Effect of metal catalysis in the electrochemical oxidation of petrol on platinum electrodes and its use in petrol brand fingerprinting

This study presents a novel study to explore the electrocatalytic characteristics of platinum electrodes to obtain the electrocatalytic oxidation of petrol in non – aqueous solutions. A petrol sample was added into an electro-chemical cell containing 0.1 M Tetrabutylammonium perchlorate (TBAP) in acetonitrile and studied using cyclic voltammetry and differential pulse voltammetry. The results demonstrated that the metal content in TBAP acted as a catalyst, which effectively interacted with substances present in petrol and formed a new electroactive product, providing an electrochemical oxidation response at + 1.2 V. Inductively coupled Plasma analyses and statistical comparison of different TBAP salt brands, indicated that Calcium, Iron and Manganese were directly response on the catalytic electrooxidation. The electrochemical response of petrol was observed as a diffusion – controlled process on platinum electrode with an irreversible reaction. The electrochemical profiles were used to discriminate different brands of petrol using Principal Component Analysis (PCA). Also, this behavior could be replicated in a simplistic manner using a model solution of Fe + 3 and petrol.


Introduction
Petrol is the most important petroleum derived liquid product used daily in transportation and industry.It is a complex mixture consisting mostly of organic hydrocarbons enhanced with a variety of special additives such as oxygenates, anti-knock, anti-oxidants, anti-corrosion, anti-icing and cetane improvers.Many instrumental techniques have been previously used to determine and classify petrol and their additives, as reported in the literature, such as gas chromatography [12345], infrared [678] and Raman spectroscopy [910].So, there is a good knowledge of the composition of this complex substance.
The electrochemical behaviour of individual components of basic nitrogen-containing molecules in petrol, such as quinolone and pyridine, has been previously studied prior to successful extraction [11].Some of these reactions have been found to be catalysed by metallic ions, such use Fe 3+ [12].Other compounds, such as dyes [1314] and color markers have also been analysed [15].The presence of inorganic contaminants such as metallic species in fuels may decrease the engine performance and damage the engine system due to the formation of abrasive and filter plugging substances [16].Several electrochemical methods for the determination of inorganic contaminants, such as Cd 2+ , Cu 2+ , Pb 2+ and Zn 2+ have also been reported [1718].Other individual sulphur compounds have also been determined by square wave voltammetry [19].Even in biodiesel determinations of metals using square-wave anodic stripping voltammetry (SWASV) have been performed [20].
Individual components determination after selective extraction is common.However, to the best of our knowledge, no studies on the electrochemical and electrocatalytic behaviour of bulk petrol have been previously performed.This could help to understand further use of the electro-conversion for this substance and other potential uses in specific areas (i.e., branding or low temperature catalytic conversion).
In this study, we focused on the electrocatalytic behaviour of petrol on platinum electrode.Due to poor water solubility of petrol, non--aqueous electrolyte solutions were used in this study.Three different common salts consisting of 0.1 M Tetrabutylammonium perchlorate (TBAP), Tetraethylammonium perchlorate (TEAP) and Tetrabutylammonium tetrafluoroborate (TBATFB) in acetonitrile were used as supporting electrolytes.

Samples and reagents
Petrol and diesel samples collected from different petrol stations in the city of Lincoln (Lincolnshire, UK) were used in the electrochemical analyses in this study.All reagents were used as purchased without further purification.All electrolytes used in the experiments were prepared in non-aqueous solvent.Acetonitrile (HPLC grade) used as a solvent was supplied by Fisher Scientific (Leicestershire, UK).Tetrabutylammonium perchlorate (TBAP) was purchased from Sigma-Aldrich (Gillingham, UK), with a purity of 99 % and from Acros Organics™ (Leicestershire, UK), with 98.5 % purity.Tetraethylammonium perchlorate (TEAP) with 98 % purity was purchased from Alfa Aesar™ (Leicestershire, UK).Tetrabutylammonium tetrafluoroborate (TBATFB) with a purity of 99 % was purchased from Sigma-Aldrich (Gillingham, UK).Sulfuric acid (Analytical reagent grade) used for electrochemical cleaning of platinum electrodes was purchased from Fisher Scientific (Leicestershire, UK).Fumed silica (particle size 0.007 μm) and aluminium oxide powder (Particle size 0.3 μm) used for polishing the platinum electrodes were both purchased from Sigma-Aldrich (Gillingham, UK).

