JEQ Grow Your Career With ASA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Related articles in JEQ
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via ISI Web of Science (1)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Gjettermann, B.
Right arrow Articles by Hansen, S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Gjettermann, B.
Right arrow Articles by Hansen, S.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Gjettermann, B.
Right arrow Articles by Hansen, S.
Related Collections
Right arrow Laboratory Column Studies
Right arrow Nonequilibrium Flow
Right arrow Unstable Flow/Fingering
Right arrow Phosphorus
Published in J. Environ. Qual. 33:2263-2271 (2004).
© ASA, CSSA, SSSA
677 S. Segoe Rd., Madison, WI 53711 USA

TECHNICAL REPORTS

Vadose Zone Processes and Chemical Transport

Transport of Phosphate through Artificial Macropores during Film and Pulse Flow

Birgitte Gjettermanna,*, Hans Christian B. Hansena,b, Henry E. Jensena and Søren Hansena,b

a The Royal Veterinary and Agricultural University, Department of Agricultural Sciences, Laboratory for Agrohydrology and Bioclimatology, Højbakkegård Allé 9, DK-2630 Taastrup, Denmark
b The Royal Veterinary and Agricultural University, Department of Natural Science, Thorvaldsensvej 40, DK-1871 Frederiksberg C, Denmark

* Corresponding author (bgj{at}kvl.dk)

Received for publication November 10, 2003.

   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Flow through artificial macropores may occur as a water film along the macropore walls (film flow) or as moving water segments separated by air bubbles (pulse flow). To investigate the effect of macropore flow pattern (i.e., film and pulse flow) on the interaction of solutes with macropore walls, we studied orthophosphate (P) transport and sorption in artificial macropores. The experimental setup consisted of a column (height = 20 cm, diameter = 20 cm) homogenously packed with glass beads and fitted at outflow with a vertical artificial macropore placed below the column. The artificial macropore consisted of ceramic tubes (3 or 8 mm i.d.; 31.5 cm long) coated on the inside with iron oxide serving as phosphate sorbents. An orthophosphate solution containing 0.04 mg P L–1 was applied at a rate of 9 to 12 mm h–1 to the column, eventually causing macropore flow. In the 8-mm-i.d. tubes only film flow occurred. Pulse flow was dominating in the 3-mm-i.d. tubes. Generally, the flow patterns were reproducible and seldom did pulse flow replaced film flow or vice versa. During film flow, a significantly larger decrease in macropore P concentration per tube was observed relative to that with pulse flow events. However, pulse and film flow lead to almost the same amounts of P sorbed per unit surface area when exposed to the same solute P concentration. Comparison with P sorption capacity experiments indicated that the sorption rate, rather than the sorption capacity, controls the amount of sorbed P during macropore flow in the studied system.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
IT HAS BEEN repeatedly demonstrated that macropore flow occurs in many soils (Ritsema et al., 1993; Flury et al., 1994; Stamm et al., 1998; Trojan and Linden, 1998; Laubel et al., 1999), which may imply rapid transport of water, solutes, and particulate matter from the topsoil to the subsoil, tile drain, or ground water. Since only a small fraction of the soil volume is involved in macropore flow, the velocity at which the water moves through the soil and the depth of penetration are much greater than when the entire soil matrix is involved in the flow process (Stamm et al., 1998; Shipitalo et al., 2000). Consequently, macropore flow makes it possible even for strongly sorbing species (e.g., pesticides and phosphate) to penetrate deep into the soil profile (Kladivko et al., 1991; Edwards et al., 1993; Gaber et al., 1995; Stamm et al., 1998). Geohring et al. (2001) performed column studies using packed soil and artificial macropores to examine the role of macropore size on P sorption to pore walls. They found that soluble P was transported through artificial macropores with diameters of 1, 2, and 3 mm by negligible P sorption to pore walls. In the absence of macropores, no measurable P was transported through the packed soil columns.

