Chem. Anal. (Warsaw), 53, 905 (2008)
Determination of Total Mercury by Vapor Generation
In Situ Trapping Flame Atomic Absorption Spectrometry+
by Henryk Matusiewicz* and Magdalena Krawczyk
Poznañ University of Technology, Department of Analytical Chemistry
ul. Piotrowo 3, 60-965 Poznañ, Poland
Keywords:
Mercury; Reference materials; In situ trapping; Hydride generation; Flame
atomic absorption spectrometry
The analytical performance of non-chromatographic coupled hydride generation, integrated
atom trap (HG–IAT) atomizer flame absorption spectrometry (FAAS) systems were evaluated for the determination of total mercury in environmental samples. Mercury, using formation of mercury vapors were atomized in air-acetylene flame-heated IAT. A new design of
vapor generation integrated atom trap flame atomic absorption spectrometry (VG–IAT–
FAAS) hyphenated technique that would exceed the operational capabilities of existing
arrangements was investigated. This novel approach enables to decrease the detection limit
down to low pg mL–1 levels. The concentration detection limit, defined as 3 times the blank
standard deviation was 0.4 ng mL–1. For a 120 s in situ pre-concentration time (sample
volume of 2 mL), sensitivity enhancement compared to flame AAS, was 750 folds for Hg,
using vapor generation-atom trapping technique. The sensitivity can be further improved by
increasing the collection time. The precision, expressed by RSD, was 9.3% (n = 6) for Hg.
Reference and real sample materials were analyzed. The accuracy of the method was verified by the use of certified reference materials and by aqueous standard calibration
technique.The measured Hg content, in reference materials, were in satisfactory agreement
with the certified values. The hyphenated technique was applied for mercury determinations in coal fly ash, sewage and water.
Opracowano metodê oznaczania rtêci pozwalaj¹c¹ na zwiêkszenie mo¿liwoœci analitycznych
p³omieniowej absorpcyjnej spektrometrii atomowej (FAAS) przez po³¹czenie techniki
generowania wodorków (HG) oraz systemu zintegrowanego ³¹cz¹cego nasadkê szczelinow¹
z kolektorem rurkowym (IAT). Pary rtêci ulega³y atomizacji w systemie zintegrowanym
ogrzewanym za pomoc¹ p³omienia powietrze-acetylen. Zaproponowano i zbadano now¹
technikê sprzê¿on¹ VG–IAT–FAAS, która zwiêksza mo¿liwoœci analityczne wczeœniej
zastosowanych systemów. Opracowana metoda pozwala na obni¿enie granicy wykrywalnoœci
* Corresponding author. E-mail: [email protected]
+ Dedicated to Professor Rajmund Dybczyñski on the occasion of his 75th birthday.
906
H. Matusiewicz and M. Krawczyk
do poziomu pg mL–1. Stê¿eniowa granica wykrywalnoœci, zdefiniowana jako trzykrotna
wartoœæ odchylenia standardowego œlepej próby, wynosi³a 0.4 ng mL–1. Wspó³czynnik
zatê¿ania Hg wynosi³ 750, po zatê¿aniu próbki o objêtoœci 2 mL, przez 120 s za pomoc¹
techniki sprzê¿onej VG-IAT. Czu³oœæ metody mo¿na poprawiæ przez wyd³u¿enie czasu
zatê¿ania. Precyzja, wyra¿ona jako wzglêdne odchylenie standardowe wynosi³a 9.3%
(n = 6). Dok³adnoœæ opracowanej metody sprawdzono oznaczaj¹c Hg w certyfikowanych
materia³ach odniesienia standardow¹ technik¹ kalibracyjn¹ . Otrzymane wyniki by³y zgodne
z wartoœciami certyfikowanymi rtêci. Opracowan¹ technikê sprzê¿on¹ zastosowano do
oznaczania rtêci w popio³ach lotnych, œciekach i wodzie.
The detection of mercury at trace levels is a complex analytical task because of
its specific physical and chemical properties and is one of the most difficult analytical
problems in atomic absorption spectrometry (AAS). The most sensitive mercury resonance line lies in the vacuum UV region and is therefore not suitable for conventional
spectrometers. The less sensitive wavelength at 253.7 nm is the only available alternative. However, the detection limit of conventional flame AAS (FAAS) is not sufficiently low to accurately determine of mercury, at trace levels, because of the limited
atom number density in the light path associated with poor nebulization and atomization. A FAAS technique gives only a characteristic concentration of 5 mg L–1 for 1%
absorption and a detection limit of 0.2 mg L–1 at 253.7 nm and therefore, it is of little
use for the determination of mercury [1]. Additionally chemical separation of mercury from the sample matrix is difficult to its high volatility. Therefore, in most cases
a pre-concentration and vapor and hydride generation procedures, before the element
can be determined, is essential. Critical reviews [2, 3] have focused on the methodologies for the determination at trace levels of mercury. Of all the methods for the
determination of mercury, atomic absorption spectrometry and atomic fluorescence
spectrometry (AFS) especially after a cold vapor generation technique and hydride
generation step, is the most useful and sensitive technique. This focused on the use of
three reductants, SnCl2, Na(K)BH4 and HCOOH and provided information on LOD
and tolerance to interferences.
It is well established that significant improvements in the limits of detection of
FAAS and graphite furnace atomic absorption spectrometry (GF–AAS) may be
achieved by chemical vapor generation, mostly hydride generation (HG) and in-atomizer (silica or graphite tube) trapping (pre-concentration) as reviewed [4–6] for the
determination of As, Bi, Ge, In, Pb, Sb, Se, Sn, Te and Tl (as well as Hg).
