| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Chemistry Department, Royal Veterinary and Agricultural University, Thorvaldsensvej 40, 1871 Frederiksberg C, Denmark
b Danish Forest and Landscape Research Institute, Hoersholm Kongevej 11, 2970 Hoersholm, Denmark
* Corresponding author (man{at}kvl.dk)
Received for publication October 24, 2000.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
Abbreviations: CEC, cation exchange capacity • E, accumulated metal in the effluent • PTFE, poly(tetrafluorethene)
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
Various materials have been used for construction of porous cups, but some of them have been found to dissolve slowly or to alter the soil solution, either by adsorption of soil solution compounds, or by release of substances from the cup material (Hughes and Reynolds, 1988; Raulund-Rasmussen, 1989). In order to minimize the reaction of cup material with soil solution, suction cups of more inert materials, such as nylon, polyethylene, or poly(tetrafluorethene) (PTFE) have been used (Beier et al., 1992; Knight et al., 1998). Even though these materials have greatly improved the performance of the suction cup method, by reducing both adsorption and desorption of soil solutes, problems may still arise, for example, when low concentrations of heavy metal ions are sampled (McGuire et al., 1992; Wenzel and Wieshammer, 1995). At low concentrations even a low adsorption capacity of the cup material may have a significant effect on the composition of the isolated solution. Wenzel et al. (1997) found that a nylon suction cup had no effect on the heavy metal ion concentration in the solution pumped through. Grossmann et al. (1990) also found that heavy metal ions did not adsorb to a nylon cup. However, Grossmann et al. (1990) used 10 L of test solution to equilibrate the cups beforehand. In a test of suction cups made of PTFE, ceramic, steel, and glass McGuire et al. (1992) found that PTFE cups were the best choice for sampling heavy metal ions at low concentrations. Although adsorption did occur on PTFE cups, it was less than for cups made of other materials. Other investigations have shown that PTFE cups are not as inert as expected. Maitre et al. (1991) found that even after thorough cleansing with 1 M HCl the PTFE cups released substantial amounts of Ca, Mg, and Si when subjected to a Na-citrate solution. Releases of Ca ions under field conditions have also been observed (Beier et al., 1992).
In the soil, the suction cups are subjected to a variety of chemical environments. The PTFE cups cannot be expected to be fully inert in all soil environments. With respect to heavy metal ions the suction cups may affect the metal ion concentrations in the sampled solution through either adsorption or desorption. In order to decide if PTFE cups constitute reliable devices for sampling low concentrations of heavy metal cations (Cd, Cu, Ni, Zn) in soil solution, laboratory test experiments were carried out. This paper reports the results of these experiments. Special attention was given to pH and ionic strength effects on metal cation adsorption to the suction cups.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
|
The cation exchange capacity (CEC) of the cups was measured at pH 3.6, 4.5, and 5.8. Four clean suction cups were saturated with 50.0 mL of 1.00 M ammonium nitrate solution at the respective pH for 2 d by pumping the solution through the cups in a recycled system. Subsequently, the cups were washed with MilliQ water in order to remove non-adsorbed ammonium ions. The cups were then flushed with 50.00 mL of 1.00 M KCl solution three times each time for 2 d and the KCl extracts were stored in a freezer until ammonium determination. The ammonium was determined by a FIAstar Analyzer and FIAstar 5023 spectrophotometer (Tecator, Sweden) and the CEC of the cup was calculated from the sum of ammonium in the three KCl extracts.
All heavy metal solutions were prepared from 1000 mg L-1 standard solutions (Merck, Darmstadt, Germany) of Cd, Cu, Ni, and Zn in 5% HNO3, and the solutions were adjusted to the desired pH using suprapure HNO3 or NaOH (Merck). The NaCl and CaCl2 salts used as background electrolytes were of analytical grade (Merck). All influent solutions were prepared in 1-L borosilicate glass flasks and left for 3 d at room temperature to equilibrate with the surface of the glass flask without the PTFE cup inserted in the solution. A preliminary experiment was made to test the heavy metal adsorption to the borosilicate flasks. Solutions of 10 and 30 µg L-1 of the metals and no background electrolyte at pH 4.5 were left at room temperature for 21 d. Samples were collected from the solutions immediately after preparation and after 3 and 21 d.
