From Molecular and Emulsified Ion Sensors to Membrane Electrodes

Apr 24, 2019 - ... University of Michigan and was on the faculty at Auburn University (AL), Purdue University (West Lafayette, IN), and Curtin Univers...
1 downloads 0 Views 1MB Size
Download PDF
Article Cite This: Acc. Chem. Res. XXXX, XXX, XXX−XXX

pubs.acs.org/accounts

From Molecular and Emulsified Ion Sensors to Membrane Electrodes: Molecular and Mechanistic Sensor Design Elena Zdrachek*,† and Eric Bakker*,†

Acc. Chem. Res. Downloaded from pubs.acs.org by IDAHO STATE UNIV on 04/25/19. For personal use only.

University of Geneva, Department of Inorganic and Analytical Chemistry, Quai Ernest Ansermet 30, Geneva 1211, Switzerland CONSPECTUS: Selective molecular ion probes are often insoluble in water and require a hydrophobic solvent environment for strong and selective binding, which runs counter to the desire of utilizing them in a homogeneous solution. This Account aims to guide the reader on how such molecules, often coined ionophores, can be harnessed to design exceptionally useful optical and electrochemical sensors. We start here with some historical context on the design of such ionophores and continue with the explanation of the response mechanism of optical and potentiometric sensors and the role of combined components to build a robust ion sensor. This Account is addressed to nonspecialist readers and for this reason avoids extensive use of equations or theoretical considerations. The interested reader should turn to the original literature for further reading. Emulsified optical sensors are introduced as an initial example. Here, multiple reagents are confined in an attoliter sensing nanodroplet of the organic phase, immiscible with the aqueous sample phase. In this case, the ionophore molecules may retain their high affinity and selectivity to the target ion and the aqueous sample phase does not have to be modified. Emulsified optical sensors allow one to achieve the selective chemical sensing of ions, even with optically silent ionophores. Such ionophore-based nanodroplets are also discussed as a useful novel class of complexometric titration reagents and optical end point indicators with unique selectivities. We then turn our attention to potentiometric sensing probes and briefly discuss the unique opportunity of a direct characterization of ion−ionophore complexation properties offered by membrane electrodes. A carbonate-selective membrane electrode containing a highly selective tweezer-type ionophore with trifluoroacetophenone functional groups is then used as an example for the construction of a robust all-solid-state sensor. This potentiometric probe, in combination with a pH electrode, can directly measure PCO2 in freshwater lakes, demonstrating a dramatically improved response time relative to traditional sensors equipped with a gas-permeable membrane. In recent years, new sensing modes and electrode designs have been introduced to expand the application scope of ionophorebased potentiometric sensors. Membrane electrodes containing ionophores are placed under dynamic electrochemistry control to give important progress in the field. We specifically highlight our recent works by membranes that are controlled by chronopotentiometry (controlled current) for speciation analysis, by ion transfer voltammetry on thin sensing films for multianalyte detection, by exhaustive coulometry for potentially calibration-free sensors and with coulometric membrane pumps for the selective delivery of reagents.

■

ION SENSING PROBES When one imagines a molecular ion sensor, a purposefully and carefully designed chemical receptor comes to mind. It should reversibly recognize its analyte ion with the appropriate selectivity within the relevant concentration range, even in a sample environment of high complexity. The binding event should give rise to a change in optical signal, often by fluorescence and ideally in a ratiometric fashion. Such reversible ion probes do exist and are indeed well-established as indicator dyes. Among those, Fura-2 and Calcium Green are perhaps best known as they are routinely used to image intracellular calcium fluctuations by biologists. In many cases, however, the water solubility of a reported molecular sensing probe is limited and its binding behavior must be studied in an organic solvent. Indeed, the detection of hydrophilic ionic species is often very difficult in an aqueous phase because water is such a strong competitor, resulting in very weak complex formation constants. This is well© XXXX American Chemical Society

established for the electrically neutral, nonchromogenic potassium ionophore valinomycin (see Figure 1 for the structure), for which the 1:1 complex formation constant in water has been determined to be negligibly small (log β = 0.431) while its value rises quite dramatically to log β = 11.62 in a hydrophobic environment such as plasticized poly(vinyl chloride). Stable complexation and recognition selectivity are therefore very difficult to achieve in an aqueous solvent, even though water remains the practically relevant solvent where the measurement must be performed. Let us use Figure 2 as a molecular ion sensor example,3 a chromogenic receptor molecule that encapsulates potassium ions. The resulting conformational change gives rise to an approach of the two pyrene units, thereby increasing the fluorescence response. As shown in Figure 2, this interaction is Received: January 31, 2019

A

DOI: 10.1021/acs.accounts.9b00056 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

Figure 1. Structures of selected ionophores: valinomycin as one of the first biological ionophores used as a receptor for potassium ion sensing; cryptand 222 as a macrocyclic compound that inspired the structural development of the highly successful calcium ionophore I, also called ETH 1001; tweezer-type carbonate ionophore of high selectivity, a rare example of a receptor reversibly binding an ion with a covalent bond.

