Multivariate optimization of dual-bed solid phase extraction for preconcentration of Ag, Al, As and Cr in gasoline prior to inductively coupled plasma optical emission spectrometric determination

Fuel 139 (2015) 285–291

Contents lists available at ScienceDirect

Fuel journal homepage: www.elsevier.com/locate/fuel

Multivariate optimization of dual-bed solid phase extraction for preconcentration of Ag, Al, As and Cr in gasoline prior to inductively coupled plasma optical emission spectrometric determination Philiswa N. Nomngongo, J. Catherine Ngila ⇑ Department of Applied Chemistry, Faculty of Science, University of Johannesburg, Doornfotein Campus, P.O. Box 17011, Johannesburg, South Africa

h i g h l i g h t s  Preconcentration of trace metals in gasoline using dual bed SPE system.  The experimental conditions of the method was optimized using factorial design.  The method had relatively high enrichment factor with low LOD and LOQ.  The column can be reused up to 200 cycles without loss of sorption performance.  Method was applied for analysis of real gasoline samples.

a r t i c l e

i n f o

Article history: Received 4 June 2013 Received in revised form 18 August 2014 Accepted 19 August 2014 Available online 30 August 2014 Keywords: Dual-bed resin Metal ions Factorial design Separation and preconcentration Gasoline

a b s t r a c t In this work, a dual-bed resin solid phase extraction (SPE) for preconcentration of Ag, Al, As and Cr prior to their inductively coupled plasma-optical emission spectroscopy (ICP-OES) determination has been developed. Dowex 50 W-x8 and Dowex 1-x8 packed in a column were used as metal ion sorbents. The optimization of the dual-bed SPE procedure was carried out using a two level full factorial design with three central points. Under optimized conditions, the limits of detection and quantification (n = 21) ranged from 0.16 to 0.22 and 0.52 to 0.76 lg L1, respectively. Enrichment factors of 100, 130, 130 and 150 and relative standard deviations (n = 15) of 1.2%, 2.0%, 1.8% and 1.3% were obtained in the determination of Ag, Al, As and Cr, respectively. The validity of the proposed method was checked by applying the standard addition method and the recoveries at the 20 lg L1 level using both inorganic and organic metal standards ranged from 95% to 99%. The proposed method presented an analytical throughput of about 18 samples per hour and was applied for the determination of metal ions in ten gasoline samples. In addition, the accuracy of the method was evaluated using microwave-assisted digestion method and the results were not significantly different (at 95% confidence level). Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction Controlling the levels of metals in petroleum products such as gasoline is a critical step in the petrochemical industry. This is because metal ions act as catalyst poisons thus cause deleterious effects on the refinery and processing operations unless they are added as additives [1,2]. Therefore, it is crucial to accurately determine metal content in liquid fuels which are also the main sources of energy for vehicles. Other effects of metal ions (even in trace concentrations) in liquid fuels are reported in the literature [1,3– 5]. These include (i) poor fuel performance (ii) decrease in the ⇑ Corresponding author. Tel.: +27 115596196; fax: +27 11559 6425. E-mail address: [email protected] (J.C. Ngila). http://dx.doi.org/10.1016/j.fuel.2014.08.046 0016-2361/Ó 2014 Elsevier Ltd. All rights reserved.

engine durability and efficiency and (iii) environmental pollution caused by the release of toxic metals into the atmosphere during fuel combustion [1,3–5]. Therefore, the development of sensitive and selective analytical techniques for the determination of metals in gasoline is one of the most important aspects of quality control in petroleum industries [1]. These techniques must be fast, simple, precise, accurate and economical to be easily employed in routine procedures. In addition, since metal ions in liquid fuel samples are usually present in trace levels, the analytical methods must be capable of resulting in high enrichment/pre-concentration factors enough to cope with the demands [1]. Analytical methods based on electrothermal atomic absorption spectrometry (ETAAS) are popular because they are associated with high sensitivity and tolerance to high organic matrix loads

