Microfabricated potentiometric sensor based on a carbon nanotube transducer layer for selective Bosentan determination

: In this work, a solid-state electrochemical sensor relying on potentiometric transduction was constructed and optimized to detect Bosentan (BOS) in its pharmaceutical dosage form and human plasma. BOS is useful in pulmonary hypertension management as a nonselective endothelin receptor antagonist. A printed circuit board has been constructed and used as a substrate for microfabricated Cu electrodes. In comparison to a microfabricated con-trol (Cu/ISM) electrode, the sensor potential signal drift was enhanced, and the response time was reduced by using multi-walled carbon nanotubes (MWCNTs) as an ion-to-electron transducer layer. According to IUPAC requirements, the suggested BOS sensors have been electrochemically characterized, and the linear dynamic range is (1.0 × 10 − 8 to 1.0 × 10 − 5 ) M with a limit of detection of 6.28 × 10 − 9 M and 6.12 × 10 − 9 M for MWCNT-based sensor (Cu/CNT-NC/ISM) and control sensor (Cu/ISM), respectively. The described sensors have been used successfully to selectively determine BOS in dosage form and human plasma without any pre-treatment steps.


Introduction
Over 100 million people are thought to be affected by the lethal condition known as pulmonary hypertension (PAH), which has several different etiologies.A mean pulmonary artery pressure (mPAP) of 25 mmHg or higher at rest is used to haemodynamically identify this disease, which can have major complications such as heart failure.Endothelin receptor antagonists, such as Bosentan (BOS), are an important class of drugs that have a role in the management of PAH [1].BOS (4-tert-butyl-N-[6-(2-hydroxyethoxy)-5-(2-methoxyphenoxy)-2-(pyrimidin-2-yl)pyrimidin-4-yl] benzenesulphonamide) (Figure 1), is an oral antagonist of endothelin A and B receptor.High amounts of endothelin, which is a potent blood vessel constrictor, have been found in plasma and lung tissue of individuals with PAH.In 2001, the US Food and Drug Administration granted BOS, as the first orally active drug, license for the treatment of PAH.Through vasodilation, antifibrotic and antithrombotic actions, it relieves symptoms, notably in people with World Health Organization Class III or IV symptoms [2][3][4].Another potential medical use for BOS is that it could also be used, in combination with other approved drugs, in the remediation of COVID-19 due to its antiviral effects [5,6].In order to determine BOS in pharmaceutical dosage forms and biological fluids, many analytical techniques such as chromatography and spectroscopy have been published.However, to the best of our knowledge, no potentiometric assay has been reported for BOS determination [7][8][9][10][11].Potentiometric approach has many advantages over other methods for sample analysis.These include fast response time, broad linear range, low energy and cost, simplicity of preparation, good sensitivity, and selectivity for many samples.In addition, potentiometric approaches do not need pre-treatment steps or the use of organic solvents, which makes them more eco-friendly.Other methods, on the other hand, are not appropriate for large-scale monitoring due to their high cost, complex usage, high energy and time consumption, and sample pre-treatment requirements.Therefore, potentiometric assay is a better option for sample analysis [12].Potentiometric ion sensors are essential members of the family of electrochemical sensors that can electrochemically measure primary ion activity in a variety of sample matrices with minimal sample pre-treatment [13][14][15][16].A new generation of ion-selective electrodes was developed: solid-contact ion-selective electrodes (SC-ISEs), which, among other benefits, were able to prevent internal filling solution leakage that occurred in the traditional liquid ion-selective electrodes [17,18].As a result of the quick development of micro/ nanofabrication processes and upraised understanding of concepts and techniques of electrochemistry, micro/ nanoelectrode array sensing has recently withdrawn a lot of attention, particularly in the field of bioanalysis [19].Recently, it has been verified that nanoparticles are valuable additions that enhance ion-selective electrode performance and minimize electrical resistance [20,21].Using copper microfabricated electrodes to create SC-ISEs has the advantages of both microfabrication and SC-ISEs, such as low cost, miniaturization, simplicity, portability, real-time analysis, energy savings, the ability to be applied on large-scale analysis, robustness, and reliability [13,22,23].This study aimed to develop a potentiometric sensor in order to determine BOS in human plasma and its dosage forms while complying with green analytical chemistry (GAC) through the fabrication of microfabricated copper solid contact-based potentiometric electrodes.The designed sensor response was improved to minimize signal drift and long response time via modification using carbon nanotubes (CNTs) as an ion-to-electron transducer interlayer.Because of CNTs' spacious nominated surface area and premium chemical stability, these CNTbased SC-ISEs have elevated potential stability and helped in improving SC-ISEs performance.To examine the importance of the transducer layer, the water layer test and signal drift were studied.Green Analytical Procedure Index (GAPI), Analytical Eco-Scale Green, and Analytical GREEnness Metric Approach and Software (AGREE) were used to assess the greenness of the suggested approach.

