Host-Tailored Sensors for Leucomalachite Green Potentiometric Measurements

A new biomimetic sensor for leucomalachite green host-guest interactions and potentiometric transduction is presented. e arti�cial host was imprinted in methacrylic acid or acrylamido-2-methyl-1-propanesulfonic acid-based polymers. Molecularly imprinted particles were dispersed in 2-nitrophenyloctyl ether and trapped in poly(vinyl chloride). e potentiometric sensors exhibited a near- Nernstian response in steady state evaluations, with slopes and detection limits ranging from 45.8 to 81.2mV decade −1 and 0.28 to 1.01 𝜇𝜇 gmL −1 , respectively. ey were independent from the pH of test solutions within 3 to 5. Good selectivity was observed towards drugs that may contaminate water near �sh cultures, such as oxycycline, doxycycline, enro�oxacin, trimethoprim, creatinine, chloramphenicol, and dopamine. e sensors were successfully applied to �eld monitoring of leucomalachite green in river samples. e method oﬀered the advantages of simplicity, accuracy, applicability to colored and turbid samples, and automation feasibility.


Introduction
Industrial aquaculture is a rapidly growing industry in many developed and developing countries such as Canada, Japan, Russia and China.A signi�cant growth at food �sh production has been observed over the past decade, due to the prevention or elimination of �sh diseases.e introduction of veterinary medicines such as antimicrobials at the food production area has born the main responsibility for this scenario.However, the wide use of antibiotics in aquaculture led to environmental and food spread of antimicrobials and may result in the emergence of antibiotic-resistant bacteria in aquaculture environments [1,2].For food safety purposes, �sh samples must be sub�ect of rigorous and frequent controls that ensure that residues of antimicrobials are below the maximum legal levels [3].
Malachite green (MG) is among the drugs �nding extensive use all over the world in the �sh farming industry since the 1930s [4].It is a dye and acts as a fungicide, ectoparasiticide, and disinfectant.Recent studies pointed out that it is a suspected carcinogen, showing highly cytotoxic effects to mammalian cells and acting as a liver tumorenhancing agent.ese observations lead to its restricted or prohibited use in aquaculture farming activities worldwide.�nce absorbed by �sh tissue, MG is metabolically reduced to leucomalachite green (LMG), its reduction product and ma�or metabolite that is found in waters and �sh tissues.us, an efficient and low-cost method for monitoring LMG would be highly appreciated [5], in order to prevent further dissemination of MG in the environment by means of aquaculture activities.
Alternative and advantageous methods should rely on expeditious and e�cient procedures providing highly speci�c and sensitive measurements.e speci�c recognition of a certain analyte may be achieved by molecularly imprinted polymers (MIPs).ese synthetic polymers are prepared by letting functional and cross-linking monomers organize around a template (to be imprinted) and copolymerize them to form a highly cross-linked polymeric structure [30,31].e template is removed aer creating vacant places where the functional groups of the monomeric forms are strictly oriented to bind this speci�c compound.e produced polymer is thus capable of rebinding the analyte with a high speci�city.In many occasions it shows binding a�nities approaching those presented by antigen-antibody systems [32,33] while offering higher stability at extremes of pH and temperature, high mechanical strength, and lower cost than their natural counterparts.More recently several papers on the synthesis of MIP materials for malachite green have been published [34][35][36][37].ese materials are always obtained by bulk imprinting, being methacrylic acid (MAA) the only functional monomer tested, and applied (or intended for) chromatographic applications.
e great potential of MIP materials has led to their inclusion as recognition elements in sensors.Transduction of different nature has been employed for this purpose [38,39], with great emphasis on electrochemisty [40].Potentiometry is included here and offers the advantages of not requiring the template to be extracted from the membrane and the species to be diffused through the membrane [40].Despite these advantages, most research work within this �led has been conducted only most recently [41][42][43][44][45][46][47][48][49][50].
erefore, the present work tries to (i) develop MIP materials designed for LMG instead MG, because LMG is the major �sh metabolite that may be found in waters, (ii) search for an alternative functional monomer, by replacing the traditional MAA by acrylamido-2-methyl-1-propanesulfonic acid (AMPSA), (iii) and open new horizons on the use of these materials, by applying these in biosensing development, something never done before for LMG nor MG. e sensory materials were thus synthesized with methacrylic acid (MAA) and acrylamido-2-methyl-1-propanesulfonic acid (AMPSA) functional monomers, cross-linked by ethylene glycol dimethacrylic acid (EGDMA) within the template molecule.e selective membranes were obtained by dispersing these materials in a PVC matrix, plasticized with onitrophenyl octyl ether (oNPOE).e potentiometric sensors were evaluated in steady-state and �owing media and applied to the analysis of contaminated water.[51] for batch and �ow mode evaluations, respectively.Both devices had no internal reference solution and epoxy graphite was used as solid contact.