286

P.N. Nomngongo, J.C. Ngila / Fuel 139 (2015) 285–291

[2]. However, ETAAS is not common in routine analysis because of its low sample throughput as compared to inductively coupled plasma-based techniques [2]. Inductively coupled plasma optical emission spectrometry (ICP-OES) is widely used in routine quantification of metal ions in different sample matrices [6–10]. This technique is attractive due to its multielement capability, relative sensitivity, wide linear range and high sample throughput. However, the direct introduction of fuels into the plasma requires special care, as the organic load may de-stabilize or extinguish the plasma [11–14]. Hitherto, different sample preparation approaches for determination of metal ions in fuels has been developed to overcome the problems associated with ICP-OES and are reported in the literature. These methods include conventional ashing and acid dissolution [14], microwave digestion [6,11,15], dilution with organic solvents [7], emulsion/microemulsion [9,10] and preconcentration using solid phase extraction [16,17]. However some of these sample preparation methods have some limitations, for instance conventional ashing and acid dissolution methods are time-consuming and also volatile elements may easily escape (be lost) [18]. Microwave digestion methods may be a good alternative to these methods and solve the problem of volatilization, but they increase the risks of explosion and cross-contamination. In addition, the use of concentrated acids could increase the blank values and cannot be supported by some analytical techniques such as ICP-OES [19]. Emulsion/microemulsion technique is one of the most promising approaches due to its short sample preparation time and the low risk of analyte losses by volatilization or sorption. However, its disadvantage is the low stability which then affects the sensitivity and reproducibility of the analytical instrumental signal [19]. Dilution with organic solvents is one of the simplest sample pretreatment procedures but does not reduce the problem of organic loading and plasma destabilization or extinction in case of the ICP techniques [13]. To overcome these difficulties associated with sample pretreatment methods, an accurate and reliable analytical procedure based on separation and preconcentration of analytes prior to analysis in fuel samples, is required. Preconcentration of the analytes from the organic matrices combines the advantages of separating the analyte from the complex fuel matrix, transferring it to an aqueous phase and enriching it at the same time [13]. Procedures based on solid phase extraction (SPE) for the separation and preconcentration of trace elements in gasoline and fuel kerosene are reported in the literature [16,17,20]. Recently, chemometric techniques have been used for optimization of different analytical methods. This is because, these techniques allow more than one variable to be optimized simultaneously [21]. The advantages of multivariate techniques include reduction in the number of required experiments, thus, lowering reagent consumption and significantly less laboratory work. They are faster to implement and more cost-effective than traditional univariate approaches [22,23]. In addition, chemometric methods are able to generate mathematical models that permit assessment of the relevance and statistical significance of factors being studied, and evaluation of interaction effects between them (factors) [21,22]. Full factorial design is one of the well-known statistical processes for multivariate optimization and is widely applied in analytical chemistry. This is due to its effectiveness in the identification of significant variables and the best conditions of an experimental procedure [21]. The aim of this work was to investigate the analytical performance and the potential applicability of a dual-bed resin column for preconcentration and determination of trace metals in gasoline via off-line SPE/ICP-MS system. The literature shows extensive focus on the methods for analysis of Cu, Zn, Ni, Pb, Cd, V, Mo and Fe, particularly in petroleum and its derivatives [9,12,15,16,20].

In the current study we focused on the selected elements due to their peculiar behavior. For instance, arsenic causes severe and irreversible catalyst poisoning, even at trace levels [24]. The rest of the elements mentioned above have been investigated in our previous studies [17,25,26]. Therefore, this study seek quantify the amount of Ag, Al, As and Cr in gasoline samples. A full two-level factorial design with a central point was used for optimization of experimental variables (pH, eluent concentration and sample flow rate) that affect the retention/desorption of metal ions. To the best of our knowledge, this is the first time that dual-bed resin column and the optimized preconcentration method are proposed for Ag, Al, As and Cr determination in gasoline. In addition, this study offers a simple system with no need of acid digestion prior to metal ion determination. 2. Experimental 2.1. Instrumentation Metal ions (Ag, Al, As and Cr) were determined using a Spectro Arcos 165 ICP-OES (SPECTRO Analytical Instruments, GmbH, Germany) equipped with Cetac ASX-520 autosampler. The ICP-OES operating conditions are listed in Table 1. Sample introduction was achieved using a pneumatic cross-flow nebulizer mounted onto a Scott double-pass spray chamber. Sample solutions were pumped to the nebulizer using a built in four channel peristaltic pump. The most prominent atomic and ionic analytical spectral lines of the metal ions were selected for investigation, that is, Ag 329.068 nm, Al 167.078 nm, As 193.759 nm and Cr 283.563 nm. Solid phase extraction was carried out in a VacMaster-24 sample SPE station (Supelco, PA, USA). The latter was used to control the sample loading and elution flow rates. The microwave digestions were carried out in an Ethos D (Milestone, Sorisole, Italy) with maximum pressure 1450 psi and maximum temperature 300 °C. 2.2. Reagents, solutions and samples All reagents were of analytical grade unless otherwise stated and Millipore water was used throughout the experiments. Absolute ethanol (99.9%) used to prepare model solutions and suprapur 30% hydrogen peroxide (H2O2) used for the acid digestion procedure were obtained from Merck, (Darmstadt, Germany). Spectrascan stock solutions (1000 mg L1) of Ag, Al, As and Cr (Teknolab, Norway) were used to prepare the working solutions for SPE at concentrations of 10 lg L1 for each analyte. Working solutions, as per the experimental requirements, were freshly prepared from the stock solution for each experimental run. A Spectrascan multielement 100 mg L1 standard solution (Teknolab, Norway) was used to prepare working standard solutions at concentrations of 0–120 lg L1 for measurements of concentrations of analytes in all model and sample solutions. The cation exchangers used in this study as packing materials, that is, Dowex 1-x8 (Chloride form) and Dowex 50 W-x8 (sodium form) as well as solutions of nitric acid at concentration range of 1.0–4.0 mol L1 used for the elution of the analytes from the columns, were prepared from ultrapure concentrated acid (65%), were purchased from Sigma Aldrich (St. Loius,

Table 1 The operating parameters of determination of elements by ICP-OES. RF power Plasma argon flow rate Auxiliary argon flow rate Nebulizer argon flow rare Sample aspiration rate Replicate measurements (n)

1400 W 13 L min1 2.00 L min1 0.95 L min1. 2.0 mL min1 3

287

P.N. Nomngongo, J.C. Ngila / Fuel 139 (2015) 285–291

MO, USA). The pH adjustments were performed with glacial acetic acid and ammonia solutions (Sigma–Aldrich, St. Loius, MO, USA). Ten gasoline samples from different local filling stations were used for method development and validation. 2.3. Preparation of a two bed column The two bed resin was prepared in a 6.5 mL polyethylene column. The preparation procedure was carried out as follows: the first bed was prepared by placing the slurry (prepared in Millipore water) of 750 mg Dowex 50 W-x8 resin in the column that has a porous frit at the bottom. Another frit was then placed on the top of the packing material. The second bed was prepared by placing the slurry of 750 mg Dowex 1-x8 resin on the top of the first bed and a porous frit was placed on top of the packing material to hold and confine the adsorbents within the designated capacity/volume. The total height of the duo sorbent bed was approximately 4 cm in the column. The columns were sequentially washed with 10 mL of 3.0 mol L1 HNO3 and Millipore water followed by conditioning with 10 mL ammonium acetate buffer (1.0 M, pH 9.0) and then 10 mL of ethanol. 2.4. Preconcentration and recovery of Ag, Al As and Cr in synthetic gasoline solution The model solutions were prepared as follows: 10 mL of synthetic gasoline was placed in a 100 mL polyethylene volumetric flask followed by addition of 1.0 mL of 1.0 mg L1 of metal ions solution and made up to the mark with ethanol to obtain 10 lg L1 of each metal ion. The mixture was homogenized by shaking. An aliquot of 50 mL of the model metal solutions were passed through an ion exchange column at an appropriate flow rate. The columns were washed with 10 mL of Millipore water to remove excess organic solution, followed by 5.0 mL of ammonium acetate buffer solution to remove major cations (Na, Ca, K, etc) [17]. Lastly the metal ions were eluted with 5 mL of appropriate concentration of HNO3 solutions. All fractions obtained during the elution stage were collected separately and analyzed by ICP-OES. The same procedure was applied to the blank solutions. In the case of real sample analysis, a gasoline–ethanol mixture was prepared according to Chaves et al. [7]. An aliquot of 1.0 mL of gasoline sample was placed in a 100 mL polypropylene volumetric flask and 500 lL of concentrated HNO3 was added to the sample which was then diluted with ethanol. In between experiments, the resins in the column were washed with 20 mL of Millipore water followed by 10 ml of 1.0 M NaOH (this was done in order to keep the resin in sodium cation and hydroxide anion forms) and stored for the next experiment. 2.5. Optimization approach The most important steps in the multivariate optimization of an analytical procedure are choice of the experimental design, statistical analysis of the data and the determination of response variables, factors and factor levels. Recently, two-level full factorial experimental design (FFD) where each variable is evaluated at two levels has been widely used [27]. This work, the separation and preconcentration method was optimized using FFD. The latter is used to conduct and plan experiments in order to extract the maximum amount of information from the collected data in the presence of instrumental noise and in few experimental runs. The basic idea is to vary all relevant factors simultaneously over a set of planned experiments and then connects the results by means of a mathematical model. The later highlights the effect of individual variable and their relative importance in the separation and preconcentration of metal ions in gasoline samples. The obtained

Table 2 Factors and levels used in 23 factorial design for separation and preconcentration of metal ions. Variable

Low level (1)

Central point (0)

High level (+1)

pH EC (mol L1) SFR (mL min1)

3.0 1.0 1.0

6.0 2.5 2.0

9.0 4.0 3.0

information is then used for interpretation, predictions and optimization [28]. In addition, full factorial design allows for the assessment of factors affecting solid phase extraction and possible multi-variable interactions. The variables/factors investigated in this study were sample pH, eluent concentration (EC) and sample flow rate (SFR). The experimental runs required for the investigation of all the effects (variables) is given by 23, where 2 stands for the number of levels and 3 are the number of factors. The FFD levels for each variable are assigned minimum (), maximum (+) and central point (0) values, as shown in Table 2. Each factor was chosen according to data from previous experiments. All the experiments were carried out in random. The experimental data was processed by using the Minitab version 15 software program. 2.6. Comparative method The Microwave acid digestion procedure was carried out according to Kowalewska et al. [29]. Briefly, 5.0 mL of the gasoline sample was placed into a Teflon vessel followed by 6 mL HNO3 (65%) and 2.0 mL H2O2 (30%). The vessels were inserted into a microwave unit and heated according to the conditions recommended by the manufacturer. The digested material was left to cool down to room temperature. After cooling, the vessels were opened and 2 mL of concentrated HNO3 and 2 mL of hydrogen peroxide were added. The heating program was then repeated. This step was done in order minimize incomplete mineralization of the organic matrix. Finally, the Teflon vessel contents were cooled down to room temperature and quantitatively transferred to a 50 ml calibration flask. The samples were spiked with 20 lg L1 followed by the addition of 1 mL of concentrated nitric acid and the flask was filled up to the mark using Millipore water. The latter was submitted to the same procedure and used as the blank. The samples were then analyzed with ICP-OES. 3. Results and discussion 3.1. Factorial design The factors affecting the performance of dual-bed solid phase extraction for separation and preconcentration of Ag, Al, As and Cr ions in gasoline samples were investigated. The variables (factors) chosen for the optimization of the preconcentration system included sample pH, eluent concentration (EC) and sample flow rate (SFR). In order to determine the main factors of the preconcentration system, a two-level full factorial design (23) with three replicates of the central point (CP) was performed. The percentage recovery of each metal ion was used as the analytical response. Table 3 shows the experimental design matrix (coded values) and the results derived from each run for Ag, Al, As and Cr, respectively. In the design matrix, the rows are experimental runs and the settings for each variable (factor) are given by plus (maximum level) or minus (minimum level) signs in the respective columns. It can be seen from this table that all variables are changed simultaneously in order to cover the entire area of interest will very few number of experiments. In addition, the central point (0, middle level) is included to detect the curvature in the experimental

