Materials based of phenylboronic acid (PBA) are widely applicable as a synthetic receptor groups recognizing 1,2- and 1,3-diol moieties. This functionality is useful for the analysis biological samples, which mostly comprise sugars [1, 2], as well as for controlled cell capture-release [3] and bacteria detection [4]. Reagentless sensors developed earlier on the basis of PBA [5] are aimed at potentiometric sugar detection [6, 7]. However, the practical application of these sensors is impossible because of the extremely low response (1.5–2.0 mV) to maximal glucose concentration, which is therefore unreliable.

A high sensitivity of electrochemical sensors can be reached in the electrochemical impedance spectroscopy (EIS). Impedimetric sensors were developed on the basis of PBA [8–10] and structurally similar boronate-substituted polyaniline [11–14]. The main drawback of these sensors is due to the use of an increase in resistance as an analytical signal, which makes impossible the discrimination between the latter and background signal, always also resulting in an increase in resistance. Hence, PBA-based sensor investigation is demanding nowadays.

Our group has developed novel sensors on the basis of aminophenylboronic acids. We demonstrated a decrease in the resistance of poly(aminophenylboronic acids) as a result of specific binding to analytes, which is opposite to the background signal. The review is aimed at the investigation of analytical characteristics of the developed electrochemical sensors on the basis of 2- and 3-aminophenylboronic acids and their applications.

Phenylboronic acids as synthetic receptors. Phenylboronic acids are monosubstituted arylboronic acids Ph−B(OH)2 having affinity to compounds bearing nucleophilic groups or, in other words, Lewis bases which are electron donors [15]. The neutral form of boronic acid includes a boron atom with sp2-hybridization, which possesses an open shell (6 valence electrons: 3 intrinsic ones and 3 from covalent bonds with neighboring atoms) and is, therefore, electron acceptor due to the vacant p-orbital. This low-energy orbital is orthogonal to the three substituents of the atom, forming a planar trigonal geometry [16].

As an electron acceptor boron atom readily interacts with Lewis bases forming a stable octet of electrons. This property allows boronic acids to reversibly form stable complexes with simple Lewis bases (fluoride, hydroxide and cyanide ions), as well as with 1,2- and 1,3-substituted Lewis bases containing hydroxyl or carboxyl groups [17]. Consequently, these properties provide an application of boronic acids as a receptor group for binding sugars, α-hydroxyacids, cyanide, and fluoride. We will further discuss boronic acids as receptors in reactions with substances containing diol fragments, for example, monosaccharides, polyols and α-hydroxyacids.

Considering the interaction of boronic acid with 1,2- and 1,3-cis-diol fragments, it is necessary to take into account factors influencing the selectivity of the sensors. First, the selectivity itself is determined by the process of analyte binding to the sensitive element. The desired selectivity is usually achieved by choosing an appropriate selective element of the sensor that thermodynamically favors binding to the target via the other components of the analyzed sample. In addition, selectivity appears as a nature of the analytical signal. Though the affinity of the target to sensing layer may not be the highest in comparison to other analytes, the host−quest interaction can result in a unique response of the sensor [18]. The maximum effectiveness of the sensor is obviously achieved in the case of a combination of both factors. For instance, the appearance of the unique response alone can be hindered by the presence of interfering compounds giving opposite response due to concurrent binding. Further we will discuss the unique response of poly(aminophenylboronic acids) to the presence of analytes.

Acid−base equilibrium of phenylboronic acid. The acid−base properties of phenylboronic acid differ from the common Brönsted acidity, resulting in ionization after the release of a proton. In proton solvents, boronic acid acidity develops in the reaction with a solvent molecule forming an anionic tetrahedral boronate and releasing a proton (Scheme 1).

Scheme 1 . Interaction of phenylboronic acid with 1,2-cis-diol.

Phenylboronic acid hydroxyl group deprotonation is determined by higher pKa than that for the interaction of the initial form of boronic acid with water or alcohol resulting in the formation of the negatively charged boronate, and the last one is difficult to deprotonate. Hereby, the acidity of boronic acid is a consequence of Lewis acidity. In aprotic solvents, boronic acid demonstrates only Lewis acidity. It can also react with different Lewis bases containing hydroxyl, fluoride and cyanide groups. In the reaction between phenylboronic acid and analytes bearing diol fragments, an ether forms, which is also acidic because of the incomplete electron shell of the boron atom.

The interaction of phenylboronic acid with the diol resulting in ether formation is described in term of the pqr-scheme:

$${\text{pA + qD + r}}{{{\text{H}}}^{ + }} = {{{\text{A}}}_{{\text{p}}}}{{{\text{D}}}_{{\text{q}}}}{{{\text{H}}}_{{\text{r}}}},$$
((1))

where A is the neutral form of phenylboronic acid, D is neutral form of the diol. P, q and r mean stoichiometric coefficients with r being positive, negative, or zero [19].

Considering the scheme in a wide pH range, one should to take into account the deprotonation of phenylboronic acid and the diol as well as two ether forms of the former: planar neutral and anionic tetrahedral [20]. In dilute aqueous solutions, the fraction of the neutral is negligible [21] and, therefore, neglected. There are three simultaneous processes during the interaction: an acid−base equilibrium of phenylboronic acid ArB(OH)2 corresponding to the dissociation constant \(K_{{\text{a}}}^{{\text{B}}}\) [equation (2)], the first stage of diol D(OH)2 dissociation with the dissociation constant \(K_{{\text{a}}}^{{\text{D}}}\) [equation (3)], and ether ArB(O)2D(OH)‾ formation according to Scheme 1 above with the equilibrium constant β11–1:

$${\text{ArB(OH}}{{)}_{2}} + {{{\text{H}}}_{2}}{\text{O}} \rightleftarrows {\text{ArB(OH}})_{3}^{ - } + {{{\text{H}}}^{ + }}(K_{{\text{a}}}^{{\text{B}}}),$$
((2))
$${\text{D(OH}}{{{\text{)}}}_{{\text{2}}}}{\text{D(OH)}}{{{\text{O}}}^{ - }}\,\,{\text{ + }}\,\,{{{\text{H}}}^{{\text{ + }}}}{\text{(}}K_{{\text{a}}}^{{\text{D}}}{\text{)}}{\text{,}}$$
((3))
$$\begin{gathered} {\text{ArB(OH}}{{{\text{)}}}_{{\text{2}}}}\,\,{\text{ + }}\,\,{\text{D(OH}}{{{\text{)}}}_{{\text{2}}}} \\ \rightleftarrows \,\,{\text{ArB(O}}{{{\text{)}}}_{{\text{2}}}}{\text{D(OH}}{{{\text{)}}}^{ - }}\,\,{\text{ + }}\,\,{{{\text{H}}}^{{\text{ + }}}}{\text{(}}{{{\beta }}_{{{\text{11}} - {\text{1}}}}}{\text{)}}. \\ \end{gathered} $$
((4))

Considering the formation of the tetrahedral complex, \({\text{ArB}}({\text{OH}}){{_{{\text{3}}}^{ - }}_{{}}},\) we can describe binding of the deprotonated acid with the neutral diol with the constant Ktetr:

$$\begin{gathered} {\text{ArB(OH)}}_{{\text{3}}}^{ - }\,\,{\text{ + }}\,\,{\text{D(OH}}{{{\text{)}}}_{{\text{2}}}} \\ \rightleftarrows \,\,{\text{ArB(O}}{{{\text{)}}}_{{\text{2}}}}{\text{D(OH}}{{{\text{)}}}^{ - }}({{K}_{{{\text{tetr}}}}}). \\ \end{gathered} $$
((5))