reversible and quite selective but requires a hydrophobic solvent environment. To make these molecular probes practically useful, one may choose to incorporate them into a separate solvent, immiscible with the aqueous sample phase. One such approach involves the use of emulsified nanoscale droplets that contain the receptor of interest. In such a manner, the receptor molecule is adequately solubilized and retains its high affinity and selectivity to the target ion, while the aqueous phase to be probed is not unduly modified. Such systems, however, must now be understood on the basis of heterogeneous phase transfer equilibria. With the data shown in Figure 2, potassium sensing is accomplished by doping the

emulsified organic phase with the sodium salt of a lipophilic cation-exchanger. A potassium ion in the sample is then exchanged with these sodium ions in order to interact with the chromogenic receptor. Sensor behavior then depends on the relative concentrations (strictly, activities) of the two competing ions and the relative affinities of the two ions with the sensing phase and receptor molecule. In Figure 2, the potassium ions from the sample are quantitatively taken up by the emulsified phase, resulting in the observed linear calibration curve.3 Droplets (or particles) of 100 nm in diameter that make up a microemulsion can be also understood as confined chemical B

DOI: 10.1021/acs.accounts.9b00056 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

depends on sample pH, sample concentration and nature of the competing ion, and the binding strength of the indicator in the emulsified phase. If no ion-exchanger is present, protonation of the indicator may still occur but requires the coextraction of hydrogen ions together with a solution anion. Thus, this sensing principle is highly flexible because the characteristics of ion binding and spectral change can be designed separately and easily adjusted to a particular sensing problem. The main drawback of this sensing approach is that the ion of interest cannot be quantified without prior knowledge or control of the solution pH. More recently, this limitation has been overcome with lipophilic solvatochromic dyes, where competitive ion-exchange and signal reporting involve an ionic dye.4−7 The chromogenic unit of the dye molecule changes color or fluorescence when expelled into the aqueous phase, while the dye itself remains anchored onto the sensing particle thanks to a lipophilic side chain (see Figure 3b).

■

IONOPHORES The sensing strategies discussed above make it possible to design optical sensors with numerous optically silent receptors, such as so-called ionophores well established in potentiometric sensing probes. While the first studied ionophores, such as valinomycin, were isolated from living cells, a wide range of new ionophore structures were subsequently synthesized that typically rely on multitopic interactions with the guest ion on the basis of ion-dipole, metal ligation, hydrogen, and covalent bonds. Macrocyclic ionophores were originally developed independently by different research groups, with Pedersen introducing crown ethers,8 circular macrocyclic compounds containing ether groups for the recognition of cations and Jean-Marie Lehn presenting the cryptands (“cryptates” in French),9 a class of very effective synthetic polycyclic multidentate ligands where metal cations can be encapsulated, as in a crypta (Latin for cavity) (see Figure 1 for structures). Later, a different class of versatile macrocyclic ionophores formed from p-hydrocarbylphenols and formaldehyde was presented. David Gutsche called them calixarenes,10 inspired by the shape of a Greek vase called “calyx krater”. As with valinomycin, the specificity of the macrocycle molecule binding is driven by the size of the cavity and the number of oxygen/ nitrogen atoms available for complexation. To be practically useful, ionophores should also provide rapid and reversible binding characteristics. For cryptands, their very strong and almost irreversible complexing ability has limited their application in chemical analysis. Still, the synthetic pathway for the realization of cryptands served as the inspiration for the development of highly successful noncyclic amide ionophores for the recognition of calcium, essentially by eliminating the macrocyclic structure to weaken the interaction. As a case in point, the structural similarity of calcium ionophore I (ETH 1001) (Figure 1) to cryptands is rather evident. This ionophore is used with great success in a wide range of calcium-selective sensors for blood analysis and exhibits exquisite selectivity. Early electrochemical blood analyzers were reported to include a trifluoroacetophenone-based plasticized membrane that showed preference for bicarbonate, even though interference from blood chloride had to be corrected.11 Owing to the electron-withdrawing character of the fluorinated substituent, trifluoroacetyl groups are understood to form

Figure 2. (a) A conformational change from potassium binding to pyreneamide-functionalized 18-crown-6 ether, triggering a fluorescence response. (b) Observed normalized fluorescence at 406 nm as a function of the concentration of different cations (λexcitation = 346 nm). Adapted with permission from ref 3. Copyright 2016 The Royal Society of Chemistry.

reactors with a volume of 4 aL that may be doped with multiple selective sensing reagents to accomplish an ion sensing task. In contrast to the former described example, organic microemulsions, part of the bulk optode family, can be doped with two competitive receptor molecules together with an appropriate ion-exchanger. One serves as the indicator dye, typically for pH, while the other is an optically silent but highly selective receptor for the ion of interest. An unsuspecting reader might interpret the pH response of such an emulsified indicator dye in complete analogy to the classical pH buffer behavior, relating signal change directly to sample pH. This would be incorrect. Hydrogen ions partition into the emulsified phase by competitive ion-exchange and bind only then to the indicator. If the ion of interest increases in concentration, it competes more favorably with the hydrogen ion, which is exchanged out of the sensing phase. This results in a visible deprotonation of the lipophilic pH indicator dye (see Figure 3a). The response of the indicator therefore

