Use of Functionalized Resin for Matrix Separation and Trace Elements Determination in Petroleum Produced Formation Water by Inductively Coupled PlasmaMass Spectrometry

1 Instituto de Quı́mica, Universidade Federal do Rio de Janeiro, Avenida Athos da Silveira Ramos 149, Cidade Universitária, 21941-909 Rio de Janeiro, RJ, Brazil 2 Divisão de Quı́mica Anaĺıtica, Instituto Nacional de Tecnologia, Avenida Venezuela 82, 20081-312 Rio de Janeiro, RJ, Brazil 3 Laboratório de Quı́mica Anaĺıtica, Universidade Estadual do Sudoeste da Bahia, Rua José Moreira Sobrinho s/n, Jequié 45206-190, BA, Brazil


Introduction
Produced formation water (PFW) is a complex mixture consisting of formation water with injection water, solids, residual hydrocarbons, naturally occurring radioactive material, metals in the form of inorganic salts, and chemical residuals from the production.These residuals are obtained when sea water is injected into the deposit rock to help force oil to the surface and ultimately increase oil production.Once it is brought to the surface, the water is separated from the hydrocarbons and treated to remove oil, and additional chemicals are added to the produced formation water as needed.The effluent is then discharged into the sea or injected back into the wells or other suitable geological formations, with or without further treatment.Production chemicals used in the oil production process include corrosion inhibitors, scale inhibitors, biocides, oxygen scavengers, demulsifiers or deoilers, antifoamers, viscosifiers, hydrate inhibitors, cleaners and emulsion breakers.These substances also contribute to the modification of the chemical composition of the original matrix [1][2][3].
The volume of PFW at offshore oil platforms is, generally, several times greater than that of oil.Thus, offshore oil production frequently results in environmental discharge of drill cuttings and produced formation water contaminated with residuals of exploration and production, chemicals, and formation chemicals.Chemical substances present in produced formation water are the focus of management activity, including total hydrocarbon content (THC), aromatics (PAHs, BTEX, and naphthalene), organic acids and phenols; alkylphenols (two groups: C1-C3 and C4-C9); toxic metals (arsenic, lead, cadmium, copper, chromium, mercury, nickel, zinc); radionuclides; and production chemical residuals.Thus, toxic metal quantification in production formation waters is of environmental concern [3][4][5][6].
Inductively coupled plasma mass spectrometry (ICP-MS) is becoming a frequently employed analytical technique for its good analytical performance characteristics, such as low limits of detection, multielemental capabilities, wide dynamic range, and capacity to perform isotopic analysis with high sensitivity for most elements.However, using this technique to directly analyze saline samples, such as produced formation waters, can be very troublesome due to severe interferences caused by matrix components.
Although the detection capacity of ICP-MS permits the direct determination of trace metals at sub-ng mL −1 levels, spectroscopic (caused by polyatomic species) and nonspectroscopic (due to the high content of dissolved solids) interferences make the direct analysis of these samples difficult [7,8].
Recent studies have presented various pretreatment techniques to overcome the problems generated by complex matrixes in ICP-MS.In all of these techniques, the aim is to separate and/or preconcentrate the analyte in order to improve the detection limit and reduce matrix content, thereby enhancing analytical conditions.Much of the literature recommends the use of offline or online separation and preconcentration systems as means of reducing or eliminating spectroscopic and non-spectroscopic interferences in ICP-MS for analysis of samples with a high saline content [8][9][10][11].
Solid-phase extraction (SPE) is a very important technique for metal separation and its determination at trace levels.SPE has specific advantages, including simplicity, reliability, good metal loading capacity, and the ability to obtain high enrichment factors [12,13].Versatile systems have been obtained by applying solid supports loaded or functionalized with complexing reagents [14][15][16].These systems have been applied for preconcentration and determination of metals in a wide range of samples [17][18][19].A notable advantage of chemically functionalizing a resin is the reduction of leaching of the ligand from the support.This increases lifetime and it maintains good analytical features for the developed systems [20].Br-TAO has been synthesized for the first time and applied in an online cloud point extraction procedure for determination of Mn [21].
In this study, a multivariate technique (Doehlert design) was used to develop a procedure based on solid-phase extraction to separate and preconcentrate trace metals from produced formation water (PFW).These PFW samples were collected from offshore platforms in the Brazilian Northeast.Finally, metals determination were made using ICP-MS.This procedure is based on solid-phase extraction of metals in an offline minicolumn.The column was packed with Amberlite XAD-4 polymeric resin chemically modified with the 4-(5bromo-2 -thiazolylazo) orcinol (Br-TAO) reagent and trace metals were re-extracted using a nitric acid solution.