Electrodes
A three electrode system was used in the experiments consisting of a single crystal platinum (Pt) working electrode with a 3 mm diameter of electrode disk, silver/silver chloride (Ag/AgCl) as a reference electrode and platinum (Pt) as a counter electrode, all were purchased from Metrohm U.K., ltd.The Ag/AgCl reference electrode was immersed in 3 M saturated potassium chloride solution when not in use.The internal solution was also 3 M KCl.

Experimental setup
An electrochemical cell consisted of Pt working electrode, Ag/AgCl reference electrode and Pt counter electrode.All voltammetric measurements were performed and controlled by an Autolab Potentiostat/ Galvanostat model PGSTAT302 (Metrohm Autolab B.V.) equipped with NOVA version 2.1 software (Metrohm Autolab B.V.).During the measurement, the electrochemical cell was put in the Faraday cage (Metrohm Autolab B.V.) which an earth terminal of the Faraday cage was connected to the Autolab in order to protect any electromagnetic interference from nearby electronic equipment such as computer monitors, other instruments or power lines.

Electrode cleaning process
Before each measurement, the Pt working electrode was polished to a mirror-like surface with aluminum oxide fine powder and fumed silica on a polishing cloth.After polishing, the electrode was rinsed with de-ionized water and dried using nitrogen gas.
Next, the polished electrode was performed electrochemical cleaning with fifty cycles of cyclic voltammetry at a scan rate of 100 mV s − 1 and a potential range from − 0.25 V to + 1.3 V in a solution of 0.5 M aqueous sulfuric acid.Dissolved oxygen was purged from the solution by bubbling nitrogen gas for 5 min.This process was performed to ensure that any absorbed deposits are completely eliminated and the electrode surface is free with any contaminants that might interfere the analysis.

Electrochemical experiments
In all the electrochemical experiments, 20 mL of 0.1 M of different supporting electrolytes was prepared in an electrochemical cell and de-aerated with nitrogen for 2 mins.The bubbling nitrogen gas was used to remove dissolved oxygen from the electrolyte solution.After degasification, a petrol/diesel sample was added to the electrolyte solution with different volumes and analysed using cyclic voltammetry (CV) technique.
After optimization of the experimental conditions, the samples were each analysed in septuplicate (7 repeats) using cyclic voltammetry (CV) and differential pulse voltammetry (DPV) techniques.

Chemometric analysis
The voltammograms obtained from all samples were analysed using principal component analysis (PCA) for sample classification.The vast number of variables contained in these voltammograms were eventually rearranged from the large number of possible correlated variables and reduced to the small number of uncorrelated variables by PCA, so that the classification patterns based on the common data features are more readily observed.
In this study, the data of the intensity of the current in a potential range from 0.0 V to + 1.5 V of each sample were performed by principal component analysis (PCA) and linear discriminant analysis (LDA) followed by leave-one-out cross validation algorithms using TANAGRA free data mining software version 1.4.50 (Ricco Rakotomalala, Lyon, France).

Cyclic voltammetric behaviour of petrol
The electrochemical behaviour of a petrol sample (100 μL) in 20 mL of each electrolyte solution was studied.Nitrogen gas was bubbled through the sample for 2 min to ensure an oxygen free system.The sample was analysed by cyclic voltammetry at a scan rate of 50 mV s − in a potential range from 0.0 V to + 1.5 V. Electrolyte blanks were also performed.Fig. 1 shows the cyclic voltammetry of the different salts showing that there were non-electrochemical signals of petrol observed in TBATFB (Fig. 1a) and TEAP (Fig. 1b) electrolytes.A cyclic voltammogram of 0.1 M TBAP/ACN exhibited two anodic peaks at + 0.70 V and + 1.01 V and two cathodic peaks occurred at + 0.51 V and + 0.93 V which was clearly a different voltammogram to those two electrolytes.When petrol was added, three anodic peaks were clearly observed.The first two anodic peaks exhibited a slight shift of peak potential observed at + 0.71 V and + 0.98 V compared to 0.1 M TBAP/ACN blank.Another extra anodic peak appeared at + 1.27 V (Fig. 1c).
With these results, we concluded that the Pt electrode is active for the anodic oxidation of petrol within the positive potential window between + 0.00 V to + 1.5 V when using TBAP as salt.

Electrochemical behavior of TBAP salt
From the previous results, peaks were observed in a cyclic voltammogram of TBAP/ACN of petrol samples.It is presumed that those peaks might be produced from an electroactive impurity contained in the TBAP salt interacting with petrol through a chemical reaction to form an electroactive compound.
A study of the relationship between peak current and scan rate is particularly useful to determine whether the electrochemical process is either diffusion or adsorption controlled (or a mixture of both).In an adsorption-controlled process, a linear plot of current versus scan rate should be obtained.This indicates that the electroactive species are being adsorbed onto the electrode surface.The linear relationship between current and the square root of scan rate is a characteristic of a diffusion-controlled process, meaning that the correlation between peak current and the concentration of analyte could be used as an analytical tool [2122].
To observe the behaviour of the system under analysis, variable scan rate ranges from 10 to 500 mV s − 1 cyclic voltammetry were performed in 0.1 M TBAP in acetonitrile (Fig. S-1; supplementary).From the figure, the cyclic voltammograms exhibited that the peak current of two anodic peaks and two cathodic peaks were gradually increased when the scan rate increased.
A correlation between peak current (I p ) versus the square root of scan rate (ν 1/2 ) of Peak 1, Peak 2, Peak 3 and Peak 4 from Fig.S-1 exhibited a linear relationship with correlation coefficients (R 2 ) of 0.9953, 0.9989, 0.9916 and 0.9835, respectively.This indicated a diffusion controlled process [2324].The electrochemical reaction on the electrode of the components found in TBAP diffuse from a region with high concentration in the bulk solution to a region with lower concentration near the surface area of the electrode.
The peak potential (E p ) was also dependent on the scan rate (notshown) regarding to a linear relationship between peak potential (E p ) and logarithm of scan rate (log ν).Peak 1, Peak 2, Peak 3 and Peak exhibited a linear relationship with a correlation coefficient of R 2 = 0.9832, 0.9841, 0.9773 and 0.9819, respectively.

Effect of sample volume
To understand the nature of peak A (petrol born), the effect of the petrol volume was studied and a series of different samples (10 -600 µL) were analysed in a 20 mL solution of 0.1 M TBAP and acetonitrile.Cyclic voltammograms in a potential interval 0 -1.5 V at the scan rate 50 mV s − 1 was performed (Fig. 2).
The CV profiles showed Peak 1 and Peak 2 as observed in electrolyte blank.It is clearly seen that the current of Peak A noticeably increased when petrol volume increased.

Differential pulse voltammetric behavior
Differential pulse voltammetry was also used in the same potential range at scan rate of 50 mV s − 1 .The resulting differential pulse voltammograms were baseline corrected using a moving average algorithm.Differential pulse voltammograms for different volume of petrol in 0.1 M TBAP and acetonitrile as solvent are shown in Fig. 3.
From Fig. 3, it is evident that the third anodic peak (Peak A) is linked to increasing petrol concentrations.Conversely, the height of Peak gradually decreased when more petrol volume added.Peak 1 slightly changed its peak height compared to the electrolyte blank but remained constant.
Additionally, peak potential was clearly observed as being shifted from higher to lower potentials in Peak 2 and Peak A. An EC mechanism is typically used to define a process in which an electron transfer step firstly occurs and is followed by a chemical reaction step.
The EC mechanism illustrates the electrochemical oxidation of the reactant (E) to form a product and subsequent chemical interaction (C) to form an electroactive species.An electron transfers immediately followed by a chemical process does sometimes change the molecule in such a way that a return wave is not observed.This seems to be the case in our study where peak 2 from the TBAP salt will react with one/some component(s) of the petrol to produce peak A. This would explain why Thus, we anticipate that an electrocatalytic reaction of petrol on platinum electrodes occurred by the interaction between some catalytic species present in TBAP salts, the platinum surface and some of the components present in petrol.To prove this point, we used a purer form of TBAP purchased from Sigma-Aldrich (99 % purity).The TBAP salt from Sigma-Aldrich did not produce any of the previous peaks.Moreover, the Acros Organics (98.5 % purity) TBAP, was washed with water and TBAP (Insoluble in water) was later dried and used.The results obtained were similar to those achieved by the Sigma-Aldrich TBAP.Furthermore, when the aqueous supernatant from the washing process using the Acros TBAP was mixed with the purer Sigma-Aldrich TBAP in the electrolytic cell, similar results to those from the initial Acros TBAP were obtained.
To understand the differences in behaviour between the TBAP purchased from Sigma Aldrich and Acros Organic™, inductively coupled plasma optical emission spectrometry (ICP-OES) was used for the screening of the aqueous solutions obtained from the two different brands of TBAP.Metals were suspected to be the reason for the differential behaviour as they have been known to previously catalyse the oxidation of hydrocarbons.Three elements, calcium (Ca), manganese (Mn) and iron (Fe) were found to be present in greater amounts in the Acros TBAP, in significantly different amounts, when compared to Sigma Aldrich TBAP, at 95 % confidence interval (p < 0.05) using paired t-test.Table 1 demonstrates the intensity and ppm values of these elements from two TBAP samples including the statistical data of two-tailed t-test.
There are multiple examples of different chemical species of these elements found in the literature presenting different forms of catalytic activity on the oxidation of hydrocarbons.Also, combinations of calcium doped manganese ferrite have demonstrated catalytic activity in the oxidation of hydrocarbons at 80 • C [25].Even the oxidation of more complex organic structures are also catalysed by Iron, as this metal is well-known to favour catalysis [26].
It is evident that the combination of petrol, metals and a catalytic surface, such as platinum, are essential for this electro-oxidation.
Information involving the electrochemical reaction can be generally obtained from the relationship between peak current and scan rate, as shown in Fig. S-2 (supplementary).
In order to prove the need for a catalytic surface as working electrode, similar experiments to the ones described in the previous section of this paper, were conducted using a glassy carbon electrode.Under these conditions no signal like those described on both Fig. 3 and Fig. 4 could be achieved.This is again evidenced that a double electrocatalytic effect that comes from the metal content in the TBAP salt and the electrode is a feasible explanation of the electrochemical behaviour in petrol.

Effect of scan rate
To determine whether the electrocatalytic process in peak A of petrol is diffusion or adsorption controlled, variable scan rate (10 -500 mV s − 1) cyclic voltammetry studies were performed on a 20 mL of 0.1 M TBAP in acetonitrile with a presence of 100 μL of petrol sample.This showed that the current of Peak A significantly increased with the increasing of scan rate.Also, the peak potential was evidently shifted to a more positive value (Fig. S-3a)-supplementary.The peak current for petrol  solutions was found to be linear to the square root of the scan rate (Fig. S-3b) which the equation can be expressed as; with a correlation coefficient of R 2 = 0.9908.The dependence indicates that the electrochemical oxidation reaction of petrol on Pt electrode is a diffusion-controlled process.
The peak potential was also dependent on the scan rate.The peak potential shifted to more positive values with increasing scan rates.This confirms the irreversibility of the oxidation process and a linear relationship with a correlation coefficient of R 2 = 0.9915 between peak potential and logarithm of scan rate ( From the results, it can be determined that a metal-aided electrochemical oxidation reaction of petrol on a Pt electrode is a diffusioncontrolled irreversible reaction.
A similar experimental protocol followed with petrol was used for diesel samples.No signals were obtained from these experiments, showing that the electrooxidation is specific for lighter compounds from the rectified fraction of petroleum, as diesel composition is richer in heavier compounds.
Moreover, when specific well-known compounds and additives present in petrol, namely, MTBE, ETBE, Isooctane, Benzene, Ethylbenzene, m-xylene and naphthalene, were individually tested under the same conditions no peak A appeared (Fig. S-4 supplementary).This clearly indicates the complex nature of the problem given the complex petrol composition and further fractionation and research is needed.

Classification of different petrol samples using cyclic volt-ammetry
In general, the composition of petrol typically depends on the refinery process, the distributor, the market demand and the quantity desired.The transportation and the storage of petrol can also make some difference in petrol compositions.The residues left inside the storage tank is another factor causing the change of chemical composition in petrol.With these factors, petrol sample is chemically unique.The unique chemical composition is useful for source identification as samples which statistically exhibit similar composition are likely from the same origin.
As shown in previous results, the electrochemical oxidation of petrol was successfully detected in a solution of 0.1 M TBAP (Acros Organic™) in acetonitrile with Pt electrode using cyclic voltammetry and differential pulse voltammetry methods.With this useful results, different petrol samples were examined in order to observe the difference of electrochemical behaviors of the samples from different sources.
Six super unleaded petrol sample from six different brands consisting of Shell, Gulf, Esso, Tesco, Sainsbury's and BP were examined using differential pulse voltammetry.The 100 μL of each petrol sample was prepared in the 20 mL of acetonitrile containing 0.1 M TBAP (Acros Organic™).DVP was performed in septuplicate (7 repeats) for each sample in a positive potential scan from + 0.0 V to + 1.5 V at a scan rate of 50 mV s − 1 .
Fig. S-5 (supplementary) demonstrates the different pulse voltammetric profiles of six different petrol samples.It is clearly seen that each sample exhibited the different DPV profiles.An anodic oxidation peak of petrol was clearly observed in three samples from Shell, Tesco and Esso, the rest three samples from Gulf, Sainsbury's and BP did not show this peak.Esso showed the highest anodic peak of petrol compared to other samples.As the peak current was proportionally associated with the concentration, the higher the concentration the higher the peak was.This indicates that the presence of an electroactive substances in each petrol sample have different concentration resulting to the difference of different pulse voltammetric profiles.Thus, this confirms that the chemical composition of petrol is unique due to different brands.
For classification purpose, principal component analysis (PCA) was applied.The DPV data of all six samples was exported to TANAGRA statistical software as a 42 × 298 data matrix, representing the 42 samples (6 samples × 7 repeats) and the 298 variables (peak currents in the potential of + 0.0 V to + 1.5 V).PCA was performed on the data set.The result showed that the most of the variability in the data was explained by the first three principal components, corresponding to 78.50 % of the variance in the data.PC1 described 38.94 % of the variance in the data, PC2 described an additional 24.40 % of the variance and PC3 described 15.16 %.This can be explained that the data matrix has effectively been reduced from 298 dimensions to 3 dimensions while still maintaining 78 % of the information.
Correlation to the principal components was also evaluated by using factor loading scores.Factor loading is the correlation coefficient between variables and principal components.Factor loadings can range from -1 to 1, the closer to -1 or 1 the stronger the effect of the variable on the PCs.In this study, PC Axis 1 showed that potential range from 0.181 V − 0.569 V and potential range from 0.957 V − 1.375 V exhibited factor loading close to − 1 where as potential range from 0.775 V − 0.796 V exhibited factor loading close to 1.The potential range from 1.073 V − 1.496 V in PC Axis 2 exhibited factor loading close to 1.In PC Axis 3, the potential range from 0.614 V − 0.690 V exhibited factor loading close to 1. Table S-1 in supplementary shows factor loadings closer to -1 or 1 of each potential in first three principal components.This confirms that the chemical composition of each petrol sample is uniquely different.The electrochemical data was able to classify all the petrol samples from different brands.Therefore, the determination of petrol not only can be analysed by chromatographic and spectroscopic techniques but also can be detected by electroanalytical methods such as cyclic voltammetry and differential pulse voltammetry.As a conclusion, there is sufficient evidence to conclude that different petrol brands have different compositions generating different voltammetric responses that could be employed as an electrochemical fingerprint'.
In order to prove the importance of the presence of the catalytic metals, 100 ppm of Fe +3 (FeNO 3 ) was added to the supporting electrolyte (0.1 M TBAP/ACN) and a cyclic voltammogram obtain at a scan rate 0.1 V s − 1 .Fig. 5 shows the typical Fe +3 /Fe +2 system from 0.2 to 0.8 V and the oxidation peak from the petrol sample at 1.4 V.In this simpler model (one metal) the clear signal of the catalytic metal is clearly observed and how the oxidation/reduction of iron is not affected by the different amounts of petrol added to the system.However, peak A increases with increasing concentrations of petrol (as previously observed) and it has been displaced to more anodic potentials (1.4 V).
This simple experiment demonstrates that the behaviour observed from the impurities in TBAP corresponded to the interaction of different metals with the platinum electrode and the petrol sample.Fig. S-6 supplementary shows the behaviour with different scan rates.

Conclusion
This study offered novel results to utilize the electrocatalytic characteristics of platinum electrodes and the presence of catalytic elements in the supporting electrolyte to achieve the electrooxidation of petrol in non-aqueous solution.It has been found that species of iron, calcium and manganese in TBAP acted as a catalyst, which effectively interacted the platinum surface and substances present in petrol to form a new electroactive product, providing electrochemical oxidation response above + 1.2 V.
The electrochemical oxidation of petrol was suppressed if none of these catalytic elements were present in solution.Moreover, studies using glassy carbon electrodes did not render any results either, indicating that the electrocatalytic nature of the platinum electrode is also needed.The results also found that the electrochemical oxidation of petrol increased as a function of petrol volume, as petrol substances involved in the electrooxidation increased their concentration.The results also showed that the electrochemical oxidation reaction of petrol was observed as a diffusion-controlled process on Pt electrode through an irreversible reaction.
In addition, the electrochemical data was successfully used to discriminate among six different petrol samples with 100 % correct classification by PCA.This indicates that each petrol uniquely exhibited the electrochemical response due to different chemical composition.Thus, electrochemistry was able to determine the presence of petrol and also classify petrol from different brands with a high degree of accuracy.
Moreover, a model solution of 100 ppm Fe +3 , petrol and the supporting electrolyte (TBAP/ACN) were subjected to the same electrochemical conditions and the subsequent voltammogram demonstrated the same behaviour than previously observed.
Although a suspected EC mechanism is foreseen, the specific electrochemical reaction mechanism for petrol on Pt electrode in non--aqueous solution could not be determined in this study due to unknown nature of the electroactive compound(s) and the complex nature of petrol.
The study opens the door to a more efficient oxidation of petrol products using low temperature electrocatalysis, compared to other hydrocarbon oxidations at high temperature reported in the literature.There are still many questions unanswered in terms of what petrol substances are mainly responsible for the electrooxidation, given that diesel did not produce any signals.Also, other questions around the specific nature of the elements found to participate/catalyze the reactions would need to be clarified, as we only concentrated in elemental and not molecular analysis.Further studies including the rectification of petrol samples and the study of the specific catalytic electrooxidation of these fractions would be also needed.These were not included in this communication as it was deemed that the findings reported in this study were urgent and needed to be immediately reported so other groups can also contribute to expand on this.It is highly probable that other catalytic metal species, different from those reported here, could also be used as catalytic substances.

Fig. 2 .
Fig. 2. Cyclic voltammograms of a series of petrol volume (10-600 μL) in 0.1 M TBAP in acetonitrile on Pt electrode at scan rate of 50 mV s − 1 .A background electrolyte presented in dashed line.

Fig. 3 .
Fig. 3. Differential pulse voltammograms of a series of petrol volume (10-600 μL) in 0.1 M TBAP in acetonitrile on Pt electrode at scan rate of 50 mV s − 1 .A background electrolyte presented in dashed line.

Fig. 4 Fig. 4 .
Fig. 4. A scatterplot of PC Axis 2 versus PC Axis 3 of six different petrol samples using different pulse voltammetric data in the potential range from 0 to 1.5 V.

Table 1
Intensities and ppm of the most significant elements from two different brands of TBAP salt.