Figure 3. Sensing mechanism of ionophore-based ion-selective optical nanosensors containing (a) a pH indicator (chromoionophore) or (b) a charged lipophilic solvatochromic dye as an optical reporter. The ion exchange with an analyte cation results in protonation/ deprotonation of the indicator or a localized repartitioning of the solvatochromic dye bearing a positive charge. C

DOI: 10.1021/acs.accounts.9b00056 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

This makes the response dependent only on the amount of analyte in the sample but not that of the reference ion. The latter means that the response of optodes based on pH indicator dye will be pH-independent when they are operated in the exhaustive sensing mode (compare Figure 5a and 5b). In Figure 5, it is apparent that another important feature of the exhaustive sensing mode is that the response is linear with the analyte concentration.

hydrates by nucleophilic attack of water in aqueous solution. Stabilization of bicarbonate was therefore thought to occur via hydrogen bonding with this hydrated form.11 A careful spectroscopic study by Meyerhoff et al. suggested that the nucleophilic attack occurs not by water, but actually by the carbonate ion itself.12 This appears to be a rare occurrence of reversible recognition by covalent bond formation. The structure shown in Figure 1 was introduced by the groups of Nam and Cha13 and is based on a cholic acid molecular frame that contains two axial trifluoroacetophenone substituents that interact with carbonate in a tweezer action. The selectivity of the resulting sensors to carbonate is exquisite and allows for carbonate detection in samples as complex as seawater without interference from chloride.14

■

DESIGN FEATURES OF OPTICAL SENSORS So, what happens when this carbonate-selective ionophore is used in conjunction with an added lipophilic indicator to make an optical sensor of the type discussed above? Here, the sensor responds by extracting carbonate together with hydrogen ions from the sample to the sensing phase, with each of the two ions interacting with their respective molecular receptor (see Figure 4). As carbonate is a divalent ion, two hydrogen ions are

Figure 5. Theoretical response curves of the calcium-selective optode used in (a) an equilibrium or (b) an exhaustive sensing mode. The response of optodes operated by an exhaustive sensing mode is linear with the analyte concentration and sufficiently independent of the pH of the sample. Adapted with permission from ref 18. Copyright 2015 Springer-Verlag Berlin Heidelberg.

Figure 4. (a) Illustration of a CO2 optode sensing mechanism. The carbonate ion is extracted together with hydrogen ions into organic sensing film containing the anion exchanger, an electrically charged pH indicator and the carbonate ionophore shown in Figure 1. The optical response is proportional to the product of carbonate ion activity and the square of hydrogen ion activity, which is directly linked to the activity of carbon dioxide. (b) Calibration curves of the CO2 optode to CO2 in a hydrogen carbonate solution, in the gas phase equilibrated with hydrogen carbonate solution, and a dry CO2 + N2 gas mixture with different levels of PCO2. For proper sensor function, humid air samples are required. Adapted with permission from ref 15. Copyright 2012 American Chemical Society.

Emulsions of ion optodes may not only function as conventional optodes but also offer new analytical opportunities. Nanosphere emulsions have been used for lightcontrolled ion concentration perturbation when doped with photoresponsive compounds such as spiropyran.19 In another application, a nanosphere emulsion was applied with a wax printed paper-based microfluidic device for the quantification of potassium in aqueous and human serum samples, using colorimetric distance-based analysis methods.20 Ion optode emulsions prepared without a lipophilic pH indicator were recently proposed as a new promising class of complexometric titration reagents, chiefly for Ca2+ and Pb2+.21−23 This versatile toolbox allows one to use ionophores as highly selective chelators despite their poor solubility in water. In comparison with traditional water-soluble chelators such as EDTA, these new reagents possess a significantly reduced pH dependence, and control of pH is often not necessary.21 Such chelating emulsions can also serve as optical indicators for complexometric titrations, using a small amount of lipophilic pH indicator or ionic solvatochromic dye as an optical reporter.22,23 Only at the end point, when the chelating nanospheres become saturated with analyte, the indicating nanospheres will be interacting with the analyte and result in a change of chromoionophore protonation or repartitioning of

extracted for each anion. The resulting sensor is therefore sensitive to the carbonate ion activity, multiplied with the square of the hydrogen ion activity. A glance at the appropriate dissociation equilibria for carbon dioxide and carbonic acid, however, shows that this is directly proportional to the carbon dioxide partial pressure. This optical sensor therefore measures carbon dioxide by the direct molecular recognition of carbonate (see Figure 4).15 Another attractive approach for ion optodes operates them in a so-called exhaustive sensing mode.16−18 While in a classic equilibrium mode, the sample concentration is usually considered to be unchanged before and after the contact with the sensor; the analyte here is completely consumed by the sensor, but this works only if the sensing phase contains more binding sites than analyte ions available in the sample. D

DOI: 10.1021/acs.accounts.9b00056 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

