limit of detection
of ISEs operating in real matrices is dictated by the
concentration of other ions present in the sample, and, therefore, by the selectivity coefficients.
Important improvements in limits of detection have emerged over the last years (including
optimization of membrane materials, incorporation of new molecular receptors, electrode
conditioning and theoretical treatments predicting experimental values),[54-59] leading to ISEs
that rather meet the practical considerations for freshwater (i.e. proton, Ca
2+
, Na
+
, K
+
, ammonium,
Cl
–
, CO
3
2
–
, sulphate) and for other cases not quite yet (i.e. phosphate, nitrite, nitrate, trace
metals). It worth mentioning that in certain scenarios (e.g. ammonium), the concentration can
largely vary from nM to µM levels (from the oxic to anoxic zone),[14] being difficult to provide a
general statement.
In an attempt to provide an overview of the required limits of detection, Table 1 summarizes the
expected levels of important ions that can be found in freshwater and seawater.[16, 60-67] As
observed, the requirements for the limits of detection range from nM to µM levels. Although limits
of detection in the nM range have been demonstrated for several environmental ions (e.g. lead,
mercury and cadmium) under laboratory bench conditions,[46, 68-71] only values at µM levels
seems to be more realistic for direct field applications (meaning without preliminary sample
pretreatment).
For alkali and alkali earth-metal cations, the impossibility of achieving better limits of detection has
been claimed as a consequence of transmembrane fluxes when inner liquid ISEs were used in
conjunction with the quality of the ionophore with respect to ion discrimination. While the most
discriminated ionophores are for trace metals with selectivity coefficients of 10 at a logarithmic
scale, there are rarely ionophores for alkali and alkali earth-metal cations with such great
selectivity coefficients (refer to Bulhmnan regarding ionophores [35]). Ideally, with the
implementation of all-solid-state ISEs, transmembrane ion fluxes should be reduced/eliminated
9
and the limits of detection may be fairly improved. Certainly, this was not strictly the case, but
slight enhancements (10-fold) have been accomplished using non-plasticizer polymeric
membranes that have led to lowering ion mobility in the membrane phase.[72, 73] Nonetheless,
with the current ISEs for Ca
2+
, K
+
, Na
+
and H
+
, freshwater and seawater samples could be
analysed (no evidence for Mg
2+
). In the case of ammonium for freshwater, there are levels that
are close to the limits of detection, but again with huge heterogeneity. For seawater, there is
empirical support for ammonium ISEs being strongly interfered by K
+
ions. Under specific ideal
assumptions, such as determination of the selectivity coefficient for ammonium and K
+
,
independent determination of K
+
(by a K-ISE) and using the Nikolsky-Einsenmman equation,
ammonium profiles were reported.[3]
For anion detection in freshwater, OH
–
represents a key important interfering ion (pH~6.5
–8.5)
exclusively related to the supramolecular interaction between an ion and ionophore (for example,
this effect is not present in crystalline or glass ISEs at mild pH [20]). This undesired interaction
limits the detection of nitrite and phosphate and make less favourable nitrate quantification. This
issue is worsened with seawater based on the high levels of Cl
–
(~600 mM) that fully mask the
discernment of nitrate, nitrite and phosphate. Conversely, a very selective carbonate ionophore
[74] permits the detection of carbonate at environmental pH both in fresh and seawater. [38, 42]
To some extent, the complexity of the matrix effect is mitigated in laboratory bench experiments
by applying standard addition calibrations, but this alternative would be much more difficult for
decentralized measurements. Accordingly, other strategies have recently been reported. For
instance, the elimination/reduction of the major interfering ions in the sample prior to the
potentiometric readout or the conditioning of ISEs with sacrificial mixtures containing the
ionophore [75] have been postulated (see Section 4). Even though the limits of detection of anion
potentiometric sensors are successfully reduced through the use of these concepts, this is not yet
sufficient for the detection of certain ions, like phosphate.[76]
10
Another aspect to be considered is the long-term stability of ISEs as it is an important issue for the
monitoring of ions over a period of several weeks. The alteration of the membrane components
(ionophore and/or ion exchanger) over time modifies the potential response of the sensor and
hence its reproducibility and lifetime [34]). Briefly, the membrane composition should be optimized
to avoid first its detachment from the solid substrate selecting appropriate polymeric matrices (e.g.
PVC, acrylic, polyurethane) and adequate membrane thickness [77], then to reduce leaching of
the components (making them sufficiently lipophilic) and finally minimizing fouling and passivation
using continuous flow systems. Additional tests in presence of redox species and high sulfide
concentrations (normally found at environmental conditions) are highly recommendable [78]
Overall, during the development and validation of the ISE, a careful evaluation of these aspects is
necessary to produce robust electrodes (in terms of mechanical and chemical robustness).