288

P.N. Nomngongo, J.C. Ngila / Fuel 139 (2015) 285–291

Table 3 Design matrix and the results of Ag, Al, As and Cr. Run

pH

EC (mol L1)

SFR (mL min1)

1 2 3 4 5 6 7 8 9 10 11

+1 +1 +1 +1 1 1 1 1 0 0 0

+1 +1 1 1 +1 +1 1 1 0 0 0

+1 1 +1 1 +1 1 +1 1 0 0 0

Recovery (%) Ag

Al

As

Cr

30.34 32.45 38.59 40.25 71.29 77.02 78.87 79.82 97.48 97.48 97.31

44.71 33.02 20.18 60.67 49.73 71.09 72.05 55.30 98.56 98.45 99.01

60.82 59.87 98.96 75.76 45.91 19.01 32.92 42.12 60.25 61.17 62.16

99.11 96.62 46.51 36.88 65.44 70.32 37.43 44.78 59.04 59.41 59.26

the optimum sample pH, eluent concentration and sample flow rate chosen for simultaneous separation and preconcentration of Ag and Al in a gasoline matrix concur with the conditions established by experimental runs 9–11. For As and Cr, the highest recoveries were observed at experiments 1 and 3, respectively. Therefore, the optimum sample pH, eluent concentration and sample flow rate for preconcentration of As were 9.0, 1.0 mol L1 and 3.0 mL min1, respectively. For separation and preconcentration of Cr on the other hand, the optimum sample pH, eluent concentration and sample flow rate were selected to be 9.0, 4.0 mol L1 and 3.0 mL min1, respectively.

EC = eluent concentration; SFR = sample flow rate.

3.2. Effect of sample volume

region [28]. The actual values of each factor in Table 3 can be seen in Table 2. The data in Table 3 was evaluated by analysis of variance (ANOVA) and p-values. The Pareto charts produced from ANOVA results, were used visualized the main effects and their interactions (Fig. 1). The bar lengths of the Pareto chart are proportional to the absolute value of the estimated effects and they help in comparing the relative importance of effects [30]. It can be seen from Fig. 1 that sample pH was highly significant for all the metal ions studied except for Al. Other factors such as sample flow rate and eluent concentration had little or no statistical significant effect on the extraction of Ag, Al, As and Cr. In view of the information obtained from Table 3 and ANOVA results, Ag and Al can be preconcentrated simultaneously. This is because they both had the highest percentage recoveries at the same experimental conditions (experiments 9–11). Therefore,

The concentrations of metals in real gasoline samples are typically at trace levels. To solve the problem of detectability of trace levels, a preconcentration procedure can be employed where large sample volumes are used to obtain high enrichment factors. In addition, the optimization of sample volume helps in evaluating the saturation point of an absorbent. Therefore, the capacity of the column was examined by loading, 50–1000 mL volumes of synthetic gasoline solutions containing 15 lg L1 of metal ions. The recoveries of the analytes from different volumes of synthetic gasoline solutions are presented in Fig. 2. The recoveries were found to be stable up to 500, 650, 650 and 750 mL for Ag, Al, As and Cr, respectively. Therefore, the experimental preconcentration factors, defined as the ratio of the analyte concentrations before and after preconcentration, were calculated to be 100, 130, 130 and 150 for Ag, Al, As and Cr, respectively. The observed decrease in recoveries for each metal ion was probably due to the excess analytes, loaded over the column capacity as the sample volume increased. As a

Fig. 1. Pareto chart of standardized effects for variables in the separation and preconcentration of silver (A); aluminum (B), arsenic (C) and Cr (D).

P.N. Nomngongo, J.C. Ngila / Fuel 139 (2015) 285–291

289

for preconcentration of triplicates was approximately 10 min. Hence, the throughput sample was approximately 18 samples h1. The analytical performances (LOD, LOQ and%RSD) of the proposed method were compared with those reported in the literature. The present method showed better analytical figures of merit for Al, As and Cr than those reported in the literature [32– 35]. On the other hand, analytical performances obtained indicated that the proposed preconcentration method has relatively higher or similar LODs when compared with reported by Refs. [36,37]. 3.5. Validation of the dual-bed SPE method

The regeneration of the dual bed column was investigated by monitoring the changes in the recoveries of Ag, Al and As through several retention–elution cycles. In each cycle, 50 mL multielement solution (20 lg L1) was passed through the column and then eluted with 5 mL of 3.0 mol L1 HNO3. The procedure was carried out 40 times per day for 5 subsequent days (a maximum of 200 runs) without any changes in the performance.