Atom trapping approaches that involve the collection and in situ pre-concentration of mercury using a dual „bent” tube atom trap (Au was used as a coating material) was first demonstrated by Ellis and Roberts [7]. Sensitivity of the technique was
0.0598 mg L–1 for a 2 min collection. A comparison of the atom trapping technique
and cold vapor method was carried out and the results were found to be in close
agreement.
Determination of total mercury by vapor generation in situ trapping FAAS
907
In situ trapping permits a significant enhancement in sensitivity for batch and
continuous hydride generation approaches used in the ultra-trace determination of
volatile hydride species. Due to its importance, in situ trapping, which allows the
coupling of hydride generation to integrated atom trap (IAT) [8–13] was chosen for
this study.
An analytical system was developed to trap and pre-concentrate Bi from the vapor
phase stream. Bismuthine formed by sodium tetrahydroborate reduction was trapped
on a tungsten coil previously heated to 270°C. The analyte species were re-volatilized by increasing the coil temperature to 1200°C and then transported to an externally heated silica T-tube by using a mixture of argon and hydrogen as the carrier gas
[14]. Korkmaz et al. [15] investigated the nature of re-volatilization from atom trap
surfaces in flame by AAS. Analytes Au, Bi, Cd, Mn and Pb were trapped on a watercooled, U-shaped silica trap or a slotted silica tube trap and re-volatilized by organic
solvent aspiration. They concluded that, although heating was not necessarily associated with re-volatilization, direct contact between the flame and the active silica surface was required. Recently [16] the analytical performance of three trap systems
(water-cooled U-shaped silica trap, water-cooled U-shaped silica trap combined with
slotted silica tube and slotted tube trap) for flame AAS were evaluated for determination of Cd and Pb in waters. Guo and Guo [17] reported SeH2 collection at gold wire
heated to 200°C situated in a quartz tube atomizer (AFS or AAS detection) with
separate inlet for argon. A successful trapping of PbH4 in a bare quartz tube was
announced by Korkmaz et al. [18], who also suggested that the same trap could be
used also for other hydrides. Recently, a preliminary evaluation of quartz tube trap for
collection of SbH3 and for volatilization of trapped analyte with subsequent atomization in a multiple microflame quartz tube atomizer for AAS was presented [19]. Kratzer
and Dedina extended their investigation of stibine trapping in quartz tube traps [20]
to stibine collection (and subsequent analyte atomization) in conventional quartz tube
atomizers [19]. They employed the simplest possible experimental arrangement; just
the commercially available externally heated quartz tube atomizer without any trap
or additional heating device. Further, a modification of the externally heated quartz
tube atomizer, making possible in situ trapping of bismuthine and subsequent analyte
atomization for Bi was described [21]. Krejèi et al. [22] investigated the collection of
Sb and Bi on a molybdenum foil strip situated in a laboratory-made quartz T-tube,
single-slot burner head, following hydride generation. Recently [23], the analytical
performance of a miniature quartz trap coupled with electrochemical hydride generator for antimony determination was described. A portion of the inlet arm of the conventional quartz tube atomizer was used as an integrated trap medium for on-line preconcentration of electrochemically generated hydrides. Very recently [24], a novel
quartz device has been designed to trap arsine and selenium hydride and subsequently
to volatilize the collected analyte and atomize it for AAS detection. The device is
908
H. Matusiewicz and M. Krawczyk
actually the multiple microflame quartz-tube atomizer with inlet arm modified to
serve as the trap and to accommodate the oxygen-delivery capillary used to combust
hydrogen during the trapping step. Ertas et al. [25] demonstrated in situ trapping of
lead hydride species on the inner walls of a flame-heated quartz tube under highly
oxidizing flame conditions followed by the release of trapped species by introducing
MIBK.
In this work we described a conceptually similar approach of trapping mercury
vapor (mercury vapor was generated in a laboratory built hydride generators) using
an integrated atom trap system for flame AAS that is applicable to determination of
total Hg in real and certified reference materials. The hydride generation technique
brings the FAAS method closer to the detection limits of HG–GF–AAS (in situ trapping technique) [26]. Finally, in order to check the accuracy of the proposed system,
analysis of certified reference materials were performed. The hyphenated technique
was satisfactorily applied to different types of samples such as coal fly ash, sewage
and water.
EXPERIMENTAL
Spectrometer
A Carl Zeiss Jena (Jena, Germany) Model AAS3 flame atomic absorption spectrometer [27] equipped
with a 10 cm air-acetylene burner head assembly and an IBM–PC compatible computer was used throughout
the study. The sampling rate for the PMT signal was 10 Hz. Signals were processed with in-house software
(Turbo Pascal Version 7) to extract the transient peak heights, area and peak time. NARVA Hg hollow
cathode lamp was used as the radiation source. No background correction was required in this mode of
operation. Operating parameters of the AAS instrument are summarized in Table 1 after appropriate optimization.
Table 1.
Instrumental operating conditions for determination of mercury by IAT–FAAS
(Continuation on the next page)
Determination of total mercury by vapor generation in situ trapping FAAS
Table 1.
a
909
(Continuation)
Nebulizer uptake rate, 5 mL min–1. Air flow rate 475 L h–1, acetylene flow rate 50 L h–1 (fuel-lean flame);
10 cm slot burner.
Hydride generation systems
Hydride generation was accomplished in the batch (discontinuous) mode and in a continuous mode
using a manually controlled two channel peristaltic pump (Gilson, Model Minipuls 3, France). A mass flow
controller with a precision pressure gauge (Models ERG 500 and ERG 2000, power supply Model ERG 2M,
DHN, Warsaw, Poland) were used to regulate the purge and transfer gas flow rates accurately and reproducibly.