Macropore flow is not necessarily initiated at the soil surface and it may terminate at any depth. Likewise, matrix flow can start where macropore flow ends. Soil macroporosity and the proportion of the precipitation moving through macropores often increase with the adoption of conservation tillage (Shipitalo et al., 2000). Although tillage creates larger total porosity, macropore continuity is reduced (Roseberg and McCoy, 1992) and a compacted layer, which often occurs at the transition zone between the plow layer and the subsoil, may cause water accumulation and thus initiate macropore flow (Andreini and Steenhuis, 1990; Petersen et al., 1997). This phenomenon has been demonstrated in several dye tracing experiments (Flury et al., 1994; Gjettermann et al., 1997; Petersen et al., 1997). However, the conditions for initiating macropore flow, the flow mechanisms of solutes, and interactions between the soil matrix and macropores were not fully explained during these studies.

It is a well-established hypothesis that macropore flow may occur when rainfall intensities exceed the saturated hydraulic conductivity of the soil matrix (Beven and Germann, 1982; White, 1985; Brusseau and Rao, 1990) and several studies related to macropore flow have been performed by applying ponding (e.g., Wilson and Luxmoore, 1988). However, Shipitalo et al. (1990) and Edwards et al. (1992) observed that macropore flow can be initiated if local, saturated areas occur in the soil matrix. Studies performed by Phillips et al. (1989) and Tofteng et al. (2002) showed that macropore flow can be maintained even under unsaturated conditions. Furthermore, these authors observed that flow through artificial macropores could occur as a water film along the macropore walls (film flow) or as moving water segments separated by air bubbles (pulse flow).

Ghodrati et al. (1999) also observed film flow in an artificial macropore system. They observed that macropore flow occurred predominantly at the matrix–macropore interface and in the matrix zone directly adjacent to the macropore. Along this interface, only film flow was observed. At higher fluxes where the system was ponded, excess flow into the macropore (i.e., running water) was observed, which was probably a result of the matrix–macropore interface film increasing in thickness to the point where excess flow became visually apparent (Ghodrati et al., 1999). Flow modes of fingered flow in unsaturated fractures and sand have also been examined (Glass and Nicholl, 1996; Su et al., 1999, 2001). Su et al. (1999) conducted flow visualization experiments on an initially dry, transparent epoxy replica of natural rock fracture to identify important physical mechanisms controlling seepage of liquids in unsaturated fractures. Highly localized and nonuniform channels consisting of broader, water-filled regions connected by thin threads of liquid were observed. Even though great care was taken to maintain steady boundary conditions, the flow generally proceeded in an intermittent manner. Threads along the flow channel would snap, drain, and then reform again. These intermittent flow events occurred periodically.

Due to the observed different flow types in fractures, macropores, and porous media, the aim of the present study was to investigate (i) the macropore flow pattern in relation to initiation and flow type; (ii) the magnitude of P sorption during macropore flow; and (iii) the effect of the flow type in the macropores (i.e., film and pulse flow) on P sorption. This paper presents results obtained from an experimental setup illustrating a simplified two-layer system of matrix and macropore using glass beads as column matrix material and long ceramic tubes coated with P-sorbing material as artificial macropores. Furthermore, the paper presents the methods developed for coating ceramic tubes with iron oxides.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Experimental Setup
The experimental setup consisted of a homogenously packed column with a single outlet at the bottom leading to an artificial macropore made of a ceramic tube (Fig. 1) . The bottom plate of the column consisted of a 1-cm-thick Plexiglas plate with a 10-mm-wide hole in the middle. The ceramic tubes were mounted in this hole with fittings so the top of the tube was in line with the surface of the plate. To prevent the glass beads from escaping the column, a metal net of stainless steel with a net spacing of 250 µm covered the hole in the Plexiglas plate. Solution was applied to the column with a peristaltic pump through an automatic sprinkling device with 90 syringe needles to ensure a uniform application rate. The pressure potentials of the water were measured at the elevation of the bottom plate of the column at 2.5, 5.0, and 7.5 cm from the center of the plate using sintered glass tensiometers (Art. 112.000-4; Struers Kebo Lab, Roskildevej, Denmark). Ceramic cup tensiometers (1.5 cm long; Gravquick, Herlev, Denmark) were installed in the column at the heights of 12.1, 14.6, and 17.0 cm above the bottom plate, protruding 7 cm into the column matrix. All the tensiometers were connected to pressure transducers (Micro Switch 26PC PK 8875 2; Honeywell, Morristown, NJ), which were calibrated before use. Two balances (Sartorius AG, Goettingen, Germany) measured the weight of the influent and effluent, respectively. A thermopile measured the air temperature. The weights of the influent and the effluent, the air temperature, and all the pressure measurements were recorded continuously by the data logger (dataTaker, Rowville, VIC, Australia) coupled to a computer. The fraction collector (Foxy Jr.; Isco, Lincoln, NE) coupled to the data logger sampled the effluent passing through the artificial macropore into a funnel placed below (Fig. 1). A computer program linked to the data logger recognized a drop in pressure measurements, by the tensiometers close to the artificial macropore at the beginning of a flow event, which then activated the fraction collector. The fraction collector sampled every 50 s during the first hour of film flow; thereafter the sample frequencies were reduced to one sample per 5 min, and after several hours it was further reduced to one per 30 min. During pulse flow, the fraction collector sampled every 20 s. For additional details of the experimental equipment reference is made to Tofteng et al. (2002).