For Ktetr we obtain:

$$\begin{gathered} {{K}_{{{\text{tetr}}}}} = \frac{{{{\beta }_{{11 - 1}}}}}{{K_{{\text{a}}}^{{\text{B}}}}}\,\,\,\,{\text{or}} \\ {\text{log}}{{K}_{{{\text{tetr}}}}} = {\text{log}}{{\beta }_{{11 - 1}}} + {\text{p}}K_{{\text{a}}}^{{\text{B}}}. \\ \end{gathered} $$
((6))

A crucial experimental parameter reflecting the degree of complexation in the given conditions is the apparent constant of binding of phenylboronic acid to diol (Kapp), which is calculated from the total concentrations of all components [equation (7)]. The letter Σ means the total concentration of both protonated and deprotonated forms of phenylboronic acid and diol, and, as the reaction predominantly follows the tetrahedral anionic complex pathway [21], its total concentration coincides with the equilibrium one.

$${{K}_{{{\text{app}}}}} = \frac{{{\text{[ArB(O}}{{{\text{)}}}_{{\text{2}}}}{\text{D(OH}}{{{\text{)}}}^{ - }}{\text{]}}}}{{{{{{\text{[(ArB(OH}}{{{\text{)}}}_{{\text{2}}}}{\text{)]}}}}_{\sum }}{{{{\text{[D(OH}}{{{\text{)}}}_{{\text{2}}}}{\text{]}}}}_{\sum }}}}.$$
((7))

Combining the equation (7) with mass balance equations and expressions for \(K_{{\text{a}}}^{{\text{B}}}\) and \(K_{{\text{a}}}^{{\text{D}}},\) we obtain \({{K}_{{{\text{app}}}}}{\text{:}}\)

$${{K}_{{{\text{app}}}}} = \frac{{{{\beta }_{{11 - 1}}}[{{{\text{H}}}^{ + }}]}}{{(K_{{\text{a}}}^{{\text{D}}} + [{{{\text{H}}}^{ + }}])(K_{{\text{a}}}^{{\text{B}}} + [{{{\text{H}}}^{ + }}])}}.$$
((8))

Hence, the Kapp vs. pH curve contains a maximum, which corresponds to pH of the most effective binding

$${\text{p}}{{{\text{H}}}_{{{\text{opt}}}}} = \frac{{({\text{p}}K_{{\text{a}}}^{{\text{D}}}) + {\text{p}}K_{{\text{a}}}^{{\text{B}}})}}{2},$$
((9))
$$K_{{{\text{app}}}}^{{\max }} = \frac{{{{{{\beta }_{{11 - 1}}}} \mathord{\left/ {\vphantom {{{{\beta }_{{11 - 1}}}} {{\text{p}}K_{{\text{a}}}^{{\text{D}}}}}} \right. \kern-0em} {{\text{p}}K_{{\text{a}}}^{{\text{D}}}}}}}{{{{{\left( {1 + \sqrt {{{K_{{\text{a}}}^{{\text{B}}}} \mathord{\left/ {\vphantom {{K_{{\text{a}}}^{{\text{B}}}} {K_{{\text{a}}}^{{\text{D}}}}}} \right. \kern-0em} {K_{{\text{a}}}^{{\text{D}}}}}} } \right)}}^{2}}}}.$$
((10))

It should also be noted that the value of Kapp depends on the stereochemistry of the diol and the amount of 1,2- and 1,3-cis-diol groups. For example, the Kapp value of boronic acid at pH 7.4 for glucose is 4.6 M−1, for galactose, 15 M−1, and for sorbitol, 370 M−1 [22].

In a neutral medium, only 30% of phenylboronic acid forms complexes with glucose [7]. The application of phenylboronic acid to analyzing samples (where sugar content is usually low and the amount of the complex is negligible) requires an increased affinity of phenylboronic acid to the diol. One of the possible ways to increasing the apparent binding constant is to insert an electron acceptor group into the benzene ring. The conducting polyaniline chain also possesses electron acceptor properties and can, therefore, also be applied to increasing the binding constant.

Another approach to enhance the affinity of the boronic acid to the diol-containing substance is to mimic multivalent binding. The latter is one of the key features of molecular recognition in biological systems. Indeed, the binding of two multivalent objects (similar to enzyme−substrate complex) displays much a higher affinity than the simple sum of monovalent interactions. This phenomenon is known in coordination chemistry as the chelate effect.

Such an approach can easily be implemented for phenylboronic acids, as they in fact form a polydentate ligand, which has multiple neighboring boronic groups capable of binding a sugar molecule. This mimics cluster the glycoside effect of the polypeptide chain of lectins with multiple binding sites [23]. Thus, inserting of phenylboronic acids into a polymer chain may increase the affinity of boronic group to diol fragments.

The idea of boronate-substituted application in sensing systems is promising from the point of view of biomimetics and electrochemical sensors. On the one hand, phenylboronic acid acts as a synthetic receptor towards the diol fragments of sugars. However, all known PBA-based sugar detection systems have several limitations: low sensitivity and stability, mediator usage, impossibility of distinguishing between the analytical and background signal [8–10, 24].

On the other hand, the formation of a charged complex involving substituents of the main chain (for example, upon binding of sugar to boronic acid moiety) affects polymer conductivity. It makes possible the use of polymer conductivity as a response and hereby to refuse mediator utilization resulting in a reagentless sensing system.

Aminophenylboronic acid electropolymerization. 3‑Aminophenylboronic acid (3-APBA) electropolymerization was carried out from acidic solutions containing a fluoride ion as described in [5].

We have chosen the optimal maximum anodic potential equal to 0.9 V (all potentials are given in respect to silver chloride reference electrode unless otherwise stated), which is 200 mV lower than in [7]. It leads to a decrease of the degradation of the polymer formed occurring at high potentials in the polymerization and, therefore, to an increase in the conductivity of the polymer.

In 3-APBA molecule B(OH)2 group in the meta position to the amino group (Scheme 2, a) is a weak electron acceptor, which prevents an electrophilic attack of the radical cation on the vicinal carbon of the benzene ring in the para position to the amino group and hereby inhibits polymer growth occurring “head-to-tail.” Introducing an excess of fluoride ions with respect to the monomer amount converts the B(OH)2 group into the electron donor \({\text{BF}}_{3}^{ - }\) (Scheme 2, b), acting as an electrophilic attack activator on the para position to the amino group [25].

Scheme 2 . 3-Aminophenylboronic acid in (a) neutral form and (b) in the excess of F–.

In the optimization of the growth solution content, we considered different fluoride and sulfuric acid concentrations. The found content optimal for the 3‑APBBA electropolymerization consists of 40 mM 3-APBA and 200 mM fluoride into 100 mM sulfuric acid [26]. According to the equilibrium constant for phenylboronic acid in [27], under the optimal conditions, the B(OH)2 group is almost entirely converted into \({\text{BF}}_{3}^{ - }\).

Conducting polymer growth is accompanied by a nucleation processes at high potentials (Fig. 1)—these are displayed at the first cycles of cyclic voltammograms as “nucleation loops” [28]. Additionally, at higher anodic potentials about 0.9 V, the oxidation current increases with the number of cycles coinciding with an increase in conducting surface during electropolymerization

Fig. 1.
figure 1

Electropolymerization of 3-aminophenylboronic acid in 0.2 M NaF in 0.1 M H2SO4; potential sweep rate 40 mV/s. Insert: cyclic voltammogram of the electrode modified with poly(3-APBA) in 0.1 M KCl in 0.1 M HCl.

Full size image

The formation of a conducting polymer does not occur upon the electropolymerization of 2-ABPA if it is conducted under the conditions optimal for 3‑APBA. As it was mentioned above, the weak electron acceptor B(OH)2 group in the meta position of 3‑APBA turns into an electron-donating group in the presence of fluoride ions, thereby activating electrophilic substitution in the para position to the amino group.