In preliminary tests the tubing used was tested for adsorption of heavy metals by pumping a 10 µg L-1 solution of Cd, Cu, Ni, and Zn at pH 5.6 and without background electrolyte through the system without any cups installed. Samples of the solution were taken before and after it passed through the tubing.
Prior to each experiment the influent solutions were analyzed in order to learn the actual metal concentrations pumped through the PTFE cups. The pumping rate was approximately 3 mL h-1 in all experiments. Samples of 10 mL each were collected in the fraction collector into polystyrene vials containing 1 mL of 1.00 M HNO3. Heavy metal concentrations were determined in the first five samples and thereafter in every fifth sample. The total volume of effluent solution pumped through the cups was measured by mass. Solution concentrations of Cd, Cu, Ni, and Zn were determined by inductively coupled plasma–optical emission spectrometry (ICP–OES) (PerkinElmer [Wellesley, MA] Optima 3000 XL axial view, AS 90). Detection limits were 3.0 µg L-1 for Cu, Ni, and Zn, and 2.0 µg L-1 for Cd. A Methrom (Herisau, Switzerland) 691 pH meter and a Methrom 6.0202.110 combination electrode were used for determination of pH in the influent solutions.
Effect of pH
Three solutions containing 10 µg L-1 of Cd, Cu, Ni, and Zn in a 0.01 M NaCl background electrolyte were prepared. The three solutions were adjusted to pH 3.6, 4.5, and 5.8, respectively. All the experiments were made in duplicates, thus two PTFE cups were lowered into each solution. In order to monitor the influent solution concentration, samples were collected from the influent container before, during, and after the experiment. The effluents were sampled using the fraction collector as described above. The pH of the influent was monitored daily in a sample from the influent container. Heavy metal concentrations were determined by ICP–OES.
The total accumulated mass of heavy metal in the effluent (E) was calculated for each suction cup as shown in Eq. [1]. The volume of effluent pumped through the cup in sample x is denoted Vx, and Cx and Cx+5 denote the effluent concentration in sample x and sample x + 5:
 | [1] |
Because the effluent concentrations were measured in the first five samples, the contributions from these samples were calculated differently, as the first part of the equation shows. Concentrations below the detection limit were included in the calculation of E with half the detection limit. The areas under the effluent curves are estimated by E and were used to compare the effect of the different pH levels by a one-way analysis of variance using the SAS procedure GLM (SAS Institute, 1999). The level of statistical significance employed was p < 0.05. Pair-wise comparisons between pH levels were conducted and means reported together with a mass balance of the total amount of metal adsorbed by the cup (Table 1). Because two cups were submerged into the same influent solution the mass balance could only be calculated as a mean value.
|
For each suction cup the total accumulated amount of heavy metals in the effluent (E) was determined the same way the total accumulated heavy metals were determined for each pH level (Eq. [1]). E estimates the area under the effluent curve and it was used to test for significant differences among the three background electrolytes using a one-way analysis of variance performed by the SAS procedure GLM (SAS Institute, 1999). The level of statistical significance employed was p < 0.05. Pair-wise comparisons of background electrolytes were conducted and mean values reported together with a mass balance of the total amount of metal adsorbed by the cup during the entire experiment (Table 2).
|
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
|
The CEC of the cups was measured at 0.049 mmolc cup-1 (0.0023), 0.044 mmolc cup-1 (0.0011) and 0.031 mmolc cup-1 (0.0014) at pH 5.8, 4.5, and 3.6, respectively, with the standard deviations in parentheses.
pH-Dependent Adsorption
The experiments to test adsorption of Cd, Cu, Ni, and Zn from a solution with an initial heavy metal concentration of 10 µg L-1 and a 0.01 M NaCl background electrolyte at pH 3.6, 4.5, and 5.8 indicated similar properties of Cd, Ni, and Zn, whereas Cu appeared to have different adsorption properties (Table 1).
The metal ions adsorb to the surface of the PTFE cup both directly from the influent solution, thus affecting the influent concentration during the experiment, and from the solution pumped through the cup. The influent concentrations of Cd, Cu, Ni, and Zn decreased during all experiments except for Cd and Zn at pH 3.6. The influent concentration of all four cations decreased more as pH was increased. At pH 5.8, less than 50% of the initial metal concentrations remained in the influent solutions after 7 d (i.e., after passage of 500 mL of solution), thus implying that the missing 50% of the total metal added to the influent solution was adsorbed to the outer surface of the PTFE cup. The influent concentration of Cu decreased more than Cd, Ni, and Zn during the experiments, even at pH 3.6.
