| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Professor, Dep. of Crop and Soil Science, Oregon State Univ., Corvallis, OR 97331 USA
will.austin{at}orst.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMethodsConclusionsREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMethodsConclusionsREFERENCES |
|---|
One of the concerns associated with permanent installation of electrodes in soils is that they might become poisoned, or contaminated, such that the readings produced are erroneous and do not accurately represent even the mixed potential of a dynamic soil system. Devitt et al. (1989) observed that Pt-tipped electrodes installed in soil columns became coated with precipitates from the soil solution, which had some effect on electrode readings. Campbell (1980), however, found that poisoning was not manifest as erratic readings that would be expected if electrode function were severely impacted. Whisler et al. (1974) found that electrode potentials above -200 mV for 6-mo periods did not affect electrode function. They concluded that longevity of Pt electrodes may be inferred by the correspondence between the amplitude of oscillations of electrode readings and the intermittent flooding and drying of soil in columns.
We used Pt electrodes to measure redox potential fluctuations in soils for 3 yr at one site (Polk County) and for 5 yr at another site (Benton County) in the Willamette Valley, Oregon. When the study was completed, electrodes were carefully removed, examined, and tested to determine the extent of poisoning, or contamination, and whether or not there was any impairment of electrode function.
It is not our intent to justify whether or not Pt electrodes or any other quantitative method should be used to assess redox potentials in soils. Our objective is limited simply to evaluating the quality control of mercury-junction Pt electrodes and their reliability after several years of use in field applications.
|
Methods |
|---|
|
TOPABSTRACTINTRODUCTIONMethodsConclusionsREFERENCES |
|---|
Construction, Testing, and Installation of Electrodes
Platinum electrodes were constructed using methods described by Faulkner et al. (1989). A 13-mm length of 18-gauge Pt wire was inserted into a small hole in one end of an 8-mm Pyrex tube. The Pyrex was heated until it was permanently attached to the wire. Jewelers wax was placed around the wire inside the tube to create a watertight seal. A Cu-wire lead was inserted into the tube from the other end, and enough Hg was placed in the tube to create a junction between the Pt tip and the Cu lead. Silicon, latex, and heat-shrink tubing were used to create a seal around the Cu lead at the top of the tube.
The accuracy of newly constructed electrodes was checked by placing them in a solution of 0.1-g quinhydrone per 50 mL of distilled water buffered to pH 7 (Jones, 1966; Yu and Ji, 1993). Electrical circuits were completed by connecting the positive lead of a millivolt meter to a Pt electrode and the negative lead to a reference calomel electrode also placed in the quinhydrone. Properly functioning electrodes should give readings within 10 mV of the reference value (47 mV at 20°C) given in Jones (1966).
Electrodes whose readings were not within the specified range were removed from the quinhydrone, rinsed with distilled water, scrubbed with household cleanser, then soaked in distilled water for at least 2 h before retesting. Electrodes whose readings still were outside the specified range were further treated by soaking for 30 min in a 1:1 mixture of concentrated hydrochloric acid and concentrated nitric acid, then soaking overnight in distilled water and retesting in fresh quinhydrone solution.
All electrodes that were deemed accurate by quinhydrone testing were considered ready for field use and were stored in a clean environment until needed.
Platinum electrodes were installed in triplicate at depths of 25, 50, and 100 cm by first making a pilot hole the same diameter as the Pyrex tube and 2 cm shallower than the desired depth. A hollow tube slid over the Cu lead of the electrode was used to push the electrode down the pilot hole and seat the Pt firmly into the soil 2 cm beyond the bottom of the pilot hole. Care was taken during installation to neither break the Pyrex nor bend the Pt wire, as both would lead to erroneous readings. After insertion, the pilot hole was packed with bentonite to within 5 cm of the soil surface and covered with a tamped soil plug.
Field Monitoring
Each of the replicate electrodes at a given depth was color-coded so that readings from an individual electrode could be identified during the term of the study. Readings were made with a volt meter and recorded as thousandths of a volt. In the field, the electrical circuit was completed by connecting the positive lead from the volt meter to the Cu lead of the electrode and the negative lead to a calomel reference electrode placed in 4 M KCl and connected to the soil with a salt bridge similar to that used by Veneman and Pickering (1983).