In contrast to 3-APBA, there is no activation of the 2-APBA molecule, because the para position with respect to the amino group, by which polymerization proceeds, is in the meta rather than the ortho position to the electron-donating group (Scheme 3).

Scheme 3 . 2-Aminophenylboronic acid (a) in the neutral form and (b) in the excess of fluoride.

Thus, unlike 3-APBA, the –B(OH)2 group of the neutral form of the 2-APBA monomer does not inactivate the growth of polyaniline in the absence of fluoride ions. To confirm this hypothesis, the electropolymerization of 2-APBA was carried out in the presence and in the absence of fluoride ions.

Cyclic voltammograms (CV) of the obtained material were recorded in the background electrolyte in the absence of the monomer (Fig. 2, insert): two intense pairs of peaks appear if the polymer is synthesized in the absence of fluoride, while the material obtained in the presence of fluoride exhibits insignificant electroactivity [29].

Fig. 2.
figure 2

Electropolymerization of 2-aminophenylboronic acid on glassy carbon disk electrode in the absence of fluoride; potential sweep rate 40 mV/s. Insert: cyclic volammogram of the electrode modified with poly(2-APBA) in 0.1 M KCl in 0.1 M HCl: solid line – in the absence of fluoride, dashed line – in the presence of fluoride.

Full size image

It should be noted that the CV of the polymer obtained by the electropolymerization of 2-APBA in the absence of fluoride is the same as the CV of the polymer obtained upon the electropolymerization of 3-APBA in the presence of fluoride: the transition potentials in the electroactivity region are the same, indicating the identity of the structure and redox states of polymers synthesized under various conditions.

Change of resistance of poly(3-aminophenylboronic acid) after specific and unspecific interactions. As described above in Scheme 1, the interaction of 1,2- and 1,3-cis-diol with phenylboronic acid results in a negatively charged tetrahedral complex. The boronic acid moiety included into the polymer chain retains its ability to form charged complex with 1,2- and 1,3-cis-diols, including saccharides, polyols or hydroxiacids. The influence of complex formation on the conducting properties of poly(3-aminophenylboronic acid) was investigated by electrochemical impedance spectroscopy. The last method is one of the most effective, informative, and non-destructive methods for the investigation of conducting properties of polymers. The most common procedure consists in the application of a sinusoidal potential E(t) with a small amplitude (usually 5 mV) at the chosen angular frequencies (ω) and measurement of current I(t) at the same frequencies. After passing current through the investigated system, phase shift ϕ between E(t) and I(t) occurs, which is reflected in total impedance Z:

$$\begin{gathered} Z(\omega ) = \frac{{E(t,\omega )}}{{I(t,\omega )}} = \frac{{{{E}_{{\text{M}}}}{\text{sin(}}\omega t)}}{{{{I}_{{\text{M}}}}\sin (\omega t + \phi )}} \\ = \operatorname{Re} (Z) + j{\text{Im(}}Z{\text{)}}{\text{,}} \\ \end{gathered} $$
((11))

where EM and IM are amplitude of voltage and current, respectively, ϕ is phase shift, and Re(Z) and Im(Z) are real and imaginary parts of impedance. Impedance is presented as a complex number, because the current behavior differs from voltage not only in terms of amplitude but also in terms of phase shift. The way of plotting imaginary part Im(Z) versus the real part Re(Z) of impedance is called the Nyquist plot or diagram.

Impedance spectra of poly(3-APBA) in Nyquist plots have stable shape (not less than three sequential spectra coincide). After the addition of glucose into a buffer solution with an electrode modified by poly(3-APBA), the diameter of high frequency semicircle decreases, which means that polymer resistance Rp also decreases (Fig. 3, ◼). The replacement of a glucose-containing solution by a fresh buffer results in restoring to the initial shape of the spectrum (Fig. 3, △). Therefore, polymer response is reversible and the sensor is reusable.

Fig. 3.
figure 3

Nyquist plots of polymer-modified disk electrode in phosphate buffer pH 7.0, Edc = 50 mV: (◻) without glucose, (◼) with 50 mM glucose, (△) after glucose washing out; solid lines—approximations of the spectra by equivalent circuit used (see insert).

Full size image

For analytical purposes the quantitative performance characteristics of the sensors are obtained by the approximation of the spectra according to a suitable equivalent scheme. The scheme is an electric scheme consisting, as a rule, of resistors and capacitors modeling the physicochemical processes occurring in the investigated system. Sometimes special elements that do not have analogues in common electric schemes imitating electrokinetic phenomena, diffusion, surface geometry, and other features are used. Equivalent schemes often contain a parallel circuit as a result of a simultaneous processes of double layer charging presented by capacitor CDL and a faradaic process in series with ohmic resistance of the solution RΩ. In case of conductive polymers, the faradaic impedance comprises ohmic resistance of the polymer Rp and diffusion impedance W [30].

After fitting the spectra with the simplest equivalent scheme valid for conductive polymers—Randles scheme with diffusional impedance (Fig. 3, insert)—we obtained values of approximation parameters (Table 1). As is seen, the most significant changes after binding of glucose to poly(3-APBA) is observed for polymer resistance Rp (ca. 1.5 times). The difference in Rp between the spectra obtained before and after glucose addition is considered as a signal.

Table 1.   Parameters of equivalent circuit used for the Nyquist diagrams approximation (Fig. 3) before and after glucose addition, and after replacing the buffer solution with a glucose free solution
Full size table

Such an increase of the polymer conductivity can be explained by an analogy with polyaniline doping upon the introduction of ionogenic groups into the benzene ring. Anionic complex formation freezes negative charge in the polymer chain and leads to poly(3-APBA) conductivity increase. Thus, specific interactions result in an increase in the polymer conductivity.

Unspecific interactions do not result in significant changes in the polymer resistance in contrast to specific ones. For example, sodium acetate does not form a cyclic ether with boronic acid, as it has one OH group, only 1,2- and 1,3-cis-diols energetically favors five- or six-membered cycle formation. In addition, in our experiments we demonstrated that common polyaniline slightly and irreversibly increases its resistance in the presence of glucose. These results indicate that the boronic acid moiety indeed cause conductivity increase upon binding to substances with 1,2- or 1,3-cis-diols.

The other research groups showed an increase in resistance because of the interaction of phenylboronic acid with polyols [11, 31]. This behavior does not allow discrimination between the response of the sensor to the analyte and either degradation or unspecific binding.

To conclude, the developed sensors capable of differentiating specific interactions from unspecific binding in contrast to the majority of other conducto- and impedimetric systems. The reversible change in resistance calculated from Nyquist plots was taken as a response of the sensor.

As for poly(2-APBA), specific binding in this case also results in conductivity increase. This means that the poly(APBA) conductivity increase as a response is universal and does not depend on whether 2- or 3-substituted aniline was used as a monomer.

Apparent constant Kapp of binding of glucose to poly(3-aminophenylboronic acid) calculation. From Nyquist plots of poly(3-APBA) on the surface of a glassy carbon disk electrode in a buffer solution with known pH it is possible to obtain polymer resistance Rp for each glucose concentration and, therefore, to derive the dependency of Rp versus concentration (Fig. 4).

Fig. 4.
figure 4

Calibration curve of Rp calculated from Nyquist plots versus glucose concentration, approximation done by equation (n = 3, P = 0.95).

