Solid phase extraction and sequential elution for pre-concentration of traces of Mn and Zn in analysis of honey by flame atomic absorption spectrometry

Abstract

A pre-concentration procedure based on solid phase extraction and two-step elution was developed for the determination of the total concentrations of traces of food safety relevant elements (Mn and Zn) in ripened honeys. Accordingly, 10% (m/v) solutions of analyzed honeys (100 ml) were passed at 20 ml min⁻¹ through resin beds of Dowex 50W × 8–400 to retain Mn and Zn ions and separate them from the glucose and fructose matrix. Afterwards, 20 ml of a 0.5 mol l⁻¹HNO₃ solution was used to elute K and Na, easily ionized interfering metals. Finally, Mn and Zn were recovered prior to measurements by flame atomic absorption spectrometry (FAAS) using 5 ml of a 2 mol l⁻¹HCl solution. This procedure was applied in the analysis of six honeys and resulted in the determination of 0.2–13.6 μg g⁻¹ of Mn and 0.2–1.2 μg g⁻¹ of Zn with the precision (n = 3) within 3–10% and an accuracy better than 5%. Detection limits of Mn and Zn with this pre-concentration/separation procedure and the FAAS detection were 4 and 3 ng g⁻¹, respectively.

---

Introduction

The knowledge about the composition of metals in honey appears to be of a primary interest since their presence affect its quality and safety.¹ Apart from the main origin of metals in honey related to the composition of the nectar and other sugar-rich liquids and secretions collected by bees, honey can be contaminated by environmental pollution present in the forage area or improper honey handling and processing practices.^2–5 The dependable evaluation of total metal concentrations is useful in order to predict not only the nutritive value of honey and its wholesomeness, but also to judge its authenticity and to trace an eventual fraud caused by forbidden practices during the production and finally, to identify potential health hazards due to the contamination exceeding admissible levels.

Flame atomic absorption spectrometry (FAAS) is commonly used for the analysis of honey due to low operational costs and reasonably good analytical performance.⁶ Measurements of alkali and alkaline earth metals, which are the most abundant mineral constituents of honey, typically impose high (from 50 to 500 times or more) dilutions of prepared honey solutions. At such conditions, chemical interferences occur in regularly used air–acetylene flames as the presence of high concentrations of cations and anions of honey can be completely eliminated or substantially reduced. As a result, such solutions could be directly analyzed by FAAS. On the other hand, minor and trace elements of food safety relevance like Cu, Fe, Mn and Zn, can be measured by FAAS only in undiluted sample solutions.^6,7 Particularly, the evaluation of the content of Mn and Zn in honey seems to be important since these elements catalyze the dehydration of fructose and glucose and the formation of hydroxymethylfurfural (HMF).⁸ Their concentrations are also highly correlated with the antioxidant activity of honey.⁹

According to the literature, due to matrix effects caused by the predominant organic matter present, fructose and glucose, and vast amounts of SO₄²⁻, PO₄³⁻ and other anions, in addition to cations of easily ionizable elements (EIEs: K and Na), honey samples are mandatorily mineralized before the measurements of Mn and Zn and other minor and trace elements.^{4,5,7,10–23} Accordingly, high temperature (450–700 °C) incinerations^{4,7,12,13,15,18,19,21–23} or wet oxidative digestions with the aid of concentrated reagents or their mixtures, i.e., HNO₃,¹⁸HNO₃ with HClO₄,⁵ or HNO₃ with H₂SO₄,¹⁷ are used. Both strategies enable the destruction of the carbohydrate-rich honey matrix and release simple ions of elements associated with this matrix into solution. As a result, non-spectral interferences impairing the aspiration, the nebulization and the transport of the honey sample solutions are eliminated.^6,7 Releasing buffers can subsequently be added to prepared sample solutions to minimize the effect of above mentioned matrix constituents in determinations of minor and trace elements.^7,12,14,23 Spectral interferences related to the occlusion of measured minor and trace elements into any refractory compounds or the formation of their less volatile species are minimized in this way as well.^{7,10,12,14,23}

Unfortunately, all the above mentioned procedures are cumbersome, tedious and time-consuming. They can also be responsible for losses of elements of interest due to the formation of their volatile compounds or gains due to different forms of contamination. The direct and non-destructive analysis of honey to determine the content of Mn, Zn and other elements using faster and less laborious and protracted approaches, are however quite uncommon.^7,10,24,25 Consequently, it seems reasonable to suppose that the development of alternative approaches to digestion procedures would be desirable for a crucial time reduction and an overall simplification of honey analysis. A possibility of the pre-concentration of trace elements prior to their accurate and precise determinations by FAAS would also be important.