View larger version (15K):
[in this window]
[in a new window]
  		Fig. 1. Experimental setup showing glass bead column and an artificial macropore installed.

 
Column Matrix
In the present study the column consisted of glass beads (Art. A553.1; Carl Roth GmbH, Karlsruhe, Germany) which were 0.25 to 0.50 mm in diameter. The glass beads were acid-washed and rinsed with double-deionized water, dried at 40°C, and packed in the column to an average bulk density of 1.53 g cm–3, giving a porosity of 0.43 and total column pore volume of 2590 mL assuming a particle density of 2.68 g cm–3. The water retention curve was measured by initially water saturating the column followed by stepwise draining to a pressure potential of –30 cm and then stepwise rewetting to saturation. The pressure potentials were measured at the bottom of the column by the installed tensiometers and transducers.

Artificial Macropores
The artificial macropores were made of ceramic tubes (Phytagoras 610; W. Haldenwanger, Munich, Germany), 31.5 cm long and coated with iron oxides based on the techniques of impregnated filter paper by Lin et al. (1991). Ceramic tubes with inside diameters of 3 and 8 mm, respectively, were used. The iron oxide coatings were produced as follows. Acid-washed tubes were dipped into a solution containing 0.65 M FeCl3·6H2O acidified with concentrated HCl (50 mL L–1). After 24 h, the tubes were removed from the solution and kept vertical until dripping stopped and then dried for a further 24 h at room temperature. This treatment was repeated once. The tubes dried for several weeks at room temperature, until completely dry. Then, the tubes were dipped vertically and smoothly into 2.7 M NH3 solution for exactly 45 s to ensure complete and homogeneous conversion from iron chloride to iron oxide. Thereafter to make sure that they were rinsed absolutely, they were dipped vertically into double-deionized water for 15 s and lifted into a second container with double-deionized water for 15 s and further into a third container with double-deionized water, where they stayed for an hour. Finally, coated tubes were dried at room temperature. The tubes were colored yellow-brown by the iron oxides, and were kept in polyethylene bags until use.

Sorption Capacity
The phosphate sorption capacity of the iron oxide–coated ceramic tubes was determined by exposing the tubes to 1000 mL aqueous solution containing 0.2 and 0.04 mg P as KH2PO4 for the 8- and 3-mm tubes, respectively. The tubes were placed in a closed circulating system comprising a reservoir, a pump, and the ceramic tube. The tubes were placed vertically and the solution was pumped at a rate of 6 x 10–4 m3 s–1 from the bottom to the top. The experiment was repeated three times for the 3- and 8-mm tubes, and each time new tubes were used. Between 20 and 40 samples of solution for analysis of P were collected during a maximum period of 9 d, with the first sample taken after 10 min. Control measurements were made using tubes with no iron oxide coating for both the 3- and 8-mm tubes; two replicates were made.

Flow Experiments
To establish a chemical steady state in the column matrix, influent containing P (0.04 mg P L–1 added as KH2PO4) was passed through the glass bead column (10 L d–1), with no tube installed, for 2 wk before the flow experiments were started.