Experimental
e Flow Injection Analysis (FIA) assembly had a Gilson Minipuls 2 peristaltic pump, �tted with PVC tubing (0.80, 1.60 and/or 2.00 mm i.d.), and a four-way Rheodyne 5020 injection valve holding a loop of variable volume.e several components were gathered by PTFE tubing (Omni�t, Te�on, 0.8 mm i.d.), Gilson end-�ttings, and connectors.e support devices for tubular and reference electrodes and the con�uence point accessory were constructed in Perspex.Aer reaching a stable baseline, the emf was recorded continuously by means of a homemade high-impedance data acquisition eight-channel box connected to a PC through the interface ADC 16 (Pico Tech., UK) and PicoLog for windows (version 5.07) soware.
When necessary, the pH was measured by a Crison CWL/S7 combined glass electrode connected to a decimilivoltammeter Crison, pH meter, GLP 22.A double-beam UV-vis spectrophotometer (Shimadzu model 1601) equipped with 10-mm quartz cells was used for absorbance measurements in binding studies.Infrared spectra were collected by a Nicolet 6700 FTIR spectrometer; surface measurements were carried out in ATR (attenuated total re�ectance) mode with the Nicolet ATR sampling accessory of diamond contact crystal.

Synthesis of Host-Tailored Polymers.
For molecularly imprinted polymers, the template (LMG, 1.0 mmol) was placed in a glass tube (14.0 mm i.d) with the functional monomer (4.0 mmol MAA), the cross-linker (EGDMA, 20.0 mmol), and the radical initiator (BPO, 0.24 mmol), all dissolved in 8 mL chloroform.A polymer of AMPSA was prepared similarly.e mixture was sonicated, degassed with nitrogen for 5 min, and cured at 70 ∘ C for 30 min.Nonimprinted polymers (NIP) were prepared in a similar way, by excluding the template from the procedure.e resulting polymers were ground and sieved to particle sizes ranging from 50 to 150 m.Extraction of the template molecule and washout of nonreacted species was carried -axis was wavenumber, ranging from 525 to 4000 cm −1 ; and -axis was % transmittance.e resolution was 4000.
2.5.Binding Experiments.e binding studies were carried out aer suitable washing of the synthesized particles.is was con�rmed by measuring the absorbance of the washout solution against blank at 253 nm; the particles were repeatedly washed with methanol until LMG was no longer detected, meaning that its concentration was below the limit of detection of the spectrophotometric readings (<0.21 g mL −1 ).e polymer was let dry at ambient temperature before use.
Binding experiments were carried out by placing 20.0 mg of MIP particles (washed and dry) in contact with 10.0 mL of LMG standard solutions of varying concentrations, ranging from 0.12 to 4.0 mmol L −1 .e mixtures were oscillated for 12 h at room temperature and the solid phase was separated by centrifugation (3000 rpm, 10 min).e concentration of free LMG in the supernatant was calculated by UV spectrophotometry at  max = 253 nm.e increase in absorbance was directly proportional to the concentration of LMG in the supernatant, obeying Beer's law from 25 to 530 g mL −1 .
e amount of LMG bound to the polymer was calculated by subtracting the concentration of LMG in solution (free LMG) from the initial concentration; the free LMG concentration was obtained by means of the previously indicated calibration data.e obtained data was used for Scatchard analysis.
2.6.Sensory Membranes.e sensing membranes were prepared by mixing 200 mg of PVC, 350 mg of plasticizer oNPOE, and 15 mg of the sensing polymer (Table 1), corresponding to a mass composition of 35.4,61.9 and 2.7%, respectively.An amount of 7 mg of TpClPB, acting as anionic additive was added to some of these membranes; in this case, the corresponding mass composition was 35.0, 61.2, 2.6, and 1.2%, respectively.e mixture was stirred until the PVC was well moistened and dispersed in 3.0 mL THF.ese membranes were placed in conductive supports of conventional or tubular shapes [51].Membranes were let dry for 24 h and conditioned aerwards in a 1 × 10 −3 mol L −1 LMG solution.e electrodes were also kept in this solution when not in use.e dry membranes were about 1.0 (±0.1) mm thick and got thicker aer aqueous conditioning.