Full size image

If we suggest a linear dependence of the signal from the concentration of glucose-bound and glucose-free boronic acid groups, we obtain:

$${{R}_{p}}([{\text{S]) = }}\frac{{{{R}_{{p{\text{,0}}}}} + {{R}_{{p{\text{,}}\infty }}},{{K}_{{{\text{app}}}}}[{\text{S]}}}}{{1 + {{K}_{{{\text{app}}}}}[{\text{S]}}}},$$
((12))

where Rp,0 is polymer resistance in the absence of ligand S, Rp,∞ is polymer resistance when all boronic group are bound in complex; Kapp is apparent constant of binding of diol to boronic group. The dependence of Rp versus [S] is a curve with saturation at high concentration of S and has a linear shape at low concentration of S.

From the approximation of the dependence of Rp on glucose concentration by Eq. (12), we obtained Kapp = 20 ± 2 M–1 (n = 3, P = 0.95). Since standard errors are less than few percent, the approximation is applicable.

From the approximation of the concentration dependencis of Rp gained from Nyquist plots for different diols, such as glucose, lactate, fructose, galactose, and sorbitol, we obtained detection range and lower limits of detection. Both are in mM: from 6 to 250, the limit equals 3 (glucose); from 2 to 100, the limit equals 1 (lactate); from 2 to 50, the limit equals 1 (galactose); from 0.5 to 35, the limit equals 0.2 (fructose); from 0.02 to 0.34, the limit equals 0.01 (sorbitol). Sensitivity depends on Kapp and Rp,∞ and was calculated as tangent of the initial slope of Rp versus [S] dependence (Fig. 4). Analytical performance characteristics for different poly(APBA) are presented in Table 2.

Table 2.   Analytical performance characteristics of sensors based on poly(aminophenylboronic acids): (a) – poly(2-aminophenylboronic acid), (b) – poly(3-aminophenylboronic acid)
Full size table

From impedance we calculated also Kapp for different polyols and poly(3-APBA) (M–1, n = 3, P = 0.95): 20 ± 2 (glucose); 45 ± 3 (lactate); 100 ± 15 (galactose); 220 ± 12 (fructose); 6500 ± 500 (sorbitol). For poly(2-APBA) the similar constants were determined as well (presented in Table 3). The trend of an increase in Kapp in the series glucose < lactate < fructose for poly(2-APBA) and poly(3-APBA) is similar to that of phenylboronic acid [22].

Table 3.   Apparent binding constants of phenylboronic acid and poly(aminophenylboronic acids) with glucose, fructose and lactate
Full size table

Apparent constant behavior indicates that the effect of polymer conductivity increase is indeed a result of the interaction of diols with boronic acid.

Apparent binding constant pH-dependences for the interaction of different substances with poly(aminophenylboronic acids). Apparent binding constants Kapp are pH-dependent because all components of the reaction, boronic acid, diol, and ether undergo acid dissociation processes [see equations (2)–(4)] and Scheme 1). Kapp for phenylboronic acid interactions with different substances are known while Kapp for polymer are not known. We investigated the interaction of poly(3-APBA) with diols with different \({\text{p}}K_{{\text{a}}}^{{\text{D}}}.\) in the pH range 5.4–7.4. The choice of the pH range is dictated by the following facts: above 7.4 polymer dedoping processes are significant and below 5.4 conductivity is too high for the reliable calculation of poly(3-APBA) resistance. The dependencies of the latter versus pH are given at Fig. 5.

Fig. 5.
figure 5

Dependences of the apparent constant of binding of poly(3-APBA) to glucose (a) and lactate (b) on solution pH: (◼)— experimental; solid line and dashed line—approximation (n = 3, P = 0.95).

Full size image

From Eq. (8) one can conclude that the curve Kapp vs. pH reaches its maximum, which is dependent on the diol acid dissociation constant \(K_{{\text{a}}}^{{\text{D}}}\) and boronic acid dissociation constant \(K_{{\text{a}}}^{{\text{B}}}.\) In the case of big difference in \({\text{p}}K_{{\text{a}}}^{{\text{D}}}\) and \({\text{p}}K_{{\text{a}}}^{{\text{B}}}\) values for poly(3-APBA), Kapp remains virtually constant at pH close to \(K_{{{\text{app}}}}^{{{\text{max}}}}\). For substances with pKa > 12 (for example, glucose), one observes a “saturation” rather than a “bell-shaped” plot. It follows from Eq. (8) that when \(K_{{\text{a}}}^{{\text{D}}} \ll \) [H+], a condition which is fulfilled for glucose at pH < 11, the expression for Kapp takes the following form:

$${{K}_{{{\text{app}}}}} = \frac{{{{\beta }_{{11 - 1}}}}}{{K_{{\text{a}}}^{{\text{B}}} + [{{{\text{H}}}^{ + }}]}},$$
((13))
$${\text{li}}{{{\text{m}}}_{{[{{{\text{H}}}^{ + }}] \to 0}}}{{K}_{{{\text{app}}}}} = {{{{\beta }_{{11 - 1}}}} \mathord{\left/ {\vphantom {{{{\beta }_{{11 - 1}}}} {K_{{\text{a}}}^{{\text{B}}}}}} \right. \kern-0em} {K_{{\text{a}}}^{{\text{B}}}}} = {{K}_{{{\text{tetr}}}}}.$$
((14))

For diols with pKa > 12 deprotonation at pH < 7.5 is negligible and one observes a flat region at pH above \({\text{p}}K_{{\text{a}}}^{{\text{B}}},\) where Kapp remains virtually constant. In case of hydroxyacid with pKa < 4 it completely deprotonated at the whole pH range under investigation and Kapp decreases with increasing pH in accordance with Eq. (8). The predicted behavior is indeed observed for Kapp calculated for diol with high pKa (glucose) as well as for hydroxyacid with low pKa (lactate).

Now assume that boron exists in both trigonal ‒B(OH)2 and tetrahedral \( - {\text{B}}({\text{OH}})_{{\text{3}}}^{-}\) states. Consequently, ether formation occurs in both ways (Scheme 4). For Kapp as indicator of the fraction of complexed boronic acid we obtain:

$${{K}_{{{\text{app}}}}} = \frac{{{\text{[ArB(O}}{{{\text{)}}}_{{\text{2}}}}{\text{D(OH}}{{{\text{)}}}^{ - }}{\text{] + [ArB(O}}{{{\text{)}}}_{{\text{2}}}}{\text{D]}}}}{{{{{{\text{[D]}}}}_{\sum }}{{{{\text{[B]}}}}_{\sum }}}}.$$
((15))

Combining Eqs. (2), (3) and (5) and the condition of electro neutrality of the medium [D(OH)O‾] = 0, we obtain:

$${{K}_{{{\text{app}}}}} = \frac{{{{{{K}_{{{\text{tetr}}}}} + {{K}_{{{\text{trig}}}}}[{{H}^{ + }}]} \mathord{\left/ {\vphantom {{{{K}_{{{\text{tetr}}}}} + {{K}_{{{\text{trig}}}}}[{{H}^{ + }}]} {K_{{\text{a}}}^{{\text{B}}}}}} \right. \kern-0em} {K_{{\text{a}}}^{{\text{B}}}}}}}{{{{1 + [{{H}^{ + }}]} \mathord{\left/ {\vphantom {{1 + [{{H}^{ + }}]} {K_{{\text{a}}}^{{\text{B}}}}}} \right. \kern-0em} {K_{{\text{a}}}^{{\text{B}}}}}}}.$$
((16))

However, pH dependence of Kapp for glucose appears to have higher slope than Eq. (16) predicts (Fig. 5a).

Experimental values are better described by the following approximation analogous to Eq. (16) (Fig. 5a, dashed line):

((17))

Proportionality to [H+] squared is possibly due to bis-boronate complex formation, which is possible for glucose in the furanose form.

Scheme 4 . Reaction between poly(3-APBA) and 1,2-cis-diol.

In this case, two boronic acid groups interact with one glucose molecule, forming a dianionic complex and revealing two protons. So \({\text{p}}K_{{\text{a}}}^{{'{\text{B}}}}\) reflects the deprotonation of both groups and should be twice as big as \({\text{p}}K_{{\text{a}}}^{{\text{B}}}\). From the approximation by Eq. (17) we indeed obtain \({\text{p}}K_{{\text{a}}}^{{'{\text{B}}}} = {\text{11}}.{\text{9}} \pm 0.{\text{1}},\)Ktetr = 20 ± 2 M–1 and Ktrig = 0.

Scheme 5 . Lactate to aminophenylboronic acid binding.

Lactate (pKa = 3.86) in the pH range given above is deprotonated and interacts with the boronic group as an anion with the equilibrium constant K’tetr (Scheme 5) and the following Kapp:

$${{K}_{{{\text{app}}}}} = \frac{{{{K_{{{\text{tetr}}}}^{'}[{{{\text{H}}}^{ + }}]} \mathord{\left/ {\vphantom {{K_{{{\text{tetr}}}}^{'}[{{{\text{H}}}^{ + }}]} {K_{{\text{a}}}^{{\text{B}}}}}} \right. \kern-0em} {K_{{\text{a}}}^{{\text{B}}}}}}}{{{{1 + [{{{\text{H}}}^{ + }}]} \mathord{\left/ {\vphantom {{1 + [{{{\text{H}}}^{ + }}]} {K_{{\text{a}}}^{{\text{B}}}}}} \right. \kern-0em} {K_{{\text{a}}}^{{\text{B}}}}}}}.$$
((18))

The Kapp dependence versus pH is consistent with Eq. (18), as Kapp decreases with [H+] decrease (Fig. 5b).

In the case of substances forming a negatively charged complex with a boronic acid moiety of the polymer, the resistance of the latter thereby decreases. On the contrary, substances without 1,2- or 1,3-diol fragments (for instance, acetate) cannot form the complex and a decrease in polymer resistance does not occur. Moreover, in the case of common polyaniline, polymer resistance in the presence of diol does not decrease as it increases irreversibly.

The Kapp dependences of the binding of saccharides and hydroxiacids with poly(2-APBA) also demonstrate the behavior similar to that of phenylboronic acid (Fig. 6). In the case of lactate, Kapp decreases when pH increases and, in case of fructose, Kapp increases when pH increases. Thus, Kapp for poly(2-APBA) and diols, as well as pH dependences, have the same trends as for phenylboronic acid. It confirms the suggestion about the observed conductivity increase of poly(2-APBA), being a result of the interaction of the latter with diol fragments of saccharides and hydroxiacids.

Fig. 6.
figure 6

Dependences of the apparent constant of binding of poly(2-APBA) to fructose (⚫) and lactate (▲) on pH (n = 3, P = 0.95).

Full size image

Despite the structure of the poly(APBA) being independent on the initial monomer (3-APBA or 2‑APBA), the resulting polymer has significantly different affinity towards saccharides and hydroxiacids. The tendency of Kapp change with respect to diol nature is similar to that for phenylboronic acid for both poly(2-APBA) and poly(3-APBA). However, the latter has from two to ten times higher Kapp than the former (Table 3). So the presence of fluoride in the polymerization of 3-APBA possibly influences the performance of the sensor by configuring the boronic group in a way favoring analyte binding. Since fluoride is relatively small, the key role may be played by the steric effect of the boronic acid group itself rather than by the cavity formed around the fluoride: an excess of the latter transforms boron atom state from trigonal sp2 into tetrahedral sp3. Apparent binding constants for the tetrahedral form exceeds those for the trigonal form and, therefore, changing boron atom state during the electropolymerization increases the affinity of the resulting polymer to the analytes.

The calculated threshold point \({\text{p}}K_{{\text{a}}}^{{\text{B}}}\) for poly(APBA) is 6.1–6.6, which is lower than that for common phenylboronic acid (≈8.9) and makes it possible to utilize poly(APBA) as a sensor material in neutral media. The fact that the discovered threshold point does not coincide with the pKa value for phenylboronic acid can be explained in terms of complicated receptor structure and electron acceptor properties of the polyaniline chain [20].

The obtained results have a great impact on the analytical application of the poly(APBA). The great majority of the known impedimetric and conductometric (bio)sensors generate their specific signals as the increase in resistance. This, however, is of doubtful practical importance, because nonspecific signals (background) are also directed to a resistance increase. Thus, one never knows whether a low concentration of analyte or a high content of interfering compounds is sensed. Accordingly, it is hard to overestimate the true reagentless analytical principle based on the transducer, which specific signal results in a resistance decrease, directed oppositely to the background.

Thermodynamic data also confirm that the observed conductivity increase because of the interaction of the electropolymerized 3-aminophenylboronic acid with polyols is indeed a result of complex formation between the boronic acid residue and saccharides or hydroxyacids.

Molecularly imprinted poly(3-aminophenylboronic acid). The lack of specificity of poly(APBA), because of the affinity of the latter to a number of important compounds (sugars, hydroxyacids, etc.), limits the applicability of sensors based on it. Increasing selectivity and sensitivity of synthetic materials to a particular substance is a key task in analytical practice. Polymers molecularly imprinted with analytes are used for this purpose. Synthesis is performed in the presence of target molecules or their structural analogs. Polymerization in the presence of template molecules with their subsequent extraction leads to the formation of cavities in the polymer matrix that are complementary in shape and size to these molecules. A molecularly imprinted polymer has an increased affinity for the compound in the presence of which it was synthesized, because of the stereochemical orientation of the reaction centers formed in the synthesis.

The use of sugars or hydroxyacids as template molecules in the polymerization of APBA will provide both the beneficial tetrahedral configuration of boronic acid and an appropriate size of cavities in the poly(APBA) matrix.

Conductive polyaniline derivatives can be synthesized only in an acidic medium (pH <4) [32], but monosaccharides bind with poly(APBA) preferably in the neutral pH range according to the pH dependences of the observed binding constants in Figs. 5 and 6. In [33] the electrochemical synthesis of poly(APBA) molecularly imprinted with fructose was proposed. The binding constant of fructose with phenylboronic acid is high enough, but at pH of the synthesis (pH 5), it decreases significantly. In addition, only nonconducting polyaniline films unsuitable for electrochemical detection by impedance spectroscopy can be obtained if polymerization is carried out in a phosphate buffer solution. For this reason, the use of sugars to create molecular imprints in boronate-substituted polyaniline is inefficient.

Unlike monosaccharides, hydroxyacids possess high binding constants with PBA in a weakly acidic medium required for the formation of conductive polyaniline. Therefore only hydroxyacids, but not sugars, can be used to create molecular imprints in poly(APBA) (Fig. 7).

Fig. 7.
figure 7

Comparison of the optimal pH values for conductive polyaniline synthesis and for binding of phenylboronic acid to saccharides and hydroxyacids.

Full size image

In [34] we have proposed the conditions for the electrochemical synthesis of conducting boronate-substituted polyaniline in the presence of fluoride and hydroxyacids (tartrate, lactate). It was shown that one can manage the analytical characteristics of the sensors increasing the sensitivity and selectivity of the sensors by varying the template molecules for the polymerization of 3-APBA. The EIS method can be used for analysis if the electrode surface is modified by a conductive polymer. We have found that it is possible to obtain conducting molecularly imprinted poly(APBA)-films in course of polymerization of 3‑APBA in presence of a hydroxy acid instead of the fluoride ion. Salts of hydroxyacids (lactate, tartrate, and citrate) capable of covalently binding with 3-APBA at pH 2–4 were used to create molecular imprints.