Even the hydride generation cells served as gas-liquid separators, in an attempt to completely reduce
water vapor, a quartz gas-liquid separator identical to that described earlier [28] was installed between
the chemifolds and the nebulizer/spray chamber.
Batch hydride generation. Analyte vapors were generated in a laboratory-made Pyrex cell (batch
system, the total volume of the glass cell was about 80 mL) [29] into which SnCl2 was introduced with
a peristaltic pump. The evolved vapors were stripped from the solution and swept into the air-acetylene
nebulizer/burner-IAT system with an air purge gas. The cell assembly and the sequence of operations have
been described in a previous paper [29].
Continuous-flow hydride generation. Vapor generation was accomplished in the continuous-flow
mode using a system similar to that described previously [30].
The operating conditions for batch and continuous-flow vapor generation atomic absorption spectrometry are summarized in Table 2. A schematic diagram of the batch and continuous-flow VG–IAT–FAAS
system with in situ pre-concentration in the IAT unit is shown in Figure 1.
Figure 1. Schematic diagram of the VG–IAT–FAAS system
910
H. Matusiewicz and M. Krawczyk
Determination of total mercury by vapor generation in situ trapping FAAS
Table 2.
911
Optimized operating conditions for determination of Hg by VG–IAT–FAAS
Atom trapping techniques
Three designs of atom trap were investigated. Since the atom trap systems (trapping medium) was
described in detail in previous papers [8, 31] this will not be discussed again here, but briefly summarized
only.
A double-slotted quartz tube (STAT) was installed on a standard 10 cm air-acetylene burner. The design
permits changeover from analysis with the IAT [8] to that for a conventional flame in a few seconds.
A water-cooled single silica tube (WCAT) atom trap was arranged, as previously described [8] and
mounted on a 10 cm burner in such a manner in order to permit the system to be vertically and laterally
adjusted to the flame. The tubes were of an O.D. 3 mm and an I.D. of 1 mm for water cooling.
An integrated atom trap (IAT) was designed and constructed in this laboratory [8] it consisted of
a combination of a WCAT and a STAT [8]. The IAT system was mounted over an air-acetylene burner on
a mounting bracket, which permitted calibrated movement both vertically and horizontally.
A modified cooling system was used for cooling the water [9]. The continuously flowing cooling water
kept the surface of the silica tube at the temperature below 100°C. This allowed the analyte atoms to condense on the surface of the tube.
912
H. Matusiewicz and M. Krawczyk
The single silica and slotted quartz tubes were coated with palladium to prevent devitrification and
to improve the silica surface properties (increasing the surface area) and adsorption efficiency by continuous
aspiration via the burner nebulizer of 0.1% palladium solution, for 5 min; then tubes were conditioned for
several times in the flame. The tubes were re-coated after approximately 100 runs.
Gases and reagents
Compressed air and argon gases of N–50 purity (99.999%) obtained from BOC GAZY (Poznañ,
Poland) were employed as the carrier gas (air) for the nebulizer/burner unit and purification agent, respectively, without further purification. Compressed medical purity acetylene (Cezal, Poznañ, Poland) was used
as the source of air-acetylene flame.
Standard solutions of Hg were prepared from a 1000 mg L–1 Hg atomic absorption standards (Titrisol grade, Merck, Darmstadt, Germany). All working standard solutions of Hg were prepared daily to prevent any possible species change, by diluting appropriate aliquots of the stock solution in high-purity water.
Sodium tetrahydroborate(III) and potassium tetrahydroborate(III) used as reducing solutions, were prepared daily, or more frequently if required, by dissolving proper amounts of NaBH4 and KBH4 (pellets)
(Alfa Inorganics, Ward Hill, USA) in high-purity water and stabilizing with 0.1% (m/v) NaOH (Suprapur,
Merck, Darmstadt, Germany) solution to decrease its rate of decomposition, and was used without filtration.
Tin(II) chloride solution: a 1% (m/v) solution of SnCl2 in 2 mol L–1 HCl was used. The solution
was gently sparged for 30 min with argon in order to minimize the Hg0 content. Only clear, colorless reducing solution was used.
The quartz tubes were coated with 1000 mg L–1 Au, Pd and Pt solutions (Titrisol grade, Merck, Darmstadt,
Germany). The mixtures of Pd (700 mg L–1) or Pt (700 mg L–1) and Au (300 mg L–1) were prepared before
use.
All mineral acids (HNO3, HCl, HF) and hydrogen peroxide 30% (v/v) of the highest quality (Suprapur,
Merck, Darmstadt, Germany) were used. High-purity water: deionized water (model DEMIWA 5 ROSA,
Watek, Czech Republic), and doubly distilled water (quartz apparatus, Bi18, Heraeus, Hanau, Germany)
were used throughout the experiments.
Certified reference materials and samples
Validation of the method described in this work was performed using three certified reference materials. The following materials were chosen: SRM 1633a (Coal Fly Ash) NIST, IAEA/W-4 (Simulated Fresh
Water) and GBW 07601 (Human Hair) from National Research Center for CRM’s, Beijing, China.
The certified reference values are available for mercury for assessment of the method accuracy. All solid
reference materials were used as bottled, without further grinding and sieving.
The following real samples were used in this study: the untreated waste water and coal fly ash was
sampled from Poznañ Coal-fired Power Plant in Poland.
To ensure homogeneity, it was necessary to grind the real, solid sample in an agate pestle and mortar by
manual grinding of coal fly ash and by a vibrational mixer mill Model S (Testchem, Pszów, Poland) equipped
with 30 mL grinding chamber and rod (6 cm diameter), all made of tungsten carbide.