Potentiometric Procedures.
All potentiometric measurements were carried out at room temperature.e emf values were always obtained in solutions of �xed pH/ionic strength.Increasing concentration levels of LMG was obtained by transferring 0.0200-10.0mL aliquots of 1.0 × 10 −3 mol L −1 LMG aqueous solutions to a 100 mL beaker containing 50.0 mL of 5.0 × 10 −2 mol L −1 of suitable buffer.Potential readings were recorded aer stabilization to ±0.2 mV.e calibration graphs plotted emf as a function of logarithm LMG concentration and were used for subsequent determination of unknown LMG concentrations.

Determination of LMG in River Water
Samples.e water samples were collected from different points along Douro River.e collected samples had no antibiotics and were subsequently spiked with two solutions of different concentrations in LMG, in order to produce a �nal concentration in LMG of 50 and 100 g mL −1 .A volume of 5.0 mL of these solutions was further diluted in 25 mL of 0.05 mol L −1 phosphate solution of pH 4.5.e direct potential method was applied to determine LMG in these samples.

Results and Discussions
e ionophore or the ion carrier is the most vital component in a polymeric membrane sensor for potentiometric transduction, in terms of selectivity and sensitivity [52].Ion exchangers and neutral macrocyclic compounds have been employed for this purpose over the past decades.Until now, some reports found in the literature describe the use of MIP as potentiometric sensing materials.[41][42][43][44][45][46][47][48][49][50].
Noncovalent interactions between the MIP active sites and the template have been considered for allowing fast and reversible binding [52].AMPSA and MAA were used as monomers as they allow electrostatic interactions mostly by hydrogen bonds with the template compound.Ion-pair interactions may also be achieved with AMPSA.

Binding Characteristic of the MIP.
In liquid phase applications of MIPs, a molecule in solution interacts with binding sites in a solid adsorbent.e free ligand concentration in the liquid phase aer equilibrium and reached is constant and is used to plot the corresponding adsorption isotherm.Adsorption isotherms plot the equilibrium concentrations of bound versus free ligand.e bound ligand is expressed in terms of binding capacity and was calculated according to the following equation: where  is the binding capacity of MIPs (mol g −1 ),   the initial concentration of LMG (mol mL −1 ),   the �nal concentration of LMG (mol mL −1 ),   the volume of the test solution (mL), and MIP the mass of dry polymer (mg).
us, the synthesized particles were let stand with a wide range of LMG concentrations, for several hours and under continuous stirring.When equilibrium was reached, the free LMG concentration was calculated by UV spectrophotometry.e corresponding binding capacities were then calculated and plotted against the initial LMG concentration (Figure 1(a)).In general, the adsorption data showed that the binding capacity of the imprinted polymers increased with the increasing of the initial concentration of LMG and had a tendency to saturation for higher analyte concentrations.e above experimental data was used to carry out the Scatchard analysis and estimate the binding parameters of MIP particles.e Scatchard equation was applied for this purpose, where  was the binding capacity;  free the free analyte concentration in equilibrium (mol L −1 );  max was the maximum apparent binding capacity;   was the dissociation constant of the binding site.
Plotting  free versus , a liner function was expected,   where   was the slope and the y-intercept the apparent maximum number of binding sites.e Scatchard plot for MIP particles made with AMPSA (Figure 1(B1)) showed a concave curve that is indicative of the presence of multiple classes of binding sites with different  values.Its x-asymptote behavior may also indicate the presence of nonspeci�c binding [53].Overall, this concave curve suggests that the interaction of the ligand is governed by a positive cooperativity phenomenon between strong and weak binding sites [54][55][56].On the contrary, the Scatchard plot of MAA MIPs displayed a linear behavior meaning that the binding sites were uniform (Figure 1(B2)) and most probably based on hydrogen-bonding interactions.e   and  max of the binding sites were equal to 1283 mol L −1 and 268 mol g −1 for dry polymer.