With respect to this problem, the objective of this work was to develop a procedure based on solid phase extraction (SPE) and two-step sequential elution and evaluate its suitability for the enrichment of Mn and Zn and their separation from matrix constituents (K and Na ions, fructose and glucose) prior to the determination of these elements in honey samples using FAAS. The sorption behavior of K, Mn(II), Na and Zn(II) ions as well as fructose and glucose toward three commercially available strongly acidic cation exchangers, namely Amberlite IRP-69, Dowex 50W × 8–400 and Dowex HCR-W2, was studied at different solution pHs and flow rates across resin beds. For a selected resin, Dowex 50W × 8–400, conditions of the sequential elution aimed at recovering K and Na ions and then Mn(II) and Zn(II) ions prior to their suitable measurements by FAAS were thoroughly optimized. The reliability of a devised SPE pre-concentration/separation procedure was verified by comparing its results with those achieved with a wet oxidative digestion procedure. In addition, spiking experiments were conducted and respective recoveries were evaluated. The procedure was applied for the determination of traces of Mn and Zn in several raw mono- and multi-flower honeys.

Experimental

Apparatus and measurements

A single-beam Perkin Elmer model 1100 B atomic absorption spectrometer was used to determine concentrations of K, Mn, Na and Zn in solutions. The instrument was equipped with a wettable plastic coated burner mixing chamber, an end cup and a drain interlock assembly. A 10 cm long Ti single slot burner head was used for an air–acetylene flame. All solutions were nebulized using a stainless steel nebulizer. Concentrations of Mn and Zn were determined in an absorption mode (FAAS). In case of measurements of K and Na, the instrument was run in an emission mode (flame optical emission spectrometry, FOES). Absorbance and emission intensity readings were made using a time-average integration mode. Three readings, integrated at 0.1 s intervals over a 1 s integration time, were averaged in each read cycle. An internally mounted deuterium arc lamp (D₂) operated in a time-shared mode served as the continuum source for simultaneous background-corrected atomic absorption measurements. All measurements were performed using a very sharp fuel-lean flame. The spectrometer was operated under parameters given in Table 1.

Table 1 Operating conditions for FAAS (Mn, Zn) and FOES (K, Na) measurements fructose and glucose were used

	Mn	Zn	K	Na
a Instrumental detection limit (3 × SD of average intensity signals for a water blank, n = 5). b Expressed as RSD of average intensity signals (n = 3) for 0.10, 0.20, 0.50 and 1.00 mg l^−1 (Mn, Zn) and 0.02, 0.05, 0.10 and 0.20 mg l^−1 (K, Na).
Air flow rate, l min^−1	8.0	8.0	8.0	8.0
C2H2 flow rate, l min^−1	1.2	1.2	1.2	1.2
Lamp current, mA	20	15	7	10
Wavelength, nm	279.5	213.9	766.5	589.0
Slit, nm	0.2	0.7	0.4	0.2
Calibration range, mg l^−1	0.05–2.00	0.05–2.00	0.01–1.00	0.01–1.00
Detection limit,^a μg l^−1	7.8	6.3	0.4	1.1
Precision,^b %	1.4–5.4	0.9–6.1	0.5–5.6	1.0–6.7

---

A single-beam Spectronic 20D+ digital visible spectrophotometer from Thermo Scientific (Bremen, Germany) was used to determine the sum of concentrations of fructose and glucose according to the arseno–molibdate (Somogyi–Nelson) method.²⁶ In brief, 0.5 ml portions of an alkaline 6.0 g l⁻¹CuSO₄ solution were added to 10 μl sample portions containing both monosaccharides and then the resulting mixtures were incubated in a water bath at 90 °C for 10 min to reduce Cu(II) ions to Cu(I) ions. In turn, an arseno–molybdate reagent, prepared by reacting a 50 g l⁻¹(NH₄)₆Mo₇O₂₄ solution with a 6.0 g l⁻¹Na₂HAsO₄ solution in medium of a 0.8 mol l⁻¹H₂SO₄ solution, was added to produce a polymolybdate complex of an intense blue color due to the reaction with resulting Cu(I) ions. Absorbance readings of final solutions containing molybdenum blue produced in this reaction were taken at 520 nm. Sums of concentrations of fructose and glucose were determined against standard solutions of glucose at concentrations ranging from 1 to 50 mg l⁻¹.

Reagents and materials

Ultra-pure water from a WIGO (Wroclaw, Poland) PRO-11G reverse osmosiswater purification system was used in all preparations. Solutions of ACS grade concentrated reagents, i.e., 30% (m/m) H₂O₂, 37% (m/m) HCl and 65% (m/m) HNO₃, were purchased from J. T. Baker (Deventer, Netherlands). D-Fructose (>98%) and D-glucose (>98%) were supplied by POCh (Gliwice, Poland). Merck KGaA (Darmstadt, Germany) AAS Titrisol single-element standards containing 1000 mg of K (KCl in water), Mn (MnCl₂ in water), Na (NaCl in water) and Zn (ZnCl₂ in 0.06% HCl) were applied to prepare 1000 mg l⁻¹ single-element stock standard solutions. These solutions were used to prepare single-element standard solutions of K, Mn, Na and Zn for the calibration as well as 100 ml working standard solutions containing 20, 0.2, 5 and 0.2 mg l⁻¹ of K, Mn, Na and Zn, respectively, in addition to fructose and glucose, both at a concentration of 20 g l⁻¹. The composition of the latter solutions represented an average content of K and Na in addition to major monosaccharides in 5.0% (m/v) water solutions of Polish honeys.^27,28 These solutions were finally adjusted to pH 3.5, 4.0 and 4.5 with a 0.010 mol l⁻¹HCl solution and used to study the cation exchange behavior of elements of interest, fructose and glucose, toward the polymeric resins applied.

Strong cation exchange styrene–divinylbenzene resins with sulfonic acid functional groups, i.e., Amberlite IRP-69 (100–500 mesh size, Na⁺ form), Dowex 50W × 8–400 (200–400 mesh size, H⁺ form) and Dowex HCR-W2 (16–40 mesh size, H⁺ form), were supplied by Sigma-Aldrich (Saint Louis, MO, USA). These resins were packed into Sigma-Aldrich glass columns (1.0 cm inner diameter) with glass coarse frits and polytetrafluoroethylene (PTFE) stopcocks. Cole-Parmer (Vernon Hill, IL, USA) 4-channel MasterFlex L/S peristaltic pumps were applied to control and measure flow rates of solutions that were loaded onto SPE columns and passed through resin beds.

SPE column preparation and operation

SPE columns were filled with water slurries of resins as received (1.0 g). Circles of tightly-fitting cellulose filter paper were placed on the top of the resin beds formed. Resin beds were washed with 10 ml of a 2.0 mol l⁻¹HCl solution at 5.0 ml min⁻¹ and afterward rinsed with 50 ml of water to remove the excess of HCl.

The cation exchange behavior of K, Mn(II), Na and Zn(II) ions in addition to fructose and glucose was studied by passing working standard solutions (pH 3.5, 4.0 and 4.5) at 2.0 ml min⁻¹ through SPE columns packed with certain resins and subsequently analyzing effluent portions (10 ml) collected at the end of the solution loading. Concentrations of K, Mn, Na and Zn retained by the resins were assessed by subtracting concentrations of these elements found in effluents (matrix matching standard solutions containing 20 g l⁻¹ of fructose and glucose were used) from their initial concentrations in working standard solutions loaded onto columns. Rationing these values to the original contents of elements in loaded working standard solutions, respective retention efficiencies of K, Mn, Na and Zn (in %) were assessed. In the case of fructose and glucose, separation efficiencies (in %) were evaluated by directly relating concentrations of these monosaccharides determined in effluents to their initial concentrations in loaded working standard solutions.

In a similar way, the effect of flow rates (in the range from 2.0 to 20 ml min⁻¹) with which working standard solutions were driven through SPE columns on retention efficiencies of K, Mn, Na and Zn, and separation efficiencies of fructose and glucose was examined. In this case, working standard solutions (pH 4.0) were passed through SPE columns and respective effluents were collected prior to the analysis.

To evaluate the suitability of 0.2 and 0.5 mol l⁻¹HCl and HNO₃ solutions for the removal of K and Na from the cation exchange resin Dowex 50W × 8–400 and the elimination of matrix effects in FAAS measurements, working standard solutions (pH 4.0) were first driven through SPE columns at 20 ml min⁻¹. Then, four 10 ml portions of the aforementioned HCl or HNO₃ solutions were passed at 2.0 ml min⁻¹ through resin beds to elute K and Na ions. 10 ml portions of respective eluates were consistently collected and concentrations of K, Mn, Na and Zn were determined using simple water standard solutions. Finally, effectiveness of 1.0, 2.0, 3.0 and 4.0 mol l⁻¹HCl solutions applied to recover retained Mn(II) and Zn(II) was tested. Accordingly, working standard solutions (pH 4.0) were loaded onto SPE columns with the Dowex 50W × 8–400 resin and then K and Na were eluted by passing through 20 ml of a 0.5 mol l⁻¹HNO₃ solution. Afterwards, 5 ml of given HCl solutions were used to recuperate Mn and Zn. Respective 5 ml portions of eluates were collected and subjected to analysis on the content of elements of interest by FAASversus matrix matching standard solutions (containing corresponding amounts of HCl). Recovery efficiencies (in %) for K, Mg, Na and Zn were calculated relating concentrations of these elements determined in respective eluates to their original concentrations in working standard solutions.

All retention, separation and recovery efficiencies were mean values for six independent replicates. Relevant column blanks were run and considered in the final results.

Honey samples and their preparation before analysis by FAAS

Six samples of freshly ripened honeys, including acacia, heather, goldenrod, multi-flower, lime and rape, were directly taken from an apiary located in suburbs of Wroclaw (Lower Silesian Province, Poland). Honeys were kept in the original glass jars used for their shipping.

Three different sample preparation procedures were applied to determine total concentrations of Mn and Zn in analyzed honeys by FAAS. For each one, three independent analyses were made and respective procedural blanks were prepared, measured, and considered in the final results.

In the longest preparation procedure (the wet oxidative digestion with a time investment of about 240 min), 2.5 g samples of honey were placed in 250 ml beakers treated with 10 ml of a concentrated HNO₃ solution and digested under watch glasses for about 150 min. After that, solutions were cooled down and 10 ml of a 30% (m/m) H₂O₂ solution was added. The digestion was prolonged for further 60 min so as to completely decompose the added H₂O₂ and reduce volumes of sample solutions to approximately 1–2 ml. Resulting aliquots were re-constituted with water and diluted to 50 ml. To avoid the splashing of sample solutions, the temperature of the hot plate was limited to 85–95 °C.