A review by Bulhmann’s group [39] summarized the drift values reported in the literature using
various transducing elements. Those numbers are around 10 µV.h
-1
when recorded under
controlled laboratory conditions (i.e., fixed ion concentrations, temperature control, stirring
solutions, constant O
2
/N
2
atmosphere). To put this into context, for a monovalent ion at 25
o
C,
drifts of 10, 100 and 1000 µV h
-1
would lead to 0.04, 0.4 and 4% precision, respectively. As
observed from these values, a reliable precision can be ensured using all-solid-state ISEs, but
within less mild conditions, many other factors may cause an increment in drift.
What is the tolerable precision and how much of that precision are we able to sacrifice rendering
less calibration steps or
in situ
ion information are remaining open questions. The flexibility in
precision is directly connected to the diversity of the analyte and its surroundings. To illustrate this
latter point, for example, it is expected that a 200-fold variation in ammonium ion concentration will
exist in lakes where steep solution and redox gradients across the oxycline are presented while
potassium ion concentrations remains practically constant.[14] As a consequence, a much more
precise K
+
ISE (over an ammonium sensor) may be required in order to monitor minor variations
in K
+
levels along the water column.
11
Traditionally, the frequent re-calibration of the ISE guarantees accuracy and repeatability of the
analytical results, but the implementation of this procedure into submersible devices requires the
integration of standard solutions and valves for fluid handling, which may result in highly complex
systems with severe restrictions regarding time resolution of the concentration profiles. Reducing
both the frequency of the calibration steps towards calibration-free sensors and the complexity of
the manifolds is an urgent necessity for
in situ
monitoring.
A key requirement for calibration-free (or, at least, less calibration points) potentiometric
measurements is the reproducibility of the standard potential (E°), and research in this direction is
being carried out. Promising advances in controlling the oxidation state of the redox probe (e.g.,
Co(II)/Co(III) species) [79] or by integrating the current needed to transfer the redox polymer to a
new state of equilibrium may lead to interesting directions towards implementation in field
applications.[80, 81]
An alternative to calibration-free ISEs is dependent on the use of dynamic electrochemistry. Here,
ion transport through the ion-selective membrane is modulated by potential or current control.[82,
83] The additional confinement of the sample into a thin-layer gap (less than 100 µm thickness)
fosters the homogeneous and complete electrochemical modification of the ion analyte.[84-86]
For this reason, rigorous maintenance, diagnostics and data correction of the electrodes is
reduced and suitability for long-term applications is expected. The concept has been successfully
applied at the laboratory bench for the simultaneous detection of halides (Cl
–
, Br
–
and I
–
) in
freshwater and seawater [87, 88] as well as for alkalinity detection in freshwater. [89] In other
direction, the group of Bobacka used constant-potential coulometry allowing amplification of the
analytical signal of solid-contact ISEs. This method is suitable for detecting small changes in ion
activities, which may be of interest in water analysis.[90]
The disposability feature of ISEs is in line with the idea of mass production and random calibration
(e.g., screen printing, paper based, etc.) exhibiting high reproducibility in sensor-to-sensor
potential responses. Unfortunately, while this concept seems to work well for new emergent point-
12
of-care platforms that minimize contamination between different patients,[29, 30, 91, 92] it is not
yet appropriate for
in situ
analysis as the change of the electrode with every single measurement
is not physically possible. Therefore, for environmental analysis, very robust sensor technology for
long monitoring is preferred.
Last but not least, ISEs and electronic instrumentation are physically small for portability. The
simplicity of the electronic circuitry and minimal power consumption for operation offer an easy
adaptation to custom-built devices, even those featuring wireless control.[23, 93-95]
Consequently, all-solid-state ISEs and acquisition instruments form an appropriate starting point
for the development of inexpensive decentralized sensing platforms that are attractive for
in situ
environmental analysis with the optional potential of implementation into submersible probes.
Such deployable systems could additionally be wirelessly controlled and take totally decentralized
measurements. Interestingly, the development of wireless sensing has evolved in parallel with the
growing trends in health care, requiring diagnostic tools to be applied in a remote fashion.[92, 95-
97] Significant advances have been made regarding the fabrication of wireless chemical-sensing
network (WCSN) devices, like AM radio, radiofrequency, WiFi, 3G/4G band or GPS.[26, 93, 98-
102]
Chia sẻ với bạn bè của bạn: |