FTIR Data.
FTIR spectra were taken by direct surface analysis of the materials, placing the MIP particles over a diamond crystal support of the ATR accessory.e resulting spectra are indicated in Figure 2. Similar spectra were observed for NIP materials because they bear the same chemical functions.e two or one peaks at about 3000-2950 cm −1 were assignable to absorption bands from sp 3 C-H stretching in -CH 3 and -CH 2 groups of atoms.e absence of similar bands close to these and just above 3000 cm −1 suggested that most monomers were suitably polymerized because no sp 2 C-H stretching was found.Absorption peaks near 1450 cm −1 were also attributed to sp 3 C-H bonds but resulted from bending the -CH 3 or -CH 2 groups.e amide function in AMPSA MIPs, the carboxylic acid function in MAA MIPs, and the ester function in the crosslinker have in common a carbonyl group, making it difficult to distinguish the nature of the carbonyl involved in this strong and typical adsorption band.In general, its stretch and bend may be identi�ed in several points of the spectra, lying within 1610-1720 cm −1 (Figure 2).Regarding AMPSA polymers, the amide function is clearly identi�ed by the broad and strong adsorption band centered in 3287 cm −1 , corresponding to the stretch of the N-H bond.e sulfonic acid group has a typical strong adsorption at about 1350 cm −1 , also present in the AMPSA MIP spectrum.Regarding MAA polymers, the carboxylic acid is identi�able because of the strong and sharp adsorption band of the carbonyl group.
In general, the observed spectra con�rmed the expected chemical functions, suggesting a high degree of polymerization and con�rming the adequate removal of nonreacted monomers as well as template.e washout from the template was also con�rmed by analyzing the MIP materials a�er along the several washing steps.

Sensor Performance. LMG sensors contained either
MIP or NIP particles as electroactive materials dispersed in plasticized PVC.eir main analytical features were obtained under batch conditions and followed the IUPAC recommendations [57].e results obtained are shown in Table 2. LMG sensors based in MIP particles displayed different behavior in terms of sensitivity and detection limit.e sensors prepared with MAA and AMPSA showed, respectively, linear responses starting at 5.0 × 10 −5 and T 2: Main analytical features of the LMG potentiometric sensors in 5.0 × 10 −2 mol L −1 phosphate buffer, pH 4.5.1.6 × 10 −5 mol L −1 LMG, cationic slopes of 50.7 and 45.8 mV decade −1 , and detection limits of 9.9 and 4.0 g mL −1 .e corresponding NIP particles displayed a linear response aer 8.9 × 10 −5 mol L −1 and 3.0 × 10 −5 LMG, cationic slopes of 55.7 and 67.7 mV decade −1 , and detection limits of 5.1 and 5.0 g mL −1 , respectively.In general, MAA MIP sensors were the only ones presenting a near-Nernst behavior while NIP AMPSA displayed a clear super-Nernstian performance (Figure 3).Overall, the MIP sensors displayed wider linear ranges than the NIP-based sensors but this feature was coupled to the lowest slopes.is behavior may be attributed to the presence of template and unreacted monomers on the MIP materials, hindering the performance of the electrodes.

ISE
e sensors based in MIP particles were subsequently washed with methanol/acetic acid.Aer this procedure, they showed linear responses starting at 5.0 × 10 −5 mol L −1 LMG, for MAA and AMPSA, respectively, cationic slopes of 81.2 and 77.8 mV decade −1 , and detection limits of 11.9 and 12.9 g mL −1 .In general terms, super-Nernstian behavior was observed for both MIP sensors made with MAA and AMPSA monomers.When compared to the corresponding sensors without washing, a signi�cant improve in terms of slope was observed for both sensors.It seems that the existence of template inside the membrane hinders the potential changes within the membrane/solution interface.is behavior could not be attributed to the mediating solvent, because a blank membrane (without ionophore) did not respond to LMG (see Table 2).
To improve the operating features of the previous membranes, the MIP-based sensors were added of an anionic lipophilic compound.Typically, the addition of ionic compounds of lipophilic nature to potentiometric sensors reduces the anionic interference and lowers the electrical resistance of the membranes [58,59].In this work, TpClPB was selected for this purpose (Table 1).Sensors based in MIP/MAA and MIP/AMPSA showed linear response ranges within 8.9 × 10 −5 and 1.6 × 10 −5 mol L −1 , 14.2 and 4.3 g mL −1 detection limits, and near-Nernstian responses of 78.3 and 67.9 mV decade −1 , respectively.When compared to the corresponding sensors without additive, a signi�cant improve in terms of slope was observed for both sensors (see Figure 3).

Response Time and
Lifetime.e time required to achieve a steady potential response (±3 mV) using the proposed sensors in 10 −6 to 10 −4 mol L −1 LMG solutions with a rapid 10-fold increase in concentration was <15 s.Replicate calibrations for each sensor indicated low potential dri, long-term stability, and negligible change in the response of the sensors.e sensors were stored and conditioned in 1 × 10 −3 mol L −1 LMG solution of pH 4.5.With all sensors examined, the detection limits, response times, linear ranges, and calibration slopes were kept within ±3% of their original values over a period of at least 7 weeks.

Potential Stability and
Lifetime/Reusability. e LMG sensors were stable for 1 month and were reused almost in a daily basis.e slope dried less than 5 mV decade −1 over this period and the linear range was unaltered.Along this period (excluding weekends), all electrodes were conditioned independently and washed with water before calibration.
e intermediate precision of the sensor was assessed by comparing three independent potentiometric measures of a speci�c concentration within calibrations made in different days.For a concentration of 1.4 × 10 −4 mol L −1 in LMG solution, the average potential dri was of ±3.5 mV, thus con�rming the good precision of the device.e intraassay precision was accessed similarly by checking three consecutive calibrations and showed potential variations of ±6 mV.Overall, these results suggested a high stability of the potentiometric device.
3.6.Effect of pH.LMG has two ionizable amine groups, as may be seen in Figure 4, with a pK1 of 6.9 [60].is feature points out the need to evaluate and control the effect of pH over the potentiometric response.is was studied for a test solution of 5.0 × 10 −4 mol L −1 of LMG, in which pH was altered by the addition of small aliquots of concentrated hydrochloric acid or saturated sodium hydroxide solution.
e Reilley diagrams were obtained by plotting the emf of this solutions against its pH (Figure 4).e results indicated that the electrode did not respond to the pH change from 2.0 to 5.0.Lower pHs rendered an increase of the analytical signal.Since the LMG selective electrode responds to a cationic species, this increase in potential was correlated to an interference of H + .Above pH 5.0 potentials started decreasing.is behavior was attributed to the formation of the free LMG base in the solution, leading to a decrease in the concentration of LMG ion.is was con�rmed by a perceptible precipitation occurring at higher pH values.A pH of 4.5 was selected for further studies, because the solubility and ionization of LMG were both promoted under this condition.