Standard solutions 2.2.1 Stock BOS standard solution
A stock solution of BOS (1.0 × 10 −5 M) was made by weighing 1.5 mg of BOS into a 250 mL volumetric flask, which was then dissolved in an adequate quantity of carbonate buffer (pH = 9.2) before filling up to the final volume using the same buffer.

Working BOS standard solutions
Freshly made solutions of various strengths (1.0 × 10 −6 -1.0 × 10 −10 M) were made by successive dilutions from stock solution using a carbonate buffer with a pH of 9.2.

Microfabricated copper electrode preparation
In this study, a microfabricated copper electrode was used as a solid-contact electrode substrate.The electrodes were created in accordance with published instructions [18,23].Using high-resolution laser printing equipment, CAD software was utilized to create a photomask, which was printed on a transparent sheet.The photomask on top of pre-sensitized photoresist-coated PCBs was subjected to UV light (365 nm) for 90 s, blocking the light away from the needed areas while excluding areas that were not covered.For removal of the exposed portion of the photoresist, 0.25 M NaOH was used (as a developer).The exposed Cu was wet etched at 40°C with 1.1 M NH 4 S 2 O 8 .Acetone was used for stripping the photoresist and revealing the patterned Cu electrodes.The electrodes were first rinsed with water and then dipped in acetic acid to eliminate surface oxides just before applying either CNTs or an ion-selective membrane (ISM).

ISM preparation
The ISM was prepared by blending 190 mg PVC, 0.4 mL DOP, and 10 mg tridodecylmethylammonium chloride in 6.0 mL of THF.

Carbon nanotubes/PVC nanocomposite preparation
The solution-blending method was used to prepare CNTs/ PVC nanocomposite (CNT-NC), as described in the literature [24].About 3.0 mL of THF and 0.2 mL of DOP were used to dissolve 95.0 mg PVC entirely.To obtain a homogeneous CNT-NC dispersion, MWCNT (10 mg) was mixed with the above-mentioned solution in an ultrasonic bath to ensure proper dispersion of nanomaterials in THF.

Solid-state ion-selective electrode fabrication
The microfabricated potentiometric sensor 1 (Cu/CNT-NC/ ISM) was developed by applying 10.0 μL of CNT-NC dispersion onto the Cu microfabricated electrode and leaving it 24 h for solvent evaporation.Then, 10.0 μL of the ion-sensing cocktail was applied to the copper-modified electrode as presented in Scheme 1 and allowed to dry overnight.As a preconditioning phase, the sensor was immersed in a 1.0 × 10 −5 M BOS solution for 1 day at 25°C before being used for the first time.CNT-NC free sensor as a control sensor (sensor 2, Cu/ISM) was prepared to examine the CNT-NC effect on potential drift and response time.

Potentiometric measurements
Potentiometric experiments were performed using a Jenway pH meter 3310 Orion, reference electrode (Ag/AgCl, double junction) model.Bandelin Sonorox, Rx 510 S, and magnetic stirrer (Budapest, Hungary) as well as Jenway pH glass electrode (UK) were used for pH adjustment.Each sensor was individually coupled with an Ag/AgCl double junction reference electrode, calibrated by being submerged in BOS drug solutions (1.0 × 10 −5 -1.0 × 10 −10 M).Solutions were allowed to equilibrate till the potentiometer's readings were steady.The observed emf from the two suggested sensors was shown on calibration graphs in relation tolog [BOS] concentrations.Regression equations were computed for the suggested sensors.The sensors' performance was validated in conformity with IUPAC recommendations [25].

Sensor selectivity
A matched potential approach was used to evaluate the potentiometric selectivity factor.The change in potential  Potentiometric sensor for Bosnian determination  3

Water layer test
The water layer test was developed by the Pretsch group [27][28][29].It is used to determine whether a layer of water exists between the ion-to-electron transducer layer and the ISM.Potentiometric readings were acquired for 1.0 × 10 −6 M BOS for 1 h; then, a more concentrated interfering solution (1.0 × 10 −5 M AMB) was placed, and the potential was measured for another 1 h and returned to 1.0 × 10 −6 M BOS for the third hour.