Readings were made at weekly intervals between November and June for 3 consecutive yr at the Polk County site and for 5 consecutive yr at the Benton County site.
Removal of Electrodes
After each site was monitored, electrodes were carefully removed to avoid breaking the Pyrex or bending the Pt. Soil was gently excavated until the top 5 cm of an electrode was exposed. Carefully wiggling and pulling on the electrode was then enough to remove it. Electrodes were preserved in their as-removed condition with as little disturbance as possible to any soil material remaining on the Pt tip or the Pyrex body.
With this procedure, 39 of the 47 electrodes installed at the Polk County site were recovered intact, and all 63 electrodes installed at the Benton County site were recovered intact. These 102 electrodes were subsequently used to assess their accuracy after several years of field use.
Evaluation of Electrode Viability
In situ Evidence
Soil redox potentials measured with Pt electrodes are at best semi-quantitative expressions of mixed potentials in a nonequilibrium environment. As a consequence, the values obtained are best used to show how redox potentials change through time and to make comparisons among different soil environments. The expectation that readings from permanently installed Pt electrodes should change as the depth to saturation, duration of saturation, and temperature of the soil environment changes provides one test of electrode function.
Such a test is given in Fig. 1 , which illustrates a pattern of a properly functioning electrode. During December and early January, when the soil at 25 cm is not saturated, electrode potentials nevertheless fluctuate somewhat in response to rainfall and changes in soil temperature (data not shown). At the onset of saturation in mid-January, redox potentials begin to decline rapidly and continue to do so throughout the duration of saturation. Electrode potentials rise again upon desaturation, and presumably re-aeration, then fall again with the onset of a second period of saturation. In the spring, as the soil warms and dries, redox potentials return to relatively high levels and remain there, suggesting the soil is in a well oxidized state.
|
Readings from only one electrode (Number 6-4 in our scheme) out of 120 installed were sufficiently erratic to suggest failure based on field performance. Failure occurred during the first year of operation, and readings were discontinued. Another electrode was installed to take its place. This line of evidence, however, suggests that more than 99% of the electrodes installed were still functioning properly after either 3 yr or 5 yr, and it was relatively easy to identify a pattern of behavior indicating improper electrode function.
Visual Evidence
Electrode poisoning is defined as any chemical or physical change in the status of the Pt electrode that prevents electrochemical reduction of oxygen at the Pt surface and leads to erroneous measurements (Devitt et al., 1989). It usually refers to the development of coatings, or precipitates, on the Pt surface composed either of salts (Rickman et al., 1968) or dark-colored metal oxides combined with carbonates and organic C (Devitt et al., 1989).
Both Rickman et al. (1968) and Devitt et al. (1989) suggest that any adverse effects due to poisoning can be minimized by removing electrodes after 2 mo of continuous operation in soil and cleaning the wire tips in acid.
Because our electrodes were in the ground for much longer than 2 mo, we wanted to determine whether or not electrode poisoning had occurred to the extent of creating erroneous readings. To do that, we examined each electrode after it was removed from the soil, both by eye and with the aid of up to 100x magnification. In all cases, the electrodes were free of conspicuous deposits and precipitates, and most were free of soil aggregates. Pt wires appeared to be lustrous and shiny, though not overly bright. We conclude from these observations, combined with in situ evidence, that electrode performance was not impaired by poisoning after 3 or 5 yr of permanent installation in the soil.
Electrode removal did reveal a type of failure that we hadn't expected, namely the presence of water inside the Pyrex tube of the electrode. Water was observed in 21 of 39 electrodes removed from the Polk County site and in 43 of 63 electrodes removed from the Benton County site. Although it seems possible that this water could interfere with electrode function, the patterns of readings did not suggest such a problem. Subsequent testing of electrode function in the laboratory also indicated that the presence of internal water did not adversely affect the readings obtained.
