| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Centre for Advanced Analytical Chemistry, CSIRO Energy Technology, Private Mail Bag 7, Bangor, NSW 2234, Australia
b EWL Sciences Pty. Ltd., PO Box 39443, Winnellie, NT 0821, Australia
* Corresponding author (rob.jung{at}csiro.au).
Received for publication May 27, 2002.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMaterials and MethodsResults and DiscussionAdvantages and Limitations of...ConclusionsREFERENCES |
|---|
Abbreviations: EVA, ethylvinyl acetate • PVC, polyvinyl chloride
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and MethodsResults and DiscussionAdvantages and Limitations of...ConclusionsREFERENCES |
|---|
An alternative is to collect and measure fluxes of contaminants from a sediment core overlain with site water. There are potential problems with this approach, including the ability to both sample a sufficiently large and representative sediment volume, and then to transfer it to a reactor while maintaining its integrity in a state representative of the field. The many samplers that have been designed for the collection of sediments have been well-summarized by Mudroch and Azcue (1995). Most corers collect cores that are <75 mm in diameter, and for contaminant flux measurements a higher surface area is desirable. An alternative is to take a core from a box-cored sample, but this is an exercise that involves considerable sample manipulation and the possibility for both physical and chemical disturbance to the sample.
This paper describes a novel reactor that overcomes these problems by functioning initially as a corer to collect a sediment core and then as a reactor without disturbance to the sediment. This apparatus has been used in our laboratory for a number of years (Jones and Jung, 1996; Jung et al., 1997) and has generated useful data that are compatible with those collected by other techniques such as benthic chambers. The device is inexpensive and the technology has already been transferred to a number of Australian laboratories.
|
Materials and Methods |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and MethodsResults and DiscussionAdvantages and Limitations of...ConclusionsREFERENCES |
|---|
|
A base-sealing cap for the corer (Fig. 1C) that could be put in place when a core had been retrieved was fabricated from polyvinyl chloride (PVC). This base was 190 mm in diameter and 45 mm deep, into which was machined a 166-mm-diameter by 30-mm-deep well. The end of the corer fits into this well.
To provide a seal between the outer wall of the corer tube and the base cap, a 165-mm-i.d. and 5.3-mm-thick nitrile O-ring (Fig. 1D; British Standard 2-363) was inserted into a 6-mm-wide groove in the interior face of the PVC located 15 mm from the base of the well. The depth of this groove was critical to allow the core tube to be pushed down onto the base with the O-ring providing a sufficiently tight seal to prevent leakage. The O-ring groove had a 174.5-mm o.d. Silicone high-vacuum grease was applied on the O-ring before each use to complete the seal.
A 2-mm hole (Fig. 1E) was drilled in the base to allow air and water to escape when the corer barrel was being inserted into the base cap. Before each use the core tube and base assembly were water pressure tested to ensure no leaks. To do this, the hole in the base was plugged with PVC tape and the reactor was filled with water.
A top collar (Fig. 1F) was constructed of acrylic (170-mm o.d.) with a double-start thread cut with a lathe. The thread was machined to easily screw on the PVC top cap. The acrylic collar was glued to the tube with acrylic glue.
The corer lid (Fig. 1G) was screwed onto the top of the corer tube to enable efficient sampling and retrieval of sediment. It consisted of a standard 152-mm (6-in) PVC plumbing screw cap. Stainless steel bolts connected a strong wooden (or PVC) handle (Fig. 1H), 550 mm long, 80 mm wide, tapering to 40 mm at the ends, and 19 mm thick. Also attached to the corer lid was a 12.7-mm (0.5-in) PVC ball valve (Fig. 1J) to allow water to flow from the corer when coring or to be sealed when withdrawing the core from the sediment.
Usually the visibility at coring sites was poor due to the presence of suspended particulate matter, which caused the position of the corer in relation to the sediment–water interface to be indistinguishable. Typically, a diver or a wading sampler would work in the dark by feel. Thus, a stopping plate was developed to assist the user in obtaining cores of consistent length.