In the shortest preparation procedure (the water dissolution with a time investment of about 15 min), 2.5 g samples of honey were placed in 100 ml beakers, dissolved in about 10 ml of water and finally diluted with water to 50 ml.

In both latter cases, the resulting honey solutions were analyzed by FAAS against simple water standard solutions.

In case of the determination of Mn and Zn concentrations using the proposed alternative to the wet digestion of SPE-based pre-concentration/separation procedure with sequential elution (a time investment of about 40 min), 10 or 5.0 g samples of honey were placed in 100 ml beakers, dissolved in 10 ml of water and finally diluted to 100 ml. The resulting 10 or 5.0% (m/v) honey solutions were loaded onto SPE columns and filled with the resin Dowex 50W × 8–400 (length of the resin bed of 2.0 cm) at 20 ml min⁻¹ to retain Mn and Zn ions and separate them from monosaccharides and anionic minerals. After that, 20 ml of a 0.5 mol l⁻¹HNO₃ solution was subsequently driven through the columns at 5 ml min⁻¹ to elute K and Na. At the end, 5 ml of a 2 mol l⁻¹HCl solution at 2 ml min⁻¹ was used to completely recover retained Mn and Zn prior to measurements of their concentrations by FAASversus matching matrix standard solutions.

Results and discussion

Cation exchange behavior of polymeric resins toward K, Mn, Na and Zn ions and fructose and glucose

The sorption behavior of strongly acidic cation exchange resins Amberlite IRP-69, Dowex 50W × 8–400 and Dowex HCR-W2 toward K, Mn(II), Na and Zn(II) ions and fructose and glucose were studied as described in the section “SPE column preparation and operation”. On the basis of measurements taken on column effluents (Table 2) it was possible to conclude that all studied cation exchangers quantitatively retain Mn(II) and Zn(II) ions and separate them from fructose and glucose which were not retained by these resins. This enables the pre-concentration of elements of interest prior to their measurements by FAAS. Unfortunately, at the studied pH 3.5–4.5, K and Na ions, present in working standard solutions at much higher concentrations as compared to concentrations of Mn and Zn ions, were also completely retained by these cation exchange resins.

Table 2 Percentage retention (K, Mn, Na and Zn) and separation (fructose with glucose) efficiencies evaluated for strong cation exchange resins under pH 3.5–4.5. 100 ml working standard solutions containing 20, 0.2, 5 and 0.2 mg l⁻¹ of K, Mn, Na and Zn, respectively, and 20 g l⁻¹ of fructose and glucose were used. Average values (n = 3) with figures behind denoting standard deviation

pH	K	Mn	Na	Zn	Fructose and glucose^a
a Sum of concentrations of fructose and glucose: 40 g l^−1. b Concentrations of Mn and Zn in effluents collected were below respective DL values.
Amberlite IRP-69
3.5	99.8 ± 0.2	>95.8^b	98.6 ± 4.6	>96.8^b	101.2 ± 3.2
4.0	99.9 ± 0.1	>95.8^b	97.5 ± 2.8	>96.8^b	99.8 ± 4.9
4.5	99.7 ± 0.2	>95.8^b	100.0 ± 0.1	>96.8^b	102.2 ± 1.3
Dowex 50W × 8–400
3.5	100.1 ± 0.1	>95.8^b	100.0 ± 0.1	>96.8^b	101.4 ± 2.2
4.0	99.9 ± 0.1	>95.8^b	100.0 ± 0.2	>96.8^b	99.1 ± 3.8
4.5	100.0 ± 0.1	>95.8^b	100.0 ± 0.2	>96.8^b	98.2 ± 1.1
Dowex HCR-W2
3.5	99.7 ± 0.2	>95.8^b	99.5 ± 0.5	>96.8^b	97.1 ± 5.7
4.0	99.5 ± 0.3	>95.8^b	99.6 ± 0.7	>96.8^b	101.6 ± 2.1
4.5	99.6 ± 0.1	>95.8^b	98.5 ± 2.0	>96.8^b	95.6 ± 2.8

---

It was also established that fructose and glucose do not affect the cation exchange of K, Mn(II), Na and Zn(II) ions on the studied resins when concentrations of both these monosaccharides in working standard solutions are higher, i.e., 40 or even 80 g l⁻¹, as could be in case of 10 or 20% (m/v) honey sample solutions, respectively. Accordingly, it was found that at the solution pH within 3.5–4.5, sums of fructose and glucose concentrations determined in collected effluents corresponded to sums of their original concentrations in solutions loaded onto SPE columns. These eluates were either found not to contain K, Mn(II), Na and Zn(II) ions or concentrations of these elements were lower than (Mn, Zn) or close to (K, Na) respective detection limits (DLs) assessed for FAAS and FOES, respectively (Table 1).