Microwave digestion system
A laboratory-built prototype of high pressure-temperature focused microwave heated digestion system,
equipped with a closed TFM–PTFE vessel (30 mL internal volume) [32] was employed for wet-pressure
sample digestion.
Determination of total mercury by vapor generation in situ trapping FAAS
913
Method development
The whole analytical procedure consists of various steps: (1) closed wet digestion of the samples,
(2) generation of the Hg vapors and its in situ trapping (collection) in an IAT system, (3) flame atomization
of collected vapors and (4) measurement by FAAS.
Microwave-assisted high pressure Teflon bomb digestion. Preparation of all standards and digestions of all samples were conducted under typical laboratory conditions. The microwave-assisted pressurized digestion technique was used for biological and environmental samples [32].
Approximately 500 mg of powdered inorganic reference material and sample (Coal Fly Ash,) were
placed in the TFM–PTFE vessel („bomb”) of the microwave digestion system and moistened by 1 mL of
30% H2O2; then 3 mL of concentrated HNO3 and 1 mL of concentrated HF were added. The samples were
heated for 15 min at 150 W. After dissolution, the clear digested solution was transferred into 10 mL calibrated flask and diluted to volume with water. When working with organic material (Human Hair), approximately 250 mg of sample was first moistened by 1 mL of 30% H2O2, then 3 mL of concentrated HNO3 was
used. 10 mL of sewage sample was moistened by 1 mL of 30% H2O2, then 3 mL of concentrated HNO3 was
added. The samples were heated for 10 min at 100 W. Before further analysis these were appropriately
diluted depending on the concentration level of the element. In all cases, a corresponding blank was also
prepared according to the above microwave-assisted digestion procedure.
Simplex optimization procedure. In the optimization that was done in this study, the one-factor-at-a-time method was used to obtain a satisfactory condition for the analysis. It took this method a long time,
however, not only to find the optimization condition but also to determine whether or not an optimization
condition better than the one that was found existed. To avoid these drawbacks the simplex optimization
approach was undertaken in this study to establish, for mercury, the best conditions for vapors generation,
transport and atomization. The parameters optimized are listed in Table 3, along with the ranges over which
optimization experiments were possible and conducted. In practice, the ranges were judiciously selected for
each parameter in turn, taking into account the practical problems of maintaining a stable absorbance signal.
Table 3.
Optimum operating conditions for VG–IAT–FAAS measurement of Hg obtained by simplex and
univariate methods
Simplex optimization experiments were performed using a software package obtained from the University of Plymouth. The optimization was carried out using aqueous standard solutions of element (mercury)
914
H. Matusiewicz and M. Krawczyk
determined. Net S/B ratio was taken as the criterion of merit. Some preliminary univariate experiments
(searches) were performed prior to the simplex optimization in order to establish the boundaries of the
values of each parameter. Three measurements for each variables were conducted at the factor of interest.
Between each experiment, a blank corrective experiment was run to ensure stable and repeatable results.
The optimum conditions obtained from this procedure were then used to run standard mercury solutions
and quantify the mercury present in the samples.
Vapor generation and procedure for in situ trapping. The procedures for vapor generation of samples
from trapping to atomization are outlined below and were not conducted in a clean laboratory environment.
Vapor generation was accomplished using two different hydride generators, the continuous unit and
the batch system (cell) [30]. In brief, the mercury vapors were generated continuously or in batch mode and
were introduced into the IAT system by the carrier gas (air) during the vapor-trapping step of the atomizer
temperature program only; this step was 120 s in duration in all experiments.
Total mercury. After preparation of solid samples, all the mercury present in the sample will be obtained in solution in the +2 oxidation state (all species will be converted to Hg(II) as the main oxidation product).
Continuous-flow generation measurements of volatile mercury vapors (Hg0) were studied using
the system shown schematically in Figure 1. 2 mL aliquot sample: 0.2 mL volume of 32% HCl solution and
1.8 mL volume of water were placed in a quartz vessel. The PVC/PTFE transfer line from the reaction cell
was placed in the nebulizer/burner system. The Hg sample was being continuously introduced at a rate of
1 mL min–1 to merge with a 1.0% (m/v) solution of SnCl2 (flow rate 1.0 mL min–1). The merging solution
feeds the gas-liquid separator to the IAT system. Mercury vapor that evolved was transferred (carrier air
flow rate, 100 mL min–1) to the IAT system, where analyte species (Hg0) were trapped and collected onto
the palladium-coated quartz tube surface. A continuous flow of cooling-water permitted Hg volatile species
to condense on the surface of the tube during vapors collection. The liquid phase was being continuously
removed to waste after neutralization with 0.1% NaOH solution.
Vapor generation was also accomplished using the hydride generation batch system as described previously [30]. In brief, the mercury vapors were generated in batch mode and were introduced into the IAT
system by the carrier gas (air) during the mercury vapor-trapping step of the atomizer temperature program
only; this step was 120 s in duration in all experiments.
Mercury vapors were generated from 10 mL volume of samples (batch mode). The SnCl2 solution (1.0% m/v) was pumped for 60 s and the vapors that evolved were transferred (carrier air flow rate,
125 mL min–1) to the IAT system, where they were collected onto the palladium-coated quartz tube.
The merging solution feeds the gas-liquid separator to the IAT system. A continuous flow of cooling water
permitted analyte vapors (atoms) to condense on the surface of the tube during vapors collection; a further
60 s air purge of the generator completed the transfer process.