Experimental
2.1.Instrumentation.An X Series II quadrupole ICP-MS (Thermo Fisher Scientific, Bremen, Germany), with a Peltiercooled impact bead spray chamber and a concentric glass To correct analyte signal fluctuations, a mixture of internal standards (Sc, Ge, Rh, and In) was added to standards and samples in 50 μg L −1 concentrations.The trace metal determinations were carried out under the conditions presented in Table 1.This mixture is used as universal internal standard in our laboratory because it allows correcting the signal fluctuation of several elements.
A peristaltic pump C-6 XV (Alitea, Stockholm, Sweden) furnished with Tygon tubes was used to pump all solutions to preconcentration through the minicolumn loaded with the solid phase.

Reagents, Solutions, and Materials.
All reagents were prepared using ultrapure water (Elix and Synergy, Millipore, Bedford, MA, USA) and analytical grade chemicals.Nitric acid was of Suprapur quality (Merck, Darmstadt, Germany).New polyethylene recipients and graduate tubes were used in all steps, to avoid contamination.These were kept overnight in 10% nitric acid solution.Before use, these flasks were rinsed with ultrapure water and dried in a dust-free environment (laminar flow cabinet).
NaCl solutions were prepared from dissolution of the solid reagent (Merck, Germany) in ultrapure water.
A 1.0 mol L −1 TRIS (tris(hydroxymethyl)aminomethane) buffer solution (pH in the range 7 to 9) was prepared by dissolving 6.88 g of its sodium salt and concentrate nitric acid (about 2.5 mL) in ultrapure water to a final volume of 50 mL and pH adjusted with the same acid before to complete the volume in a volumetric flask.Acetate (pH 5.0) and citrate (pH 6.0) buffers also were used in the optimization step.
A 1.0 mol L −1 HNO 3 for re-extraction of trace metals from minicolumn was made from Suprapur quality.

Synthesis of the 4-(5 -Bromo-2 -thiazolylazo) Orcinol (Br-TAO).
For Br-TAO preparation, a modified procedure was performed, based on the synthesis of similar reagents [22,23].2-Amino-5-bromothiazole (2.0 g) was dissolved in 50 mL of a 6.0 mol L −1 hydrochloric acid solution at 0-5 • C. A solution of 0.53 g of sodium nitrite in 20 mL of water was added dropwise.The mixture was stirred and kept at 0-5 • C for 45 min.For coupling, 0.96 g of orcinol was added to 20 mL of 1.0 mol L −1 sodium carbonate solution and the mixture was cooled to 0-5 • C.This solution was added dropwise to the above diazotized compound at vigorous stirring.The system was allowed to stand overnight in a refrigerator at 0-5 • C. The product was filtered and rinsed with cold water.A dark-red precipitate resulted and was purified by recrystallization in ethanol with activated carbon [21].The solution of Br-TAO was prepared by dissolving 0.5 g of reagent in 250 mL of a 5% (w/v) NaOH solution.

Functionalization of the XAD-4 Resin with Br-TAO
Reagent.Amberlite XAD-4 (5 g) was treated with 10 mL of concentrated HNO 3 and 25 mL of concentrated H 2 SO 4 and stirred at 60 • C for 1 h in a water bath.The reaction mixture was then poured into an ice-water mixture.The nitrated resin was filtered, washed repeatedly with water until free acid was formed, and treated with a reducing mixture of tin (II) chloride, concentrated hydrochloric acid, and ethanol.The system was refluxed for 12 h at 90 • C. The precipitate was filtered and washed with water, 2 mol L −1 HCl and 2 mol L −1 NaOH.The amino resin was first washed with 2 mol L −1 HCl and finally with ultra-pure water to remove excess HCl.This was suspended in an ice-water mixture (150 mL) of 1 mol L −1 HCl and 1 mol L −1 NaNO 2 .The diazotized resin was filtered, washed with ice-cold water, and reacted with an alkaline solution of Br-TAO (as prepared in Section 2.3) at 0-5 • C for 24 h.The resulting dark-red resin was filtered, washed with water, and dried in air. Figure 1 present the proposed resin structure.