The effect of varying flow rates with which working standard solutions are driven through SPE columns was examined in the range of 2–20 ml min⁻¹. Unfortunately, in case of the resin Amberlite IRP-96, it was not possible to apply flow rates higher than 14 ml min⁻¹ because a strong resistance in the flow was observed, possibly due to a very small particle size of this resin, i.e., 25–180 μm. From analyzing column eluates collected in this experiment, it was established that all studied resins maintain the exhaustive retention of Mn(II) and Zn(II) ions and their complete separation from fructose and glucose during the sample loading step. Correspondingly, concentrations of both elements of interest in analyzed column eluates were found to be lower than respective DL values, while sums of the concentrations of both monosaccharides were practically the same as in working standard solutions. The retention of K and Na in these conditions was also quantitative and changed from 99.5 to 100.0% as referred to their original concentrations in working standard solutions.

In comparison to the other studied cation exchange resins, i.e., Dowex 50W HCR-W2 and Amberlite IRP-96, the resin Dowex 50W × 8–400, with a particle size of 38–74 μm, had the most uniform particle distribution that supported the most uniform flow of solutions through SPE columns. In addition, the fine particle size of this resin reduced the time required for the equilibration as compared to Dowex 50W HCR-W2 but was not too short as in case of Amberlite IRP-96 that could result in the resistance in the solution flow at higher solution flow rates. For these reasons this resin was selected for further experiments.

Conditions of sequential elution: separation of K with Na and recovery of Mn with Zn

Because EIEs, i.e., K and Na ions, are retained on the resin Dowex 50W × 8–400 together with Mn(II) and Zn(II) ions, the next step of this work was to find out the optimal conditions for the sequential elution of these two groups of elements. Considering the cation exchange behavior of monovalent metal ions on strong cation exchange resins that was assessed in a former work,²⁹ it was presumed that diluted solutions of common mineral acids (HCl and HNO₃) would enable the recovery of K and Na ions without stripping divalent metal ions, which are normally retained on this type of resins much more strongly.^30,31

Accordingly, 0.2 and 0.5 mol l⁻¹HCl and HNO₃ solutions were passed in four separate 10 ml portions through resin beds of Dowex 50W × 8–400 (at 5.0 ml min⁻¹) with retained K, Mn(II), Na and Zn(II) ions to elute K and Na ions. Respective 10 ml eluates were collected and analyzed on concentrations of K, Mn, Na and Zn. Results of these experiments are given in Table 3 and relate to percentage elution efficiencies of elements of interest. As can be seen, the most desirable conditions of the quantitative elution of K and Na ions from resin beds of Dowex 50W × 8–400 were ensured by using a 0.5 mol l⁻¹HNO₃ solution. With 20 ml of this solution, 97.8 ± 6.1% of total K and 100.8 ± 5.0% of total Na retained were recovered from SPE columns and separated from Mn and Zn, which under these conditions remained unaffected. It was found that concentrations of the latter elements in all eluates collected were lower that their DLs assessed with FAAS.

Table 3 Percentage elution efficiencies evaluated for the strong cation exchange resin Dowex 50W × 8–400. 100 ml working standard solutions (pH 4.0) containing 20, 0.2, 5 and 0.2 mg l⁻¹ of K, Mn, Na and Zn, respectively, in addition to 20 g l⁻¹ of fructose and glucose were loaded onto resin beds. 10 ml portions (I, II, III and IV) of eluting HCl and HNO₃ solutions were applied. Average values (n = 3) with standard deviations in brackets

K					Mn				Na					Zn
I	II	III	IV	Sum	I	II	III	IV	I	II	III	IV	Sum	I	II	III	IV
a Concentrations of Mn and Zn in effluents collected were below respective DL values.
0.2 mol l^−1HCl
0.1 (0.1)	0.1 (0.1)	3.5 (3.1)	22.6 (3.4)	26.3 (4.6)	<0.4^a	<0.4^a	<0.4^a	<0.4^a	1.3 (0.7)	74.8 (3.8)	21.7 (3.7)	0.4 (0.3)	98.2 (5.4)	<0.3^a	<0.3^a	<0.3^a	<0.3^a
0.5 mol l^−1HCl
4.1 (2.3)	85.0 (4.1)	7.6 (2.1)	0.4 (0.1)	97.1 (5.1)	<0.4^a	<0.4^a	<0.4^a	<0.4^a	94.4 (3.3)	8.4 (6.8)	1.3 (0.1)	0.9 (0.8)	105.0 (7.6)	<0.3^a	<0.3^a	<0.3^a	<0.3^a
0.2 mol l^−1HNO3
0.5 (0.4)	2.6 (2.2)	28.5 (1.8)	27.3 (7.4)	58.9 (7.9)	<0.4^a	<0.4^a	<0.4^a	<0.4^a	15.0 (3.1)	64.3 (4.4)	20.2 (0.9)	7.0 (3.8)	106.5 (6.6)	<0.3^a	<0.3^a	<0.3^a	<0.3^a
0.5 mol l^−1HNO3
25.5 (5.8)	72.3 (2.0)	0.8 (4.0)	0.3 (0.2)	98.9 (7.3)	<0.4^a	<0.4^a	<0.4^a	<0.4^a	95.3 (2.3)	5.5 (4.4)	0.3 (0.2)	0.5 (0.4)	101.6 (5.0)	<0.3^a	<0.3^a	<0.3^a	<0.3^a