Flame atomization of collected vapors. After collection, vacuum water pump was then turned on
to rapidly remove the cooling water from the quartz tube. The tube rapidly heated in the flame, generating
a transient atomic absorption signal as the analyte atoms were released from the surface. Finally, the analytes
were atomized for 7 s at feasible temperature of about 1330°C.
VG–IAT–FAAS analysis. After completion of the vapor generation and collection stage, the analyte
was vaporized and atomized for 7 s by heating the quartz tube up to ca 1330°C. The integrated transient
absorbance signals peak area were measured for the Hg line. Both peak height and peak area signals were
recorded. Peak area absorbance signals were used for calculations. A simplex optimization approach was
undertaken to establish, the best conditions for vapors generation, transport, in situ trapping and vaporiza-
Determination of total mercury by vapor generation in situ trapping FAAS
915
tion/atomization. Analytical blanks were also carried through the entire procedure outlined above, in order
to correct possible contaminants in the reagents that were used for the sample preparation. The mean blank
value, if necessary, was substracted from the sample value after all calculations. The system was manually
operated during the experiments. Quantification of Hg was based on aqueous standard calibration curves
(external calibration). All detection limits were calculated for raw unsmoothed data based on a 3σ criterion
of the blank counts.
RESULTS AND DISCUSSION
The study included an investigation of the vapor generation with in situ trapping
(pre-concentration, collection), the thermal vaporization-atomization into the FAAS,
and its application to practical analysis. The optimization parameters affecting the
efficiency of the vapor generation, trapping, atomization and analysis technique will
be discussed separately.
Effect of palladium on the trapping efficiency of mercury vapors
When mercury was determined without any stabilizing agents the sensitivity was
found to be very low. Therefore, application of a modifier should be realized by
coating modifier on to the quartz tube. Initial experiments were carried out to evaluate the different coating materials. The gold, palladium, platinum and their alloys
have been used to improve the sensitivity of a mercury. These coating elements were
examined for their suitability to determine Hg. It was thought that the gaseous mercury may become trapped in the matrix of the coating material allowing Hg to be
determined. This did occur and improvement in sensitivity was observed (Fig. 2).
A possible explanation is that enough amalgam is formed and therefore sufficient Hg
is trapped for an increase in sensitivity to be observed. With all investigated metals
and alloys, the lowest sensitivity was observed in the presence of gold; the best sensitivity was achieved in the presence of Pd, when introduced into the quartz tube alone.
Therefore, the trapping efficiency of vapors onto the Pd-coated quartz tube was found
to be optimum for Hg vapors.
Based on the above experiments it is supposed that palladium is not removed
from the quartz surface and remains on the quartz tube surface acting as an efficient
permanent chemical modifier. The amount of Pd remaining on the surface decreases
rapidly with increasing number of firings up to about 10 firings. After 10 firings
palladium remained in the quartz tube as a permanent modifier (adsorber agent).
The lifetime of the Pd coating was ca 100 quartz tube firings without significant
changes in reproducibility of Pd absorbances.
916
H. Matusiewicz and M. Krawczyk
Figure 2. Collection efficiency of mercury on quartz tube surface. 1000 mg L–1 Au, Pd and Pt solutions
were used; the mixtures of Pd (700 mg L–1) or Pt (700 mg L–1) and Au (300 mg L–1). 10 ng mL–1 Hg
Continuous-flow system vs batch mode for vapor generation
First, the generation efficiency greatly depends on whether sodium or potassium
tetrahydroborate and tin chloride is used. It was found that Na(K)BH4 were a weaker
reductants compared to SnCl2. For this reason, the tin chloride was ultimately preferred because a better generation efficiency and detection limit were obtained (Fig. 3).
Therefore, SnCl2 was selected for further experiments in order to ensure, with greater
certainty, that all of the volatile species had been stripped from solution and transferred into the atom trap unit.
The hydride vapor generation systems that we were using for these investigations
allow us to select from two basic modes of operation, a continuous-flow and a batch
mode. Preliminary experiments were performed by FAAS using aqueous standard
solutions of mercury. It was determined that the generation efficiency of the volatile
Hg vapors did not greatly depend on the modes and the type of reactors used. However, the continuous-flow system appeared to offer some advantages, such as easy
automation and, in combination with trapping of the generated mercury vapor in
a IAT, the use of an almost unlimited sample volume simply by pumping sample
solution for a correspondingly long period of time, increasing the sensitivity accordingly. In addition, continuous-flow vapor generation ensures rapid, efficient mixing
of the sample and reductant, thereby reducing the required reaction time, so lengthy
purge times are not necessary as the system is a closed one. Traditional „batch” type
VG methods require large purge and/or reaction times to ensure the system is flushed
of air and an efficient hydrogenation reaction.
Determination of total mercury by vapor generation in situ trapping FAAS
917
), NaBH4 (
) and KBH4 (
) concentration on the peak area
Figure 3. Influence of SnCl2 (
signals of Hgtot (5 ng mL–1) in batch system. The experimental conditions employed are detailed
in the Experimental Section
Process blank
Blanks using the continuous mode system were determined using the same treatment procedure as for samples (microwave-assisted sample digestion procedure).
Although these experiments were conducted in an ordinary laboratory, the detectable
source of the blank was determined experimentally to be the reductant solution and,
in particular, the hydrochloric acid used to stabilize the tin chloride. Even though
the chemicals used were of the best quality available, trace amounts of the analytes in
the reagent were pre-concentrated in the silica tube during the in situ trapping, resulting in blank signals. Using the continuous mode system and SnCl2 as reductant,
absolute blank of 0.6 ng for Hgtot, was achieved.