As phenylboronic acid—diol complexes predominantly form in the anionic form (stability constants of charged tetrahedral complexes are higher than neutral trigonal ones [20]), the equilibrium of the reaction between 3-APBA and hydroxy acid shifts toward the formation of a tetrahedral negatively charged boronic ether. As in the case of fluoride a negative charge on boron atom promotes the orientation of electrophilic substitution in the 3-APBA aromatic system favorable for the growth mechanism of conductive polyaniline (Scheme 6).

Scheme 6 . The formation of tetrahedral boronate ether upon the interaction of 3-aminophenylboronic acid with hydroxyacid.

For this reason, only the meta isomer of APBA can be used for the synthesis of molecularly imprinted poly(APBA), from which the synthesis of a conducting polymer is possible only if there is a negative charge on the boron atom. In contrast, the binding of the ortho isomer with a saccharide in the synthesis prevents the polymerization of 2-APBA; therefore, a molecularly imprinted polymer cannot be obtained in this way.

We have synthesized a conducting polymer from 3-APBA in the presence of templating agents: sodium tartrate, potassium lactate [34], and sodium citrate. The formation of a conducting polymer coating the electrode surface confirms the covalent binding of the monomer (3-APBA) with the hydroxyacid in the polymerization, and, therefore, the inclusion of the template molecule into the polymer matrix.

According to scheme 6, the binding of aminophenylboronic acid with hydroxy acids depends on the ratio of equilibrium monomer concentration to the deprotonated hydroxy acid form in the solution for polymerization. The fraction of the uncharged form of 3‑APBA at the pH of synthesis (2–4 units) can be considered equal to 100%, because the pKa of aminophenylboronic acid is 8.9. However, the concentration of the ionized form of hydroxy acid at the corresponding pH is determined by its pKa value.

An optimal pH value for the synthesis of conductive molecularly imprinted poly(3-APBA) was found experimentally. The pH-optimum corresponds to the maximum growth rate of the conductive polymer that was estimated by the height of the sharp cathode peak on the CVA of modified electrodes. It is seen that the pH-profiles of the growth rate are bell-shaped with the maximum nearby the first pKa of the hydroxyacid.

The properties of molecularly imprinted poly(APBA) were investigated by electrochemical impedance spectroscopy Similarly to poly(2-APBA) and poly(3-APBA), the conductivity of polymers synthesized in the presence of hydroxy acids increases upon reaction with analytes. This is reflected in the Nyquist diagram by a decrease in the diameter of the high-frequency semicircle corresponding to Rp. The binding constants of sugars and hydroxyacids with molecularly imprinted polymers, determined from the concentration dependencies, are given in Table 4.

Table 4.   Apparent binding constants (M–1) of imprinted and non-imprinted poly(3-aminophenylboronic acid) with glucose, fructose and lactate
Full size table

As the polymerization of the ortho isomer doesn`t require any molecules capable of binding to boronic acid groups at the synthesis pH for the formation of conductive poly(APBA), only the polymer obtained by the fluoride-free polymerization of 2-APBA should be considered as a non-imprinted one. Comparing the observed binding constants of poly(APBA) films prepared under different conditions with the analytes, one can estimate the gain in molecular imprinting. Table 4 shows that the binding constants of molecularly imprinted polymers with glucose increase by about 10–14 times and with fructose, by 2–4 times. The binding constants of poly(APBA) synthesized in the presence of lactate and tartrate ions with glucose are 2 times higher than that of poly(APBA) obtained in the presence of fluoride ions. Poly(APBA) imprinted with citrate and poly(APBA) imprinted with fluoride have similar binding constants with glucose and fructose. Thus, molecular imprinting with hydroxyacids provides a higher affinity of poly(APBA) to monosaccharides. The analytical characteristics of electrodes modified with hydroxyacid imprinted polymers for the determination of glucose and lactate are given in Tables 5 and 6. The limit of glucose detection is lower and the dynamic range shifts to lower concentrations of glucose in the case of hydroxyacids imprinted poly(APBA). However, sensitivity to glucose does not depend on the nature of the hydroxyacid used for the synthesis of imprinted poly(APBA). Nevertheless, the sensitivity coefficients for hydroxyacid imprinted polymers are 2 times higher in comparison with that for poly(APBA) obtained in the presence of fluoride.

Table 5.   Analytical performance characteristics of electrodes modified with poly(aminophenylboronic acid) for glucose detection
Full size table
Table 6.   Analytical performance characteristics of electrodes modified with poly(aminophenylboronic acid) for lactate detection
Full size table

The highest affinity gain due to molecular imprinting is observed in the lactate detection. It can be seen from Table 4 that the binding constant of citrate imprinted poly(APBA) with lactate increased by 2.5 times compared with non-imprinted poly(APBA). The polymer imprinted with lactate displays an eight times higher binding constant.

The opposite effect was obtained for a tartrate imprinted polymer: the binding constant with lactate did not change and, therefore, was two times lower than that of poly(3-APBA) obtained in the presence of fluoride. The limit of detection for lactate was reduced from 2 to 0.3 mM for the polymer with lactate imprints (Table 6).

Thus, one can control the affinity of the polymer to the analytes and achieve the desired sensitivity and selectivity of the sensor depending on the task by synthesizing hydroxyacid- imprinted poly(3-APBA).

Different sensitivities of hydroxyacid-imprinted polymers allow the selective detection of sugars and hydroxyacids in a mixture, which was demonstrated for the mixture of lactate and fructose. Two sensors with the most differing selectivity: poly(3-APBA) imprinted with lactate and fluoride were used for analysis of mixtures of lactate and fructose in following ratios: 1 : 1, and 3 : 2. In the case of no cooperative effect of analytes on the impedimetric response of sensors, the polymer resistance in the mixture can be represented as a sum of expressions describing the calibration dependence according to Eq. (12). The parameters of nonlinear calibration curves for two sensors (\(R_{p}^{0},\)\(R_{p}^{\infty },\)Kapp) were used to calculate the concentrations by solving a system of two equations with two unknowns. The recovery values are not lower than 89% (Table 7).

Table 7.   Analysis of model mixtures of fructose and lactate
Full size table

Thus, the ability of adjusting the analytical characteristics of sensors based on poly(aminophenylboronic acids) by varying the template molecules for electropolymerization, opens up prospects for the selective non-reagent detection of sugars and hydroxyacids using synthetic receptors.

Non-enzymatic sensor based on poly(aminophenylboronic acid) for human sweat analysis. The successful imprinting of poly(APBA) allows the use of these nonezymatic sensors for medical purposes. For instance, it is possible to apply poly(APBA) sensors for the detection of lactate, which is one of the most important low molecular weight analytes in clinical diagnostics. It is known that the hypoxia (oxygen starvation) of organs and tissues leads to an increase in the lactate concentration in human blood. In addition, high interest in non-invasive diagnostics calls for the analysis of other excretory fluids in addition to blood [35, 36]. A poly(APBA)-based sensor for the detection of lactate in sweat has been developed. Blood lactate content correlates with its content in sweat, moreover, sweat gland concentrates lactate, so that its concentration in sweat is an order of magnitude higher than that in the blood.