Because the Pyrex tubing was not broken or cracked, water must have entered the electrode either through the wax seal at the Pt end or the silicon–rubber seal at the Cu-lead end, or both. To determine if the wax seal had become defective, we took the electrodes removed from the soil to the laboratory, removed the heat-shrink tubing and silicon seal, cleaned the inside of the body with distilled water, and applied a fluctuating positive air pressure of 85 to 100 kPa to the inside of the electrode body. The electrode tip was then submerged in water. Emergence of air bubbles indicated a defective seal.
Results of this testing indicated that only one (Number 1-8 in our scheme) of the 102 electrodes removed had a defective wax seal. Inspection of this electrode revealed that the joint between the Pt and the Pyrex had broken, and the wire could be moved partially through the opening in the Pyrex. We don't know whether this breakage occurred during construction or during installation, but in any case the breakage and entry of water did not appear to have impaired electrode performance.
Inspection of the silicon–rubber seals at the other end of the electrodes revealed that on some electrodes, the heat-shrink tubing appeared to be torn, which probably occurred during installation with the hollow tube. Torn heat-shrink tubing may have been the reason for water entry into some electrodes, but it is not the sole reason, as some electrodes with torn tubing did not contain interior water, and many other electrodes that did contain water had intact heat-shrink tubing. We also observed that on some electrodes, the silicon seal appeared to have pulled away from the Pyrex tubing, creating another potential source of water entry.
In general, we conclude that the silicon–heat-shrink seal was not always watertight for a variety of reasons, and water entered many of the electrodes through the upper end of the Pyrex tube. Had we used a different electrode design, such as the welded design of Faulkner et al. (1989), it is quite possible that water entry into the electrode could have been precluded entirely.
Laboratory Testing: Quinhydrone
Recovered electrodes were tested in quinhydrone using the same procedure as was used to conduct pre-installation quality checks. Only the Pt tips and a small area of the adjoining Pyrex tubes were immersed in the test solution. Care was taken to prevent soil particles adhering to the electrode from getting into the test solution.
Both pre-installation and post-removal quinhydrone test results are shown in Fig. 2 for Polk County and Fig. 3 for Benton County. Means and standard deviations calculated from the data are given in Table 1 . In both the figures and the table, values reported are for deviations of the observed redox potentials from the standard value given by Jones (1966), rather than for the actual redox potentials measured.
|
|
|
Post-removal data show more variation, but 99 of the 102 electrodes tested were still within the acceptable window of tolerance. These data suggest that more than 97% of our electrodes were still functioning properly and giving valid readings even after continuous installation in the soil for up to 5 yr. After removal, we observe an increase in the mean value of the deviation from the standard, and the standard deviation increases markedly. Clearly there is more variation among electrodes after several years in field soil, yet the vast majority of these electrodes are still functioning within acceptable limits.
The Polk County quinhydrone data (Fig. 2) are particularly interesting because of the presence of one obvious outlier, electrode Number 23 in Fig. 2, which was removed from the Dayton soil. This electrode's quinhydrone reading deviated from the standard value by -43.9 mV, well outside the 10-mV range. This electrode, however, was the very same electrode Number 6-4 whose behavior had been observed as erratic during the first year of use, readings from which had been discontinued accordingly. The magnitude of its deviation from the standard value explains the rather large standard deviation (9.30) of the post-removal test values for the Polk County electrodes (Table 1). Though it's always dangerous to censor data after the fact we think that because we discontinued field readings upon observing erratic behavior, it may be appropriate in this case to exclude this electrode from the data set. Had we done that, the mean post-removal test value would have been 3.23, and the standard deviation would have been 4.20. Excluding this outlier also would have meant that only two electrodes remain just outside the 10-mV range, and 98% of the electrodes could be said to function adequately after 5 yr.
The real value of this outlier in the data set is its confirmation of both field and laboratory methods for detecting improperly functioning electrodes. The quinhydrone test is sufficiently sensitive to identify a failed electrode, and the fact that the electrode identified by the lab test corresponds exactly to the one we suspected had failed in the field reinforces the validity of using expected patterns of field readings to determine whether or not electrodes are functioning properly.