In the effluent solution no Cd, Ni, or Zn was detected at pH 5.8 and the calculated E values in Table 1 are solely based on half the detection limit, which demonstrates strong adsorption of the metal cations (Table 1). At pH 4.5 effluent concentrations of Cd and Zn increased over time, and some metal was detected in the effluent, whereas the Ni concentration in the effluent did not exceed the detection limit at this pH. At pH 4.5 the extent of adsorption was in the order Ni > Cd > Zn based on the accumulated effluent. At pH 3.6 the accumulated effluent concentrations of Cd, Ni, and Zn were higher than at pH 4.5, indicating less adsorption (Table 1).
Adsorption of Cu to the cup was also strongly affected by pH. There is some scatter in the data for effluent concentrations of Cu, but the adsorption of Cu to the cup material was significantly stronger at pH 4.5 and pH 5.8 than at pH 3.6 (Table 1). Adsorption directly from the influent solution was especially important for Cu.
Generally, this experiment demonstrates that the suction cups investigated have a high affinity for adsorption of heavy metals ions in the near-neutral pH range even in the presence of 0.01 M NaCl.
Effect of the Ionic Composition
The metal ion concentrations of Cd, Cu, Ni, and Zn in both the influent solution and in the effluent solution were affected by the ionic composition at pH 4.5. The initial influent concentration of 10 µg L-1 of Cd and Ni was unchanged during Period 1 using 0.01 M CaCl2 as background electrolyte (Fig. 3)
. A small decrease in the influent concentration was observed in Period 1 when using 0.01 M NaCl. In the experiment where no background electrolyte was added, more than 50% of the Cd and Ni in the influent solution was adsorbed to the cup during Period 1, after 500 mL of solution was passed through the cup. In Period 2, when the influent concentration was raised to 30 µg L-1, a small decrease in the influent concentration of Cd and Ni was observed during the experiment for all three electrolytes. The Cd and Ni influent concentration of 10 µg L-1 remained unchanged during Period 3 when 0.01 M CaCl2 and 0.01 M NaCl were used as electrolytes, but a decrease was seen when no electrolyte was used.
The effluent concentrations of Cd resembled that of Ni. In 0.01 M CaCl2, almost no adsorption of Cd and Ni was found during the entire experiment and increasing adsorption was found in the 0.01 M NaCl and no-electrolyte experiments (Table 2). After the initial 200 mL of 10 µg L-1 Ni influent solution in Period 1 had passed the cup no further adsorption of Ni was found during the following periods. Equivalence between effluent and influent in the 0.01 M NaCl background electrolyte experiment was observed for both Cd and Ni in Period 2 when the 30 µg L-1 solution was pumped through the cups (Fig. 3). No adsorption nor desorption of Cd or Ni to or from the cups was observed when the concentration was decreased from 30 to 10 µg L-1 between Periods 2 and 3 and during Period 3 of the experiment. However, when no background electrolyte was used neither Cd nor Ni was detected in the effluent in Period 1. A sharp increase in the effluent concentration of Cd and Ni was seen in Period 2 after 100 mL of 30 µg L-1 metal solution was pumped through the cups and equivalence between influent and effluent was achieved. When the influent solution of Cd and Ni was changed from 30 to 10 µg L-1 from Period 2 to 3 the effluent concentration of Cd and Ni decreased to below the influent concentration. Some desorption of Cd to the effluent was seen in Period 4 when lowering the influent concentration from 10 to 0 µg L-1 in the no-electrolyte experiment, whereas the Ni concentration was below the detection limit. The accumulated effluents (E) of Cd and Ni in all background electrolytes were significantly different (Table 2).