Phenylboronic acids based non-enzymatic sensors can be used for the analysis of sweat, since lactate is the predominant component of sweat (Table 8). The impedimetric response of electrodes modified by boronate-substituted polyaniline is observed exclusively upon the binding of polymer with polyols, sugars and 1,2-hydroxyacids. Amino acids, carboxylic acids, 1,3-hydroxyacids, ketoacids, and inorganic anions (except fluoride ion) do not affect the sensors based on the phenylboronic acid signal, and, therefore, do not interfere the determination of lactate. Among sweat constituents only glucose and the fluoride ion can bind to poly(APBA) under the conditions of sweat analysis, but their concentration is 2–3 orders of magnitude lower than the lactate concentration. In addition, at pH < 7, the binding constant of poly(APBA) with glucose decreases, while it increases for lactate. Poly(APBA) imprinted with lactate has the best analytical performance characteristics for lactate detection. The selectivity of such a polymer to lactate compared to glucose can be estimated as the ratio of the corresponding binding constants: selectivity coefficient at pH 6.0 is \(K_{{{\text{lactate}}}}^{{{\text{glucose}}}}\) < 3 × 10–2.

Table 8.   Composition of human sweat [36]
Full size table

In [36] a non-enzymatic sensor based on lactate-imprinted poly(aminophenylboronic acid) synthesized on planar three-electrode structures was elaborated for lactate detection in human sweat. The impedimetric sensor based on molecularly imprinted poly(APBA) has the following analytical characteristics: sensitivity 17 ± 3 M–1, limit of detection 1.5 mM, dynamic range 3–100 mM. The stability of the analytical characteristics on storage in air at room temperature for 6 months is advantageous over that of biosensors.

The non-enzymatic sensor was applied to the analysis of 17 sweat samples from 7 volunteers. The concentrations of lactate found in the sweat are within the working range of the sensor from 10 to 40 mM. The accuracy of lactate detection for all sweat samples was confirmed by means of a highly selective biosensor based on the enzyme lactate oxidase (Fig. 8).

Fig. 8.
figure 8

Correlation of the sweat lactate concentration determined by polymer-based sensor and by biosensor.

Full size image

Thus, the non-enzymatic sensor based on molecularly imprinted poly(APBA) for the detection of lactate in sweat has been elaborated. The broad specificity of phenylboronic acids does not limit the applicability of the sensor. Firstly, lactate is the main component of sweat, and its concentration is two orders of magnitude higher than the concentration of interfering polyols, particularly glucose. Secondly, the molecular imprinting of poly(aminophenylboronic acid) with lactate improves the sensor selectivity towards lactate (the selectivity coefficient to glucose versus lactate is reduced by one order of magnitude). The sensor exhibits high stability upon storage in air at room temperature. In addition, there is no preliminary sweat sample preparation, except for two-fold dilution with buffer solution, as well as for biosensors. Thereby non-enzymatic molecularly imprinted poly(APBA)-based sensors can be proposed for the non-invasive clinical diagnostics of hypoxia.

Microorganism detection on the base of poly(3-aminophenylboronic acid). To complete the electoanalytical cell, a polymer-modified working electrode should be accompanied by an external reference and a counter electrode. Such a bulky setup has limited applicability outside the laboratory and does not allow the miniaturization and fabrication of an embedded system. To overcome these circumstances, we proposed an alternative electrode structure for sensor development. Instead of glassy carbon disk electrodes, we chose interdigitated gold microelectrodes and then used modification by poly(3-APBA) [37].

Interdigitated construction offers some advantages over classic disk electrodes. First, the interdigitated ones consist of at least two independent microelectrodes and allow two-electrode scheme application [38]. After modification by a sensing layer, these microelectrodes become the whole sensor structure. Second, microelectrodes are easy to miniaturize compared to classic electrodes and, therefore, open up a possibility of embedding the developed microsensor into a working or a living room or a ventilation system.

After changing the electrode design, we investigated poly(3-APBA) resistance change in an aqueous solution in the presence of diol (glucose). Upon the addition of glucose, the diameter of the high-frequency semicircle decreases (Fig. 9a), indicating the resistance decrease similarly to the case of polymer-modified disk electrodes (Fig. 3). Changing the electrode construction therefore does not influence the observed effect of polymer conductivity increase upon binding to glucose.

Fig. 9.
figure 9

Nyquist plots in aqueous solution (a) or in aerosol (b) for microsensors: (◻)—without glucose, (◼)—with glucose, (△)— after glucose washing out.

Full size image

From the practical point of view, aerosol analysis is among the most promising ones. Microorganism detection by conductive polymer-based sensor is difficult in aqueous solutions in long-term and in situ experiments. As the water evaporates, polymer conductivity changes in an uncontrollable manner, which makes impossible a correlation between the microorganism concentration and the change in conductivity. Commonly, aqueous media in electrochemical systems acts as a conductor between the electrodes providing the electric connection, which is necessary for the operation of the whole system. Modifying interdigitated microelectrodes with a conductive polymer makes it possible to refuse application of liquid media and to maintain electric connection even in air. As seen from micrographs of modified microelectrodes (Fig. 10), a conductive polymer forms fibers between two interdigitated microelectrodes providing electric connection and hereby operation of the sensor without a liquid electrolyte.

The applicability of microsensors for aerosol analysis was evaluated by measuring the impedance change after introducing glucose into air flow. The resulting Nyquist plots are presented in Fig. 9b.

So, the effect of 3-APBA based polymer conductivity increase as a result of specific binding occurs at interdigitated microelectrodes as well as for disk electrodes indicating the universal character of the phenomenon. In case of microsensors, this effect appears in an aqueous medium as well as in air, which allows the application of the microsensor for aerosol and water analysis.

Fig. 10.
figure 10

Scanning electron micrographs of interdigitated microelectrodes: (a) before; (b) after polymerization.

Full size image

Microsensors for microorganism detection in aqueous solutions. Microbial control is a key task of occupational safety and health. The problem of microbial air contamination is of vital importance in all human activities and everyday life. Microorganisms cause mycoses, concomitant allergic reactions, and bronchial asthma [39–41]. In addition to the threat to human health, they negatively affect processes of food, medicinal, and microelectronic production. Penicillium and Aspergillus genera are among the most wide-spread microorganisms in the domestic air and working environment [42, 43]. Currently, several basic methods are used for microbiological control: cultural method, microscopic studies [44], a DNA-based method like polymerase chain reaction [45, 46], ATP-bioluminescence [47], mass spectrometry [2] and flow cytometry [48]. The main disadvantages of the existing methods, i.e., the need in using reagents, biorecognition elements, time-consuming analysis, and expensive equipment , limit their practical application and stimulate interest in creating simple, reagentless, and rapid methods for detecting microorganisms. In addition, the methods listed above are most often not applicable to long-term monitoring. To eliminate these drawbacks, tone can apply electrochemical methods that do not require sample preparation and are applicable outside the laboratory, using simple and miniature, but sensitive and inexpensive equipment. As part of this review, we consider the use of electrochemical methods for the determination of microorganisms using reagentless sensors based on poly(3-APBA).

The cell wall of most microorganisms, including the Penicillium chrysogenum spore, consists of more than 60% of oligo- and polysaccharides containing 1,2- or 1,3-cis-diol fragments [49, 50]. The use of boronate-substituted polyaniline or poly(3-APBA) as a sensing layer with affinity to the components of the cell wall surface will allow the use of a reagentless sensor for the detection of microorganisms.

For experiments with microorganisms, aqueous suspensions of Penicillium chrysogenum (phosphate buffer solution with pH 7.0) were prepared. The content of microorganisms in the suspension was determined by cultivation on Petri dishes with a selective nutrient medium, followed by counting the colonies. The suspension with a content of 50 000 colony-forming unit per mL (CFU/mL) was investigated using an automatic laser diffraction particle analyzer. The suspension mainly contains particles of about 3 µm in size, which corresponds to the size of a single spore Penicillium chrysogenum [51]. Large-size particles (10–100 µm) were not found; particle size distribution during incubation in the analysis remained unchanged. This means that the main component of the suspension (95%) is the comprised by spores of mold.