---

Next the conditions of the exhaustive stripping of Mn(II) and Zn(II) ions prior to measurements of their concentrations by FAAS were evaluated. For that purpose, 100 ml working standard solutions (pH 4.0) were loaded onto resin beds of Dowex 50W × 8–400 at 20 ml min⁻¹ and then 20 ml of a 0.5 mol l⁻¹HNO₃ solution were passed through them at 5.0 ml min⁻¹ to recover K and Na ions. Finally, 1.0, 2.0, 3.0 and 4.0 mol l⁻¹HCl solutions were tested with the purpose of the complete recovery of Mn and Zn. Because 5 ml portions of these solutions were used, their flow rates were limited to 2.0 ml min⁻¹.

Except for a 1.0 mol l⁻¹HCl solution, it was established that remaining solutions produce quantitative recoveries of both elements of interest, i.e., 99.0 ± 2.8% (Mn) and 101.0 ± 3.7% (Zn) in case of a 2.0 mol l⁻¹HCl solution, 103.4 ± 4.0% (Mn) and 103.2 ± 6.1% (Zn) for a 3.0 mol l⁻¹HCl solution and 106.3 ± 1.8% (Mn) and 102.0 ± 6.7% (Zn) in case of a 4.0 mol l⁻¹HCl solution. Considering the results obtained, 20 ml of a 0.5 mol l⁻¹HNO₃ solution and 5 ml of a 2.0 mol l⁻¹HCl solution were applied for the two-step elution in the devised SPE pre-concentration/separation procedure. This procedure was more than 3 times faster than the one recently described for the pre-concentration of traces of Li using Dowex 50W × 8–200 before the determinations of this element in honey by FOES.²⁹

Analytical characteristics

Accuracy of the proposed procedure was evaluated by the recovery of known amounts of Mn(II) and Zn(II) ions added to 10% (m/v) solutions of rape, goldenrod and heather honeys.

Three different levels of concentrations were used, i.e., 0.02, 0.05 and 0.10 mg l⁻¹ of Mn and Zn, respectively, corresponding to additions of 0.20, 0.50 and 1.00 μg g⁻¹ of these elements. Recoveries obtained were in the range of 95.2–102.3% for Mn and 96.7–104.8% for Zn which proves the accuracy of the devised pre-concentration/separation procedure. Respective DL values for Mn and Zn assessed with the proposed procedure and FAAS detection were expressed as concentrations corresponding to absorbance signals of 3 standard deviations (3σ criterion) of blanks, i.e., 100 ml working standard solutions only containing K, Na, fructose and glucose. These DLs were found to be 4 and 3 ng g⁻¹, respectively, in case of Mn and Zn. The usual reusability of the resin bed was 15 times.

Application

The developed SPE pre-concentration/separation procedure with two-step sequential elution was applied for the determination of Mn and Zn in six different mono- and multi-flower honeys. Accordingly, 10 or 5.0% (m/v) solutions of analyzed honeys were prepared and subjected to the whole procedure described. Two other procedures of sample preparation, namely the decomposition in concentrated HNO₃ with the admixture of 30% (m/m) H₂O₂ and the dissolution in water, were both followed by the analysis of the resulting 5.0% (m/v) honey solutions, were executed on the same honey samples and results of their analyses are compared in Table 4.

Table 4 Concentrations (in μg g⁻¹) of Mn and Zn in raw honeys determined by FAAS in sample solutions resulting from different sample preparation procedures. Values of statistical tests calculated for compared mean results achieved with the wet digestion or the water dissolution and the proposed SPE with sequential elution are in brackets. Average values (n = 3) with figures behind denoting standard deviations (SDs). NA Not applied

Honey	Wet digestion with HNO3 and H2O2	Dissolution in water	Dissolution in water and SPE with sequential elution
a The C-test was applied (Ccritical = 4.303). b The t-test was applied (tcritical = 4.303).
Mn
Acacia	0.34 ± 0.11 (+0.632)^a	<0.16 (NA)	0.29 ± 0.02
Goldenrod	0.52 ± 0.23 (+0.858)^a	0.64 ± 0.12 (+3.022)^a	0.38 ± 0.02
Heather	13.39 ± 0.26 (−0.780)^b	13.12 ± 0.16 (−2.066)^b	13.59 ± 0.36
Lime	0.32 ± 0.09 (+0.767)^a	0.39 ± 0.11 (+1.518)^a	0.27 ± 0.02
Multi-flower	0.28 ± 0.08 (+1.930)^a	<0.16 (NA)	0.17 ± 0.01
Rape	<0.18 (NA)	<0.16 (NA)	0.25 ± 0.01
Zn
Acacia	<0.14 (NA)	<0.12 (NA)	0.28 ± 0.02
Goldenrod	0.64 ± 0.17 (−0.410)^a	0.59 ± 0.11 (−1.519)^b	0.69 ± 0.03
Heather	1.78 ± 0.41 (+1.862)^a	1.37 ± 0.12 (+1.681)^b	1.23 ± 0.08
Lime	0.67 ± 0.20 (−0.555)^a	0.72 ± 0.12 (−0.411)^b	0.75 ± 0.04
Multi-flower	0.80 ± 0.15 (+0.092)^a	0.75 ± 0.10 (−0.664)^b	0.79 ± 0.03
Rape	0.25 ± 0.10 (+0.555)^a	<0.12 (NA)	0.21 ± 0.02