Analysis of BOS in pharmaceutical tablet formulation
Ten BOS tablets were precisely weighed, the tablet mean weight was determined, and the tablets were then pulverized.Of this powder, a precise quantity was weighed out, equal to 1.5 mg of BOS, then put into a 250 mL volumetric flask and diluted to the appropriate concentration with carbonate buffer, pH 9.2.The resultant solution was claimed to have a concentration of 1.0 × 10 −5 M, then another two concentrations were prepared (1.0 × 10 −6 , 1.0 × 10 −7 M).Following Section 2.3.5, the measurements were carried out, and the recovery percentages were determined using the corresponding equations.

Analysis of BOS in spiked plasma
One millilitre of human plasma was put into a set of three volumetric flasks (10 mL); then, 1.0 mL, 550 µL, and 100 µL of BOS stock solution were added and completed to the final volume with carbonate buffer (pH 9.2) to prepare 1.0 × 10 −6 , 5.5 × 10 −7 , and 1.0 × 10 −7 M. Following Section 2.3.5, the measurements were carried out, and the recovery percentages were determined using the corresponding equations.

Results and discussion
For analysts, building SC-ISEs with consistent and repeatable potentials has always been a big challenge.The aim of this study was the fabrication of portable, affordable, miniature-sized solid-contact sensors with highly sensitive, reproducible, and consistent readings.

Performance of the investigated sensors
Since BOS has a sulphonamide group (pk a = 5.8) that carries a negative charge in a basic medium, this study was performed at an alkaline pH (9.2) [30].The necessity of an anion exchange mechanism is required for BOS ion-selective electrode membranes; therefore, tridodecyl methyl ammonium chloride was used to introduce a lipophilic anionic exchanger, where the exchangeable counter ion (Cl − ) was initially replaced with BOS by conditioning the membrane for 1 day in 1.0 × 10 −5 M BOS [31,32].PVC was used as a polymer matrix in the fabrication of the ionsensing cocktail.In this study, MWCNT-based sensor performance (sensor 1, Cu/CNT-NC/ISM) was evaluated compared to the control CNT-NC free sensor (sensor 2, Cu/ ISM).According to IUPAC recommendations, the performance characteristics of the suggested sensors were assessed [25] and are listed in Table 1.The calibration plots' slopes for sensors 1 and 2 exhibited near-Nernstian response and were found to be 61.24mV/concentration decades and 58.49mV/concentration decades, respectively.The suggested sensors covered a dynamic linear concentration range of 1.0 × 10 −5 -1.0 × 10 −8 M, as shown in Figure 2. From the point where the two curves' linear extrapolated portions meet, the limit of detection (LOD) in accordance with IUPAC recommendations was determined and found to be 6.28 × 10 −9 M for sensor 1 and 6.12 × 10 −9 M for sensor 2 [25].Short dynamic response time is a key parameter for increasing the number of samples that can be quickly analysed and hence contributes to the analytical application of the developed sensors [33].For concentrations higher than 1.0 × 10 −7  M, sensor 1 responds quickly in 31 s, while sensor 2 takes 1.37 min.Meanwhile, a comparatively longer time is needed for more diluted concentrations to achieve a stable potential that is 56 s for Cu/CNT-NC/ISM sensor 1 and 1.50 min for Cu/ ISM sensor 2. As expected, the lowest concentrations give the longest response times due to the longer time needed to attain equilibration.The shorter response time of sensor 1 is attributed to the presence of MWCNTs as a transducer layer, whereas the transduction behaviour of MWCNTs can be attributed to a high double-layer capacitance in the CNT's ion selective sensor.In contrast with conducting polymers, where the transduction mechanism is based on ionexchange mechanisms and redox reactions, MWCNTs are more robust against any redox side reactions that might occur in the electrode [34].The transducer layer (CNT) has been previously characterized using both Raman spectroscopy and SEM imaging techniques [35].