Due to the absence of certified reference material (CRM) that is similar to the investigated samples, the validity of the proposed dual-bed SPE system was examined by using standard addition method. Gasoline sample (G1) was spiked with organic and inorganic standard solutions. Additionally, the purpose of spiking the diesel sample with organic and inorganic standard solutions was to evaluate the ion exchange efficiency of resins to different metal species in fuel samples. This is because the speciation of trace elements in petroleum products is not fully known and different species may display different adsorption behaviors [16]. The recoveries of analytes spiked into the gasoline sample are presented in Table 4. It can be seen from this table that the recovery values calculated for the standard additions (organic and inorganic forms) for the investigated metal ions were greater than or equal to 95%. These results confirmed the accuracy of the proposed method and insignificant matrix effects, taking into consideration that the recoveries were in the range from 95–99%. In addition, since similar percentage recoveries were obtained for organic and inorganic forms, this implies that a dual-bed SPE system may be used for preconcentration of trace elements in their inorganic or organic forms.

3.4. Analytical Performances of the dual-bed SPE method

3.6. Analysis of real samples

Calibration solutions were prepared with multi-element standards containing 0, 5, 20, 40, 80, 100 and 120 lg L1 in 100 mL volume. Each solution was passed through the column and collected in 5 mL of HNO3. The calibration curve was linear (r2 = 0.9981–0.9994) for Ag, Al, As and Cr. The IUPAC limit of detection (LOD) and limit of quantification (LOQ) under optimized conditions were calculated from the signals of 21 successive measurements of the blank (100 mL) and the slope (m) of the calibration curve. The LOD was defined as the lowest concentration of an analyte giving signals equal to three times standard deviation (SD) of blank signal divided by the slope of the calibration curve that the analytical technique can detect (3SD/m). The LOQ, on the other, was defined as the to the smallest concentration of an analyte giving signals equal to ten times the standard deviation of blank signal divided by the slope of the calibration curve which can be accurately and precisely measured with an analytical procedure (10SD/m). Under the optimum conditions, the LOD were found to be 0.17, 0.16, 0.18 and 0.22 lg L1, for Ag, Al, As and Cr, respectively. On the other hand, the LOQ values were determined as 0.57, 0.52, 0.59 and 0.76 lg L1 for Ag, Al, As and Cr, respectively. The precision of the preconcentration system, calculated as the relative standard deviation (% RSD; n = 15), was less than or equal to 2% with mean recoveries of 97.3 ± 1.2%, 99 ± 2.0%, 98.7 ± 1.8% and 97.6 ± 1.3%, for Ag, Al, As and Cr respectively. The time required for preconcentration of 100 mL of sample was obtained to be 6 under the following conditions: percolation for 300 s at a flow rate 2–3 mL min1; elution for 30 s at a flow rate 3.0 mL min1; washing and conditioning for 30 s. However, the sample preconcentration was performed in triplicate and they were all carried out at the same time. Therefore, the overall time

The proposed method was applied for determination of Ag, Al, As and Cr in ten gasoline samples. The obtained metal ion concentrations are presented in Table 5. In general, the concentrations of Al were higher in all the samples compared to other studied metal ions. The higher concentration might be due to the abundance of Al in the Earth’s crust. It was observed that Al and As were always present in all the samples, whereas Cr was not detectable in G7 and G8 samples. It can be seen from Table 5 that the concentrations of Ag in most of the samples are below the LOQ (calculated using 10SD/m) of the proposed method. The use of dual-bed SPE method prior to ICP-OES determination showed an improvement for both sensitivity and LOD for metal ions those that were in ultra-trace levels. This can be an important feature in the analysis of the fuel samples. In view of the fact that there is no gasoline or similar reference material available with certified values for the studied metal ions, it was crucial to use an independent sample pretreatment technique for further validation (microwave-assisted digestion). It should be noted that ICP-OES after microwave-assisted digestion was used as a reference method. The results obtained for gasoline samples by the proposed method did not differ significantly from the values obtained by the reference method according to the paired Student’s t-test at the 95% confidence level (tcal = 0.98, 0.46, 0.16 and 1.40 for Ag, Al, As and Cr, respectively). In all the cases, tcal was lower than tcrit = 3.18, n = 4 for Ag; tcrit = 2.26 for Al and As, n = 10; tcrit = 2.44 for Cr, n = 7. In addition, the statistical F-test showed that the precisions of the proposed analytical methods were not significant at 95% confidence level. The overall results are satisfactory and show that the dual-bed SPE method has provided accurate results.