The stopping plate was a large PVC annular-shaped disk (Fig. 1K; 200-mm i.d., 350-mm o.d., and 10 mm thick) attached to the lid by two PVC rods (Fig. 1L; 310 mm long x 20 mm in diameter) and four hexagonal PVC nuts (Fig. 1M). An annulus was fitted around the core tube. The rods were threaded and nuts attached, enabling adjustment of the stopping position. The lid prevented the bottom nuts from moving significantly once they were set.
For use as a reactor, an alternative lid (Fig. 1N) was employed. This was the same size as the corer lid, but was fitted with a 12-v direct current motor (Fig. 1P) mounted centrally. Two types of motor, geared down to give 60 rpm at 12 v direct current, have been used effectively. The motor must have sufficient torque to operate a paddle stirrer at low (<15 rpm) speeds and low voltages. Heat generation was not a problem with the motor running a paddle stirrer at 15 or 40 rpm for weeks at a time. Motor speed control to within 5% of the set speed over a period of one week was provided by dual-channel model electric train speed controllers (e.g., Type CDA 223; Hobby Co., Sydney, Australia). The controllers were slightly modified (with assistance from the manufacturer) to give an easier adjustment in the lower speed range. The stirrer speeds were set to the required value by timing the revolutions of the paddle blade and adjusting the controller voltage.
To gently mix the water overlying the collected core, a stirrer blade (Fig. 1R) was attached to the motor shaft via an acetal (polyoxymethylene) collar (Fig. 1Q). The stirrer blades were made from acrylic. The paddle part was rectangular and was 35 mm wide and 8 mm high. This size works effectively in the speed range provided by the motor. Larger paddle blades, such as those the size of microscope slides, were found to cause significant disruption and resuspension of sediment.
The paddle shafts had a typical length of 110 mm. Variable-length paddle shafts were also used. These were manufactured in two sections, with the bottom fixed-length part containing the paddle blade. The variable-length top part was joined to the lower part with a screw thread. The latter were deployed when different-length sediment cores had been collected. This enabled a similar paddle blade to sediment distance to be maintained to standardize mixing conditions.
The reactor lids were fitted with portholes (Fig. 1S) for sampling and for inserting measurement probes. Typically, PVC-pipe adapters (e.g., Vinidex Catalogue no. 2, male valve take-off adapters, 25 x 25 size, 24-mm i.d.) were used (Vinidex Tubemakers Pty. Ltd., Smithfield, NSW, Australia). For some other applications, other sizes, such as 20 x 20 and 15 x 15, which have an i.d. of 19 mm and 14.5 mm, respectively, were used. All except 5 mm of the adaptor end of these fittings was cut off and the fitting secured to the lid with PVC glue.
Panel-mounted barbed and Luer-lock fittings (Fig. 1T) were fitted to the lid to facilitate the easy connection of purging gases such as air. Cole-Parmer (Vernon Hills, IL) polypropylene fittings were used (Catalog no. EW06359-50, 1.59-mm [0.0625-in] barbs and EW06359-60, 3.18-mm [0.125-in] barbs). These Luer-lock fittings were carefully connected to avoid overtightening. Similar fittings are now available in harder polycarbonate. It was necessary to fix the panel fittings in place with nuts and glue to avoid unscrewing when the Luer fittings were being connected and being removed. Nylon lock nuts (6.35 mm [0.25 in] x 28 UNF; Cole-Parmer Catalog no. C040708051) secured the fittings and the glue used was super-strength epoxy.
Sediment Sampling Using the Corer
In shallow waters, the corer was deployed by an operator wearing waders. In deeper waters (>1.5 m) a diver and boat were required. In our studies, cores were usually sampled at shallow depths (<4 m), one at a time, with the diver returning to the boat after each coring. The procedure used to obtain a core was similar whether the core was acquired by wading or diving.
To collect sediment, the corer was pushed into the sediment with the ball valve open, using a gentle clockwise rotation action to avoid unscrewing the lid. Care was taken to avoid major sediment translocation during this operation. When deployed by a diver, extra weight on the diver's weight belt was useful to assist the pushing. When the stop collar reached the sediment, the valve was closed. However, even with the stopping plate, it was frequently difficult in soft sediment to establish when the sediment stop position had been reached. In these situations, it was necessary to determine the correct position by feeling the relative positions of the sediment stop plate and the sediment surface with one hand. To avoid losing the equipment in murky waters, the corer handle was attached to the sampler with nylon cord.
