SSSAJ Grow Your Career with SSSA
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 Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (11)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Giesler, R.
Right arrow Articles by Persson, P.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Giesler, R.
Right arrow Articles by Persson, P.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Giesler, R.
Right arrow Articles by Persson, P.
Related Collections
Right arrow Forest Soils
Right arrow Biogeochemical Processes
Right arrow Nutrient Cycling
Right arrow Phosphorus
Right arrow Soil Chemistry
Published in Soil Sci. Soc. Am. J. 69:77-86 (2005).
© 2005 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA

Division S-2—Soil Chemistry

Phosphate Sorption in Aluminum- and Iron-Rich Humus Soils

Reiner Gieslera,*, Tove Anderssonb, Lars Lövgrenb and Per Perssonb

a Dep. of Forest Ecology, Swedish Univ. of Agricultural Sciences, S-901 83 Umeå, Sweden, and Dep. of Ecology and Environmental Science, Umeå Univ., S-901 87 Umeå, Sweden
b Dep. of Chemistry, Inorganic Chemistry, Umeå Univ., S-901 87 Umeå, Sweden

* Corresponding author (reiner.giesler{at}eg.umu.se).


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Groundwater discharge areas in boreal forest ecosystems can potentially be areas where Fe and Al accumulate in the forest floor and affect the solubility of P. This study was conducted to determine phosphate sorption properties using humus soils containing high native amounts of Al and Fe. Humus soils were collected from two different forested groundwater discharge areas varying in pH and amount and distribution of Al and Fe species. The phosphate sorption capacities were 265 and 216 mmol P kg–1 dry wt. soil, respectively, for the two humus soils. Pyrophosphate extractable Al and Fe dominated in the first soil, 738 mmol kg–1 dry wt. The concentration of pyrophosphate extractable Al and Fe in the second soil was 317 mmol kg–1 dry wt., whereas citrate-dithionite (CD) extracted 548 mmol Al and Fe kg–1 dry wt. Using 0.1 mol dm–3 NaCl as ionic medium gave a higher phosphate sorption compared with deionized water at the midrange additions, but no difference at the highest addition. Phosphate sorption was pH independent when 0.1 mol dm–3 NaCl was used, whereas deionized water gave a sorption minimum at approximately pH 6. Dissolved organic carbon (DOC) release in solution was positively correlated with phosphate sorption, especially in 0.1 mol dm–3 NaCl ionic medium. However, dissolved organic P (DOP) was unaffected by the phosphate sorption. Attenuated total reflectance Fourier transform infrared spectroscopy (ATR-FTIR) indicated that precipitation of Al and Fe phosphates dominated at higher phosphate additions in the soil with mainly organically bound Al and Fe, whereas both surface sorption and precipitation occurs in the soil with a larger amount of CD-extractable Fe.

Abbreviations: ATR-FTIR, attenuated total reflectance Fourier transform infrared spectroscopy • CD, citrate-dithionite • DOC, dissolved organic carbon • DOP, dissolved organic phosphorus • FTIR, Fourier transform infrared


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
PHOSPHATE SORPTION is associated with the occurrence of reactive surface sites in the mineral soil and is extensively described in the literature. However, the effect of possible sorption sites such as Al and Fe precipitates or organically bound Fe and Al in humus soils on phosphate sorption is less known. Previous studies have shown that Al and Fe are likely to accumulate in the humus layer of groundwater discharge areas (Mulder et al., 1991; Norrström, 1993, 1995; Giesler et al., 1998, 2002; Pellerin et al., 2002). Thus, potential sorption sites for phosphate are created in these humus soils, which can affect the availability of P for plants and microorganisms (Pare and Bernier, 1989; Giesler et al., 2002, 2004) and the transfer of P from terrestrial to aquatic ecosystems (Lyons et al., 1998).

Previous studies on peat soils have shown that accumulation of Al and Fe in organic soils increases the capacity of phosphate sorption (Cuttle, 1983; Richardson, 1985; Nieminen and Jarva, 1996). Furthermore, model experiments have been performed where metal–organic matter complexes were created and used for phosphate sorption studies (Bloom, 1981; Gerke, 1993). These results indicated that metal–organic matter complexes in surface soils could be an important contribution to the total phosphate sorption. Gerke and Hermann (1992) showed that the phosphate sorption in Fe-humic substance mixtures was six to seven times larger than amorphous Fe oxide, and suggested that this may be due to the formation of ternary complexes between the Fe-humic substances and phosphate. They also found that the mole fraction between the sorbed P and Fe in Fe-humic substances was close to 1, whereas the same ratio for amorphous Fe oxide was much lower. It was hypothesized that Fe in Fe-humic substances is more finely distributed on the organic surface and not present as Fe polymers, hence creating a larger number of surface sites for phosphate binding per Fe. Both ternary complexes involving metal-organic matter complexes and phosphate or Al or Fe precipitates with phosphate may occur in humus soils; however, more conclusive evidence is needed to distinguish among the different P forms. This can be accomplished by means of spectroscopic techniques such as Fourier transform infrared (FTIR) spectroscopy. The FTIR technique allows in situ analyses, and both solutions and soil samples from batch sorption experiments can be studied. In the particular case of phosphate sorbed to the soil, FTIR spectra can be used to study molecular speciation by analyzing the absorption bands that originate from the P–O stretching vibrations, which appear in the region between 900 and 1300 cm–1. These are known to be sensitive to the chemical and structural environment of the phosphate ions (Tejedor-Tejedor and Anderson, 1990; Persson et al., 1996).

In humus soils with an inherent high content of Al and Fe, phosphate sorption may be affected by competition for sorption sites by DOC. For instance, soil solution concentrations of DOC as well as low-molecular weight organic acids in humus soils in boreal forests can be 10-fold higher than those found in the mineral B-horizon (Giesler et al., 1996; van Hees et al., 2000). Kaiser and Zech (1996)( 1997) have shown that phosphate decreased sorption of DOC in mineral soils. In soil column flow-through experiments, Beck et al. (1999) showed that the release of organic C coincided with the P retention in Andisols and suggested ligand exchange between the two components on allophanic surfaces. Studies using citrate, oxalate, phthalate, or tartrate and a defined surface (Violante et al., 1991; Violante and Gianfreda, 1993; Nilsson et al., 1996; Geelhoed et al., 1998; De Cristofaro et al., 2000; Liu and Huang, 2000) have shown that phosphate does compete with these substances for surface sites. Phosphate as well as the model substance could be outcompeted, but pH, the concentration of ligand, and the order in which the ions were added to the suspension affected the competition. Competition between phosphate and other anions may also affect the sorption capacity. A number of studies have shown that organophosphorus compounds such as inositolphosphates may compete for the same surface sites as phosphate (Anderson and Arlidge, 1962; Anderson et al., 1974; McKercher and Anderson, 1989; Shang et al., 1990; Ognalaga et al., 1994; Celi et al., 2001, 2003).

This study was initiated to investigate in detail the sorption of phosphate in humus soils from forested groundwater discharge areas. We selected two humus soils, with known high inherent Al and Fe as well as organic P concentrations. We assume that phosphate sorption in these organic-rich soils may affect the solubility of the inherent DOC and organic P compounds. Although past studies have indicated that Al and Fe can accumulate in humus soils in groundwater discharge areas, few studies have addressed how this relates to phosphate sorption and none how it affects DOC and organic P. Our objectives were (i) to study the sorption capacity of phosphate in these humus soils, (ii) to study the effect of phosphate sorption on inherent DOC and DOP, (iii) to study the effect of increased ionic strength and pH dependence on phosphate sorption, and (iv) to use FTIR spectroscopy for molecular speciation of the sorbed phosphate.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Field Sites
Two forested groundwater discharge areas with known high native amounts of Fe and Al in the humus layer (Giesler et al., 2002) were selected. The sites, Flakastugan (64°25' N, 19°25' E) and Betsele (64°39' N, 18°30' E), are located in northern Sweden. The mean annual temperature and precipitation in the area are 1°C and 570 mm, respectively. On average, the sites are snow covered from late October to early May.

The forest stands are dominated by Norway spruce [Picea abies (L.) H. Karst.] with individual trees of Scots pine (Pinus sylvestris L.). The field layer vegetation is dominated by tall herbs such as Cicerbita alpina (L.) Wallr. at Flakastugan and Aconitum lycoctonum L. subsp. lycoctonum, Rubus idaeus L., and Actaea spicata L. at Betsele.

The soils are loamy-sandy till soils and are classified as Aquic or Oxyaquic Haplocryods (Soil Survey Staff, 1998) since saturated conditions within 100 cm from the soil surface most likely occur during part of the year. The humus layer (O horizon) at Betsele was about 7 cm thick and that at Flakastugan was about 10 cm. The organic matter was highly decomposed (Oa) at both locations, and the C content was about 34% (Giesler et al., 2002). The Al, Fe, and P concentrations determined for the two soils are given in Tables 1 and 2. Aluminum and Fe extractions were performed on separate samples using pyrophosphate, oxalate, and CD extractions (Table 1). The pyrophosphate-extraction is assumed to extract mainly organically bound Al and Fe; the oxalate extraction also includes amorphous forms, and the CD also includes crystalline Fe. These extractions are, however, only operationally defined and do not necessarily reflect the actual composition of the Al and Fe forms in the humus.


View this table:
[in this window]
[in a new window]
  		Table 1. Average concentrations of Al, Fe, and P extracted from the two humus soils with pyrophosphate (Pyr), ammonium oxalate (Ox), and citrate–dithionite (CD). All extractions were made on separate samples. The SD is in parentheses (n = 3).

 

View this table:
[in this window]
[in a new window]
  		Table 2. Average cation exchange capacity, exchangeable cations and acidity, and inorganic and organic P in the two humus soils. The SE is in parentheses (n = 3).

 
Sampling
The sampling of humus soil at the two sites was performed in September 1999. The O-horizon was sampled randomly with a 0.10-m-diam. soil auger. The content of seven to eight humus samples, taken approximately within a 25-m2 area, were bulked into one composite sample for each site. The samples were kept at 5°C during transportation. Within 24 h after sampling, the samples were passed through a 2-mm mesh sieve, gently homogenized, and then stored in polyethylene bags at –20°C. The water content was measured on a subsample (105°C, 24 h).

Phosphate Sorption
Batch experiments were performed to determine the phosphate sorption capacity of the humus soils. Two parallel series of sorption experiments were performed, one with 0.1 mol dm–3 NaCl as ionic medium and one with deionized water. Field-moist soil samples (equivalent to 2 g dry wt.) were reacted in 50-cm3 centrifuge tubes with various concentrations of NaH2PO4 obtained by adding different volumes from a stock solution of 48 mM NaH2PO4 (soil to solution ratio of 1:20). The phosphate additions ranged from 0 to 560 mmol P kg–1 soil (28 mmol L–1). The samples were reacted during end-over-end rotation for 17 h in darkness at 25°C. After reaction, pH and conductivity were measured. The soil pH determined in the soil suspension after 17 h reaction time (deionized water, soil to solution ratio of 1:20) was 5.0 for the Flakastugan sample and 6.0 for the Betsele sample. In the batch experiments, the pH was adjusted to these values by additions of NaOH at the start of the experiment so that these pH values were achieved after 17 h of reaction time. The samples were centrifuged at 3500 rpm for 15 min, and the supernatant was filtered through 0.45-µm membrane filters (Millipore HA membrane filter, Millipore Ireland B.V., Cork, Ireland). The supernatant was kept at 4°C for further analyses. The kinetics of phosphate sorption using deionized water was tested in batch experiments as above, using different reaction times with an initial phosphate concentration of 14 mmol L–1. The reaction times ranged from 2.5 min to about 20 h, not including the centrifugation time of 15 min. The first sample (2.5 min) was shaken by hand. Inorganic phosphate, total P, pH, and conductivity were measured in all samples. Metal cations, anions, and DOC were analyzed on a subset of samples. The DOP was calculated as the difference between total P and inorganic phosphate.

To study the pH dependence of phosphate sorption, an additional series of batch experiments was performed. One phosphate concentration was used (6.0 mmol P L–1), and pH was varied between 4 and 7.5 by the addition of HCl or NaOH.

To test if biological activity affected the sorption, an additional series of batch experiments was performed where a biocide, NaN3, was used to inhibit microbial growth. Phosphate sorption experiments were performed as above, with and without the addition of 0.02% NaN3. Four P concentrations were used: 0, 1, 6, and 12 mmol P L–1, with three replicates for each concentration. The experiment was performed using only the humus soil from Betsele.

ATR-FTIR Spectroscopy
The ATR-FTIR spectra were collected for the Betsele and Flakastugan soil samples in both the absence and presence of phosphate. Batch experiments using 0, 14, or 28 mmol L–1 NaH2PO4 were performed as above. Before IR analysis, the soil particles and the solutions were separated by means of centrifugation as above, and IR spectra were recorded for both the wet soil samples and the supernatants. The ATR-FTIR spectra were collected using a Bruker IFS 66v/S FTIR spectrometer fitted with a deuterated triglycine sulfate detector. The spectra were recorded with a horizontal ATR accessory and a diamond crystal as the reflection element (SensIR Technologies), and all measurements were performed under vacuum (about 3 kPa). The soil samples and the supernatants were applied to the diamond crystal surface directly and a steel lid was placed over the sample and pressed tightly against a rubber gasket, which sealed the sample from the vacuum during data collection. One hundred scans were collected for each sample over the range of 370 to 7800 cm–1 with a resolution of 4 cm–1. Spectra of the soil samples were interpreted after subtracting spectra of the empty cell (collected for each sample) and the corresponding supernatant.

Analyses
Analyses of phosphate from sorption experiments were performed with a colorimetric method (ammonium-molybdate/stannous chloride) using a flow injector analyzer (5020 Analyser, Tecator, Höganäs, Sweden). Total P was determined as above after potassium peroxodisulphate digestion. Total P was also determined on a subset of samples using inductively coupled plasma–mass spectrometry (ICP-MS, PerkinElmer, SCIEX, Norwalk, CT). Organic P was calculated as the difference between total P and inorganic P. Analyses of Fe, Al, Ca, Mg, Na, and K were performed using inductively coupled plasma–atomic emission spectrometry (Plasma II emission spectrometer, PerkinElmer). Anions in the sorption experiment were analyzed using high-performance liquid chromatography (Dionex model 4000i, Dionex Corporation, Sunnyvale, CA). A total organic C analyzer (TOC-5000, Shimadzu Company, Tokyo) was used to analyze DOC in the sorption experiment.

Solution pH was measured using a pH combination electrode (U402-M6-S7/100, Mettler Toledo AG, Schwerzenbach, Switzerland) in suspensions with deionized water and using an Ag/AgCl pH combination electrode (9103SC, Orion Research, Boston, USA) in suspensions with ionic medium. The conductivity was determined with a conductivity electrode (Konduktometer E382, Metrohm AG, Herisau, Switzerland).

Statistics
A two-way ANOVA was used to evaluate the effect of NaN3 addition with added NaN3 and P as fixed factors. Multiple comparisons were performed with Tukey's test. Significant differences refer to the P < 0.05 level unless otherwise stated.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Phosphate Sorption
The phosphate sorption capacity was larger for the Flakastugan humus soil than for the Betsele humus soil (Fig. 1) . The highest phosphate addition gave a sorption of about 216 mmol kg–1 for the Betsele humus soil whereas the same phosphate addition gave a sorption of 265 mmol kg–1 for the Flakastugan soil. No difference in sorption capacity was found using deionized water or 0.1 mol dm–3 NaCl with the highest phosphate addition (Fig. 1). Using 0.1 mol dm–3 NaCl as ionic medium gave a higher phosphate sorption in the midrange of the phosphate additions for both soils compared with deionized water



View larger version (15K):
[in this window]
[in a new window]
  		Fig. 1. Adsorption isotherms for the two soils, (a) Betsele and (b) Flakastugan. The batch sorption experiments were performed in both deionized water (unfilled squares) and 0.1 mol dm–3 NaCl (crosses). The error bar denotes the maximum confidence interval for the Betsele (n = 3–5) and Flakastugan soil (n = 2). The total amount of sorbed phosphate, q, has been defined as the sum of the native amount of inorganic phosphate (2.9 and 7.2 mmol kg–1 for the Betsele and Flakastugan soil, respectively) and the sorbed inorganic phosphate.

 
Most of the phosphate added was sorbed within 5 h of reaction time with only small differences between the two soils (Fig. 2) . The sorption kinetics for the two soils could be described by the empirical linear relationship between [PO4]0 – [PO4]t = k x ln(t) + A, where the suffix 0 indicates the start concentration at time 0 and the suffix t the concentration at reaction time t. The parameter k was 0.41 and 0.39, and A was 9.2 and 10.8 (r2 = 0.98 and 1.0, P > 0.001) for the Betsele and Flakastugan soils, respectively.



View larger version (15K):
[in this window]
[in a new window]
  		Fig. 2. The relationship between reaction time and phosphate sorption in deionized water.

 
Changes in solution pH due to phosphate additions differed at the most 0.11 pH units from the initial soil pH in 0.1 mol dm–3 NaCl. Use of deionized water gave a similar, small difference for the Flakastugan soil (0.09 pH units), but pH decreased gradually with increasing phosphate additions in the Betsele soil (maximum change of 0.5 pH units) despite adjustments with NaOH. This would give an approximate 6% overestimate of the phosphate sorption compared with the target pH of 6 (see below).

pH Dependence and Effect of Biocide
The phosphate sorption decreased with increasing pH between about pH 4 and 6 in deionized water (Fig. 3) . At pH values higher than 6, the phosphate sorption increased again. No obvious pH-dependent phosphate sorption was found using 0.1 mol dm–3 NaCl in the two soils.



View larger version (12K):
[in this window]
[in a new window]
  		Fig. 3. The pH dependence of phosphate sorption. The amount of phosphate sorbed as a function of pH in solution for humus soils from (a) Betsele and (b) Flakastugan. The batch sorption experiments were performed in both self-medium (unfilled circles) and 0.1 mol dm–3 NaCl (filled circles).

 
Treatment with the biocide, NaN3, had a small effect only on the phosphate sorption in deionized water (Two-way ANOVA, P = 0.052; Table 3). There was a slight increase in phosphate sorption, at most 6%. In contrast, no significant effect of NaN3 additions in solutions with ionic media was observed (Two way ANOVA, P = 0.71; Table 2).


View this table:
[in this window]
[in a new window]
  		Table 3. Adsorption data from the biocide NaN3 experiment performed on the Betsele humus soil.{dagger}

 
ATR-FTIR Spectra of Soil Samples
As seen in Fig. 4 , the typical P–O stretching bands (Laiti et al., 1996; Persson et al., 1996) overlap with strong bands of the native soils (no addition of phosphate). In the spectra of the Betsele samples, however, small changes around 1050 cm–1 are observed as the phosphate concentration is increased (Fig. 4). Small changes are also observed for the Flakastugan samples, but these are even subtler. It should be pointed out that these changes are not due to phosphate remaining in solution since the spectra of the supernatants have been subtracted from the soil spectra. The changes therefore are most likely due to the presence of sorbed phosphate. To better resolve the IR features of the sorbed phosphate fractions, the spectra of the native Betsele and Flakastugan samples were subtracted from the spectra of the corresponding phosphate-reacted samples (Fig. 5) . In the Betsele spectra, three bands appear at the lower phosphate addition (1100, 1060, and 1000 cm–1), but the band at about 1060 cm–1 disappears at the higher phosphate addition, and the two peaks remain (1100 and 1020 cm–1). In the Flakastugan soil, only two peaks appear at 1110 and 1040 cm–1, independent of phosphate addition.