Simplex optimization of operational variables
The optimized vapor generation conditions are given in Table 1 and 2.
The stability of the flame is obviously controlled by the gas flow rate, it is therefore essential to keep the flame stable in the section of the atomizer. The effects of
flame conditions on the trapping and release of the mercury were studied by varying
the fuel flow rate. The influence of the flame condition on the signal intensity was
investigated by fixing the air flow rate (475 L h–1) and altering the acetylene flow
rate. The best sensitivity was obtained by using a 50 L h–1 of flow rate (lean flame) for
acetylene. The absorbances of Hg was not very different for collecting or releasing
in a lean, full-rich and stoichiometric flame.
918
H. Matusiewicz and M. Krawczyk
The water-cooled silica tube trap position was not optimized in our experiments,
but was selected based upon previous experience [8]. The optimum position of the
trap tube (single silica tube and STAT) corresponded to the distance (gap) of 5 mm
above the burner, and the position of the silica tube corresponded to obscuration of
about one-third of the light beam by the upper part of the tube. No significant differences were found in the absorbances when a coolant water flow rate of 1–4 L min–1
were used during the collection cycle of Hg vapors.
The trapping time is one of the most important factors concerning the sensitivity
of VG–IAT–FAAS technique. In time-based pre-concentration system the sample
loading time value indicates the pre-concentration time of the method and reflects the
enrichment factor. It was demonstrated (although not shown) that, although the relationship is not in general linear, a longer trapping time increased the analytical signal.
Although sensitivity is not linear with respect to trapping time the calibration graphs
are linear for any single trapping time. A reasonable trapping time per sample in
a routine laboratory would be about 2 min, in order to ensure, with greater certainty,
that all of the Hg vapors had been stripped from the solution and transferred into the
atomizer and as a compromise between medium sample consumption, high sensitivity and sufficient sampling frequency. This time was chosen to investigate the analytical performance of the VG–IAT–FAAS system with respect to linearity, sensitivity,
precision and detection limit.
To optimize the sample and reductant flow rate, first the optimum flow for Hg
was estimated in the range of 0.5–3 mL min–1 (Fig. 4a). In this study a 1 mL min–1
sample and reductant flow rate was chosen.
The continuous hydride generation system was optimized by varying the concentration of SnCl2, HCl and the air carrier flow that regulate volatilization and transport.
However, this information was available from a review [4] pertaining to trace element detection by atom trapping and in situ pre-concentration for FAAS and on our
own observations. All of the factors show more or less significant effect.
The efficiency of the generation of mercury vapors in hydrochloric acid and
nitric acids was investigated. When using HNO3 the Hg signals were found to be
about 20 to 30% lower than those obtained with HCl which is the most appropriate
acid to use. The effect of HCl concentration on the peak area absorbance is illustrated
in Figure 4b, and 1 mol L–1 portion of HCl was chosen. The concentration of SnCl2
has been recognized as one of the most critical variables in VG. Concentrations in the
range 0.5–3.0% of SnCl2 were assayed using 3 mol L–1 HCl (Fig. 4c). The higher its
concentration the higher the signal, so the faster should be the reaction and more
active intermediates should be formed, but relative standard deviations also increased.
On the other hand, higher SnCl2 concentrations than 1% (m/v) would result in a violent reaction (more hydrogen generated) in the gas-liquid separator and eventually
lead to an unstable signal. High SnCl2 concentrations had to be avoided and an opti-
Determination of total mercury by vapor generation in situ trapping FAAS
919
mum value of 1% was chosen for further experiments. The highest VG efficiency
was obtained when the concentration of HCl was 2 mol L–1; and the reagent blank
was kept as low as possible.
Figure 4. Influence of sample and reductant uptake rate (a); concentration of HCl (b); concentration of
SnCl2 (c) and flow rate of carrier air (d) on the peak area signals of Hgtot (5 ng mL–1) in continuous flow system. The experimental conditions employed are detailed in the Experimental
Section
920
H. Matusiewicz and M. Krawczyk
The air carrier gas flow through the apparatus is one of the basic parameters
influencing the transport of the Hg0 into the trap-flame system, but also the mixing
effect of vapor-forming reaction solutions, and thus it can affect the determination of
Hg markedly. The vapors were stripped and trapped in the silica tube (atomizer) at air
flow rates in the range of 50–150 mL min–1 with a 120 s collection time. It was evident that the use of low carrier gas flow can successfully reduce high analyte losses
caused by sorption on the inner surfaces of the apparatus (transport tubing), and leads
to a slight improvement in the analytical signals. This may be connected with the
more efficient separation of the vapors from the reaction solution. On the other hand,
higher carrier gas flow rates (> 100 mL min–1) can result in a slightly decrease in
trapping efficiency. The decrease observed is probably due to the diluting effect of
the air flow. An air flow of 100 mL min–1 was therefore chosen as optimum and used
throughout the experiments (Fig. 4d). In the atomization step, signal with regular
shape was observed for trapping temperature and was not significantly influenced by
sample introduction time in the tested range between 1 and 3 min. It should be stressed
that multiple absorption peaks were not observed.
The length of the transfer tubing was not optimized, but was selected based upon
previous experience [9–13]. Therefore, for practical reasons, a transport tubing length
of 10 cm was selected for this study.
Validation of the method by analysis of certified reference materials
Validation of the technique proposed included analysis of three standard reference materials (SRMs): coal fly ash, hair and water. These reference materials were
chosen as they were the closest available to biological samples and are certified for
the analyte of interest to be determined in the environmental and biological samples.