Ionophore-based ISEs have found applications in many different fields including environmental science, medicine, biochemistry, neuroscience, and corrosion studies, but indisputably their largest success has been in clinical diagnostics where blood-gas analyzers with ionophore-based potentiometric sensors for detection of Na+, K+, Ca2+, and Cl− have completely displaced the previously used flame photometers from the market. Owing to their simplicity, ease of miniaturization, and low energy consumption, ISEs are also used for aquatic in situ analysis. Several examples of developing and deploying submersible potentiometric multisensor platforms for in situ monitoring of ion species in natural waters have been reported by our group.28−32 Let us use the tweezer-type carbonate ionophore discussed above as an example to develop an environmental sensor. As mentioned, the recognition selectivity of membranes doped with this receptor is so high that it may detect carbonate directly in environmental samples, without concern for interference. However, to use such materials for submersible in situ probes, the inner aqueous solution at the back side of the membrane should be replaced with a solid support, owing to the associated pressure fluctuations. All-solid-state membrane electrodes require a lipophilic ion-to-electron transducing layer, which are now well established for cationic analytes but less so for anionic analytes. The conducting polymer poly(octylthiophene), for example, placed between electrode and ion-selective membrane, has been shown to partially oxidize with cation-selective membranes.33 For complete oxidation, the cation-exchanger in the membrane must migrate to the transducing layer to form the counterion of the oxidized, anionic form, but it is not able to do so, chiefly for reasons of steric hindrance. The polymer therefore exhibits a mix of reduced and oxidized forms, giving the desired redox buffer. With anion-selective membranes, however, this mechanism is not possible because the membrane contains an anion-exchanger that cannot play this role. A stable potentiometric response behavior for both cationic and anionic analytes was, however, achieved with highly capacitive and lipophilic nanomaterials, making it now possible to fabricate potentiometric probes that can be deployed in situ. An interesting characteristic of potentiometric probes is that the reference electrode can be replaced by another indicator electrode to gain a different type of information from the sample. A carbonate probe, for example, can be combined with a pH electrode. In some analogy to the PCO2 optode discussed above, by considering the acid dissociation constants of CO2/ carbonic acid, such a detection cell has been shown to directly measure carbon dioxide partial pressure.34 Compared to other known CO2 probes that require gas-permeable separation membranes, one expects dramatically faster sensor responses. This was recently confirmed during deployment in a stratified Swiss lake, see Figure 6, where a traditional membrane-based Severinghaus probe failed to provide the same level of detail in the depth region where steep gradients of dissolved gases occur.32

solvatochromic dye, giving a color change. The titrimetric detection of alkali metal ions has recently also been demonstrated (K+, Na+, and Li+).24

■

MEMBRANE ELECTRODES The optically silent molecular ion receptors discussed above and shown in Figure 1 have their origin in ion-selective membrane electrodes (ISEs) that are typically interrogated by potentiometry, at zero current, just as one would operate an established pH glass electrode. In many ways, they are complementary to optical ion sensors. The sensing material containing the ionophore and added ion-exchanger is termed a membrane because of its permselective properties, which mean that only cations or anions may pass through the membrane upon application of a small current. This property is essential to observing a Nernstian electrode response (at 25 °C, 59.2 mV/zi change for each 10-fold ion activity change; zi is the charge of the analyte ion) that characterizes these sensors. The first ion-selective membranes were prepared as a solution of an electrically neutral ionophore in an organic solvent infused into porous glass filters as a support, starting with the group of Simon.25 In 1967, Bloch and co-workers introduced the first ionophore-based solvent polymeric membrane made of polyvinyl chloride,26 which is still commonly used for ISE fabrication despite the fact that many other polymers, such as silicone rubbers, polyurethanes, acrylates, and perfluoropolymers, have been suggested over the years. In order to lower the glass transition temperature below room temperature and form homogeneous elastic films, polyvinyl chloride and other polymers are often blended with plasticizers, such as 2-nitrophenyl octyl ether, bis(2-ethylhexyl) sebacate, or dibutyl phthalate. The choice of plasticizer influences the polarity of the membrane phase and subsequently the ISE response. An important feature of ionophore-based membrane electrodes is that the composition of an ideally functioning membrane electrode does not change as the analyte ion concentration varies in the contacting sample solution. The potential response has its origin in a free energy balance of the interfacial concentration gradient and electrical potential difference where the ion to be measured may freely transfer into and out of the membrane. If this process is sufficiently selective and no other ions can compete with the ion of interest, a Nernstian response slope is observed. Unlike most other sensing methods, such potentiometric probes are quite robust, and changes in size and shape, in principle, bear no influence on the potential readout. Membrane electrodes can also be used to conveniently characterize the binding properties of ion−ionophore complexes. For this purpose, membrane symmetry must be overcome, which otherwise cancels the contribution of the ion binding event within the membrane. One method relies on the potentiometric characterization of ion-exchange selectivity if a reference ion can be assumed not to interact with the ionophore.27 For cations, this is often the tetrabutylammonium ion, which is water-soluble but typically too bulky for complexation. An ion-exchange selectivity study with and without ionophore gives the desired information. Another technique performs potentiometric measurements by fusing two membranes of a different composition, one with and one without ionophore.27 It can also be used to evaluate to what extent any reference ion in the first method binds to the ionophore.