3.7.
Sensor Selectivity.e selectivity of the chemical sensors was evaluated by calculating potentiometric selectivity coefficients ( Pot ) by the separate solution method (SSM) [61].e obtained values of log  Pot were plotted in Figure 5 and indicated the degree of preferential interaction for LMG over different organic and inorganic species that are common in biological and food samples.e former group includes several antibiotics such as oxycycline (Oxy + ), doxycycline (Doxy + ), enro�oxacin (ENR) and trimethoprim (TMP), and creatinine (Crea + ), chloramphenicol (CHML), and dopamide (DPM).e sensors with AMPSA materials showed a similar selectivity pattern.e log  Pot relative order for MIP AMPSA was LMG ≫ DOXY > OXY > ENR > TMP > CREA > DPM > CHML.e corresponding washed and NIP sensors displayed similar patterns, excluding the higher interference from DPM and CHML and the lower one from DOXY.e additive on the selective membrane did not alter the general relative order of selectivity but the absolute values of log  Pot were much higher.us, the additive deteriorated the selectivity of the MIP AMPSA-based sensor.
For the potentiometric sensors with MIP MAA-based particles the relative order of log  Pot was LMG ≫ ENR > DOXY > TMP > OXY > DPM > CHML > TMP > CREA.e corresponding washed and NIP sensors displayed completely different patterns when compared to the previous sequence.e MIP washed sensor showed the relative sequence LMG ≫ ENR ≈ TMP ≈> OXY > DOXY ≫ CREA > DPM ≫ CHML.e NIP sensor had a relative order of selectivity equal to LMG ≫ ENR > DOXY ≫ OXY > TMP > DPM > CHML ≫ CREA.Unlike in AMPSA, the additive on the MAA-based sensors improved the selectivity properties, with the exception of CHML.
In general, all sensors displayed a good selectivity for LMG, with DOXY, OXY, ENR, and TMP species interfering a bit more than DPM, CHML, and CREA.e additive on AMPSA sensors decreased the selectivity properties of the detector, for which its use in routine applications is not recommended.MG was not tested as an interfering species but it is expected to contribute signi�cantly to the total response of the device.So, the analytical response under real conditions may be assumed as a measure of a global amount from LMG and/or MG.

Optimization of Flow Injection
System.Routine analysis procedures require an expedite method.In this case, a continuous mode of operation is of regular selection and may be achieved by means FIA systems.ese are particularly attractive in view of their versatility, simplicity and suitability for large-scale analyses.
e �ow assembly had a double-channel, allowing the on-line adjustment of pH and ionic strength.A �ow cell had a tubular con�guration, as described by �amel and coauthors [51].It had the same graphite/epoxy conductive material as the conventional electrodes.is material �lled the inner gap of a cylindrical tube (1 cm length, made of Perspex) and was machined aer dry to have a 1.0 mm diameter central hole.is inner hole was coated later by the selective membrane.e resulting working electrode was of simple fabrication and allowed full membrane/sample contact, maintaining the general features of conventional con�guration ISEs in terms of homogeneity, thickness, and �xed area.e reference electrode was placed aer the working electrode, while the auxiliary (required to eliminate the noise created by the �owing stream) was placed before.
To take full advantages of this FIA system, �ow rate and injection volume were optimized in terms of sample dilution, sensitivity, sampling rate, reagent consumption, and wastewater generation.e dispersion is amongst the most important FIA parameters to estimate the degree of sample dilution, because it depends on the capability of the electrode to reach maximum response (which relates the response time and the �ow rate of the carrier).ere are several ways to calculate dispersion, one of which compares the signal produced by introducing the sample as carrier (steady-state signal,  0 ) or by injecting it through the loop ().e resulting dispersion () for the potentiometric response is given by log    0 −   , where  is the slope of the �ow-potentiometric system, evaluated exactly under the same experimental conditions.e percentage of steady state is another way to infer about the dilution of the sample with the carrier inside the �ow tubes.It corresponds to the percentage value of  0 .Unlike , it is however independent from the slope of the electrode under the underlying conditions.e sample loop was varied within 100 and 500 L (Figure 6).For each injection volume, a set of LMG standards ranging 1.0 × 10 −4 to 1.0 × 10 −2 mol L −1 and prepared in water was injected into the buffer carrier stream.e sensitivity of the response increased markedly with the injection volume up to 250 L.For higher volumes the slope of the potentiometric response remained almost constant (less than 5% increase).is observation was coupled to decreased sampling rates, sample consumption, and waste generation.
e effect of �ow rate was examined from 4.0 to 12.0 mL min −1 , recording for each condition of the above indicated calibration.No signi�cant changes were observed in terms of slope and dispersion, but the peak width and peak height decreased with increasing �ow rates.is good feature was however coupled to the production of higher amounts of wastewaters.On the other hand, for lower �ow rates the sensor required long time to recover to baseline, lowering the number of sample outputs and broadening the peaks.us, as a compromise, a �ow rate of 9 mL min −1 was selected for further studies.
e sensor of washed MIP/MAA was calibrated under the optimum conditions.is electrode was selected for showing the best compromise between selectivity and sensitivity.is sensor gave slopes of 45.0 mV decade −1 with detection limits of 10.7 g/mL and lower limits of linear range of 1.0 × 10 −5 mol L −1 .e sampling-rate was approximately 170 runs per hour.
3.9.Analytical Application.Considering the detection capabilities of the proposed device, one of its applications could be analytical sample handling prior the chromatographic analysis.Alternatively, it could be used to monitor MG applications in local farming units where its restricted use is allowed.So, the potentiometric analysis was conducted in steady state over river water.e collected samples had no antibiotics and were subsequently doped with LMG to produce contamination levels of 50 and 100 g mL −1 (Table 3).ese solutions were subsequently diluted 5 times with buffer to carry out the potentiometric analysis.e obtained recoveries were of 97.3 (±6.6%) or 106.5 (±0.1%), respectively, suggesting that the analytical results were accurate.e relative errors were −2.6 and +6.5%, respectively, also accounting for the good accuracy of the method.e relative standard deviation was also small, con�rming the good precision of the analytical data.e t-student test con�rmed that there were no signi�cant differences between the means of the added amount and the potentiometric set of results.e calculated t-student was 0.63, below the theoretical one (12.7).