Fig. 2. Effect of sample volume on the recoveries of metal ions.

compromise, 100 mL was chosen for further investigation. This was done in order to speed up sample analysis. 3.3. Column regeneration

290

P.N. Nomngongo, J.C. Ngila / Fuel 139 (2015) 285–291

Table 4 Determination of Ag, Al, As and Cr (lg L1) in gasoline sample spiked with inorganic and organic standard solutions (mean ± standard deviation, n = 3). Addedc

Ag Found

ISa MOSb a b c

0 5 10 5 10

Al c

27.00 ± 0.12 31.82 ± 0.95 36.71 ± 1.01 31.76 ± 1.10 36.57 ± 0.89

As

Cr

Recovery

Found

Recovery

Found

Recovery

Found

Recovery

– 96.41 ± 1.51 97.10 ± 1.44 95.24 ± 1.21 95.73 ± 1.12

57.00 ± 0.10 61.89 ± 0.55 66.83 ± 1.32 61.91 ± 0.76 66.87 ± 1.61

– 97.83 ± 1.25 98.31 ± 1.33 98.22 ± 1.18 98.70 ± 1.80

156.81 ± 1.80 161.76 ± 2.10 166.69 ± 1.87 161.58 ± 1.53 166.49 ± 2.03

– 99.00 ± 2.31 98.82 ± 1.82 95.38 ± 1.67 96.78 ± 1.15

60.58 ± 0.38 65.45 ± 1.16 70.49 ± 1.31 65.42 ± 0.81 70.47 ± 0.75

– 97.44 ± 2.12 99.10 ± 2.51 96.83 ± 1.78 98.87 ± 1.13

IS: Inorganic standard. MOS = metallo-organic standard. Concentration in lg L1.

Table 5 The determination of Ag, Al, As and Cr in different gasoline samples using dual-bed SPE/ICP-OES and MAD/ICP-OES methods. Samples

G1 G2 G3 G4 G5 G6 G7 G8 G9 G10 a b c

Al (lg L1)

Ag (lg L1)

As (lg L1)

Cr (lg L1)

DB-SPEa

MADb

DB-SPE

MAD

DB-SPE

MAD

DB-SPE

MAD

27.00 ± 0.12 ND 4.67 ± 0.16 4.14 ± 0.22 22.82 ± 0.42 NDc ND ND ND ND

26.57 ± 0.14 ND 4.48 ± 0.12 3.83 ± 0.34 23.18 ± 0.65 ND ND ND ND ND

1469.37 ± 20.80 537.70 ± 4.65 799.14 ± 10.34 709.70 ± 8.31 626.06 ± 8.65 1020.50 ± 6.97 1088.60 ± 19.40 748.43 ± 9.77 1017.81 ± 13.86 853.54 ± 11.75

1481.17 ± 21.31 581.77 ± 5.12 801.08 ± 12.26 696.38 ± 7.56 631.12 ± 7.87 1018.14 ± 10.83 1091.38 ± 20.23 753.54 ± 8.58 1015.22 ± 14.01 848.13 ± 10.61

156.81 ± 1.80 133.33 ± 2.12 112.18 ± 3.33 98.35 ± 0.95 47.86 ± 0.36 69.31 ± 0.89 83.59 ± 0.54 52.38 ± 0.65 79.33 ± 0.72 57.88 ± 0.28

162.13 ± 2.14 130.77 ± 2.57 110.93 ± 2.91 98.56 ± 1.14 48.13 ± 0.47 70.01 ± 0.78 82.76 ± 0.83 51.85 ± 0.49 78.77 ± 0.56 58.21 ± 0.33

60.58 ± 0.38 75.11 ± 0.15 12.08 ± 0.21 64.11 ± 0.45 ND 3.28 ± 0.10 ND ND 31.08 ± 0.15 20.48 ± 0.31

59.88 ± 051 75.48 ± 043 11.86 ± 017 63.93 ± 0.63 ND 2.91 ± 0.34 ND ND 31.34 ± 0.24 19.71 ± 0.25

DC-SPE = dual column solid phase extraction. MAD = microwave- assisted digestion. ND = not detectable.

Although the results obtained by the proposed method were not significantly different to those obtained by the comparative method, dual-bed SPE method displays more advantages compared to microwave-assisted digestion method. For instance, the proposed method does not require the sample to be subjected to any drastic pretreatment such as concentrated acid heating [17]. Furthermore, the use of concentrated acids in microwave-assisted digestion could increase the blank metal values. In addition, the use of concentrated acids is not suitable for use ICP-OES. Therefore, a subsequent step could be necessary to dilute or remove the excess acid [15,31]. In contrast, the samples prepared by the proposed method are compatible with ICP-OES without further dilutions. In view of the above limitations of microwave digestion method, the dual-bed preconcentration method is advantageous because it minimizes the risks of incomplete mineralization of the organic matrix and cross-contamination [19]. Furthermore, in terms of sample through put, the dual-bed SPE method had a higher throughput (18 samples h1) compared to microwaveassisted digestion method (10 samples h1).

4. Conclusion A full factorial design used for optimization of dual-bed SPE column system allowed the establishment of optimum conditions for separation and preconcentration of metal ions in gasoline samples. In addition, factorial design helped in evaluating the interaction between the investigated factors and their effect on the analytical response (recovery). The optimum conditions for retention and elution of metal ions with respect to sample pH, eluent concentration and sample flow rate were as follows: (i) for Ag and Al-6.0, 2.5 mol L1 and 2.0 mL min1; (ii) for As-9.0, 1.0 mol L1 and 3.0 mL min1 and for Cr- 9.0, 4.0 mol L1 and 3.0 mL min1. The optimized dual-bed SPE procedure proved to be suitable for total