The sampler also carried in a mesh diver's bag at least two core disks made from 20-mm-thick ethylvinyl acetate (EVA) closed-cell foam (Fig. 1U). The foam core disks were 150 mm in diameter and the EVA had a density of 75 kg m-3. This was about the optimum diameter for a disk when sampling at shallow depths (<4 m), so it could be squeezed into the base of the core tube quite easily, but not fall out with the weight of the sediment inside. Disks of at least 155 mm in diameter were required to support the sediment core when sampling at an 8-m depth. Because the disks were compressed by the pressure, larger-diameter disks or more rigid closed-cell foam were required at greater depths. More rigid (denser) grades of EVA closed-cell foam are available (up to 350 kg m-3).
For soft silt and sand–silt sediments, a core could be retrieved easily, provided that a pressure seal was applied at the top when the sediment core was removed from its bed (i.e., the top O-ring seated and sealed properly and the ball valve was closed). As soon as the base of the core tube cleared the sediment–water interface, an EVA disk, described above, was slid under the core and the ball valve was then opened slightly. The disk was then gently pushed up into the end of the corer to plug the base of the core and prevent material from falling out. With the ball valve closed and the bottom plugged, the unit was secured for travel to the water surface. To maintain integrity of the core during transit to the surface, the core was kept upright.
For unconsolidated coarse, sandy sediments, it was necessary to dig around the base of the collected core and give support at the base of the core with a closed-cell foam disk, even before removing the core from the sediment bed. A stainless-steel cutter, 300 mm long, 200 mm wide, and 0.4 mm thick, with a semicircular end, was used.
The corer has limited application in firmly compacted sediments and sediments with tightly bound root systems. The acrylic base was not designed to cut into such media and the tube is not strong enough to be hammered. The stainless-steel cutting blade was sometimes used to cut through a surface mat of plant roots.
Removing the corer assembly from the water was a critical step. The captured large-diameter core must be supported at the base during this phase, otherwise it will fall out (Burnham, 1988). Provided the EVA foam plug was held in place and the valve (Fig. 1J) was shut, transfer through the water–air interface was straightforward. Next, the outside of the corer (Fig. 1B) was cleaned of sediment and the PVC base cap (Fig. 1C) aligned with the bottom edge of the core tube. The base cap was pushed slowly onto the core base, to allow the air trapped inside the base cap to exit through the vent hole (Fig. 1E) and not upward through the sediments. Once the base cap was secured, the vent hole was sealed with duct or PVC electrical tape.
It was important to have the outer part of the core-tube base as clean as possible before sliding on the base cap. The presence of grit in the gap between the PVC cap and the acrylic core tube not only caused scoring of the surface and enhanced the likelihood of future leaks, but it also made it more difficult to remove the base cap at the end of the trials.
The top valve was then opened and the corer lid unscrewed. A plain lid was used to cap the corer, to stop the water from spilling out or swirling around too much during transport. The captured core was then placed in a foam-lined plastic box for security in transit back to the receiving laboratory.
With practice and a favorable sediment type, a diver was able to collect one core approximately every 20 min. Favorable material to core ranged from silty sediments to unconsolidated sands as discussed above. The coring operation was performed by one person with assistance from another when putting on and removing the corer lid, and when capping and sealing the core base.
Core Handling and Preparation for Reactor Experiments
Characterization of the overlying water (in particular for dissolved oxygen content, temperature, and pH) was done at the time of core collection so that ex situ reactor conditions duplicated the field conditions as closely as possible. Water samples were collected from adjacent to the sampling location just before the collection of the sediment core. This water was used in experiments on the intact cores and to measure whether there were any changes in water parameters before the start of experiments. This water was used when additional water was required.
Field lighting regimes were noted as well. The reactor design, using clear acrylic walls, permits light sources such as ambient or banks of daylight-type fluorescent tubes to be used. Most frequently, the site water was quite turbid and/or deep and the ambient light intensity at the sediment interface was low. For these cases, the reactors were kept dark for the duration of the experiments. This was achieved by wrapping the outside of the reactors in black polyethylene sheet or aluminium foil.
The sediment comprising the collected core was also characterized. This was done after subsampling the reactor core with a 50-mm-i.d. acrylic tube, immediately after the completion of the test work.