■

OTHER DYNAMIC ELECTROCHEMISTRY EXAMPLES Ionophore-based membrane electrodes can also be placed under dynamic electrochemistry control by applying a current or potential. This allows one to obtain complementary information about the sample and result in new readout modes that expand the scope of application of these chemically E

DOI: 10.1021/acs.accounts.9b00056 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

Figure 6. In situ profiles of dissolved CO2 in Lake Rotsee obtained with a commercial Severinghaus CO2 probe and the coupled hydrogen ion and carbonate selective electrodes with a profiling speed of 5 cm/s. In contrast to the Severinghaus CO2 probe exhibiting a response time of minutes, the setup with two coupled ISEs provides a response time of a few seconds. This allows for CO2 measurements at a high temporal and spatial resolution. Adapted with permission from ref 32. Copyright 2018 American Chemical Society.

selective materials. One example is the use of chronopotentiometry, where an applied current imposes a net ion flux across the membrane that can in many cases be understood as an imposed concentration gradient of the analyte ion near the membrane surface. The diffusion theory predicts that the diffusion distance over which the gradient occurs will grow with time, eventually resulting in ion depletion at the membrane surface at a so-called transition time that manifests itself by a potential change (Figure 7a). The square root of transition time is ideally proportional to concentration and can therefore be used to determine the analyte concentration in the sample. ISEs operated by chronopotentiometry are attractive for speciation analysis because they give information about the kinetically labile fraction of ion concentration, while zero current potentiometry measures so-called free ion concentration (strictly, activity). Recent examples include the speciation analysis of calcium in blood samples to distinguish the total calcium from the so-called ionized calcium.35 A sequential zero current potentiometric and chronopotentiometric readout with a single H+-selective electrode can give information about pH and acidity/alkalinity in water samples.36,37 In contrast to the above-described examples, a current pulse with a reversed sign should impose an ion flux in the direction of the sample phase. Thus, if the membrane contains a significant concentration of the analyte ion, it can be used as an electrochemical titration pump where the amount of delivered titrant is controlled by the imposed charge in agreement with Faraday’s law. This was demonstrated for the coulometric delivery of calcium38 for EDTA titrations and protons39 for the determination of alkalinity in natural water samples. Here, the electrochemical titration can be considered as nearly nondestructive because chemical perturbation occurs only in the

Figure 7. Different readout modes of membrane ISEs placed under dynamic electrochemistry control: (a) chronopotentiometry; (b) thin-layer coulometry; (c) ion transfer voltammetry at a high concentration of the analyte; (d) ion transfer voltammetry at a low concentration of the analyte; (e) coulometry with a capacitive layer; (f) colorimetry.

diffusion layer of the aqueous phase where the analyte concentration can be re-equilibrated after stirring the sample. If one reduces the sample volume to a thin layer with dimensions of the diffusional layer, exhaustive turnover becomes possible. In this case, the ion transfer current observed upon applying an appropriate potential can be integrated to give the overall charge of the ion transfer process (Figure 7b). This charge is used to calculate the analyte concentration for a known sample volume according to the Faraday’s law. An indisputable advantage of this sensing mode is a calibration-free application of ISEs. An electrochemical F

DOI: 10.1021/acs.accounts.9b00056 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

practically useful and novel sensing probes, as shown here with the carbonate-selective ionophore as a guiding example.

protocol for compensating non-Faradaic charging currents by performing a second excitation pulse whose charge was subtracted from that of the first pulse was proposed.40 The successful coulometric determination of potassium,40−43 calcium,44 and nitrite45 with ionophore-based ISEs was demonstrated. Different approaches have been proposed to decrease the sample phase dimensions, such as sampling with a paper support43 or microfluidic devices with a tubular shaped ISE membrane.42,44,45 Alternatively, membranes thinner than 1 μm have allowed for the establishment of ion transfer voltammetry as an alternative readout of ionophore-based ISEs. It was shown that, by sweeping the applied potential, the oxidation/reduction of the conducting polymer layer46,47 placed underneath the sensing membrane or redox molecular probes48,49 incorporated into the membrane phase may trigger ion transfer across the membrane/sample interface. Such voltammetric ISEs can be exploited in two sensing modes depending on the analyte concentration level.50 At relatively high concentrations, the position of the ion transfer peak becomes the analytical signal that is related to the logarithmic analyte activity in complete analogy to potentiometric measurements at zero current (Figure 7c). Interestingly, it was shown that ISE membranes doped with three different ionophores in this sensing mode allows one to detect and quantify the sample activity of three ions (Li+, K+, and Na+) within one potential scan.47 Instead, at low concentration levels, the peak current linearly increases with analyte concentration in complete analogy with stripping voltammetry measurements (Figure 7d). Another interesting coulometric readout principle was recently proposed for solid-contact ISEs with a capacitive transducer layer. For an electrochemical cell held at a constant potential, any change of the indicator electrode potential (caused by a sample concentration change) is compensated with an opposite change at a capacitive layer placed beneath the ion-selective membrane. This results in a transient current response where the accumulated charge is proportional to the logarithmic analyte activity change (Figure 7e). The applicability of this sensing mode has been demonstrated for potassium51 and chloride52 detection. It was recently shown that a closed bipolar electrode configuration with an ISE on one pole and an indium tin oxide electrode covered with a redox indicator in a thin-layer cell on another pole allows one to convert the potential change at the ISE into a color change in the detection compartment (Figure 7f).53 An electrical potential imposed across the bipolar electrode is used to trigger a color change within the desired concentration change and can be tuned to match the application. In this manner, optical ion sensors can be realized with membrane electrodes, fundamentally and physically separating optical detection from chemical sensing. This approach was successfully applied for calcium detection.53 To conclude, molecular ion recognition for chemical sensing may preferentially occur in a separate sensing phase, rather than in a homogeneous sample solution. This means that phase transfer equilibria must also be taken into account to understand how such sensors must work. This gives rise to a rich world of sensing principles and materials that can be used to design a wide range of exciting sensing systems, from nanoscale optical probes to membranes that can be addressed by dynamic electrochemistry. This is a field where advances in molecular recognition can be quite readily translated into