The matrix effects on the VG–IAT–FAAS analytical signals were evaluated by comparing the conventional calibration with the standard additions slopes. No significant
differences were found between the slopes obtained by both calibration procedures
when using VG–IAT–FAAS. Therefore, to determine total mercury in all samples by
this analytical technique, the conventional calibration mode was used. Results obtained for the analysis of SRMs by VG–IAT–FAAS method using aqueous standard
calibration technique are summarized in Table 4. The short-term precision is expressed
as the RSD of six replicate measurements of each sample. The results obtained by
external calibration technique do agree with certified values for any reference material indicating that calibration against aqueous solution could produce accurate
results. No significant differences were found between experimental and certificate
values. The concentrations reported for this element in reference materials are significantly higher than the detection limits that are attained for the measurement.
Determination of total mercury by vapor generation in situ trapping FAAS
Table 4.
921
Determination of inorganic mercury (content ± SD) in reference materials using VG–IAT–FAAS
technique (n = 6)
Determination of total Hg. Mercury contents in a wide variety of SRMs and
CRMs were determined by VG–IAT–FAAS following pre-treatment procedures,
although Hg is certified only in its total content. Table 4 gives the total mercury
concentration in samples subjected to microwave-assisted preparation procedures by
VG–IAT–FAAS.
Sample preparation is without any doubt a critical stage in Hg determination in
biological and environmental samples. Microwave-assisted acid digestion in closed
pressurized PTFE vessels is a well-established tool for sample decomposition and/or
dissolution prior to trace element determination. The Hg concentrations in the certified reference materials were determined using HNO3 and H2O2 reagents (for organic
samples) and HNO3, H2O2 and HF reagents (for inorganic samples). These „classical” procedures have been considered well suited to Hg determination by VG–FAAS.
Almost quantitative recoveries were obtained (in general, better than 96%) when certified reference materials were subject to a high pressure treatment in a TFM–PTFE
vessel at high temperature. These conditions must be employed to destroy completely
the stable organic molecules: in particular, the temperature is the determining parameter; the higher the temperature, the better the quality of decomposition. The high
pressure/temperature focused microwave heated digestion system [32] allows an effective digestion temperature up to 300°C, essential to assuring the quantitative determination of mercury in samples.
Analytical figures of merit
A comparison of the detection limits of the present procedure are summarized in
Table 5 for conventional FAAS, VG–GF–AAS (in situ trapping) and with four atom
trap designs. While direct comparison of detection limits is often misleading owing
to the use of different systems, operating conditions and modes, it is clear that the
detection limits that can be achieved with in situ atom trapping FAAS are two to three
orders of magnitude superior to those obtained with direct conventional flame AAS.
g
f
e
d
c
b
a
Comparison of detection limits (LOD)a for mercury using various atom traps (ng mL–1)
2 min collection time.
Detection limit defined as „3 blank” criterion (n = 6).
Enhancement (improvement) factor.
Enhancement factor shows the ability of the analytical system regarding period of time required for analysis: X/minutes of sampling (min–1).
Enhancement factor shows influence of sample volume: X/milliliters of sample (mL–1).
Ref. [26], CV–AAS, in situ concentration, for sample volume of 50 mL, LOD (ng L–1)
Ref. [7], FAAS, in situ concentration, 2 min collection time.
Table 5.
922
H. Matusiewicz and M. Krawczyk
Determination of total mercury by vapor generation in situ trapping FAAS
923
The analytical performance characteristics were evaluated for mercury. The limit
of detection (LOD) were obtained by use of optimized operating conditions. Since no
detection limits obtained by identical techniques are available, the results are compared by the quartz atom trap and flame atomic absorption spectrometry (QAT–AAS)
(in situ trapping) and by the CV–GF–AAS (in situ trapping) technique (Tab. 5).
The detection limits of the developed procedure are evidently better, as compared
with other atom traps. Detection limits obtained with a 120 s collection time were
in the ppt (pg mL–1) range and are the best ones; VG–IAT–FAAS procedure afforded
significant improvements in the detection limit with respect to those reported for
QAT–AAS (59.8 ng mL–1; in situ trapping) methodology for the determination of this
vapor forming trace element [7]. The achieved limit of detection (LOD): 0.4 ng mL–1
for Hg (sensitivity enhancement compared to FAAS is 750 folds), is at least several
orders of magnitude worse to the LOD achieved for Hg in situ trapping in graphite
furnace [26]. Six replicate measurements of the total procedure (reagent) blank solution were carried out and the relative standard deviation (RSD) of the background
values for the raw, unsmoothed data were calculated. Precision was in the range of
9.3% (as peak area); this reflects the cumulative imprecision of all of the sample
handling, vapor generation, trapping, atomization and detection steps. The peak height
precision was always slightly, but significantly, worse.
Total mercury determination in selected real samples
Finally, in order to evaluate the usefulness of the proposed method in determining total mercury contents in some real samples were analyzed using the optimized
experimental conditions (Tab. 6). In all cases, the calibration was achieved using the
aqueous standard calibration curves. The proposed method was validated by spiking
the samples with known amount of Hg(II). The recoveries from spiked solutions
were varied in the range 92–96%. The precision of replicate determinations is typically better than 10% RSD.
Table 6.
Inorganic mercury concentration in real samples (n = 6)
924
H. Matusiewicz and M. Krawczyk
CONCLUSION
This work is a continuation of previous efforts to employ an integrated silica
tubes as a trap for species volatilized in a hydride generation system act, in a sense,
like an enrichment device [8–13, 30, 31]. The results presented in this work confirm
the idea that the present hyphenated technique using a continuous mode vapor generation gas phase in situ trapping on an integrated silica tubes trap, followed by atomization in acetylene-air flame with simultaneous direct thermal heating of the atomizer, can be used for the determination of trace amounts of total mercury (Hgtot)
in certified reference materials and real samples. Following the trapping stage,
the performance of the device and related problems are quite similar to the case of
hydride generation-graphite furnace atomization (in situ trapping) HG–GF–AAS.