Changes in the influent concentration of Zn during the experiments were small (Fig. 3). Direct adsorption from the influent solution of Zn ions to the cup surface was only observed in Period 2 when using 0.01 M CaCl2 and 0.01 M NaCl as background electrolytes. The effluent concentration of Zn followed the influent concentration when using the 0.01 M CaCl2 electrolyte. The total accumulated effluent Zn in the 0.01 M CaCl2 and 0.01 M NaCl experiment differed significantly from the no-electrolyte experiment (Table 2). Equivalence between effluent and influent occurred in Period 1 after 400 mL of 10 µg L-1 Zn solution had passed through the cup with the 0.01 M NaCl (Fig. 3). In the experiment without electrolyte some adsorption of Zn was observed in Period 1 and 2 until a total of 800 mL of solution had passed the cup. On return to an influent concentration of 10 µg L-1 from Period 2 to 3, the effluent followed the influent closely. In the no-electrolyte experiment some desorption of Zn was observed when the influent concentration was lowered to zero in Period 4. This indicated that the Zn-loaded suction cup had a memory effect at low concentrations, that is, the concentration of the solution pumped through the cup was affected by the concentration of solutions pumped through at prior stages.
In all experiments Cu was strongly adsorbed to the suction cup (Fig. 3). In the experiments with 0.01 M CaCl2 and the 0.01 M NaCl electrolytes the influent concentration of Cu decreased by approximately 50% during Period 1 with an influent concentration of 10 µg L-1 Cu. During Period 2 at the 30 µg L-1 influent concentration of Cu the influent concentration of Cu decreased by approximately 30% when 0.01 M CaCl2 was used as background electrolyte and more than 50% with 0.01 M NaCl as background electrolyte. The influent concentrations of Cu in the experiment without background electrolyte showed large variations.
In the 0.01 M CaCl2 electrolyte experiment an increase of Cu concentrations in the effluent was observed during Period 2 after 100 mL of the 30 µg L-1 Cu solution had been pumped through the cups, but equivalence between influent and effluent solutions was not obtained. In the 0.01 M NaCl and no-electrolyte experiments, Cu was adsorbed strongly by the cup. No significant differences of the total accumulated effluent (E) of Cu between the different background electrolytes were found, indicating strong adsorption in all the electrolytes used (Table 2). The low total amount of Cu adsorbed per cup in the no-electrolyte experiment (Table 2) is explained by the lower influent concentration caused by the higher adsorption by the container walls prior to the start of the experiment.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
It is expected that the cup wall of these new cups also consists of two phases: PTFE and a mineral residue from the quartz flour. The surfaces of the mineral residues are not expected to be as inert as the PTFE surfaces and the presence of a mineral surface in the cup material would provide adsorption sites capable of binding the heavy metal cations, and thus may explain the adsorption capacity of the otherwise inert PTFE material. The cups have a pH-dependent CEC approximately 10 times lower than ceramic and almost half of the CEC of nylon cups as reported in a review by Wenzel and Wieshammer (1995). Even though the PTFE cups have a low CEC they will be capable of ion exchange reactions with metals in solution, which might significantly affect low heavy metal concentrations in solution.
Only in the no-electrolyte experiment was a decrease relative to the originally added metal concentrations observed in the influent reservoir flask during the 3-d equilibration period prior to submerging of the cups. This indicates that the decrease in the influent concentration observed during the experiments can be attributed to the presence of the suction cups in the solutions. The outer surfaces of the cups are directly in contact with the influent solution, which would decrease the influent concentration.
Effects of pH
The pH range covered in this study (pH 3.6–5.8) is within the range often found in soils. The pH experiments were carried out in a 0.01 M NaCl solution, which exceeds the ionic strength in most natural soil solutions (Berggren, 1992; Jallah and Smyth, 1995; Lehnardt, 1998). At lower levels of ionic strength metal adsorption will be even stronger.
The use of PTFE suction cups for heavy metal sampling has been shown in this study to be very dependent on pH. Wenzel and Wieshammer (1995) found that adsorption of Cd, Cu, and Ni to ceramic and nylon suction cups also was very pH dependent. The seasonal variation of pH in soil water has been found to vary up to one unit (Pedersen and Bille-Hansen, 1995). This will affect the adsorption of Cd, Cu, Ni, and Zn to the suction cup. This implies that the heavy metal concentrations in soil solutions sampled by suctions cups that possesses some adsorption capacity may rather depend on the pH than on the actual concentrations of heavy metal ions in the soil solution. The degree of saturation of the suction cups with heavy metal ions and other cations may also affect heavy metal adsorption, and consequently affect the heavy metal concentration in the sampled solution.