---

As can be seen, the dissolution of honey samples in water, although very easy and fast to handle, was inadequate for the measurement of very low concentrations of Mn and Zn in acacia, multi-flower and rape honeys. Concentrations of both elements in these honeys were lower than respective DLs assessed to be 0.16 and 0.12 μg g⁻¹ for Mn and Zn, respectively. Previously, the suitability of the direct determination of several heavy metals, including Mn and Zn, in the routine analysis of honey using inductively coupled plasma optical emission spectrometry (ICP-OES) was reported by Ioannidou et al.³² In the cited work, a 2% (m/v) concentration of honey in analyzed water sample solutions was considered to be a compromise for the sufficient sensitivity and the lack of matrix effects in the plasma related to the presence of carbohydrates. However, since a high temperature excitation source such as ICP was used, the detectability of the direct analysis of honey described in this paper was better than assessed here for FAAS, e.g., for Mn it was 0.2 μg l⁻¹ in the sample solution and 0.01 μg g⁻¹ in the solid sample of honey. Unfortunately, the analysis of more concentrated solutions by FAAS, e.g., 10% (m/v), was found to suffer from strong matrix effects.

The most commonly applied technique in the literature is wet oxidative digestion and the analysis of solutions of diluted digests has been established to be prone to low precision (expressed as relative standard deviation, RSD), which in case of very low concentrations of studied elements in analyzed honeys typically ranged from 28 to 44% (Mn) and from 19 to 40% (Zn). By contrast, the proposed SPE pre-concentration/separation procedure offered RSD values within 3–10%, likely as a consequence of the 50- or 100-fold pre-concentration of Mn and Zn as compared to the wet digestion procedure.

According to the F-test applied at the 95% significance level, standard deviations (SDs) of results obtained for Mn and Zn by the wet oxidative digestion of all honeys (the only exception was Mn in heather honey) were higher than those of the results achieved using the devised SPE pre-concentration/separation procedure with the sequential elution. For that reason, the significance of differences between mean concentrations of Mn and Zn determined using these two sample preparation procedures was tested by the Cochran and Cox test (C-test) with the critical value (C_critical) at the 95% significance level of 4.303.³³ In case of the results obtained for Mn in heather honey, the t-test was used with the critical value (t_critical) at the 95% significance level of 4.303 as well. In a similar way, results obtained with the aid of the direct analysis of water dissolved honey samples were compared with those obtained with the proposed SPE pre-concentration/separation procedure. Depending on whether the difference between SD values of individual results was statistically significant or not, the C-test or the t-test were correspondingly applied. Calculated values of both statistical tests are given in Table 4. The lack of the difference between compared values of tests used (C_calculated < C_critical, t_calculated < t_critical) suggests that all sample preparation procedures provide identical results, although the devised SPE pre-concentration/separation procedure with the sequential extraction is much faster, more precise and enables the accurate determination of very low concentrations of both studied elements.

In addition, it was verified that the acidification of prepared honey solutions, aimed at dissociating complexing ions of Mn and Zn with different organic ligands and releasing their simple ions, was not necessary in this procedure. Analyzing acidified (to a final concentration of 0.1 mol l⁻¹HNO₃) and boiled, and non-acidified samples of honeys, it was found that differences between the results of total concentrations of Mn and Zn are statistically insignificant according to the t-test carried out at the 95% significance level. Such a behavior could be explained by strong interactions of sulfonic functional groups of the strong cation exchange resin Dowex 50W × 8–400 with complexing ions of Mn and Zn, as described earlier in the case of tea infusions,^34–36 or even the mechanical retention of the aforementioned ions on the gel-type cation exchange resin applied.

Conclusions

This paper reports on the development of an SPE-based procedure for the pre-concentration of traces of Mn and Zn and their separation from fructose and glucose in addition to K and Na ions in the analysis of honey. The proposed procedure combines the cation exchange on a gel-type strongly acidic cation exchanger with a two-step plain elution and has been demonstrated to be a faster and more straightforward alternative to the time-consuming and inconvenient wet oxidative digestion procedures. It takes about 40 min and this includes the preparation of resin beds and honey sample solutions, the subsequent resin-based pre-concentration and the sequential elution before measurements of concentrations of elements of interest by FAAS in the absence of main honey matrix constituents. With this procedure proposed, 10 and 5.0% (m/v) honey solutions can be analyzed and the elements of interest can be enriched by 20 times before measurements of their concentrations. Hence, it is possible to precisely and accurately determine Mn and Zn in different mono- and multi-flower honeys at levels within 10⁻¹–10¹ μg g⁻¹.

