In all of the membrane electrodes described so far, an internal reference solution and an internal reference electrode have been used to complete the circuit. The internal conditions
k k were always kept constant so that the internal membrane/solution interface potential
would also be constant. However, there are limitations imposed by using a liquid filling solution. For example, reference solution volume changes caused by temperature varia- tions or solution evaporation can lead to a drift in the signal. It is also difficult to minia- turize electrodes with internal volumes below one milliliter. Large differences in ionic strength lead to osmotic pressure that can damage the sensor membrane [37]. Many people have attempted to eliminate the internal solution by bonding the inside of the membrane directly to an electrical conductor that is connected to the outside measurement circuit.
Even though many of these devices use liquid membranes, they have been referred to as
“solid contact ISEs” or “all solid-state ISEs.”
How do all solid-state ISEs work? To answer that question, one needs to consider what role that the inner solution of the conventional electrode plays in the operation of the sensor.
3.7.1. The Function of the Inner Reference Electrode
The inner reference solution is a key ingredient to the proper function of a reference elec- trode. The reference electrode serves as a transducer for charge carriers. That is, it is a device that enables current to move between a wire where electrons carry the charge and a solution where ions carry the charge. Why is that necessary? In order to measure electri- cal characteristics of a device, the device must be included in an electrical circuit, a closed loop through which current can pass. The act of measuring the potential of an electrochem- ical cell draws a finite current through all parts of the circuit. Fortunately, the amount of current required can be very small using modern electronics (on the order of 10−12A or less with very good equipment). Keeping the measuring current small is important so that the charge separation across the membrane is not disturbed and, therefore, the voltage is not distorted by the measurement process. However, some charge must move throughout the circuit in order to communicate the value of the potential to the measurement equipment.
Electrons carry the current in the voltmeter, but ions carry the current in the electrochemi- cal cell. Consequently, there must be a mechanism that translates between these two types of charge carriers. Something must transform the ion current in the cell into electronic cur- rent in the wire leading to the meter. In a conventional ISE experiment, a reference electrode plays this role on both sides of the membrane. For example, when negative charge is mov- ing (as electrons) from the meter into the cell, electrons move through the silver wire to the silver chloride coating on the electrode surface where they reduce some of the silver ions in the coating to form silver metal atoms and release an equal number of chloride ions into the solution (as in Eq. (3.60)):
AgCl+e⇌Ag+Cl− (3.60)
Negative charge is pumped into the solution in the process. At the reference electrode on the other side of the membrane, the reverse process occurs so that negative charge moves out of the solution (in the form of chloride ions) and into the silver wire (as elec- trons). Silver atoms donate the electrons to form silver ions that bind to chloride ions from the solution (Eq. (3.61)).
Ag+Cl−⇌AgCl+e− (3.61)
k k Of course, only the analyte ions move into and out of the sensor membrane in order
to carry the current there.
Because the reference electrode has a relatively large reserve of silver atoms and sil- ver chloride solid, it can afford to transfer electrons with the outside circuit for the small amount of charge involved without the system deviating significantly from its rest poten- tial (described by Eq. (3.62)). The original activity of the chloride must also be high enough that the small number of moles of chloride ion released or precipitated makes a negligible change in activity:
EAg∕AgCl=EoAg∕AgCl−0.059 16 log(aCl−) (3.62) Under these conditions, the reference electrode is said to be “well poised” or to have a high redox capacitance. This mechanism allows charge to continue across the boundary between the solution where the current is carried by ions to the metal where the current is carried by electrons. In this sense, the reference electrode plays the role of an
“ion-to-electron transducer.” Providing this sort of transducer in the absence of a liquid electrolyte is the challenging part of eliminating the internal solution [38].
3.7.2. All Solid-State Reference Electrodes
The purpose of the filling solution is to support the redox reaction that translates between electronic current and ionic current. One way of eliminating the aqueous solution would be to find a redox reaction that produces ions that are soluble in the sensor membrane along with the ionophore. One such system was made by dissolving Co(II)/Co(III) tris(4,4′-dinonyl-2,2′-bipyridyl) complexes as salts of the hydrophobic anion, tetrakis(pentafluorophenyl)borate (TPFPB−), along with the ionophore and polyvinyl chloride (the polymer matrix) in a volatile solvent [38]:
Co(III)L3+3 +e−⇌Co(II)L2+3 (3.63)
where L=
N
Co(III)L33+ + e– ⇌ Co(II)L3 2+
N
and the counter ion, TPFPB−=
F F
F
F F F
F F F
F
F F
F F
F
F –B
F F F
F
The membrane was formed by applying the mixture directly to the top of a carbon electrode and evaporating the solvent. In this manner, the membrane brings the redox cou- ple and ionophore into direct contact with the inert electrode. This device was successful for determining K+ ions with very good reproducibility from electrode to electrode [38].
Another approach is to use a redox polymer to replace the aqueous electrolyte and refer- ence electrode. For example, poly(3-octylthiophene) is a highly conjugated linear polymer that can be oxidized or reduced [39].
k k
R R R
R R
–2e
R R
S S S S S S S
n R R
R
R R R R
S S S S
S+ S
S n +
where R= .
3.7.3. Eliminating the Inner Reference Electrode
A third strategy for eliminating the internal reference electrode is to exploit the electrical capacitance at an inert electrode/solution interface. Consider, for a moment substituting a platinum wire for the internal reference electrode of a conventional potassium ISE. How would the platinum respond to the tiny current inherent to the measurement process as described above? With no oxidizable or reducible (“redox active”) species near the plat- inum electrode to use for transferring charge to or from the external circuit, no charge would cross the boundary there. However, as charge moves into the solution from the membrane, the excess ions would move to the platinum surface. The interface would behave as a capacitor where excess charge would build up as a layer of ions on the OHP.
Electrons inside the platinum would move toward or away from the interface as needed to balance the charge on the solution side. The electrical potential energy difference across this double layer,Edl, would be proportional to the charge that accumulates on either layer:
Edl= Q
Cdl (3.64)
whereCdlis the capacitance of the double layer.
The cell potential is the sum of all of the electrical potential energy transitions, includ- ing this double layer potential.
Emeas=Ecell =Edl−Eext.Ref+Ejnc+Ememb (3.65)
=Edl−Eext.Ref+Ejnc+0.059 16 log(aK+spl) (3.66) where Eext.Ref and Ejnc are constants. The measured potential is Nernstian (that is, the potential changes in proportion to log(aK+spl)with a slope near 59 mV) ifEdlis constant.
The problem with this system is that the double layer voltage,Edl, is not necessarily con- stant. Any net movement of charge within the cell, such as analyte ion leakage across the membrane, leads to a change in the charge at the double layer and, therefore, a change in the double layer voltage at the platinum/solution interface. That change in voltage creates an error. Fortunately, there is a remedy. If the capacitance in Eq. (3.64),Cdl, is very large, then for a fixed charge,Q,Edlis small and changes inEdlare even smaller. The strategy,
k k then, is to makeCdlvery large. One way of doing that is to use electrodes with very large
surface area because the capacitance is proportional to the electrode surface area:
C= 𝜀A
4πd (3.67)
where𝜀is the dielectric constant for the medium,Ais surface area of the electrode, andd is the distance between the charged layers.
Tiny particles have large surface area per gram of material. Carbon nanotubes and other porous carbon materials, such as three-dimensionally ordered macroporous car- bon (3DOM) [13], can serve as high capacitance electrical contacts. ISEs using these elec- trodes in place of an internal reference electrode (and without an internal filling solution) have been demonstrated with very low (<1 mV) error in signal voltage. A diagram of a solid-state ISE based on a high capacitance 3DOM contact is shown in Figure 3.22.