The extent of preservation and preparation of the sampled cores for reactor experiments depended largely on the processes being tested and modeled and on the chemistry of the sediments and overlying water. If four cores were being collected with a field station close by, it was usually 1 to 2 h before the reactor experiments were commenced. This delay was even greater for a more distant laboratory. If the overlying waters were oxic and the sediment was highly anoxic (frequently the case with fine silty sediments), it was necessary to keep the water column in the reactor waters oxic before starting the experiments. It has been found for sediments having a high oxygen demand that there is rapid depletion of oxygen in the water column above the core. The consequence of oxygen depletion is a change in the redox potential differential across the sediment–water interface and this typically leads to release of reduced chemical species [e.g., ammonium, Mn(II), and Fe(II) present in the pore water] into the overlying water column. Such releases would result in an incorrect measurement of flux rates from the sediment.
The problem of oxygen depletion was overcome by aerating the samples from the time of collection, using a battery-powered aquarium pump to bubble air through the solution. The PVC lids used during reactor transportation were fitted with Cole Palmer barbed-Luer connectors (EW06359-50), described previously. The air hose was attached to the Luer side and a 0.8-mm-i.d. Teflon tube to the barbed side to introduce air to the overlying water.
Temperature baths and/or temperature-controlled rooms were used to control the temperature of the corer–reactor vessels. In field laboratory situations, polypropylene tanks (490 mm long, 440 mm wide, and 410 mm deep) were used as water baths. One tank accommodated four reactor assemblies. Temperature regulation was provided by a Type F temperature controller (Haake, Berlin, Germany). Both heating and refrigeration were possible with this controller.
In many water bodies, there is significant flow-driven mixing in the water column. The stirrer used in the corer–reactor ensures mixing, but the radial flow pattern means that it does this in a way that is not readily related to the flow and turbulence regimes in the water body from which the core was obtained. This issue is discussed further by Huettel and Gust (1992). Stirring rates of 20 to 40 rpm were chosen for the work reported here to provide mixing without resuspending the sediments or permitting too much vortex pumping of solutes from the fine sediments. A stirring speed of 40 rpm was found in some cases to maintain mild turbidity similar to that in the source water body. Two sources of water mixing occur in the corer–reactors. These are stirring and bubbling air at a moderate rate (approximately 2 bubbles s-1) from a 0.8-mm-i.d. tube, using an aquarium pump.
The overlying water used for sediment experiments with the corer–reactor was typically the water that was collected with the sediments. In some situations the original water was replaced using site water with the same dissolved oxygen concentration, pH, and temperature. This was done with minimal disturbance to the sediment by carefully siphoning off most of the water and then slowly pumping in replacement water with a peristaltic pump.
Procedure for the Reactor Experiments
The procedure described below was for anoxic sediments with dissolved oxygen saturated overlying water. Typically 3 L of sediment (18-cm depth) was collected. The height of water above the cores was adjusted by removing excess water from above the cores, so that the solution volumes were similar for all reactors (3–3.5 L). The reactor lid with working stirrer and air bubbler was then screwed on and the reactor experiments commenced. Note that there is an air gap at the top, so that water does not contact the lid or metal shaft of the stirrer motor.
Under normal operation, the first sample was taken while starting the timer. Samples (30 mL) were then taken at regular time intervals, typically every 2 h over an 8- to 12-h period. They were filtered through 0.2-µm cellulose acetate filters. The samples were withdrawn through the porthole with a plastic (disposable) 50-mL Luer-lock syringe fitted with a 150-mm length of 4-mm-i.d. polyethylene tubing. Immediately after each use, the syringe–tubing assembly was thoroughly rinsed four times with Milli-Q water (Millipore, Bedford, MA). Immediately before being used to collect the next sample, the syringe assembly was rinsed at least once with the reactor liquid, which was carefully discharged back into the reactor, so as not to stir up the sediments. A separate syringe was used for each reactor.
Analysis of Physical and Chemical Parameters
Monitoring of the overlying waters in the reactors for temperature, dissolved oxygen, pH, conductivity, and turbidity was undertaken at regular intervals by inserting an appropriate probe through the porthole. When required, continuous measurements were made using probes inserted through portholes in the reactor lid.