preconcentration of metal ions in gasoline samples. In addition, the preconcentration step permitted the elimination of the organic matrix, thus, avoiding the need for digestion of the samples before ICP-OES determination. The proposed method was applied in the determination of Ag, Al, As and Cr in ten real gasoline samples purchased from different filling stations in Johannesburg, South Africa. The dual-bed SPE method can be considered as a alternative sample pretreatment technique because it combines relatively low LOD and LOQ values obtained in the range 0.16–0.22 and 0.52– 0.76 lg L1, respectively, with higher sample throughput of 18 samples h1 and preconcentration factor. In addition, the current procedure avoids the use of concentrated acids, incomplete mineralization of organic matrix and it significantly reduces the laboratory wastes and analysis time, which is an important aspect for routine analysis. Acknowledgements The authors wishes to thank Sasol – South Africa (Grant Number IF 021/11-3) and National Research foundation (NRF) – South Africa (Grant Number SFH20110713000020772) for financial assistance and University of Johannesburg (Spectrau) for providing ICP-OES facilities. References [1] Cassella RJ, Brum DM, Lima CF, Caldas LFS, de Paula CER. Multivariate optimization of the determination of zinc in diesel oil employing a novel extraction strategy based on emulsion breaking. Anal Chim Acta 2011;690:79–85. [2] Donati GL, Amais RS, Schiavo D, Nobrega JA. Determination of Cr, Ni, Pb and V in gasoline and ethanol fuel by microwave plasma optical emission spectrometry. J Anal At Spectrom 2013;28:755–9. [3] Saint’Pierre TD, Dias LF, Maia SM, Curtius AJ. Determination of Cd, Cu, Fe, Pb and Tl in gasoline as emulsion by electrothermal vaporization inductively

P.N. Nomngongo, J.C. Ngila / Fuel 139 (2015) 285–291

[4]

[5]

[6]

[7]

[8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

[17]

[18]

[19]

[20]

coupled plasma mass spectrometry with analyte addition and isotope dilution calibration techniques. Spectrochim Acta Part B 2004;59:551–8. Sousa JKC, Dantas ANDS, Marques ALB, Lopes GS. Experimental design applied to the development of a copper direct determination method in gasoline samples by graphite furnace atomic absorption spectrometry. Fuel Process Technol 2008;89:1180–5. Collins GE, Morris RE, Wei J-F, Smith M, Hammond MH, Michelet V, et al. Spectrophotometric detection of trace copper levels in jet fuel. Energy Fuels 2002;16:1054–8. Sant’Ana FW, Santelli RE, Cassella AR, Cassella RJ. Optimization of an openfocused microwave oven digestion procedure for determination of metals in diesel oil by inductively coupled plasma optical emission spectrometry. J Hazard Mater 2007;149:67–74. Chaves ES, de Loos-Vollebregt MTC, Curtius AJ, Vanhaecke F. Determination of trace elements in biodiesel and vegetable oil by inductively coupled plasma optical emission spectrometry following alcohol dilution. Spectrochim Acta Part B 2011;66:733–9. de Souza RM, Meliande ALS, da Silveira CLP, Aucélio RQ. Determination of Mo, Zn, Cd, Ti, Ni, V, Fe, Mn, Cr and Co in crude oil using inductively coupled plasma optical emission spectrometry and sample introduction as detergentless microemulsions. Microchem J 2006;82:137–41. Souza RM, da Silveira CLP, Aucelio RQ. Determination of refractory elements in used lubricating oil by ICP-OES employing emulsified sample introduction and calibration with inorganic standards. Anal Sci 2004;20:351–5. Bettinelli M, Spezia S, Baroni U, Bizzarri G. Determination of trace elements in fuel oils by inductively coupled plasma mass spectrometry after acid mineralization of the sample in a microwave oven. J Anal At Spectrom 1995;10:555–60. Teixeira LSG, Rocha RBS, Sobrinho EV, Guimarães PRB, Pontes LAM, Teixeira JSR. Simultaneous determination of copper and iron in automotive gasoline by X-ray fluorescence after pre-concentration on cellulose paper. Talanta 2007;72:1073–6. Korn MDGA, dos Santos DSS, Welz B, Vale MGR, Teixeira AP, Lima DDC, et al. Atomic spectrometric methods for the determination of metals and metalloids in automotive fuels – a review. Talanta 2007;73:1–11. Ekanem EJ, Lori JA, Thomas SA. The determination of wear metals in used lubricating oils by flame atomic absorption spectrometry using sulphanilic acid as ashing agent. Talanta 1997;44:2103–8. Pereira JSF, Moraes DP, Antes FG, Diehl LO, Santos MFP, Guimarães RCL, et al. Determination of metals and metalloids in light and heavy crude oil by ICP-MS after digestion by microwave-induced combustion. Microchem J 2010;96: 4–11. Santos DSS, Korn MGA, Guida MAB, dos Santos GL, Lemos VA, Teixeira LSG. Determination of copper, iron, lead and zinc in gasoline by sequential multielement flame atomic absorption spectrometry after solid phase extraction. J Brazil Chem Soc 2011;22:552–7. Nomngongo PN, Ngila JC, Kamau JN, Msagati TAM, Moodley B. Preconcentration of molybdenum, antimony and vanadium in gasoline samples using Dowex 1–x8 resin and their determination with inductively coupled plasma–optical emission spectrometry. Talanta 2013;2013(110): 153–9. Wang T, Jia X, Wu J. Direct determination of metals in organics by inductively coupled plasma atomic emission spectrometry in aqueous matrices. J Pharm Biomed Anal 2003;33:639–46. Aguirre MA, Kovachev N, Hidalgo M, Canals A. Analysis of biodiesel and oil samples by on-line calibration using a Flow Blurring[registered sign] multinebulizer in ICP OES without oxygen addition. J Anal At Spectrom 2012;27:2102–10. Roldan PS, Alcântara IL, Padilha CCF, Padilha PM. Determination of copper, iron, nickel and zinc in gasoline by FAAS after sorption and preconcentration on silica modified with 2-aminotiazole groups. Fuel 2005;84:305–9. Escudero LA, Cerutti S, Olsina RA, Salonia JA, Gasquez JA. Factorial design optimization of experimental variables in the on-line separation/