Each flow experiment was conducted as follows. Influent was applied to the column through the sprinkling device at a fixed intensity. When the effluent had passed out of the column bottom hole with no tube installed for about 10 min, four replicate samples of the outflow solution were taken for P analysis. The average P concentration was subsequently used as a measure of input P concentration to the macropore. Before starting a new flow experiment the column needed to be drained before installing a new artificial macropore, otherwise pulse or film flow could occur before the data logger and fraction collector were activated. After draining the column, which was done by holding a 3-mm tube tightly to the outlet of the column, the actual tube was installed and the experiment started. Thus, each flow experiment was started using a new, dry, artificial macropore. Now, the data logger program was started to begin receiving outputs from the transducers, the thermopile, and the two balances, which measures the weight of the influent and the effluent, respectively. As mentioned previously, it was the output from the transducers that the data logger program recognized and thereby controlled the activity of the fraction collector. However, to ensure collection of the very first effluent, manual sampling was performed at the beginning of pulse flow events. The weight of samples and fraction numbers were manually listed for later use in the calculation of outflow rate.

A total of eight flow experiments were conducted: Experiments 1 to 4 with 8-mm tubes (Table 1) and Experiments 5 to 8 with 3-mm tubes (Table 2). For simplicity the influent intensity was kept at 9 to 12 mm h–1 in the experiments, except for one experiment (Experiment 4), where the intensity was increased in an attempt to change the flow pattern. Furthermore, the experiments differed in duration and number of pulse flow events (Tables 1 and 2). The influent P concentration for all experiments was set at 0.04 mg P L–1 added as KH2PO4.


View this table:
[in this window]
[in a new window]
  		Table 1. Characteristics of flow pattern and P sorption during film flow in 8-mm ceramic tubes for Experiments 1 to 4.

 

View this table:
[in this window]
[in a new window]
  		Table 2. Characteristics of flow pattern and P sorption during pulse flow in 3-mm ceramic tubes for Experiments 5 to 8.

 
Chemical Analysis
Samples from the flow and sorption capacity experiments were collected in polypropylene tubes and stored in a refrigerator (5°C) until P analyses took place. The P determinations were performed within 24 h of sampling by flow injection analysis (Ruzicka and Hansen, 1988). The determination was based on formation of molybdophosphate blue using hydrazine and tin(II)chloride as reducing agents (Janse et al., 1983). The detection limit of the procedure was 5 µg P L–1. All glassware used for analysis and all equipment that was in contact with the column matrix or the tubes were acid-washed in 6.5 M HCl (approximately 15 min) and 1.45 M HNO3 (approximately 1 h) and carefully rinsed with water before use. Double-deionized water was used throughout.

Output Data
Data recorded by the data logger and the measured effluent P concentrations were gathered according to the logged output signal to the fraction collector. The weight of effluent samples for P analysis was added to the logged weight of the effluent and subsequently the outflow rate was calculated. The repeated filling and emptying of the funnel (Fig. 1) during sample collection caused some noise on the effluent weight and outflow rate. Therefore, the outflow rate data were recalculated using moving average around every single estimated outflow rate between two sample collections.

To test the effect of flow type on the amount of phosphate sorbed to the artificial macropores, one-dimensional analysis of variance was applied to the calculated amount of sorbed P data for film flow and pulse flow events. Furthermore, a t test was applied to determine if there were any significant differences in P concentration between the first effluent samples and the later samples when the P concentration was almost constant.


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Water Retention
Figure 2 shows the water drainage and wetting curves for the glass bead column in the range of 0 to 30 cm tension. The hysteresis reflects effects of air entrapments and water adsorption to the glass beads in the column. From the drainage curve and the capillary rise equation (assuming circular tube and zero contact angle) it was estimated that approximately 28% of the pore volume was made up of pores larger than 110 µm in diameter. Approximately 7 to 8% of the pore volume consisted of pores with diameters in the range of 110 to 120 µm.



View larger version (15K):
[in this window]
[in a new window]
  		Fig. 2. Measured water drainage and wetting curves for the glass bead column used in the experiments.