Data Analysis
The concentration versus time data from reactor experiments was placed into Excel spreadsheets (Microsoft Corporation, 2000). The graphing facility was used to assess data and calculate slopes using a linear-least-squares fit. A flux was calculated from a slope as follows:
 |
|
Results and Discussion |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and MethodsResults and DiscussionAdvantages and Limitations of...ConclusionsREFERENCES |
|---|
Figure 2 shows an example of the use of the corer–reactors for measuring ammonia fluxes over 8 h from three replicate corer–reactors containing sediment from Rickabys Creek, a tributary of the Hawkesbury River (NSW, Australia). For three flux measurements there is high precision, a very high correlation coefficient (R2) for a linear fit to the data, and a low 95% confidence limit of the slope. Cores 1, 2, and 3 have 95% confidence limits of the slopes of 0.6, 0.5, and 0.5 NH3–N µg L-1 h-1, respectively, all <3% of the slope. However, there is quite a large variation in the slopes (fluxes) between each core. The variability is 30% of the average of these three measurements, even though the cores appeared visibly similar and were obtained from locations less than 3 m apart. This variability between "replicate" cores is typical of sediment flux measurements and reflects the short-range heterogeneity of the chemical environment in sediments (Jorgensen and Revsbech, 1985; Aller and Yingst, 1985; Davison et al., 1991; Bird et al., 1999). Thus it is critical when measuring sediment fluxes to use relatively large-diameter cores and replicate deployments, to account for the extent of environmental variability at the location being investigated.
|
A ruler is normally used to measure the reactor water column height, which is required to calculate flux (see the Data Analysis subsection). This is simpler and more robust than the procedure for benthic chambers. In that case a known volume and concentration of inert tracer is added at the start and the solution is sampled a short time after complete mixing has occurred. The volume of the chamber is determined from the measured concentration of the tracer mixed with the chamber solution and the known amount added. The water-column height is then determined from the known surface area of the chamber. This approach can also be used for determining column height for the reactors.
Withdrawal of sample from the reactors incrementally reduces the water volume or column height during the measurements. Our procedure is to measure the height at the beginning and end of the measurements and use the average of these two heights. It can be shown that provided the change in volume is small, the increase in the apparent flux, measured by the change in concentration with time, is cancelled by using the average column height. For an overall decrease in solution volume of 10% of the initial volume, the error in flux is <0.5%, while for a 20% decrease in solution volume, the error in the flux is about +3% in magnitude. These errors are small compared with environmental variability.
An additional error in the height measured by ruler is the uncertainty of position of the sediment bottom where it is uneven, which is frequent. The approach is to take an average for the uneven surface. Typically this means an uncertainty of ±5 mm in height, which for a 3-L volume (height = 18 cm) is ±3%. In the context of the measurements this error is small.
|
Advantages and Limitations of the Corer–Reactor |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and MethodsResults and DiscussionAdvantages and Limitations of...ConclusionsREFERENCES |
|---|
The main limitation of the system lies in the difficulty associated with sampling very firmly compacted materials and matted materials such as reedy water plants. The broad plastic base rim makes it impossible to push down into such sediments. The collection of cores from consolidated coarse sands is also quite difficult with this technique and it is best used with sediment types from silts to unconsolidated sands. In unconsolidated sands, however, pore water flows may affect the flux of solutes, so the reactor results may be less reliable in this sediment type. For more compacted sediments, corers made from stainless steel are more likely to be successful.
The radial mixing action of the stirrer limits the ability of the reactor to simulate accurately the effect of water flow velocity and turbulence on solute fluxes. The small paddle blade and low speeds do, however, permit efficient mixing of the overlying water without vortexing or severe disturbance of fine sediments.
More detailed comparisons of flux measurements obtained using corer–reactors with those obtained from peepers and benthic chambers in concurrent deployments will be the subject of a future publication.
|
Conclusions |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and MethodsResults and DiscussionAdvantages and Limitations of...ConclusionsREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and MethodsResults and DiscussionAdvantages and Limitations of...ConclusionsREFERENCES |
|---|
Related articles in JEQ:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Journal of Natural Resources and Life Sciences Education |
Vadose Zone Journal | ||||
| Soil Science Society of America Journal | Journal of Plant Registrations | The Plant Genome | |||