We also used the post-removal quinhydrone test to independently evaluate the effect of water entry into electrodes on electrode performance (Table 1). For both Polk and Benton Counties, the mean post-removal test value of the water-containing electrodes was closer to the zero reference point than the mean of the electrodes that did not contain water, and the standard deviation of the water-containing electrodes was lower than the standard deviation of the dry electrodes. While we would not want to claim, based only on this evidence, that water-containing electrodes outperformed the dry ones, these data do corroborate other lines of evidence that the presence of water inside the electrode tube was not detrimental in any way to electrode performance.
Post-removal quinhydrone testing also allowed further evaluation of the extent of poisoning of electrodes. Values from two electrodes (Numbers 24 and 27 in Fig. 2) were just above the +10-mV allowable deviation from the standard value. These two electrodes were cleaned with a mixture of HCl and HNO3, soaked in distilled water, then retested in quinhydrone. Both then tested within the ±10-mV range. This evidence suggests that, despite the lack of visual evidence, there were sufficient coatings on the Pt wire to poison the electrodes to the point of impairing their performance. Such poisoning as did occur does not appear to be a function of the length of time in the soil, since the two poisoned electrodes had been in the ground only 3 yr, whereas none of the electrodes in the ground for 5 yr showed similar signs of impairment.
Nine additional Polk County electrodes were selected randomly for acid cleaning and retesting. Upon retesting, seven of the nine gave test results a little closer to the zero point than previously. This, too, is evidence of a mild poisoning effect, but it was not sufficient to impair electrode performance.
Laboratory Testing: Slurry
One final post-removal test measured Eh values in a 1:5 slurry of silt loam soil–water that was continuously homogenized by stirring (G. Brown, personal communication, 1996). The soil used for this test was taken from the Ap2 horizon (8–20 cm) of a moderately well-drained Woodburn soil similar to the Woodburn soil at the Benton County study site. Whereas the quinhydrone solution is highly poised, the slurry should more nearly represent real soil conditions, except that by constant stirring, much of the in situ microsite variability that affects characterization of soil redox potentials should be removed.
Data were collected from 14 sets of replicated electrodes from the Benton County transect. Each set consisted of between three and six electrodes that had been installed at a given depth (25, 50, or 100 cm) in one of the soils along the transect. After removal and following quinhydrone testing, the electrodes in each set were inserted through holes in a cardboard template that, when placed on top of a 1000-mL beaker containing the slurry, allowed for uniform immersion into the slurry. The beaker was placed on a magnetic stirrer and was stirred continuously for 18 h before reading the Eh values. Then 10 g of sucrose was added to each slurry to create anaerobic conditions, and after an additional 30 h of stirring, a second set of readings was taken.
Slurry test data are summarized in Table 2 and compared with electrode data from both actual field readings and quinhydrone testing. Both slurry and field data are reported as percentage of pairs within 20 mV. A set of four electrodes, for example, would generate six possible pairs for comparison (1-2, 1-3, 1-4, 2-3, 2-4, 3-4). For the slurry data, we simply calculated the percentage of all possible pairs in a set for which the difference in the readings was <20 mV. For the field data, we had readings from all of the electrodes in a set on each of 29 dates throughout the 1995 to 1996 field season. We calculated the percentage of pairs within 20 mV on each date, then averaged over all 29 sampling dates to arrive at the data reported in Table 2.
|
Slurry readings, by removing sources of variation due to microsites and weather, are less variable than the corresponding field data, and the percentages of all possible pairs within a set whose values are within 20 mV are higher than for the field data. Ideally, all of the slurry readings in a set would have been within 20 mV of each other, as they are in the quinhydrone solution. Only 3 of the 14 electrode sets, however, met this criterion in the 18-h slurry readings, and none of the sets met this criterion after 30 h of induced anaerobic conditions.