Penicillium chrysogenum spores in an aqueous medium were detected by impedance spectroscopy using a three-electrode scheme with a working electrode modified with poly(3-APBA). After immersing the electrode in a phosphate buffer solution with pH of 7.0, a sinusoidal potential sweep was used with a constant current potential of 0.05 V versus a silver–silver chloride electrode and amplitude of 5 mV. Nyquist diagrams for sensors in a buffer solution were consistently recorded until they became statistically indistinguishable. Then, the initial solution was replaced with a suspension of mold with different concentrations determined by the cultural method. After that, Nyquist diagrams were again recorded for each concentration of microorganisms.

We chose polymer resistance Rp calculated from the diameter of the high-frequency semicircle in the diagram as an analytical signal (Fig. 11a). After prolonged impedance recording in a phosphate buffer solution, a slight increase in Rp is observed – about 5% over 1 h, i.e. background processes lead to an increase in resistance (or decrease in conductivity). In contrast to unspecific processes, the presence of microorganisms causes an increase in the conductivity of the polymer, which is important for practical use.

With an increase in the concentration of microorganisms in suspension, the resistance of the polymer Rp decreases. By calculating the resistance of the polymer, it is possible to plot a calibration curve of Rp versus the concentration of mold. As in the case of compounds containing 1,2- or 1,3-cis-diol fragments, the dependence is approximated by Eq. (12), as shown in Fig. 11b.

The detection of microorganisms in an aqueous medium was carried out with microsensors operated according to two-electrode scheme; the second microcomb served as an auxiliary electrode and a reference electrode. The microsensor was immersed in a solution of a phosphate buffer solution with a pH 7.0 and then Nyquist plots were recorded for various concentrations of microorganisms until successively taken diagrams became statistically indistinguishable. From the values of frequencies given in the diagrams, it can be seen that, in the absence of microorganisms as well as at concentrations of 300 and 600 CFU/mL, a fragment of a high-frequency semicircle is observed, which is fully manifested at higher concentrations (Fig. 12a). With an increase in the concentration of mold fungi in suspension, the diameter of the high-frequency semicircle decreases. This indicates a decrease in Rp in the presence of mold fungi for microsensors functioning according to a two-electrode circuit, as well as upon the detection of microorganisms according to a three-electrode circuit.

A comparison of the calibration curves for the sensor and the microsensor shows that the sensitivity of microorganism detection is much higher for the latter: the mold content at 1200 CFU/mL causes the resistance of the modified microelectrode to decrease by 5 times (Fig. 12b), while for the common electrodes the same suspension causes a decrease resistance by 5% (Fig. 11b).

An increase in conductivity in the presence of microorganisms is likely due to the binding of boronic acid groups of the polymer to diol groups of the cell walls of the microorganisms. This effect appears for different types of electrodes and detectable analytes (monosaccharides, polyols, hydroxyacids, microorganisms); therefore, the phenomenon of increasing the conductivity of the polymer upon binding to 1,2- or 1,3-cis-diol groups is universal.

Microsensors for microorganism detection in aerosols. An aerosol containing microorganisms (or, in other words, bioaerosol) was created by passing an air flow through a glass vessel with an aqueous suspension of Penicillium chrysogenum. To detect microorganisms in the aerosol flow, the microsensor was fixed in the path of the air flow. In the beginning of the experiment, the flask contained only phosphate buffer solution. For each concentration of microorganisms (including phosphate buffer solution), the impedance was measured until successively recorded spectra became statistically indistinguishable. As in other experiments, the resistance of poly(3-APBA) Rp was used as an analytical response.

During the detection of microorganisms, Nyquist plots were recorded in an aerosol flow created from suspensions with concentrations from 14 000 to 46 000 CFU/mL (Fig. 13a). With an increase in the concentration of mold in a saturating solution for the aerosol, the diameter of the high-frequency semicircle decreases (Fig. 13b). This indicates a decrease in the resistance of poly(3-APBA) in the presence of microorganisms in an aerosol, similarly to experiments on the detection of microorganisms in a aqueous solution (Figs. 11b and 12b).

Fig. 11.
figure 11

(a) Nyquist plots of the sensor at different mold concentrations (CFU/mL, from right to left and from top to bottom): 0, 1200, 7000, 22 000,  26 000 (n = 3, P = 0.95). (b) Calibration curve of Rp versus mold concentration approximated by Eq. (12).

Full size image
Fig. 12.
figure 12

(a) Nyquist plots of the microsensor at different mold concentrations (CFU/mL): (◻) 0, (◆) 300, (◼) 600, (⚫) 1200, (▲) 2100 (n = 3, P = 0.95). (b) Calibration curve of Rp versus mold concentration approximated by Eq. (12).

Full size image
Fig. 13.
figure 13

(a) Nyquist plots of the microsensor in different mold concentrations in aerosol (CFU/mL): (⚪) 0, (◻) 14 000, (△) 26000, (⚫) 46 000 (n = 3, P = 0.95). (b) Calibration curve of Rp versus mold concentration in saturating solution approximated by Eq. (12).

Full size image

So, experiments on the detection of Penicillium chrysogenum in a bioaerosol demonstrated that poly(3-APBA) decreases its resistance after specific interactions with the analytes, whereas during the background processes associated, for example, with degradation and dedoping, the resistance increases.

Based on the cultivation of aerosol samples with different concentrations of microorganisms, the volumetric concentrations of molds in the aerosol were calculated. The concentrations determined in air ranges from 200 to 800 CFU/m3. This makes the developed microsensor suitable for detecting microorganisms in concentrations meeting the hygienic standard for atmospheric air in populated areas (500 CFU/m3).

Thus, a reagentless electrochemical microsensor was created, suitable for the simple and rapid detection of microorganisms (for example, Penicillium chrysogenum) directly in air [52]. It does not require additional sample preparation associated with the destruction of the cell wall or membrane, or reagents. The whole procedure, including the calibration curve, takes about 3 hours. The reliability of microorganism detection is ensured by the generation of electrochemical signals: in the case of specific interactions with cell wall components, the resistance polymer decreases, and in the case of background processes, the resistance increases. The operation of the microsensor does not require additional electrodes or a special cell, which will allow the design of an electroanalytical system to be used in situ.

CONCLUSIONS

The developed reagentless sensors based on 2- and 3-aminophenylboronic acid are promising for analytical practice. In the majority of other impedimetric and conductometric (bio)sensors, the analytical signal is assigned to a decrease in conductivity. However, unspecific processes also lead to a decrease in conductivity, thus making it impossible to reliably distinguish, for example, low concentrations of an analyte from a high concentration of interfering compounds. To overcome these problems, it is necessary to mask interfering compounds by creating high ionic strength, extreme pH values, introducing surfactants, etc. At the same time, the analysis ceases to be “reagentless,” which limits the practical use of such systems by introducing additional stages of sample preparation, using additional reagents, and attracting qualified personnel. For the developed sensors based on poly(APBA) materials, the analytical signal is an increase in the polymer conductivity, directed opposite to the signal from unspecific background processes leading to a decrease in conductivity. This principle has been successfully applied to the analysis of a mixture of diol-containing compounds, human sweat, and the detection of microorganisms in the aqueous solution and in the aerosol. In addition, you can adjust the analytical characteristics of the developed sensors using the molecular imprint method. Accordingly, it is hard to overestimate the true reagentless analytical principle based on the transducer, which specific signal is resulted in resistance decrease, directed oppositely to the background. In the present review, the effect of polyaniline self-doping in neutral media is used for the generation of the specific signal. This, however, will not limit a search for similar analytical principles, which obviously opens new horizons for analytical chemistry.