Conclusions
Molecular imprinting technique was employed to produce LMG host-tailored sensors for potentiometric transduction.MAA and/or AMPSA were used as monomers to produce different MIP materials.Both MAA-and AMPSA-based sensors offered good analytical features.ey were capable of discriminating LMG from other amine drugs in aqueous media.e advantages of these sensors include the simplicity in designing, short measurement time, good precision, high accuracy, high analytical throughput, low limit of detection, and good selectivity.e MIP/MAA sensors were successfully applied to the analysis of river water samples, both in steady state and in �owing media.e proposed method is simple, of low cost, precise, accurate, and inexpensive regarding reagent consumption and equipment involved.Wastewaters discharged are of small concern to environment regarding its volume and composition.
e tubular devices are particularly suitable for the routine screening control of LMG in �sh food.ey produce quicker responses for LMG than those provided by microbiological methods and are much less expensive than the chromatographic methods that are used for routine purposes.

F 1 :
Binding isotherm for LMG/MAA and LMG/AAMPSO imprinted polymer  is the amount of LMG bound to 20.0 mg of the corresponding polymer;    ∘ C;  = 10.0 mL; binding time: 20 h.(A2) Scatchard plot to estimate the binding characteristic of LMG imprinted polymer.

F 6 :
Double-channel FIA assembly with phosphate buffer carrier (a) and the effect of loop (b) and �ow rate (c) in the dispersion, percentage steady state, and sampling rate of the reported signals.Samples/standards are prepared in water.T 3: Potentiometric determination of LMG in river water using MAA/MIP-based membrane sensor.SampleAdded (g LMG mL −1 ) Found(a) (g LMG mL −1 ) Recovery (%) Relative standard deviation (%) Relative error (± standard deviation. All potential measurements were made by a Crison pH 2002 decimilivoltammeter (±0.1 mV sensitivity), at room temperature, and under constant stirring, by means of a Crison, micro ST 2038.e output signal in steady state evaluations was redirected to a commutation unit linked to six ways out, enabling the simultaneous reading of six ISEs.e assembly of the potentiometric cell was as follows: conductive graphite | LMG selective membrane | buffered sample solution (phosphate, 5 × 10 −2 mol L −1 , pH 4.5) || electrolyte solution, KCl | AgCl(s) | Ag. e reference electrode was an Orion Ag/AgCl double junction (Orion 90-02-00).e selective electrode was prepared in conventional or tubular con�gurations 2.1.Apparatus.
T 1: Membrane composition of LMG sensors cast in 200 mg of PVC.
C  : coefficient variation if weekly calibration for 2.72 × 10 −4 M;  2 : correlation coefficient; LOD: limit of detection; LLLR: lower limit of linear range; : standard deviation of the analytical measure for a concentration in the middle of the calibration curve, considering 3 consecutive calibrations.