The goal with all-solid-state devices is to make these sensors more robust and repro- ducible so that they have negligible error, require minimal calibration, and need little or no maintenance [13]. As such, they hold great promise for remote sensing for environmental analysis and, possibly, for sensors in wearable personal health monitors. One would also expect that the ISEs with solid contacts would have very low detection limits, since they
Electrode body
O-ring
Copper wire
PVC sheet
Fluorous sensing phase Silver paste
3DOM carbon
Screw cap Porous teflon support
FIGURE 3.22 Assembly for an “all-solid-state ISE” using three dimensionally ordered macroporous (3DOM) carbon contact between the copper lead to the measurement circuit and the sensor membrane.
This material provides a huge surface area and, therefore, a high electronic capacitance for transducing the signal from ion charge carriers to electrons. Source: Reproduced with permission from Lai et al.
[13]. Copyright 2010, American Chemical Society.
k k have no internal electrolyte with analyte that can leach out into the sample solution. This
type of work is still in its early stages as of this writing. However, progress is encouraging and may lead to an expansion of applications for ISE sensors.
3.7.4. Super-Hydrophobic Membranes
Another improvement in performance can be made by substituting a long-chain fluori- nated solvent for the usual hydrophobic medium in the membrane. Not only are fluori- nated solvents hydrophobic, but they are also immiscible with hydrocarbon solvents. They are the least polar molecules known. They prevent moisture and organic molecules from penetrating the membrane. Proteins, lipids, and other biomolecules in clinical or environ- mental samples are prone to adsorption on the surface of conventional, hydrocarbon-filled sensors. The accumulation of these nontarget molecules blocks the surface and may hold charged organic groups in the electrical double layer causing drift in the voltage or a loss of response to the analyte. Fluorinated media do not have this problem. Of course, the ionophores that are used in these membranes must also be fluorinated in order to dissolve there. Figure 3.23 shows several components suitable for constructing a silver ISE [13].
F FF F F F F
FF
F F F
F F
1 2
Ag-1
Ag-3
3 F F
F F F
F F F FF F F FF F
F F F F O
F F
F
Rf6 Rf6
Rf6
Rf8
Rf8 Rf8
Rf8 S
S S
Rf6
Rfn = CnF2n – 1 B
Rf6 Na Rf6
Rf6 Rf6
F F F F O n
n = 14.3 F F
– +
Ag-2
Cu-I Ag-4
Rf8
Rf10 Rf10
Rf8
R2N NR2
S
S S
S S S S
FIGURE 3.23 Components for fluorinated membrane ISEs. Structures 1 and 2 have been used as solvent matrices and the fluorinated derivative of tetraphenyl borate (structure 3) provides the “ionic sites” that compensate for charge when the analyte cation binds to the ionophore. Structures Ag-1, Ag-2, Ag-3, and Ag-4 are examples of fluorinated ionophores for silver. Cu–I is a selective agent for copper ion. Source: Reproduced with permission from Lai et al. [13]. Copyright 2010, American Chemical Society.
k k The fluorous side chains were attached to the ligands to make them soluble in the flu-
orinated matrix. However, hydrocarbon linkers between the fluorous chains and sulfur atoms were needed, since fluorine atoms closer than a distance of two carbon atoms from the sulfur atoms were too strongly electron withdrawing to permit strong coordination of the ionophore with the silver ion. The membranes were constructed from Teflon™ (a perfluoronated polymer) soaked in either compounds 1 or 2, that acted as solvents, plus one of four silver-binding ionophores (Ag-1, Ag-2, Ag-3, or Ag-4). In this particular paper [13], the authors dispensed with an inner filling solution and used a high capacitance con- tact. The selectivities for Ag+over many alkaline and heavy metal ions were much better than most other Ag+ ISEs reported in the literature (e.g. using ionophore Ag-4 the selec- tivities wereKAgKPOT=2.5×10−12 for potassium versus silver, KPOTAgPb=6.3×10−11 for lead versus silver,KPOTAgCu=1×10−13for copper versus silver, andKPOTAgCd=6.3×10−14for cad- mium versus silver) [13].