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. ORCID

Eric Bakker: 0000-0001-8970-4343 Author Contributions †

E.Z. and E.B. contributed equally.

Notes

The authors declare no competing financial interest. Biographies Elena Zdrachek received both her B.Sc. (2010) and M.Sc. (2011) degrees in chemistry from the Belarusian State University in Minsk. In 2015, she obtained her Ph.D. from the same university, working with Prof. Egorov V. V. Currently, she is a postdoctoral researcher at the University of Geneva. Her research focuses on new electrochemical sensing concepts based on membrane electrodes. Eric Bakker was educated at ETH Zurich in Switzerland (Ph.D. in 1993), pursued postdoctoral studies at the University of Michigan and was on the faculty at Auburn University (AL), Purdue University (West Lafayette, IN), and Curtin University (Perth, Western Australia) before assuming his current position in 2010.

■

ACKNOWLEDGMENTS The authors thank all of the co-workers and collaborators who contributed to the cited research. The authors gratefully appreciate the financial support of the Swiss National Science Foundation for supporting research in this field.

■

REFERENCES

(1) Feinstein, M. B.; Felsenfeld, H. The Detection of Ionophorous Antibiotic-Cation Complexes in Water with Fluorescent Probes. Proc. Natl. Acad. Sci. U. S. A. 1971, 68, 2037−2041. (2) Qin, Y.; Mi, Y.; Bakker, E. Determination of complex formation constants of 18 neutral alkali and alkaline earth metal ionophores in poly(vinyl chloride) sensing membranes plasticized with bis(2ethylhexyl)sebacate and o-nitrophenyloctylether. Anal. Chim. Acta 2000, 421, 207−220. (3) Jarolimova, Z.; Vishe, M.; Lacour, J.; Bakker, E. Potassium ionselective fluorescent and pH independent nanosensors based on functionalized polyether macrocycles. Chem. Sci. 2016, 7, 525−533. (4) Xie, X.; Zhai, J.; Bakker, E. Potentiometric Response from IonSelective Nanospheres with Voltage-Sensitive Dyes. J. Am. Chem. Soc. 2014, 136, 16465−16468. (5) Xie, X.; Gutierrez, A.; Trofimov, V.; Szilagyi, I.; Soldati, T.; Bakker, E. Charged Solvatochromic Dyes as Signal Transducers in pH Independent Fluorescent and Colorimetric Ion Selective Nanosensors. Anal. Chem. 2015, 87, 9954−9959. (6) Wang, L.; Xie, X.; Zhai, J.; Bakker, E. Reversible pH-independent optical potassium sensor with lipophilic solvatochromic dye transducer on surface modified microporous nylon. Chem. Commun. 2016, 52, 14254−14257. (7) Xie, X.; Szilagyi, I.; Zhai, J.; Wang, L.; Bakker, E. Ion-Selective Optical Nanosensors Based on Solvatochromic Dyes of Different Lipophilicity: From Bulk Partitioning to Interfacial Accumulation. ACS Sens 2016, 1, 516−520. (8) Pedersen, C. J. Cyclic polyethers and their complexes with metal salts. J. Am. Chem. Soc. 1967, 89, 2495−2496.

G

DOI: 10.1021/acs.accounts.9b00056 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

