, NA. Chaniotakis/Analytica

Chimica Acta 320 (1996) 53-61

~57.0

-7

-6

-5

mV

-4

-3

-2

log &+I

Fig. 5. Calibration

curves of K+ electrodes.

0 = SC-ISEs; 0 = commercial

electrodes.

200 s-52.2

-7

-4

-5

-6

mV

-3

-2

1% D703-1

Fig. 6. Calibration

curves of NO;

electrodes.

0 = SC-1%~;

0 = commercial

electrodes.

M. Vamuakaki, N.A. Chaniotakis/Analytica

sented here were obtained the first day of testing but the same results were obtained at the end of the test period of 30 days (data not shown). 4.3. Calibration

curve and detection limits

The results from the calibration curve of the electrodes provide information on the slope as well as the detection limit of the systems. Figs. 5 and 6 shows the calibration curves for K+ and NO, of the SC-ISE and commercial symmetric electrodes, while the slopes and detection limits are also indicated. The slopes of the K+ electrodes were identical at the end of the one-month test period, while that of the NO_; was a little lower (5-10%) for both the SC-BE and the Orion electrodes (data not shown). 4.4. Response

time and baseline return

The response time of the SC-ISEs is at least as fast as that of the symmetrical type, as can be seen from Figs. 3 and 4. The baseline return is also very fast, and reproducible for all electrodes examined Investigations are still under way for the evaluation of the SC-ISE matrix presented here for use in enzyme-based biosensors. At the same time the possibilities for use of these sensors in deep-sea research, in a micro-sensor array, as well as substrates for biosensors are under way.

5. Conclusion The use of a novel matrix, i.e., a highly conductive porous carbon rod, enables the construction of a new type of SC-ISEs with characteristics at least as good as the state of the art commercially available symmetrical ISEs over both short and long periods of time. The elimination of the potential drift due to the stable internal reference electrode, coupled with the possibility to use such systems in environments where symmetrical ISEs are not able to operate, may lead to their application in many new areas of ion analysis.

Chimica Acta 320 (1996) 53-61

61

Acknowledgements This work evolved from discussions with Steve West, and for this we are grateful. We would also like to thank ATI Orion Research and Xenon for their gifts.

References

[II

J. Wang, in Electroanalytical Techniques in Clinical Chemistry and Laboratory Medicine, VCH, New York, 1988. I21 GA. Junter (Ed.), Electrochemical Detection Techniques in the Applied Biosciences, Ellis Horwood, Chichester, 1988. t31 W. Selig, Anal. L&t., 15 (1982) 309. [41 R.W. Cattral and I.C. Hamilton, Ion Selective Electrode Rev., 6 (1984) 125. bl B.P. Nikolsky and E.A. Materova, Ion Selective Electrode Rev.. 7 (1985) 3. [61 L. Cunningham and H. Freiser, Anal. Chim. Acta, 180 (1986) 271. 171 L. Durselen, U. Oesterle, S. Schuppisser, H.-V. Pham, Y. Miyahara, W.E. Morf and W. Simon, Chimia, 44 (1990) 214. [81 J. Ruzicka, K. Rald, Anal. Chim. Acta, 53 (1971) 1. [91 C.P. Hauser, D.W.L. Chiang and G.A. Wright, Anal. Chim. Acta, 303 (1995) 241. [lOI M.M. Khalil, Anal. Lett., 26 (1993) 55. [ill J.D. Harrison, L.L. Cunningham, X. Li, A. Teclemariam and D.J. Permann, Electrochem. Sot.: Electrochem. Sci. Technol., t 1988) 2473. [121 F. Winquist and B. Danielsson, Biosensors (Practical Approach Series), Oxford, 1990, Chap. 7. 1131 A. Hulanicki, M. Trojanowicz and E. Pobozy, Analyst, 107 (1982) 1356. ]141 C.A. Stevens and H. Freiser, Anal. Chim. Acta, 248 (1991) 315. ]151 R.W. Cattrall and H. Freiser, Anal. Chem., 43 (1971) 1905. It61 T.A. Fjeldly, K. Nagy and J.S. Johanson, J. Electrochem. Sot., 126 (1979) 793. [17] M.L. Iglehart, R.P. Buck and E. Pungor. Anal. Chcm., 60 (1988) 290. [18] K. Cammann. Anal. Chem., 50 (1978) 936. [19] AT1 Orion Research Incorporated Model 93-07 Nitrate Electrode, Instruction Manual, 1991. [20] AT1 Orion Research Incorporated Model Y3-19 Potassium Electrode, Instruction Manual, 1991. [21] R.L. McCreery, in A.J. Bard (Ed.), Electroanalytical Chemistry, A Series of Advances, Vol. 17. Carbon Electrodes: Structural Effects on Electron Tranfer Kinetics, Marcel Dekker, New York, pp. 255-280.