Monitoring of signal drift and water layer test
The sensor's response time and signal drift can both be affected by a water layer that might be formed between the electrode substrate and the ISM; hence, the water layer test was performed to detect its presence [27][28][29].The test relies on spotting potential drift when switching between a primary ion solution (BOS; 1.0 × 10 −6 M) and a potent interfering ion solution (AMB; 1.0 × 10 −5 M), then switching back to the primary ion solution (BOS; 1.0 × 10 −6 M).Although the control Cu/ISM (sensor 2) showed good potentiometric properties, it had a long response time (1.37 min) and observable drift (25 mV•h −1 ), which is attributed to the formation of the water layer.In contrast to sensor 2, Cu/CNT-NC/ISM (sensor 1) had a lower response time (31 s) and less drift (1.9 mV•h −1 ), as shown in Figure 3a and b, which is attributed to the hydrophobicity and high capacitance of CNTs that prevent the creation of a water layer between ISM in sensor 1 and the transducer layer.

pH effect on sensor response
To attain the optimum experimental conditions, the impact of pH on the effectiveness of the introduced sensors was evaluated.There was no noticeable change in the sensors' response in the pH range of 7.5-10.5,which is due to the presence of the sulphonamide group that has a negative charge in the basic medium.Below pH = 7.5, the drug began to precipitate, so the emf is no longer stable, as shown in Figure 4a and b.

Determination of sensors' selectivity coefficients
The selectivity of an ion-selective electrode is one of its most crucial features.It is frequently used to judge whether a valid measurement in the sample of interest is possible.Assessment of selectivity coefficients was done utilizing the matched potential method, which completely ignores the Nernst equation in order to circumvent the challenge of accurately determining selectivity coefficients in the presence of unequally charged ions [26].Selectivity coefficients were assessed for structurally linked compounds such as AMB, co-administered drugs such as losartan and paracetamol, and other interfering ions such as chloride and citrate.Table 2 provides values for selectivity coefficients.
The results revealed that both electrodes showed good selectivity towards BOS against other studied interferents.Sensor 1 (Cu/CNT-NC/ISM) showed relatively higher selectivity for BOS in the presence of ambrisentan, losartan, citrate, and chloride ions than sensor 2 (Cu/ISM).In the presence of losartan, sensor 1 showed the highest selectivity for BOS, while the lowest selectivity was observed in the presence of paracetamol.As K MPM BOS/interferent value decreases, it means that the sensor is more selective to the drug under study (BOS) [26].

Analysis of BOS in pharmaceutical tablet formulation
BOS in its pharmaceutical preparation (pulmivort tablets ® ) was directly measured without pre-treatment or extraction steps using the investigated sensors.As shown in Table 3, the suggested sensors were able to get precise and accurate recoveries.For the quantification of BOS in the formulation of pharmaceutical tablets, sensor 1 (Cu/CNT-NC/ISM) displayed the lowest SD values.

Analysis of BOS in spiked human plasma
C max of BOS reaches a concentration of 1.12 × 10 −6 M, after receiving a single 62.5 mg dose in healthy male volunteers on the first day [36].This concentration is within the sensors' linear range.Thus, the proposed SC-ISE's ability to detect BOS in biological fluids was assessed using spiked human plasma directly with no preliminary treatment or extraction.The suggested sensors were successful in obtaining precise and accurate recoveries based on the data in Table 4.

Greenness evaluation of the developed approach
The greenness of the developed approach was assessed using AGREE greenness assessment tools, GAPI, along with Analytical Eco-scale.

Analytical eco-scale
The analytical eco-scale [37] is calculated by deducting penalty points from a base of 100 for each analytical method component.In conformity with its standards, the method that is ideally green scores an eco-scale of 100, the excellent green method scores an eco-scale of more than 75, and the acceptable green method scores an eco-scale of more than 50 [38].When the method yields an eco-scale score of less than 50, it is believed to be an inadequately green analytical method.The calculated penalty points for the suggested assay, revealing that the suggested approach has an excellent green Eco-Scale score, are illustrated in Table 5.

GAPI
The main benefit of GAPI is that it evaluates the greenness from sample collection till determination of its concentration,   giving a comprehensive perspective of the suggested strategy [39].Five pentagrams, each having a three-colour scale.Green indicates "low environmental impact," whereas yellow denotes "medium environmental impact," while red indicates "high environmental impact."Since there was no sample preparation step in this assay, its pentagram was eliminated.GAPI pentagram has green dominance, indicating that the procedure is of low risk, with only one yellow coloured part due to using THF during manufacture, and only one red coloured fraction denoting the total NFPA score of all the solvents.The GAPI pentagram is illustrated in Figure 5.