Schafer, J.; Bakker, E. In Situ Detection of Macronutrients and Chloride in Seawater by Submersible Electrochemical Sensors. Anal. Chem. 2018, 90, 4702−4710. (31) Athavale, R.; Dinkel, C.; Wehrli, B.; Bakker, E.; Crespo, G. A.; Brand, A. Robust Solid-Contact Ion Selective Electrodes for HighResolution In Situ Measurements in Fresh Water Systems. Environ. Sci. Technol. Lett. 2017, 4, 286−291. (32) Athavale, R.; Pankratova, N.; Dinkel, C.; Bakker, E.; Wehrli, B.; Brand, A. Fast Potentiometric CO2 Sensor for High-Resolution in Situ Measurements in Fresh Water Systems. Environ. Sci. Technol. 2018, 52, 11259−11266. (33) Veder, J.-P.; De Marco, R.; Patel, K.; Si, P.; GrygolowiczPawlak, E.; James, M.; Alam, M. T.; Sohail, M.; Lee, J.; Pretsch, E.; Bakker, E. Evidence for a Surface Confined Ion-to-Electron Transduction Reaction in Solid-Contact Ion-Selective Electrodes Based on Poly(3-octylthiophene). Anal. Chem. 2013, 85, 10495− 10502. (34) Xie, X. J.; Bakker, E. Non-Severinghaus Potentiometric Dissolved CO2 Sensor with Improved Characteristics. Anal. Chem. 2013, 85, 1332−1336. (35) Afshar, M. G.; Crespo, G. A.; Bakker, E. Direct Ion Speciation Analysis with Ion-Selective Membranes Operated in a Sequential Potentiometric/Time Resolved Chronopotentiometric Sensing Mode. Anal. Chem. 2012, 84, 8813−8821. (36) Crespo, G. A.; Afshar, M. G.; Bakker, E. Direct Detection of Acidity, Alkalinity, and pH with Membrane Electrodes. Anal. Chem. 2012, 84, 10165−10169. (37) Afshar, M. G.; Crespo, G. A.; Xie, X. J.; Bakker, E. Direct Alkalinity Detection with Ion-Selective Chronopotentiometry. Anal. Chem. 2014, 86, 6461−6470. (38) Afshar, M. G.; Crespo, G. A.; Bakker, E. Coulometric Calcium Pump for Thin Layer Sample Titrations. Anal. Chem. 2015, 87, 10125−10130. (39) Afshar, M. G.; Crespo, G. A.; Bakker, E. Thin-Layer Chemical Modulations by a Combined Selective Proton Pump and pH Probe for Direct Alkalinity Detection. Angew. Chem., Int. Ed. 2015, 54, 8110−8113. (40) Grygolowicz-Pawlak, E.; Numnuam, A.; Thavarungkul, P.; Kanatharana, P.; Bakker, E. Interference Compensation for Thin Layer Coulometric Ion-Selective Membrane Electrodes by the Double Pulse Technique. Anal. Chem. 2012, 84, 1327−1335. (41) Grygolowicz-Pawlak, E.; Bakker, E. Thin layer coulometry ion sensing protocol with potassium-selective membrane electrodes. Electrochim. Acta 2011, 56, 10359−10363. (42) Shvarev, A.; Neel, B.; Bakker, E. Detection Limits of Thin Layer Coulometry with lonophore Based Ion-Selective Membranes. Anal. Chem. 2012, 84, 8038−8044. (43) Ding, J.; Cherubini, T.; Yuan, D.; Bakker, E. Paper-supported thin-layer ion transfer voltammetry for ion detection. Sens. Actuators, B 2019, 280, 69−76. (44) Dorokhin, D.; Crespo, G. A.; Afshar, M. G.; Bakker, E. A lowcost thin layer coulometric microfluidic device based on an ionselective membrane for calcium determination. Analyst 2014, 139, 48−51. (45) Afshar, M. G.; Crespo, G. A.; Dorokhin, D.; Neel, B.; Bakker, E. Thin Layer Coulometry of Nitrite with Ion-Selective Membranes. Electroanalysis 2015, 27, 609−615. (46) Crespo, G. A.; Cuartero, M.; Bakker, E. Thin Layer IonophoreBased Membrane for Multianalyte Ion Activity Detection. Anal. Chem. 2015, 87, 7729−7737. (47) Cuartero, M.; Crespo, G. A.; Bakker, E. Ionophore-Based Voltammetric Ion Activity Sensing with Thin Layer Membranes. Anal. Chem. 2016, 88, 1654−1660. (48) Jarolimova, Z.; Bosson, J.; Labrador, G. M.; Lacour, J.; Bakker, E. Ion Transfer Voltammetry at Thin Films Based on Functionalized Cationic 6 Helicenes. Electroanalysis 2018, 30, 650−657. (49) Jarolimova, Z.; Bosson, J.; Labrador, G. M.; Lacour, J.; Bakker, E. Ion Transfer Voltammetry in Polyurethane Thin Films Based on