Effects of Ionic Composition
The ionic strength of the 0.01 M CaCl2 and 0.01 M NaCl solutions used in our experiments exceeds the ionic strength found in most soil solutions. The environment found in most noncalcareous soils will have an ionic strength between the no-electrolyte and the 0.01 M NaCl experiment (Berggren, 1992; Jallah and Smyth, 1995; Lehnardt, 1998). The differences in ionic strength of the solutions used in our investigations are very high, and cause a significant effect on the adsorption patterns for Cd, Ni, Zn, and Cu. The seasonal changes of ionic strength and soil solution composition in the field are not expected to change as much and therefore the adsorption properties of the suction cup are not expected to change. The fact that adsorption of Zn, Ni, and Cd is reduced considerably at the higher concentrations of Ca and Na cations suggests that at least some of the adsorption observed is due to ion exchange reactions or some kind of ion competition at specific adsorption sites in the cup material. The effect of the higher concentrations of electrolyte cations on the amount of total adsorbed Cu was nonsignificant, which indicates that Cu ions are adsorbed strongly to the PTFE cup surface and presumably by specific bonding rather than by cation exchange (Table 2). During the experiments a maximum total of approximately 11 µg Cd, 30 µg Cu, and 12 µg Ni was adsorbed to each cup (Table 2). This corresponds to less than 2 µmolc per cup, which is much less than the CEC of the cups. Thus, all the adsorption can be explained by cation exchange, but adsorption of a more specific character cannot be excluded, especially for Cu.
A memory effect would be expected to influence the concentration of metal cations in the effluent, if the Prenart cups are used for sampling of soil solution at low concentrations of metal ions. The memory effect observed for Cd was negligible in this experiment (Fig. 3). For Zn, the memory effect continued until the end of the experiment, and therefore it is not possible to predict for how long the release of Zn would continue (Fig. 3). The sharp drop in the effluent concentration from Period 2 to Period 3 indicates that the memory effect only exists at very low concentrations of Zn. The memory effect would damp out the differences in concentration between samples taken at different times. Other cup types having a greater adsorption capacity exhibit this effect as well (Grossmann and Udluft, 1991). Because of the relatively low adsorption capacity of the PFTE cups this effect is only important at low heavy metal concentrations and at low ionic strengths. Even under these circumstances only Zn seems to be significantly affected.
At the ionic strengths found in soil solutions significant changes of solution concentration of Cd, Ni, Zn and especially Cu can be expected to occur due to adsorption reactions between the PTFE suction cup and the metal ions. These investigations indicate that the pH of the soil solution rather than its composition will influence the extent to which heavy metal concentrations are changed by the PTFE cup. There is no evidence that the cup irreversibly adsorbs the metal cations and the metals will most likely desorb if the pH of the influent is decreased. Therefore, Cd, Cu, Ni, and Zn adsorbed to the cup during the equilibration between the cup surface and the soil solution could be released by a pH change in the soil solution and thus contaminate the sample.
Under field conditions the presence of dissolved organic carbon (DOC) in the soil solution could be expected to affect the adsorption of the metal cations to the PTFE cup. Metal–organic complexes in solution may reduce adsorption of heavy metal ions. On the other hand, DOC could adsorb to the PTFE surface with the nonpolar parts of the molecule and thus enhance the adsorption of metal cations by formation of metal–organic complexes at the cup surface. In subsoils where the soil solution has a high pH, a low ionic strength, and a low DOC concentration, the use of PTFE suction cups is foreseen to affect sampling of low concentrations of heavy metal ions.
When deciding on the use of suction cups for isolation of soil solution it is important to consider the effects on the solution concentration of the adsorption capacity of the cup material. Even though PTFE suction cups adsorb smaller amount of Cd, Cu, Ni, and Zn than other cup materials (McGuire et al., 1992), other materials like nylon have proven not to adsorb heavy metals at pH values below 5 (Wenzel et al., 1997), suggesting that nylon might be superior to PTFE when heavy metal adsorption is considered.
|
CONCLUSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  D. Rais, B. Nowack, R. Schulin, and J. Luster Sorption of Trace Metals by Standard and Micro Suction Cups in the Absence and Presence of Dissolved Organic Carbon J. Environ. Qual., January 3, 2006; 35(1): 50 - 60. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Vadose Zone Journal | Journal of Plant Registrations | ||||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal |
