Electrochemical determination of theophylline using a nickel ferrite/activated carbon-modified electrode

In this work, a nanocomposite based on nickel ferrite/activated carbon (NiF/AC) was used to modify a highly sensitive electrochemical sensor for the quantification of theophylline (TPL) in pharmaceutical tablets. The synthesized materials were characterized using x-ray diffraction, Fourier transform infrared spectroscopy, scanning electron microscopy, energy dispersive x-ray spectroscopy-elemental mapping and surface area analysis via the Brunauer–Emmett–Teller method. Cyclic voltammetry was employed to study the electrocatalytic properties of the NiF/AC-GCE toward the oxidation of TPL. The dependence of the electrochemical response on the scan rate and pH was also investigated, and the working parameters were optimized. The linear range of the established electrochemical biosensor was from 0.5 to 5 μM (R2 = 0.997), with a detection limit of 0.21 μM. The present method was tested using three pharmaceutical formulation standard samples with good accuracy and acceptable recovery. Thus, it is a promising candidate for the determination of TPL in pharmaceutical formulations.


Introduction
Theophylline (1, 3-dimethylxanthine, TPL), a xanthine derivative, is naturally found in beverages such as coffee, cocoa beans, coffee or tea [1].This alkaloid has been utilized extensively as a bronchodilator for many years to treat chronic obstructive pulmonary disease, asthma, emphysema and infant apnea [2].The acceptable therapeutic dosage of TPL is in the range of 5-20 mg.ml −1 .Concentrations greater than 20 mg.ml −1 are considered to be toxic and can cause adverse side effects such as vomiting and, in severe cases, seizures, cardiac arrhythmias, and death [3,4].Therefore, a simple, rapid, highly sensitive and accurate analytical technique for the quantitative determination of TPL has received great attention because of the critical function of drug monitoring for public health.
Among the available methods for TPL determination such as an electro membrane extraction combined with high-performance liquid chromatography-ultraviolet [5], liquid chromatography-mass spectrometry [6], liquid chromatography-tandem mass spectrometry [7], chemical sensor [8], electrochemical approaches have the potential to detect drugs in body fluids and pharmaceutical formulations due to their high selectivity and sensitivity, in situ measurement, simple sample preparation procedures, low cost and simple operation compared with other approaches [9,10].Various electrochemical sensors modified with hybrid nanocomposite materials have been employed for the electrochemical sensing of electroactive compounds of interest, such as metal ions [11] and pharmaceutical compounds, owing to their novel functional properties [12,13].For instance, graphene oxide (rGO), poly (3,4-ethylenedioxythiophene)-poly(styrenesulfonate), and poly (glutamine)/carbon nanotubes [14,15] have been used as electrode modifiers to quantitatively detect the antihistamines piroxicam and nimesulide in river water, pharmaceutical formations, and human urine samples.For electrochemical analysis, the critical component is the electrode modifier, which demands the selection of relevant materials to enhance the analytical performance.
Activated carbon usually possesses a high surface area and chemical stability.It is an ideal carrier for several applications, such as adsorption [16], catalysis [17], and electrochemistry [18].In particular, its application in the field of electrochemistry has attracted considerable attention.Ayadi et al reported electrochemical oxidation using activated carbon-ZnO for the degradation of phenolic compounds [19].SnO 2 , CuO, and Fe 2 O 3 /activated carbon nanocomposites were applied in electrochemical hydrogen storage [20].Pereira et al reported the electrochemical detection of promethazine hydrochloride using activated carbon/silver nanoparticle-modified carbon paste electrodes [21].In parallel, magnetic spinel ferrite nanoparticles (MFe 2 O 4 , M: Ni, Co, etc) exhibit excellent electrochemical activity toward some organic compounds.NiFe 2 O 4 nanoparticle-modified electrochemical sensors were reported for the simultaneous determination of folic acid and paracetamol [22].A CoFe 2 O 4 nanoparticle-modified carbon paste electrode was used for determining oxycodone and codeine [23].However, magnetic nanoparticles are heavily agglomerated due to their magnetic nature, which limits their application as electrode modifiers.Combining two types of activated carbons and magnetic spinel ferrite nanoparticles is expected to overcome these disadvantages and promote the advantages of each individual material.
In the present work, activated carbon was first prepared from rice husks by activation with potassium hydroxide.NiFe 2 O 4 /activated carbons were synthesized by a hydrothermal process.The electrochemical oxidation signal of TPL significantly increased due to the excellent adsorption of NiFe 2 O 4 /activated carbon.The determination conditions were optimized.The proposed method was further used to determine TPL in pharmaceutical preparations.

Preparation of nickel ferrite/activated carbon (NiF/AC)
Activated carbons (ACs) were fabricated from rice husks according to previous work [24].To investigate the mass ratio of NiF/AC, a (mol) of NiCl 2 •6H 2 O, 2a (mol) of FeCl 3 •6H 2 O and b (g) of AC from rice husk were mixed in a 250 ml beaker containing 130 ml of distilled water.c(g) NaOH was dissolved in 20 ml of distilled water, and then, the NaOH solution was added dropwise to the container containing the above mixture while stirring continuously for approximately 30 min.The amount of NaOH used is sufficient to react with (mol) Ni 2+ and 2a (mol) Fe 3+ .The mixture from the beaker was placed into a Teflon-lined hydrothermal autoclave heated at 180 °C for 10 h and then allowed to cool to room temperature.The experimental composition is given in table 1. Sodium hydroxide was added dropwise to the mixture salts, followed by a hydrothermal process.The ionic reaction equations can be described as follows:

+ 
Nickel ferrite (NiF) was also prepared via a similar procedure without AC for comparison.

Characterization
X-ray diffraction (XRD) patterns were obtained by a Bruker D8 Advance x-ray diffractometer (Bruker, Germany).The morphology of the materials was determined by scanning electron microscopy (SEM) using a Nova NanoSEM 450 scanning electron microscope (USA).Elemental analyses were performed through EDX mapping spectroscopy using a TEAM Apollo XL EDS system (USA).Fourier transform infrared (FT-IR) spectra of the samples were recorded with an IR Affinity-1S spectrophotometer (Shimadzu).The nitrogen adsorption/ desorption isotherms were obtained on a Micromeritics Tristar 3000.

Preparation of glassy carbon electrode
Before each modification, the bare GCE surface was thoroughly polished using an alumina slurry on sandpaper and repeatedly cleaned with deionized water.The cleaned GCE was then dried at room temperature.Five milligrams of the obtained material were ultrasonicated in 5 ml of water for approximately three hours to obtain a NiF/AC suspension.Five microliters of the NIF/AC dispersion were drop-cast onto the surface of the cleaned bare GCE and allowed to dry naturally at 40 °C in an oven for 12 h for ready use-preparation of the TPL standard solution and BR buffer solution.TPL standard solution (0.01 M) was prepared by accurately weighing the calculated amount in a flask and diluting it to the calibration mark with DI water.BR buffer was obtained from pH 2 to pH 8 by mixing 0.04 M boric acid, 0.04 M phosphoric acid and 0.04 M acetic acid that had been adjusted to the desired pH with 0.2 M sodium hydroxide.All the prepared solutions were stored in a refrigerator at 4 °C during the course of the study.

Real sample preparation
All the electrochemical experiments were performed using CPAHH5 (Vietnam), which uses a three-electrode cell, namely, an Ag/AgCl reference electrode (Model RE-5, BAS), a platinum wire as an auxiliary electrode, and a 2.8-mm-diameter glassy carbon electrode (GCE) as a working electrode.DPV was used for the quantification of TPL in real samples via the standard addition method.size, calculated using the Scherrer formula, is approximately 56 nm for NiF.The XRD pattern of the AC material presents a large and wide peak in the range of 10-30 degrees because of the enlarged gap between the layers of graphite.The XRD pattern exhibits the shape of typical amorphous carbon and presents broad asymmetric peaks at ∼26 and ∼41 degrees, indicating a highly disordered microcrystalline structure with randomly oriented graphitic microcrystals [25].The XRD patterns of NiF/AC with a mass ratio of NiF to AC are presented in figure 1(b) in which in which no significant peak was observed for AC.It is explained by the fact that AC is mainly in amorphous form and the characteristic peaks of AC at around 26 and 41 degrees are overlapped with the peaks of nickel ferrite.The intensity of the characteristic diffractions of NiF increases with increasing NiF amount.The average crystallite size of NiF (32.7-34.3nm) is smaller than that of pristine NiF (see table 2).This reduction in crystallite size may be attributed to surface tension, which can lead to crystal growth.These results support the formation of the NiF/AC composite.The textural properties of the obtained materials were investigated by the nitrogen isothermal adsorption/ desorption method, and the results are presented in figure 2(a)-(b) and table 2. All the curves are type IV with H1 types, which are characteristic of mesoporous materials.The condensations occur at a high relative pressure (0.35-1), indicating that NiF/AC comprises meso-and microporous structures.The dispersion of NiF into AC causes the specific surface area to decrease monotonically from 1231.9 m 2 .g−1 for AC to approximately 688.5 m 2 .g−1 for NiF/AC.For composites, the surface area tends to decrease with increasing NiF amount due to the accumulation of NiF nanoparticles within the pores of AC.The microarea in activated carbon is 63.3%, while it is only 24.7% in NiF.When mixed with the NiF/AC composite, the mesoarea percentage decreases by approximately 52.5%-56.6%,while the microarea percentage increases from 75.2% to 43.4%-47.5%.Therefore, in the composites, the decrease in surface area is attributed to the decrease in the mesoarea.

The electrochemical response of the modified GCE to TPL
The electrochemical behaviour of the modified GCE on 10 μM TPL was studied using the CV method (figure 5(a)).The electrochemical response of TPL on the bare GCE was very weak.However, the signals improve significantly as the electrode is modified by NiF or AC, especially when combining NiF and AC.A welldefined anodic peak was recorded for the NiF/AC-GCE surface in comparison to that of the bare GCE.The peak current increases with increasing NiF amount in the composite, peaks at 40% NiF ((4/6)NiF-GCE) and  decreases with further increasing NiF (figure 5(b)).Therefore, NiF/AC(4/6)-GCE was selected for further experiments.The advantages of this modified electrode may be attributed to the large surface area of the activated carbons providing abundant active sites for the adsorption of TPL as well as the presence of ionic metals with multiple oxidation states, which act as electrocatalytic sites that promote electron transfer.These results indicate the successful preparation of the working electrode for the sensitive determination of TPL.There was also a distinct increase in the peak current for each step of electrode fabrication, which can be justified via the Randles-Sevick equation (figures S1, 2).

= ´´´´´ẃ
here A is the electroactive surface area of the electrode, I p is the anodic peak current, n is the number of electrons transferred, C 0 is the concentration of TPL, n is the scan rate and D is the diffusion coefficient [27].The electroactive surface area of the (4/6)NiF/AC-GCE was found to be 0.070 mm 2 , whereas those of the bare GCE, NiF-GCE and AC-GCE were 0.043, 0.024 and 0.047 mm 2 , respectively.
Therefore, the modified electrode provides a very large increase in the anodic current, as observed.There was an essential enhancement in the current response, as shown in figure 5.
The effect of pH on the oxidation of TPL was studied by calculating the peak currents and potentials of TPL in 0.1 M BR buffer over the pH range of 2-8 by CV.In the absence of TPL, the GCE did not present any peak under the scanned potential window.With the addition of 40 ml of 0.1 M TPL, an irreversible oxidation peak was recorded in the pH range of 2-8, as shown in figure 6(a).The peak current decreased monotonically with increasing pH (figure 6(b)).At pH 6.0-8.0, the electrochemical signals were very poor, indicating that in alkaline media, TPL oxidation requires more protons.Moreover, the anodic peak potential decreased with increasing pH, indicating that this process may involve proton and electron transfer during the electro-oxidation of TPL.
The linear E p-pH plot is presented in figure 6(c).The linear regression equation was E p = (1.5874± 0.0125) + (−0.0555 ± 0.0023)pH, with R 2 = 0.9894.The slope of 0.0555 is very close to the Nernstian value of 0.059, as an equal number of protons and electrons were transferred.TPL protonation is favorable at lower pH values, as indicated by the broad and stable peaks.However, based on the best response in terms of the peak shape and stability of the peak currents (SD = 0.2871 for pH 2 and 0.06189 for pH 3), pH 3 was selected for further study.
The variations in the anodic peak current and peak potential with changes in the scan rate during the oxidation of 10 μM TPL were studied (figure 7(a)).The oxidation peak currents change linearly in the studied scan rate range, with a high determination coefficient of 0.9602 (figure 7(b)), indicating that the electrochemical reaction was controlled by the adsorption process.The relationship between the peak potential and the natural logarithm scan rate is described by the Lavion equation [28].
The experimental data were regressed following the Lavion equation as follows: =  +  = According to the slope and intercept of the regression equation, the value of nα is 1.05 (figure 7(c)).The acceptable value of α varies from 0.3 to 0.7.For the irresponsible system, α is set to 0.5.Therefore, n is 2.1 or 2. Based on the pH effect mentioned above, the oxidation process involves an equal number of electrons and protons.The proposed mechanism of TPL is based on a two-electron two-proton transfer oxidation reaction.NiFe 2 O 4 possesses an inverse spinel structure in which the cations are distributed as (Fe 3+) Td(Ni 2+ Fe 3+) OhO 4 (Oh and Td: the octahedral and tetrahedral sites, respectively) [29].On the other hands, NiFe 2 O 4 , a p-type conductivity has the hole hopping during the oxidation step [30].In order to retain the charge neutrality of the lattice, Ni 2+ ions have the propensity of constant conversion between trivalent and divalent units in which the final cations are distributed as (Fe 3+ )Td(Ni 2+ Ni 3+ Fe 3+ )OhO 4 [31].The Ni 3+ is responsible for the redox behavior at the modified electrode.From the figure 6(b), the anodic peak current declines with increasing pH indicating that the chemical reactions occur according to the electrochemical mechanism, in which it is assumed that.Ni 2 FeO 4 reveals an electrocatalytic activity as (Fe 3+ )Td(Ni 2+ Ni 3+ Fe 3+ )OhO 4 .Then, the (Fe 3+ )Td(Ni 2+ Ni 3+ Fe 3+ )OhO 4 is directly reduced at the electrode according to reaction as follows: Fe  8(a) and figure S3 a.The E P of TPL does not change when E acc changes from −0.4 V to +1.0 V, with an average value of 1.358 ± 0.003 V (n = 8).On the other hand, the I P of TPL tends to increase from −0.4 V to 0.0 V and then gradually decreases from 0.0 V to 1.0 V.The increase and decrease in I P are not large.This shows that E acc does not affect the accumulation of TPL on the (4/6)NiF/AC-modified electrode surface.However, a matter of concern is that when a solution contains a certain substance whose anodic stripping peak is close to the anodic stripping peak potential of TPL, we also need to pay attention.The appropriate E acc is 0.0 V against 3 M Ag.AgCl.KCl reference electrode.An investigation of the t acc required to obtain information regarding the amount of TPL on the electrode surface needed to achieve a higher oxidation peak current was carried out from 0 s to 30 s in 0.1 M B-R buffer at pH 3 (figure 8(b) and figure S3b).The anodic peak current of TPL at the modified GCE gradually increased with increasing accumulation time and reached a maximum current response at 25 s.With increasing accumulation time, there was almost no change in the current response.This could be attributed to the saturated adsorption of TPL on the electrode surface of the GCE at 25 s.Accordingly, 25 s was chosen for further investigations.As the pulse amplitude increased, the base of the dissolution peak of TPL became wider (figure 3(S) (c).This result will greatly affect the selectivity of the analytical method.Additionally, the E P moves more toward the negative side.This phenomenon can be explained by the fact that during the stripping stage, the potential on the working electrode is more positive than the potential set by the pulse amplitude.A more positive potential will promote the stripping of the accumulated analyte.On the other hand, when the pulse amplitude varies from 0.020 V to 0.080 V, there is a correlation between the I P and pulse amplitude as follows: 1.736 2.751 90.33 50.22 E V , R 0.9677 However, when increasing from 0.080 V to 0.120 V, the I P almost does not change significantly.At a pulse amplitude of 0.080 V, the I P is 8.605 ± 0.072 μA (n = 3), and the relative standard deviation (RSD) is 0.834%, which was selected for the next experiments.The E step in the DPV method determines the potential scanning speed while fixing the time of each potential step.This affects the stripping signal, especially the stripping current.As shown in figure S3 (d), when the time of each potential step was fixed to 0.3 s, the potential scanning speed changed from 13.3 mV.S -1 to 33.3 mV.S -1 , which is not high, only 2.5 times.On the other hand, in the DPV method, it usually only ranges from 10 to 50 mV.S -1 .Therefore, the I P changes insignificantly from 7.894 μA to 8.500 μA at a step voltage of 9 mV and a potential scan rate of 30 mV.For S -1 , the I P value is large, at 8.488 μA, and has the smallest RSD value of 0.859%, which is appropriate in this study.
3.2.3.Linearity and limit of detection (LOD) for TPL DPV was utilized for the quantitative assay of TPL concentration at the proposed (4/6)NiF/AC-GCE.Figure 9(a) shows the DPVs at different concentrations of TPL with the 4/6)NiF/AC-GCE.In the absence of the TPL standard, there was no clear electrochemical response in the scanned potential range.This indicates that the 4/6)NiF/AC-GCE working electrode is not electrochemically active in this potential range in 0.1 M BR buffer at pH 3. A well-resolved anodic peak at approximately 1.35 V versus Ag/AgCl was observed, and the obtained results showed that under optimum conditions, the anodic peak current increased with increasing TPL concentration, as shown in figure 9(a).The plotted calibration curve of the peak current (I p /μA) against the TPL concentration (μM) (figure 9(b)) exhibited two linear segments between the peak current and TPL concentration over the range of 0.5-5.0μM with a high determination coefficient, R 2, of 0.997 and 5.0-19.6 μM with an R 2 of 0.991.The regression equations are as follows:

Repeatability, reproducibility and long-term stability
The repeatability of the DPV measurements was tested by conducting 10 successive runs for TPL at three concentrations (1 μM, 5 μM, and 10 μM) at the same electrode.The relative standard deviation (RSD) values were 4.91%, 3.04% and 1.90% at 1 μM, 5 μM, and 10 μM, respectively, which were less than the corresponding 1/2RSD H (Horwitz function) values of 10.4%, 8.1%, and 7.3%, respectively, indicating that good repeatability was achieved with no need to apply any regeneration procedure (figures 10(a)-(c).
Reproducibility is important in estimating the precision of a method.DPV measurements were performed under the optimized conditions for the 10 electrodes fabricated via the same procedure.The peak current response of each electrode did not change significantly (figure 10   Both %RSD values are less than 5%, which indicates that the proposed electrodes have a fairly good level of precision and stability.

Interference study
To study the selectivity of the (4/6)NiF/AC sensor, the DPV response of 5 μM TPL was studied in the presence of various inorganic salts and organic compounds as interfering substances (figure 11 and table S1).The addition of NH 4 NO 3 , NH 4 Cl, NaCl, Na 2 SO 4 , KNO 3 , Ca(H 2 PO 4 ) 2 , glucose (GLC), saccharide (SAC), sodium glutamate (SOD), and paracetamol (PRA) influenced the oxidation peak current of TPL.It is assumed that an absolute relative deviation of less than 5% means that this agent has a negligible effect.When the ratio of the interfering agent (NaCl, Na 2 SO 4 , KNO 3 , SOD or PRA) to TPL is less than 50, these species are considered to have no effect.Similarly, NH4Cl, Ca(H 2 PO 4 ) 2 , GLC, and SAC did not affect the peak current when the ratio was less than 40.These results revealed that the NIF/AC sensor exhibits phenomenal selectivity for TPL over other coexisting interfering species, suggesting good selectivity for the proposed method.

Analytical application of the proposed DPV method to real samples
The proposed DPV method was used to quantify TPL in three pharmaceutical formulations to evaluate its applicability (table 4).The prepared real sample solution was analyzed by using the standard addition method.A known concentration of TPL standard solution was spiked into a sample of human urine, and recoveries were obtained (table 2).Accordingly, the TPL recoveries of 97.12%-103.73%were satisfactory, and the reproducibility of the results was confirmed by an RSD < 5%.The results of this work indicate that the developed method using NiF/AC as an electrode modifier is promising for the determination of TPL.

Conclusion
A novel nickel ferrite activated carbon composite was synthesized by a hydrothermal process.Based on the performance of the NiF/AC-modified electrode, it may be employed in qualitative and quantitative analyses of pharmaceuticals and other electroactive target compounds in various sample matrices.Under suitable conditions, the modified electrode exhibited good accuracy, stability, and improved selectivity.Acceptable  recoveries were achieved in the analysis of real specimens, indicating that it is possible to apply the modified electrode for practical analyses.

Figure 1 (Scheme 1 .
Scheme 1.The proposed mechanism of TPH oxidation at the NiF/AC-GCE.
SEM images of the rice husk-derived AC and NiF/AC are shown in figure3(a).The macropores present in the AC solids are clearly visible with pores in the range of 1-5 μm in diameter, as shown in the SEM images.Furthermore, smaller pores were also visible in the SEM image of AC.The NiF quasipyramic-shaped nanocrystals are clearly visible in figure3(b).The particles seem to aggregate heavily due to their magnetic nature.The grain size was found to be 50-100 nm (estimated from 100 particles).
Figure 3(c) shows an SEM image of the (4/6) NiF/AC.The incorporation of NiF significantly changed the surface morphology of the AC, and the pores were less visible after the introduction of NiF to the AC.According to the elemental mapping images (figure 4), C, O, Ni, and Fe were distributed uniformly throughout the NiF/AC sample.In addition, a small amount (less than 1%) of impurities (N and S), probably from rice husks, was observed.The atomic ratio of Fe to Ni of approximately 2 is close to the stoichiometric ratio of NiFe 2 O 4 , demonstrating that the NiFe 2 O 4 nanoparticles were successfully loaded onto the surface of the AC.
An LOD of 0.21 μM was obtained from the calibration curve using the following equation (3 SD/(b): SD is the standard deviation, and b is the slope of the regression line.The obtained LODs are comparable to those reported in the literature (table3) for different developed sensors for determining TPL in pharmaceutical formulations and other matrices.

Figure 11 .
Figure 11.Effect of possible interfering compounds on the peak current of TPL (5 μM) in 0.1 M BR buffer at pH 3. The ratio on the horizontal axis represents the concentration ratio of the interferent to the analyte.

Table 2 .
Some physical parameters of the obtained materials.
*Crystallinity size from the Scherrer equation (311 peak), n/a: not applicable.
Some operational parameters were optimized, including the accumulation potential (E acc ), accumulation time (t acc ), pulse amplitude (E pulse ) and step voltage (E step ), as shown in figure8 and figure S3.The effect of E acc on the analyte stripping signals (I P (peak current) and E P (peak potential)) is shown in figure and as a result, TPL is converted to 1, 3-dimethyluric acid (Scheme 1) [32, 33]:3.2.2.Optimizing the operational parameters

Table 3 .
Comparison of the efficiency of some modified electrodes used in TPL determination.