(9) Dietrich, B.; Lehn, J. M.; Sauvage, J. P. Diaza-polyoxamacrocycles et macrobicycles. Tetrahedron Lett. 1969, 10, 2885− 2888. (10) Gutsche, C. D.; Muthukrishnan, R. Calixarenes. 1. Analysis of the product mixtures produced by the base-catalyzed condensation of formaldehyde with para-substituted phenols. J. Org. Chem. 1978, 43, 4905−4906. (11) Scott, W. J.; Chapoteau, E.; Kumar, A. Ion-selective membrane electrode for rapid automated determinations of total carbon dioxide. Clin. Chem. 1986, 32, 137−141. (12) Meyerhoff, M. E.; Pretsch, E.; Welti, D. H.; Simon, W. Role of trifluoroacetophenone solvents and quaternary ammonium salts in carbonate-selective liquid membrane electrodes. Anal. Chem. 1987, 59, 144−150. (13) Lee, H. J.; Yoon, I. J.; Yoo, C. L.; Pyun, H.-J.; Cha, G. S.; Nam, H. Potentiometric Evaluation of Solvent Polymeric CarbonateSelective Membranes Based on Molecular Tweezer-Type Neutral Carriers. Anal. Chem. 2000, 72, 4694−4699. (14) Choi, Y. S.; Lvova, L.; Shin, J. H.; Oh, S. H.; Lee, C. S.; Kim, B. H.; Cha, G. S.; Nam, H. Determination of Oceanic Carbon Dioxide Using a Carbonate-Selective Electrode. Anal. Chem. 2002, 74, 2435− 2440. (15) Xie, X.; Pawlak, M.; Tercier-Waeber, M. L.; Bakker, E. Direct Optical Carbon Dioxide Sensing Based on a Polymeric Film Doped with a Selective Molecular Tweezer-Type Ionophore. Anal. Chem. 2012, 84, 3163−3169. (16) Xie, X.; Zhai, J.; Bakker, E. pH Independent Nano-Optode Sensors Based on Exhaustive Ion-Selective Nanospheres. Anal. Chem. 2014, 86, 2853−2856. (17) Xie, X.; Zhai, J.; Crespo, G. A.; Bakker, E. Ionophore-Based Ion-Selective Optical NanoSensors Operating in Exhaustive Sensing Mode. Anal. Chem. 2014, 86, 8770−8775. (18) Xie, X. J.; Bakker, E. Ion selective optodes: from the bulk to the nanoscale. Anal. Bioanal. Chem. 2015, 407, 3899−3910. (19) Xie, X.; Bakker, E. Light-Controlled Reversible Release and Uptake of Potassium Ions from Ion-Exchanging Nanospheres. ACS Appl. Mater. Interfaces 2014, 6, 2666−2670. (20) Gerold, C. T.; Bakker, E.; Henry, C. S. Selective Distance-Based K+ Quantification on Paper-Based Microfluidics. Anal. Chem. 2018, 90, 4894−4900. (21) Zhai, J.; Xie, X.; Bakker, E. Ionophore-based ion-exchange emulsions as novel class of complexometric titration reagents. Chem. Commun. 2014, 50, 12659−12661. (22) Zhai, J.; Xie, X.; Bakker, E. Ion-Selective Optode Nanospheres as Heterogeneous Indicator Reagents in Complexometric Titrations. Anal. Chem. 2015, 87, 2827−2831. (23) Zhai, J.; Xie, X.; Bakker, E. Solvatochromic Dyes as pHIndependent Indicators for Ionophore Nanosphere-Based Complexometric Titrations. Anal. Chem. 2015, 87, 12318−12323. (24) Zhai, J.; Xie, X.; Cherubini, T.; Bakker, E. Ionophore-Based Titrimetric Detection of Alkali Metal Ions in Serum. ACS Sens 2017, 2, 606−612. (25) Š tefanac, Z.; Simon, W. Ion specific electrochemical behavior of macrotetrolides in membranes. Microchem. J. 1967, 12, 125−132. (26) Bloch, R.; Shatkay, A.; Saroff, H. A. Fabrication and Evaluation of Membranes as Specific Electrodes for Calcium ions. Biophys. J. 1967, 7, 865−877. (27) Bakker, E.; Pretsch, E. Modern Potentiometry. Angew. Chem., Int. Ed. 2007, 46, 5660−5668. (28) Pankratova, N.; Crespo, G. A.; Afshar, M. G.; Crespi, M. C.; Jeanneret, S.; Cherubini, T.; Tercier-Waeber, M. L.; Pomati, F.; Bakker, E. Potentiometric sensing array for monitoring aquatic systems. Environ. Sci.-Process Impacts 2015, 17, 906−914. (29) Cuartero, M.; Pankratova, N.; Cherubini, T.; Crespo, G. A.; Massa, F.; Confalonieri, F.; Bakker, E. In Situ Detection of Species Relevant to the Carbon Cycle in Seawater with Submersible Potentiometric Probes. Environ. Sci. Technol. Lett. 2017, 4, 410−415. (30) Cuartero, M.; Crespo, G.; Cherubini, T.; Pankratova, N.; Confalonieri, F.; Massa, F.; Tercier-Waeber, M. L.; Abdou, M.; H

DOI: 10.1021/acs.accounts.9b00056 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research Functionalised Cationic 6 Helicenes for Carbonate Detection. Electroanalysis 2018, 30, 1378−1385. (50) Yuan, D. J.; Cuartero, M.; Crespo, G. A.; Bakker, E. Voltammetric Thin-Layer lonophore-Based Films: Part 1. Experimental Evidence and Numerical Simulations. Anal. Chem. 2017, 89, 586−594. (51) Vanamo, U.; Hupa, E.; Yrjana, V.; Bobacka, J. New Signal Readout Principle for Solid-Contact Ion-Selective Electrodes. Anal. Chem. 2016, 88, 4369−4374. (52) Jarolimova, Z.; Han, T. T.; Mattinen, U.; Bobacka, J.; Bakker, E. Capacitive Model for Coulometric Readout of Ion-Selective Electrodes. Anal. Chem. 2018, 90, 8700−8707. (53) Jansod, S.; Cuartero, M.; Cherubini, T.; Bakker, E. Colorimetric Readout for Potentiometric Sensors with Closed Bipolar Electrodes. Anal. Chem. 2018, 90, 6376−6379.

I

DOI: 10.1021/acs.accounts.9b00056 Acc. Chem. Res. XXXX, XXX, XXX−XXX