Column Preparation.
A cylindrically shaped polyvinyl chloride minicolumn (3.50 cm × 4.0 mm i.d.) was packed with 100 mg of the XAD-4/Br-TAO resin.Plastic foams were fixed at both ends of the minicolumn to prevent material losses.The minicolumn was initially washed with 2.0 mol L −1 nitric acid solution and deionized water until the effluent was free of acid.It was found that these minicolumns could be used repeatedly for at least 500 operational cycles without efficiency damage of extraction capacity.The efficiency of the column was controlled by verification of its extraction performance after this column was used by a 50 extraction/elution cycle.

Sample Collection.
Samples of produced formation waters were directly collected from offshore platforms in the Brazilian Northeast.Polyethylene flasks were used that had been previously decontaminated with nitric acid solution (10%) and rinsed with ultra-pure water.After sampling, the samples were filtered through a membrane (pore diameter of 0.45 μm) using a vacuum system to remove suspended particulate material.Afterwards, samples were acidified to pH 2 with nitric acid and stored at 6 • C.

Solid-Phase Extraction Optimization.
The solid-phase extraction optimization procedure for the five trace metals studied was carried out using a Doehlert design [24][25][26] involving three variables: sample pH, sampling flow rate, and buffer concentration.These variables were chosen due their great effect on the solid-phase extraction system as verified by previous experiments.The analytical response was counts/s.Experimental errors were estimated by triplicate analysis of the experiment, corresponding to the central point.5.0 μg L −1 Co, Cd, Pb, Ni, and Cu solutions with 10‰ salinity were used in all optimization experiments.
Responses for experimental design were obtained by evaluating the analytical signal of each element.To optimize a procedure that satisfies an efficient simultaneous extraction of all elements of interest, an overall response (OR) approach was adopted.The overall response is given by the following expression: where R(X n ) is the analytical signal to element n in a particular experiment and LS(X n ) is the maximum signal in the set of experiments for element n.The statistically significant effect of the variables was studied and interactions among them were evaluated by applying analysis of variance (ANOVA) using the Statistica software for Windows [27].2.8.General Procedure after the Optimization Process.For solid-phase extraction, 25 or 50 mL of the sample was adjusted to pH 8.0 using a TRIS buffer solution for a final concentration of 0.040 mol L −1 .This solution was then pumped through a minicolumn loaded with XAD-4/Br-TAO resin at 3.6 mL min −1 .After this step, the minicolumn was washed with TRIS buffer (5 mL) to eliminate residual salt.
Trace metals adsorbed on this solid-phase were re-extracted with HNO 3 solution (1 mol L −1 , 2 mL) and collected in graduated polyethylene flasks.A solution of internal standards (Sc, Ge, Rh, and In) was added to the eluate for a final concentration of 50 μg L −1 .The volume was set to 5 mL and the determination of the target elements was performed by ICP-MS.

Evaluation of Sample Salinity on the Solid-Phase Extraction Efficiency.
The sample salinity was then evaluated with respect to its influence on the solid-phase extraction and the residual salinity of the eluate on the nebulization process and its consequence on the measured signal.5 μg L −1 solutions of Co, Cd, Pb, Ni, and Cu, in the presence of different NaCl concentrations (salinities of 0, 10, 20, 30, 40, 50, 60, and 100‰), were submitted to the separation/preconcentration procedure described in Section 2.8 and recovery was evaluated.

Solid-Phase Extraction Optimization.
Optimization of the extraction process was carried out using Doehlert design as a multivariate methodology to obtain individual results for each element.There was an observed tendency for efficient solid-phase extraction in a common region for all metals of interest.For feasible optimization of extraction conditions for multiple elements in a single experiment, overall response was used.The overall response can be related to the five individual responses according to (1).The maximum overall response (OR) obtained was 4.96, in the set of thirteen experiments that compose the Doehlert matrix (Table 2).
A second-order polynomial model was fit to the overall response data, in order to depict the experimental region and determine the optimum conditions.Analysis of variance (ANOVA) was used to evaluate the resulting mathematical model, using a 95% confidence interval (Table 3).Comparison of the tabulated F value (Fischer distribution) and the value obtained by division of the media of square  of regression (MS reg ) by the media of square of residual (MS res ) confirms that the regression is significant (F tabulated > F calculated ).Additionally, comparison of the tabulated F value and the value obtained by division of the media of square of lack of fit (MS lof ) by the media of square of pure error (MS pe ) shows that the model is well adjusted to experimental data (F tabulated > F calculated ).
All variables showed a significant effect for the elements studied; however, sample pH and sampling flow rate were most significant.Specifically, the extraction procedure is favored when the sample pH is increased to a sufficiently high level and when the flow rate is decreased to the minimum level.Figure 2 shows profiles of these surfaces (as contour graphs) after fitting to a quadratic model for the overall Separation procedures, such as solid-phase extraction, can eliminate the complex matrix by transferring target analytes to a simpler solution before determination.Thus, solutions with salinities ranging from 0 to 100‰ (NaCl solutions) were extracted at the optimum conditions and the recovery for each trace metal was evaluated.The results suggest that salinities as high as 100% do not affect trace metals extraction showing that matrix interference has been eliminated in the separation step.