The detection limit of this VG–IAT–FAAS system for Hg is considerably improved
compared with those reported for measurements of Hg by any flame AAS approach
and was significantly lower than that for QAT–AAS (59.8 ng mL–1; in situ trapping).
However, there are still many unknowns regarding mechanisms of trapping (collection, pre-concentration) and re-volatilization (atomization). This very simple and cheap
technique constitutes an attractive alternative to VG-in situ trapping-GF–AAS system at significantly lower cost. Although, the achieved very low concentration detection limit for Hg, 0.4 ng mL–1, is worse than those for the in situ trapping of Hg vapors
in a commercial graphite furnace: 0.628 ng L–1 [26]. The determinations were equivalent within the precision of the methods. It is expected that detection limits could be
further improved with the use of high purity reagents. However, in most of the cases
blank values and/or saturation of trap surface by analyte and interferent species put
a practical limit to sample size and sampling period; therefore, the enhancement can
never be infinitely improved.
The proposed experimental speciation approach, offers interesting perspective
and good prospects for the determination of other hydride forming elements, in the
range of pg mL–1, e.g., Ge and Sn. This is subject of on-going research. A simple AA
flame spectrometer, the low cost, easy operation, and high sensitivity of the present
system make it very attractive for laboratories not equipped with any graphite furnace apparatus. Although time required for each measurement is longer than that
required for the classic hydride generation technique.
Acknowledgements
Financial support by the State Committee for Scientific Research (KBN), Poland, Grant No. 4 T09A 10
30 is gratefully acknowledged.
Determination of total mercury by vapor generation in situ trapping FAAS
925
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
Welz B. and Sperling M., Atomic Absorption Spectrometry, Wiley-VCH, Weinheim 1999.
Clevenger W.L., Smith B.W. and Winefordner J.D., Crit. Rev. Anal. Chem., 27, 1 (1997).
Wu L., Zheng Ch., Ma Q., Hu Ch. and Hou X., Appl. Spectrosc. Rev., 42, 79 (2007).
Matusiewicz H., Spectrochim. Acta, Part B, 52, 1711 (1997).
Matusiewicz H. and Sturgeon R.E., Spectrochim. Acta, Part B, 51, 377 (1996).
Dìdina J., Spectrochim. Acta, Part B, 62, 846 (2007).
Ellis L.A. and Roberts D.J., J. Anal. At. Spectrom., 11, 1063 (1996).
Matusiewicz H. and Kopras M., J. Anal. At. Spectrom., 12, 1287 (1997).
Matusiewicz H. and Krawczyk M., Anal. Sci., 22, 249 (2006).
Matusiewicz H. and Krawczyk M., Microchem. J., 83, 17 (2006).
Matusiewicz H. and Krawczyk M., J. Braz. Chem. Soc., 18, 304 (2007).
Matusiewicz H. and Krawczyk M., Spectrochim. Acta, Part B, 62, 309 (2007).
Matusiewicz H. and Krawczyk M., J. Anal. At. Spectrom., 23, 43 (2008)
Cankur O., Ertaº N. and Ataman O.Y., J. Anal. At. Spectrom., 17, 603 (2002).
Korkmaz D., Kumser S., Ertaº N., Mahmut M. and Ataman O.Y., J. Anal. At. Spectrom., 17, 1610
(2002).
Ertaº N., Korkmaz D.K., Kumser S. and Ataman O.Y., J. Anal. At. Spectrom., 17, 1415 (2002).
Guo X-m. and Guo X-w., J. Anal. At. Spectrom., 16, 1414 (2001).
Korkmaz D.K., Ertaº N. and Ataman O.Y., Spectrochim. Acta, Part B, 57, 571 (2002).
Korkmaz D., Dìdina J. and Ataman O.Y, J. Anal. At. Spectrom., 18, 255 (2004).
Kratzer J. and Dìdina J., Spectrochim. Acta, Part B, 60, 859 (2005).
Kratzer J. and Dìdina J., J. Anal. At. Spectrom., 21, 208 (2006).
Krejèi P., Doèekal B. and Hrušovská Z., Spectrochim. Acta, Part B, 61, 444 (2006).
Menemenlioglu I., Korkmaz D. and Ataman O.Y., Spectrochim. Acta, Part B, 62, 40 (2007).
Kratzer J. and Dìdina J., Anal. Bioanal. Chem., 388, 793 (2007).
Ertas N., Arslan Z. and Tyson J.F., J. Anal. At. Spectrom., 23, 223 (2008).
Yan X-P. and Ni Z-M., Anal. Chim. Acta, 272, 105 (1993).
Matusiewicz H., J. Anal. At. Spectrom., 4, 265 (1989).
Veber M., Èujes K. and Gomišèek S., J. Anal. At. Spectrom., 9, 285 (1994).
Sturgeon R.E., Willie S.N. and Berman S.S., Anal. Chem., 57, 2311 (1985).
Matusiewicz H., Kopras M. and Sturgeon R.E., Analyst, 122, 331 (1997).
Matusiewicz H., Sturgeon R., Luong V. and Moffatt K., Fresenius’ J. Anal. Chem., 340, 35 (1991)
(and references cited therein).
Matusiewicz H., Anal. Chem., 66, 751 (1994).
Received May 2008
Revised November 2008
Accepted November 2008