 
Observed Characteristics of Film and Pulse Flow
Two types of macropore flow were observed, film and pulse flow. Typical examples of the flow patterns during film and pulse flow are illustrated in Fig. 3 and 4 , respectively. During film flow, the flow was stationary. On the contrary, during pulse flow intermittent drainage occurred in the column. The pulse flow events were short and there was no evidence of correlation between in- and outflow rates.



View larger version (32K):
[in this window]
[in a new window]
  		Fig. 3. Example of film flow in an 8-mm tube, Experiment 3. (a) Pressure head at the bottom of the column, cumulative amount of outflow. (b) Sorbed orthophosphate estimated as decrease in soluble P concentration during transport through the macropore, and outflow rate with time.

 


View larger version (36K):
[in this window]
[in a new window]
  		Fig. 4. Example of pulse flow in a 3-mm tube, Experiment 6. (a) Pressure head at the bottom of the column, cumulative amount of outflow. (b) Sorbed orthophosphate estimated as decrease in soluble P concentration during transport through the macropore, and outflow rate with time for the first pulse flow event.

 
Pulse flow did not occur in the 8-mm-i.d. tubes at all, only film flow was observed. Pulse flow as well as film flow was observed in the 3-mm-i.d. tubes. However, most of the flow events in 3-mm tubes occurred as pulse flows. At the time when film flow occurred in the 3-mm tubes, the system was not set up for sampling by the computer. Therefore, during sampling only pulse flow occurred in the 3-mm tubes. Generally, the flow patterns were very reproducible and seldom did pulse flow replace film flow or vice versa in a flow event or experiment.

Generally, the pressure potential at the bottom of the column was positive when film flow occurred (Fig. 3a; Table 1). Film flow could continue for several hours, and was recorded up to 15 h. In the middle of the column the pressure potentials did not show a systematic pattern. In Experiment 1 negative pressure potentials occurred at heights of 17.0, 14.6, and 12.1 cm above the artificial macropore during film flow (not shown). However, in Experiments 2, 3, and 4, positive pressure potentials occurred at 17.0 cm above the artificial macropore and negative pressure potentials at 12.1 and 14.6 cm above the macropore (Fig. 5a) .



View larger version (30K):
[in this window]
[in a new window]
  		Fig. 5. Pressure head at different heights in the soil column during (a) film flow in Experiment 2 and (b) pulse flow in Experiment 5. Height = 0 cm is set at the bottom, while height = 20 cm is the surface of the glass bed column.

 
In most cases, pulse flow was initiated when the pressure potential at the bottom of the column was close to zero (Fig. 4a; Table 2). Pulse flow events could repeatedly go on for several hours. A pulse event lasted approximately 10 to 12 min and resulted in drainage of 560 to 720 mL; however, during that time only 50 to 65 mL of influent was applied. The drained outflow during a pulse event accounts for approximately 18 to 25% of the total pore volume of the column plus the applied influent during one pulse event. The part of the column that drained the most was located in the zone between the bottom of the column and 12 cm above the artificial macropore, as the pressure potential decreased to approximately –26 cm at 12.1 cm above the macropore during a pulse flow event (Fig. 5b). A tension of 26 cm corresponds to a pore diameter of approximately 120 µm. Pores larger than 120 µm in diameter make up approximately 21% of the total pore volume in the column in reasonable agreement with the drained volume. However, in Experiments 5 and 6, the upper 5 cm of the column were not drained (Fig. 5b). In these experiments, the drained part of the column accounted for a maximum of three-quarters of the total column volume, which probably had pressure potentials decreasing even further than –26 cm in some areas (Fig. 5b). In Experiments 7 and 8, only the top 3 cm of the column were not drained (data not shown). Thus, during pulse flow most of the water was located at the top of the column, and the lower part of the column (0–12 cm above the macropore) was drained.

In general, film and pulse flow was initiated at positive pressure potentials in our studies. However, macropore flow was maintained during unsaturated conditions, due to the water storage capacity of the porous medium in the column, when hydraulic contact was established with the macropore. Other authors have observed film and pulse flow in artificial macropores (Tofteng et al., 2002). The flow patterns observed by these authors were similar to the patterns observed in this study. However, more variations related to flow type in a single experiment were observed. These authors used a similar experimental setup to investigate water flow in the system, but the artificial macropores consisted of 50-cm-long Duran glass pipes and the column matrix was packed sand having particle diameters in the range of 0.1 to 0.3 mm. They often observed that film flow replaced pulse flow in a flow event, or that film flow and pulse flow occurred in two separate events but in the same experiment. The causes of the different flow type events in this study compared with the observations by Tofteng et al. (2002) are difficult to elucidate. However, it could probably be related to the different surface properties of the macropore walls or the length of the macropores.

Su et al. (1999) also observed flow proceeding in an intermittent manner in artificial rock fractures using transparent epoxy cast as fracture replica. Threads along a flow channel would snap, drain, and then reform again as water accumulated behind the advancing air–water meniscus and then formed a thin thread behind the advancing front. These intermittent flow events occurred periodically. In one experiment, they observed two channels forming; however, only one channel continued to flow at a pressure of –3.5 cm water. By increasing the pressure from –3.5 to –2.5 cm flow also occurred in the other channel. As the pressure was further increased in the subsequent days, water continued to flow intermittently through the two flow paths without entering any other regions of the fracture. At a pressure of +2 cm, flow in the one channel became steady while flow in the other channel remained intermittent. So far, the experimental observations from our study and the studies of Tofteng et al. (2002) and Su et al. (1999) indicate that the intermittent flow may occur when the hydraulic contact breaks, somewhere between the column matrix and the end of the macropore, leaving advancing water menisci. But why does the hydraulic contact in the macropore loose its continuity?

Manger et al. (1995) and Wu et al. (1993), who studied flow of solids in hourglasses, give a further point of view to intermittent flow. Using a two-phase fluid flow model, the oscillatory flow of sand in an hourglass was simulated. The oscillations appeared to be caused by air bubbles forming in the solid phase at the orifice of the hourglass. Like the measurements by Wu et al. (1993), the simulations indicated an oscillatory flow of sand through the orifice of the hourglass, when the particle diameter of the sand was small (diameter of 10–100 µm), and for a certain ratio of diameter of the orifice to the particle diameter (Manger et al., 1995). In our study it was observed that air bubbles were spluttering from the outlet of the artificial macropore in the beginning of a pulse flow event, indicating that air bubbles have to be forced before flow takes place.

The above-referred observations may be the explanation of factors at the microscale level influencing the curvature of solute–solid menisci promoting hydraulic contact from water-filled areas to fractures/pores with larger apertures. Small cracks at the surface of the cross-sectional area of the artificial macropores, and small spatial and time heterogeneity in the packing of sand or glass beads around the artificial macropore, enclosed air bubbles, as well as small areas of hydrophobic surfaces in the artificial macropore or fracture, may be the explanation of change in flow type from film/steady to pulse/intermittent flow or vice versa.

Phosphate Sorption during Film and Pulse Flow
The P sorption capacity determinations for the 3- and 8-mm tubes are shown in Fig. 6a and 6b , respectively. The sorption curves have a clear maximum, indicating that no long-lasting diffusion of phosphate from solution into the tube wall is taking place. Furthermore, the control experiments with uncoated tubes sorbed no phosphate, verifying the effect of the iron oxide coating on phosphate sorption. The initial P concentration used in the sorption experiments was less in the 3-mm tubes (0.04 mg P L–1) than that in 8-mm tubes (0.2 mg P L–1). However, the resulting equilibrium solute P concentrations obtained at the leveling off of the curves in Fig. 6 were close to 0.02 mg P L–1 for both tube dimensions. After approximately 20 h, a sorption capacity of 3 mg P m–2 was reached for the 3-mm tubes, whereas it took at least 50 h to reach a sorption capacity of approximately 11 mg P m–2 for the 8-mm tubes (Fig. 6). The sorption rates determined from the linear curves within the first 20 h of the sorption experiments ranged between 0.2 and 0.5 mg P m–2 h–1. Thus, the rate of phosphate sorption to the iron oxide coatings of ceramic tubes is rather slow and not instantaneous.



View larger version (18K):
[in this window]
[in a new window]
  		Fig. 6. Sorption capacities of the iron oxide–coated ceramic tubes when (a) applying 0.04 mg P L–1 in tubes with a 3-mm i.d. and (b) 0.2 mg P L–1 in tubes with an 8-mm i.d. in a circulation system. Open symbols: tree replicates of sorption experiments to iron oxide–coated ceramic tubes. Filled symbols: two replicates of control experiments using uncoated tubes.

 
The P concentrations of the first effluent samples during pulse and film flow were significantly smaller than the samples collected after 3 min, indicating that more P was sorbed on the beginning of a flow event than later. For film flow the differences were significant at the 0.1% level, whereas in the pulse flow experiments the differences were only significant at the 5% level. A dry tube exhibits a superior suction and does not have a diffusion film attached to the walls. The increased amount of P sorbed in the beginning of the flow event could be due to suction into a dry tube. It could also be caused by the relatively small outflow rate in the beginning of a film flow event (Fig. 3b) increasing the contact time between the adsorbent and sorbate. Finally, a gradual accumulation of solute inside the tube just before a pulse event could also cause the increased amount of P sorbed in the beginning of a pulse flow event, increasing the time for diffusion through the attached diffusion film.

Generally, the differences between influent and effluent P concentrations were larger during film flow than pulse flow and these differences were significant at the 0.1% level (Fig. 3b and 4b). During film flow the concentration of P in the outflow solution was (except from the first samples) almost constant at a steady value below the input concentration (Fig. 3b), whereas during pulse flow relative large variations occurred in outflow P concentrations (Fig. 4b). During film flow 15 to 25% of the applied P (calculated as decrease in solute concentration) was sorbed to the iron oxide coating of the 8-mm tubes, while only 3 to 10% of the applied P was sorbed during pulse flow to the 3-mm tubes. However, larger differences in outflow rates during pulse flow compared with film flow were observed. Throughout film flow events, the outflow rate was 9 to 11 mm h–1 with limited fluctuations; an outflow rate of 19 mm h–1 was observed in a single experiment in an attempt to get a pulse flow event in the 8-mm tube (Table 1, Experiment 4). During pulse flow the outflow rate exhibited large fluctuations and reached maximum rates of 225 to 300 mm h–1, which declined to approximately 40 to 60 mm h–1 before the flow stopped (Table 2).

Generally, large variations were found in the cumulative amount of sorbed P per m2 between different experiments (Fig. 7a) . The largest variations were found during pulse flow experiments and there seemed to be no correlation with outflow rates. The duration of film flow in Experiment 1 was 6.6 h. According to the sorption rates determined from the curves in the sorption experiments (Fig. 6) a maximum amount of 3.5 mg P m–2 (rate 0.5 mg P m–2 h–1) could sorb, which is to be compared with the actual sorbed amount of 1.1 mg P m–2 (Fig. 7a). Similarly, for the remaining film flow experiments the actual sorbed amount of P was considerably lower than predicted from the sorption rates. This discrepancy may be explained by the slightly lower solute P concentrations used in the flow experiments (0.04 mg L–1) than in the sorption experiments (0.2 and 0.04 mg L–1 for the 8- and 3-mm tubes, respectively), or more likely, that sorption during the flow experiments is far from equilibrium.



View larger version (27K):
[in this window]
[in a new window]
  		Fig. 7. Cumulative sorbed P per unit surface area in the artificial macropores, as a function of (a) outflow volume and (b) outflow rate during pulse and film flow. Open symbols: pulse flow experiments. Filled symbols: film flow experiments.

 
If P sorbed to the macropore walls according to equilibrium conditions, then at least 11 mg P m–2 and 3 mg P m–2 should sorb to the 8- and 3-mm tubes, respectively, when 5 and 1 L of 0.04 mg P L–1 solution had passed through the 8- and 3-mm tubes, respectively. However, only 0.2 to 0.5 mg P m–2 had sorbed in the 3-mm tubes after 1 L of solution had passed through. For the 8-mm tubes, extrapolation of the curves in Fig. 7a suggests that a maximum of 2.5 mg P m–2 had sorbed after passing 5 L of solution through the macropore. Hence, only 20% of the sorption capacities have been used confirming that sorption is far from equilibrium in the flow experiments, indicating that the sorption rate rather than the sorption capacity controls the amount of sorbed P during macropore flow.

When the cumulative sorbed P per unit surface area was plotted as a function of the cumulative outflow volume, no significant effect of flow pattern on sorbed P could be distinguished (Fig. 7a), indicating that pulse and film flows are similarly effective in removing P from the solution. Illustrating cumulative sorbed amount of P per tube as a function of the cumulative outflow volume obviously showed the effect of flow pattern on P sorption (not shown) since pulse flow only took place in the 3-mm tubes with a relatively small surface area compared with 8-mm tubes. However, this only confirms that the differences between influent and effluent P concentrations were larger during film flow than pulse flow and these differences were significant. It does not tell how effective pulse and film flows actually removed P from the solute. The cumulative sorbed P per unit surface area was also plotted as a function of the outflow rate (Fig. 7b). Even though there were high outflow rates during pulse flow events, the cumulative amount of P increases to the same level as at lower flow rates during film flow events. Despite very different outflow rates, and hence times to establish contact between the adsorbent and adsorbate, the cumulative sorbed amounts of P per unit surface area reach the same levels at both flow types when exposed to similar amounts of P. The fact that pulse and film flow lead to almost the same amounts of P sorbed per unit surface area at similar P solution exposure indicates that the sorption of P to macropore walls is mainly governed by kinetics of the sorption reaction and to a less extent by P transport from bulk solution to the walls of the small macropores. Hence, sorption kinetics seems to be the rate-limiting step controlling the amount of sorbed P during macropore flow.

Hypothesizing that the system in this study could be transferred to soil conditions, the amount of leached P from a structured soil when exposed to a fixed load of P would depend on the size distribution of the macropores in the soil. Thus, when exposed to the same P loads in solution, a distribution of many small macropores in the soil would be more effective to retain P by sorption than a distribution of fewer, but wider macropores, due to the relative larger surface area of small macropores. However, the transport of P through the relatively small macropores would be much faster by pulse flow than through the larger pores by film flow making the time scale important in quantifying the efficiency of retaining P from leaching in the two scenarios.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
Both film and pulse flows were initiated at a pressure potential close to zero at the upper boundary of the artificial macropore. During pulse flow a temporary negative pressure potential developed at the entrance level of the artificial macropore. Generally, the flow patterns were very reproducible and very seldom did pulse flow replace film flow or vice versa. The pulse flow events were short and there was no evidence of correlation between in- and outflow rates during pulse flow. Approximately 18 to 25% of the total pore volume was drained during a pulse flow event. That part of the column that probably drained most in a pulse event was located in the zone between the bottom of the column and 12 cm above the macropore.

During macropore flow less than 25% of the P flowing through the artificial macropores was sorbed. Generally, the differences between influent and effluent P concentrations were larger during film flow than pulse flow and the differences were significant at the 0.1% level. Consequently, 3- and 8-mm tubes gave rise to significant differences with respect to effluent P concentrations signifying that film flow is more effective in removing P from the solution than pulse flow. However, pulse and film flow lead to almost the same amounts of P sorbed per surface area when exposed to the same solute P concentration.

Comparison with P sorption capacity experiments indicated that the sorption rate rather than the sorption capacity controlled the amount of sorbed P during macropore flow. The fact that pulse and film flow lead to almost the same amounts of P sorbed per surface area, despite large differences with respect to flow pattern, indicates that the sorption of P to macropore walls is mainly governed by kinetics of the sorption reaction and to a less extent by P transport from bulk solution to the walls of the macropores. Hence, sorption kinetics seems to be the rate-limiting step controlling the amount of sorbed P during macropore flow.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 


Related articles in JEQ:

This Issue in Journal of Environmental Quality

JEQ 2004 33: 1947-1953. [Full Text]  




This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Related articles in JEQ
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via ISI Web of Science (1)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Gjettermann, B.
Right arrow Articles by Hansen, S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Gjettermann, B.
Right arrow Articles by Hansen, S.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Gjettermann, B.
Right arrow Articles by Hansen, S.
Related Collections
Right arrow Laboratory Column Studies
Right arrow Nonequilibrium Flow
Right arrow Unstable Flow/Fingering
Right arrow Phosphorus

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