Analytical Features of the Method.
The analytical features of the proposed procedure were assessed after a variable optimization.The concentrations of the analytical standards for the preconcentration procedure ranged between 0.05 and 5 μg L −1 (unless for copper which was between 0.5 and 5 μg L −1 ).Without this preconcentration step, the concentrations of the calibration standards ranged between 1 and 100 μg L −1 .Parameters of analytical curve obtained after preconcentration/separation procedure are presented in Table 5.
The limit of detection (LOD) and the limit of quantification (LOQ) are defined as LOD = 3 σ/S and LOQ = 10 σ/S, respectively, where S is the slope of the analytical curve and σ is the standard deviation of ten consecutive measurements of the blank signal.For 50.0 mL of sample solution used in the preconcentration step, limits of detection of 0.72, 4.2, 3.0, 3.3, and 105 ng L −1 , limits of quantification of 2.4, 14, 10, 11, and 350 ng L −1 , and precision (repeatability) expressed as relative standard deviation (n = 10, 5.0 μg L −1 ) of 1.4, 1.3, 2.8, 0.9, and 1.9% were obtained for Co, Cd, Pb, Ni, and Cu, respectively.
The accuracy of the proposed procedure was assessed by analysis of the following certified reference materials: NRCC-SLEW-3 (Estuarine Water) and NRCC-SLRS-4 (Riverine Water).Results are presented in Table 4.The values obtained in this study agree well with the certified value reported for these certified reference materials.t-test paired was applied to evaluation of these results.At the 95% confidence level, the difference of the population means is not significantly different for both certified and found results.Recoveries between 88 and 109% were found for studied meals after spike test (Table 6).

Application.
The developed procedure was applied to four production formation waters obtained from offshore oil platforms in the Brazilian Northeast.These results are summarized in Table 6.The determination of trace metal concentrations is of environmental concern due to the large volume of water produced with oil (as much as 10 times the oil production) that is ultimately discharged into the sea.Brazilian legislation (CONAMA Resolution 357/2005) [29] establishes that the maximum concentrations allowed to be discharged in saline water class I (waters for more restrictive uses) are Cd 5 μg L −1 , Ni 25 μg L −1 , Pb 10 μg L −1 , and Cu 5 μg L −1 (Co concentration is not yet considered in the legislation).From Table 6 it is observed that the trace metal levels in the production formation waters analyzed did not exceed the Brazilian guidelines.The proposed method allows determining these five metals in lower limit of quantification than other methods previously published [4,28,30].Also an online solid-phase extraction offers some additional advantage, in comparison to other methods, such as reutilization of solid phase for several cycles, simplicity, elimination of matrix effect, and low reagent consumption.

Conclusions
Through application of Doehlert design conditions, the simultaneous separation and preconcentration of Co, Cd, Pb, Cu, and Ni were optimized.The influence of the sample salinities in ICP-MS was suppressed by matrix separation, employing a minicolumn of XAD-4/Br-TAO for solid-phase extraction.
The developed procedure for separation/preconcentration and determination of trace metals in saline waters was applied for both CRMs and produced formation waters, showing good accuracy.The produced formation water samples analyzed showed very low levels of trace metals of environmental concern.15-120 0-3.5 0-1.9 0.057-0.800 -3.7 0-5.5 * Analytes added for a final concentration of 0.100 μg L −1 in the sample unless for copper which final concentration was 0.5 μg L −1 ; * * Recovery from spike test in the PFW 4 sample; * * * Concentration and salinity range found in previous works [4,28].

Table 1 :
Instrumental parameters used for the analytes determination by ICP-MS after solid-phase extraction.

Table 2 :
Doehlert experimental design for solid-phase extraction optimization.
OR, overall response.

Table 3 :
Analysis of variance.
SS: sum of square; df : degree of freedom; MS: media of square; F: Fischer distribution.

Table 5 :
Parameters of analytical curve obtained after application of the developed preconcentration/separation procedure.

Table 6 :
Results for the petroleum produced formation waters analyzed by using the solid-phase extraction and ICP-MS measurements (μg L −1 , N = 3, mean ± standard deviation).