Acknowledgements

This project is implemented with the co-financing of the European Union within the European Social Fund. The National Science Centre is acknowledged for receiving the funding for the research project carried out by a pre-doctoral researcher starting a scientific career.

Notes and references

1. N. Bilandzic, M. Dokic, M. Sedak, B. S. Kolanovic, I. Varenina, A. Koncurat and N. Rudan, Food Chem., 2011, 128, 1160.
2. A. Pisani, G. Protano and F. Riccobono, Food Chem., 2008, 107, 1553.
3. M. Tuzen, S. Silici, D. Mendil and M. Soylak, Food Chem., 2007, 103, 325.
4. I. Bratu and C. Georgescu, J. Cent. Eur. Agric., 2005, 6, 467.
5. M. N. Rashed and M. E. Soltan, J. Food Compos. Anal., 2004, 17, 725.
6. P. Pohl, I. Sergiel and H. Stecka, Crit. Rev. Anal. Chem., 2009, 39, 276.
7. I. Lopez Garcia, P. Vinas, C. Blanco and M. H. Cordoba, Talanta, 1999, 49, 597.
8. P. Vit, A. Rodriguez-Malaver, C. Rondon, I. Gonzalez, M. L. di Bernardo and M. Y. Garcia, Arch. Latinoam. Nutr., 2010, 60, 405.
9. O. O. Anam and R. K. Dart, Anal. Proc., 1995, 32, 515.
10. J. S. dos Santos, N. S. dos Santos, M. L. P. dos Santos, S. N. dos Santos and J. J. de Jesus Lacerda, J. Braz. Chem. Soc., 2008, 19, 502.
11. M. Kucuk, S. Kolayli, S. Karaoglu, E. Ulusoy, C. Baltaci and F. Candan, Food Chem., 2007, 100, 526.
12. J. Lachman, D. Kolihova, D. Miholova, J. Kosata, D. Titera and K. Kult, Food Chem., 2007, 101, 973.
13. K. A. Osman, M. A. Al-Doghairi, S. Al-Rehiayani and M. I. D. Helal, Int. J. Food Agric. Environ., 2007, 5, 142.
14. G. Downey, K. Hussey, J. D. Kelly, T. F. Walshe and P. G. Martin, Food Chem., 2005, 91, 347.
15. F. Erbilir and O. Erdogrul, Environ. Monit. Assess., 2005, 109, 181.
16. O. M. Hernandez, J. M. G. Fraga, A. I. Jimenez, F. Jimenez and J. J. Arias, Food Chem., 2005, 93, 449.
17. V. Nanda, B. C. Sarkar, H. K. Sharma and A. S. Bawa, J. Food Compos. Anal., 2003, 16, 613.
18. O. Celechovska and L. Vorlova, Acta Vet. Brno, 2001, 70, 91.
19. A. M. G. Paramas, J. A. G. Barez, R. J. G. Villanova, T. R. Pala, R. A. Albajar and J. S. Sanchez, J. Sci. Food Agric., 2000, 80, 157.
20. M. J. Latorre, R. Pena, S. Garcia and C. Herrero, Analyst, 2000, 125, 307.
21. M. J. Latorre, R. Pena, C. Pita, A. Botana, S. Garcia and C. Herrero, Food Chem., 1999, 66, 263.
22. H. Yilmaz and O. Yavuz, Food Chem., 1999, 65, 475.
23. A. Uren, A. Serifoglu and Y. Sarikahya, Food Chem., 1998, 61, 185.
24. P. Pohl and I. Sergiel, Microchim. Acta, 2010, 168, 9.
25. I. Sergiel and P. Pohl, J. Agric. Food Chem., 2010, 58, 7497.
26. E. Fournier, Mono- and Oligosaccharides, in Handbook of Food Analytical Chemistry. Water, Proteins, Enzymes, Lipids and Carbohydrates, ed. R. E. Wrolstad, E. A. Decker, S. J. Schwartz and P. Sporns, John Wiley & Sons, New York, 2005.
27. M. Chudzinska and D. Baralkiewicz, Food Chem. Toxicol., 2010, 48, 284.
28. M. Madejczyk and D. Baralkiewicz, Anal. Chim. Acta, 2008, 617, 11.
29. H. Stecka and P. Pohl, J. Braz. Chem. Soc., 2010, 22, 677.
30. C. E. Harland, Ion Exchange. Theory and Practice, Royal Society of Chemistry, Cambridge, 1994.
31. A. A. Zagorodni, Ion Exchange Materials. Properties and Applications, Elsevier, Amsterdam, 2007.
32. M. D. Ioannidou, G. A. Zachariadis, A. N. Anthemidis and J. A. Stratis, Talanta, 2005, 65, 92.
33. J. N. Miller and J. C. Miller, Statistics and Chemometrics for Analytical Chemistry, Pearson Education, Harlow, 2005.
34. K. E. Odegard and W. Lund, J. Anal. At. Spectrom., 1997, 12, 403.
35. Y. Ozdemir and S. Gucer, Food Chem., 1998, 61, 313.
36. P. Pohl and B. Prusisz, Food Chem., 2007, 102, 1415.

---