The quinhydrone solution is a highly poised system, and the readings from all possible electrode pairs in each set easily fell within 20 mV of each other (Table 2). Because the soil slurry is not a strongly poised system, it is reasonable to expect wider ranges of variation than in the quinhydrone. However, many of the ranges in the 18-h data do fall between 20 and 40 mV; so if we were to use a 40-mV range as a standard, rather than 20-mV, 10 of the 14 sets have electrodes that could be assumed to be providing reliable readings. We might then also conclude that each of the other four sets contains one or more electrodes that may not be providing reliable data, despite quinhydrone test readings that fall within acceptable ranges.
Further observations of the data in Table 2 substantiate the value of the slurry test. First, the decrease in variability from the field data to the slurry data to the quinhydrone data is obvious and agrees with known reasons for variability in each medium.
Second, the slurry data suggest that electrode readings are intrinsically more variable under reducing conditions than under oxidizing conditions. Constant stirring of the slurry after adding sucrose should eliminate microsite variation, yet the millivolt range and the coefficient of variation within most sets of electrodes are higher, and the percentage of all possible pairs whose values are within 20 mV is lower under anaerobic conditions (48-h data) than under aerobic conditions (18-h data). Whether this means that field-installed electrodes are intrinsically less reliable under soil reducing conditions than under fully aerobic conditions remains to be evaluated.
Third, the slurry data do provide strong evidence that some electrodes are not functioning properly, despite quinhydrone data to the contrary. The electrodes from Wi-B 50 illustrate this effect. The millivolt ranges (Table 2) from both the 18-h reading (66) and the 48-h reading (152) seem anomalously high in relation to all other values, and the coefficients of variation are also high. The actual slurry data for these four electrodes (Table 3) show that one electrode, which was installed after the original three, gives a substantially lower reading under aerobic slurry conditions and a substantially higher reading under anaerobic slurry conditions. This suggests that the electrode may be recording mid-range soil potentials properly, but is not properly sensing potentials under either strongly reducing or strongly oxidizing conditions.
|
|
|
Conclusions |
|---|
|
TOPABSTRACTINTRODUCTIONMethodsConclusionsREFERENCES |
|---|
Water was observed to have entered more than half the electrodes, presumably through the heat-shrink seal around the Cu lead, but there was no evidence that the presence of water in any way interfered with the proper functioning of the electrodes.
Electrode testing in quinhydrone solution after removal from the soil showed that 97% of the electrodes still performed within acceptable limits of accuracy after either 3 or 5 yr of continuous use. The quinhydrone test accurately identified an electrode suspected of being faulty based on field observations and is considered a reliable test of electrode viability following a period of field use.
Post-removal testing in a soil slurry provided results that were more variable than results from the quinhydrone test but were substantially less variable than data from the actual field readings. Although the slurry test is less definitive than the quinhydrone test, it does provide a means of identifying electrodes that may be giving faulty readings when installed in the soil, even though the quinhydrone test would indicate they are viable electrodes.
Received for publication September 11, 1998.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMethodsConclusionsREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. Vorenhout, H. G. van der Geest, D. van Marum, K. Wattel, and H. J. P. Eijsackers Automated and Continuous Redox Potential Measurements in Soil J. Environ. Qual., July 1, 2004; 33(4): 1562 - 1567. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. V. D'Amore, S. R. Stewart, and J. H. Huddleston Saturation, Reduction, and the Formation of Iron-Manganese Concretions in the Jackson-Frazier Wetland, Oregon Soil Sci. Soc. Am. J., May 1, 2004; 68(3): 1012 - 1022. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Clausnitzer, J. H. Huddleston, E. Horn, M. Keller, and C. Leet Hydric Soils in a Southeastern Oregon Vernal Pool Soil Sci. Soc. Am. J., May 1, 2003; 67(3): 951 - 960. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. van Bochove, S. Beauchemin, and G. Theriault Continuous Multiple Measurement of Soil Redox Potential Using Platinum Microelectrodes Soil Sci. Soc. Am. J., November 1, 2002; 66(6): 1813 - 1820. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. V. D'Amore, S. R. Stewart, J.H. Huddleston, and J.R. Glasmann Stratigraphy and Hydrology of the Jackson-Frazier Wetland, Oregon Soil Sci. Soc. Am. J., July 1, 2000; 64(4): 1535 - 1543. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