[21]

[22]

[23]

[24]

[25]

[26]

[27] [28]

[29]

[30] [31]

[32]

[33]

[34] [35]

[36]

[37]

291

preconcentration of copper in water samples using solid phase extraction and ICP-OES determination. J Hazard Mater 2010;183:218–23. Bezerra MA, Santelli RE, Oliveira EP, Villar LS, Escaleira LA. Response surface methodology (RSM) as a tool for optimization in analytical chemistry. Talanta 2008;76:965–77. Tarley CRT, Silveira G, dos Santos WNL, Matos GD, da Silva EGP, Bezerra MA, et al. Chemometric tools in electroanalytical chemistry: Methods for optimization based on factorial design and response surface methodology. Microchem J 2009;92:58–67. de Jesus A, Zmozinski AV, Damin ICF, Silva MM, Vale MGR. Determination of arsenic and cadmium in crude oil by direct sampling graphite furnace atomic absorption spectrometry. Spectrochimica Acta Part B 2012;71– 72:86–91. Nomngongo PN, Ngila JC, Msagati TAM, Moodley B. Chemometric optimization of hollow fiber-liquid phase microextraction for preconcentration of trace elements in diesel and gasoline prior to their ICP-OES determination. Microchem J 2014;114:141–7. Nomngongo PN, Ngila JC. Determination of trace Cd, Cu, Fe, Pb and Zn in diesel and gasoline by inductively coupled plasma mass spectrometry after sample clean up with hollow fiber solid phase microextraction system. Spectrochim Acta Part B 2014;98:54–9. Kendüzler E, Yalçınkaya Ö, Baytak S, Türker AR. Application of full factorial design for the preconcentration of chromium by solid phase extraction with Amberlyst 36 resin. Microchim Acta 2008;2008(160):389–95. Gabrielsson J, Lindberg N-O, Lundstedt T. Multivariate methods in pharmaceutical applications. J Chemometrics 2002;2002(16):141–60. Kowalewska Z, Ruszczyn´ska A, Bulska E. Cu determination in crude oil distillation products by atomic absorption and inductively coupled plasma mass spectrometry after analyte transfer to aqueous solution. Spectrochimica Acta Part B 2005;60:351–9. Ferreira SLC, Queiroz AS, Fernandes MS, dos Santos HC. Application of factorial designs and Doehlert matrix in optimization of experimental variables associated with the preconcentration and determination of vanadium and copper in seawater by inductively coupled plasma optical emission spectrometry. Spectrochim Acta B 2002;57:1939–50. Stewart II, Olesik JW. Steady state acid effects in ICP-MS. J Anal At Spectrom 1998;13:1313–20. Becker E, Rampazzo RT, Dessuy MB, Vale MGR, da Silva MM, Welz B, et al. Direct determination of arsenic in petroleum derivatives by graphite furnace atomic absorption spectrometry: a comparison between filter and platform atomizers. Spectrochim Acta Part B 2011;2011(66):345–51. Brandão G, de Campos R, Luna A, de Castro E, de Jesus H. Determination of arsenic in diesel, gasoline and naphtha by graphite furnace atomic absorption spectrometry using microemulsion medium for sample stabilization. Anal Bioanal Chem 2006;385:1562–9. Vieira EG, Soares IV, Nl Dias Filho, Da Silva NC, Garcia EF, Bastos AC, et al. Preconcentration and determination of metal ions from fuel ethanol with a new 2,20 -dipyridylamine bonded silica. J Colloid Interf Sci 2013;2013(391): 116–24. de Oliveira A, de Moraes M, Neto J, Lima E. Simultaneous determination of Al, As, Cu, Fe, Mn, and Ni in fuel ethanol by GFAAS. At Spectrosc 2002;23:39–43. Cassella RJ, Brum DM, Robaina NF, Rocha AA, Lima CF. Extraction induced by emulsion breaking for metals determination in diesel oil by ICP-MS. J Anal At Spectrom 2012;27:364–70. Trindade JM, Marques AL, Lopes GS, Marques EP, Zhang J. Arsenic determination in gasoline by hydride generation atomic absorption spectroscopy combined with a factorial experimental design approach. Fuel 2006;85:2155–61. Saint’Pierre TD, Frescura VLA, Curtius AJ. The development of a method for the determination of trace elements in fuel alcohol by ETV-ICP-MS using isotope dilution calibration. Talanta 2006;68:57–962.