AGREE
Out of all the greenness evaluation techniques, only the AGREE approach uses all 12 GAC principles and yields an easily interpretable and informative result [40].The analytical method is believed to be green for drug analysis if the AGREE analytical score are greater than 0.75.Additionally, a score of 0.50 indicates that the method is acceptable for drug analysis.Scores less than 0.50 show that the proposed analytical process is unacceptable.The proposed GAC electrodes in this work received an AGREE score of 0.78, as shown in Figure 6, indicating that the approach is extremely green and may be used safely in routine analysis.

Statistical analysis
The variance ratio F test and Student's t-test were used to examine the validity of the proposed approach.A statistical comparison of the suggested sensors and the reported method [41] for BOS is presented in Table 6.The results show that there is no statistically significant difference between the reported method and the proposed sensors; the calculated t and F values were less than the theoretical ones at p = 0.05.

Comparison of other analytical techniques for analysis of BOS
Various techniques, including spectrophotometry, voltammetry, LC-MS, and HPLC, have been published for determining BOS and are summarized in Table 7.All these approaches, however, lack the mobility and affordability of the suggested potentiometric approach in addition to being slower and less straightforward compared to potentiometric analysis.

Conclusion
In this study, the described sensors were applied for the direct analysis of BOS in pharmaceutical formulation and human blood plasma.They showed many advantages over the traditional ion-selective electrodes in terms of stability and durability (30 and 23 days), as well as several advantages over other complicated methods such as HPLC, such as being eco-friendly, portable, a real-time analysis approach, easy to miniaturize, of low cost and energy saving.The proposed sensors demonstrated good, reproducible, accurate (a mean of 101.47 and 101.26 with SD < 2.0), precise (%RSD < 0.5), and selective potentiometric results in the presence of different interferents.They can be applicable for routine analysis without tedious pre-treatment procedures, due to their manufacturing simplicity and low cost.The addition of CNTs as an ion to electron transducer resulted in a shorter response time (31 s), lower signal drift, and almost no water layer is formed between the polymeric and transducer layers, as opposed to a blank ion-selective electrode free of ion to electron transducer, which had a longer response time (1.37 min) and a higher signal drift.
Acknowledgement: Sincere gratitude and appreciation are extended by the authors to EVA Pharma and AUG Pharma (Cairo, Egypt) for supplying gift samples of pure BOS and Ambrisentan.The authors would also like to thank Ain-Shams University's Department of Analytical Chemistry in the Faculty of Pharmacy for providing the tools and facilities needed to complete the research.
Funding information: Authors state no funding involved.
(E) related to increased activity of the analyte from a A = 1.0 × 10 −8 M (reference solution) to a′ A = 1.0 × 10 −5 M is recorded.Then, an interfering ion solution a B with a concentration range of 5.0 × 10 −8 -1.0 × 10 −5 M is mixed with another 1.0 × 10 −8 M (reference solution) until the same potential change (ΔE) is observed [26].The next equation was used to determine the selectivity coefficient K for each interferent .7 pH effect on sensor performance Over a pH range of 3.0-11.0regulated by adding 0.1 N HCl and 0.1 N NaOH, the impact of pH on the response of the studied sensors was observed on 1.0 × 10 −6 M and 1.0 × 10 −7 M BOS solutions.At each pH level, potential values were acquired and recorded.

Scheme 1 :
Scheme 1: A schematic illustration of a microfabricated solid-contact sensor 1.

− 9 a
Average of three determinations.b Intra-day precision (n = 9), an average of three different concentrations repeated three times within the same day.c Inter-day precision (n = 9), an average of three different concentrations repeated on three successive days.d Limit of detection (according to the IUPAC definition, measured by the intersection of the extrapolated arms of non-responsive and the Nernstian segments of the calibration plot).

Figure 5 :
Figure 5: GAPI green profile assessment of the proposed method.

Figure 6 :
Figure 6: Results of AGREE green profile assessment software.

Table 7 :
Comparison of other published analytical techniques for analysis of BOS

Table 1 :
Performance characteristics of the investigated sensors for BOS determination

Table 3 :
Determination of BOS in pharmaceutical dosage form a Average of three determinations.Potentiometric sensor for Bosnian determination  7

Table 4 :
Determination of BOS in spiked plasma a Average of three determinations.

Table 5 :
Penalty points for the greenness assessment of the proposed potentiometric approach

Table 6 :
Statistical comparison between the proposed assay and the reported reference method for the determination of BOS