View larger version (19K):
[in this window]
[in a new window]
  		Fig. 4. Attenuated total reflectance Fourier transform infrared spectroscopy spectra of, from top to bottom, native Betsele soil, Betsele soil + 14 mmol dm–3 phosphate, Betsele soil + 28 mmol dm–3 phosphate, native Flakastugan soil, Flakastugan soil + 14 mmol dm–3 phosphate, and Flakastugan soil + 28 mmol dm–3 phosphate. All spectra are plotted on the same ordinate scale in absorbance units.

 


View larger version (16K):
[in this window]
[in a new window]
  		Fig. 5. Difference attenuated total reflectance Fourier transform infrared spectroscopy spectra obtained by subtracting the spectrum of the corresponding native soil from the spectrum (a) Betsele soil + 14 mmol dm–3 phosphate, (b) Betsele soil + 28 mmol dm–3 phosphate, (c) Flakastugan soil + 14 mmol dm–3 phosphate, and (d) Flakastugan soil + 28 mmol dm–3 phosphate. All spectra are plotted on the same ordinate scale in absorbance units.

 
Effects of Phosphate Sorption on DOC, DOP, Cations, and Anions
Sorption of phosphate increased DOC in the solution, though the trend depended on the medium used (Fig. 6) . In 0.1 mol dm–3 NaCl, sorption of phosphate caused a linear increase of DOC in the solution; the slopes were 0.10 and 0.07 for the Betsele and Flakastugan soils, respectively. The molar ratio of DOC release in relation to sorbed phosphate was 1.9 and 1.5 for the Betsele and Flakastugan soils, respectively. When deionized water was used, DOC concentrations decreased in the solution initially but increased at higher phosphate additions (Fig. 6). Compared with deionized water, addition of 0.1 mol dm–3 NaCl without phosphate additions decreased the DOC concentration from 27 to 7 mmol L–1 in the Flakastugan soil. The decrease in the Betsele soil was less, from 15 to 14 mmol L–1. In the Betsele soil, the DOC concentrations in ionic medium exceeded those in deionized water at the highest phosphate additions, but this was not the case for the Flakastugan soil.



View larger version (25K):
[in this window]
[in a new window]
  		Fig. 6. Phosphate sorption in relationship to (a, b) release of dissolved organic carbon (DOC), (c, d) conductivity, (e, f) release of Ca and Mg, and (g, h) release of Al. Filled and unfilled squares represent the Betsele humus soil and the Flakastugan soil, respectively.

 
In contrast, no relationship between phosphate additions and release of DOP into the solution was observed. Only a small amount of DOP, equivalent to <7 mmol kg–1 dry wt., was released at higher phosphate additions in 0.1 mol dm–3 NaCl. In deionized water, no effect was observed. Estimating the amount of DOP using the difference between molybdate-reactive phosphate in oxidized and nonoxidized solutions gave the same results as the difference between total P determined by ICP and phosphate determined by ion chromatography. With phosphate additions, no change was found for the inorganic anions NO–3, Cl–, and SO2–4. The concentration of Cl– was about 250 µmol L–1 in the deionized water (data not shown). Sulfate concentrations in both deionized water and ionic medium were about 80 µmol L–1 (data not shown). Nitrate concentrations were below detection limits in the Flakastugan soil and about 40 and 10 µmol L–1 in deionized water and 0.1 mol dm–3 NaCl, respectively, for the Betsele soil.

Phosphate additions increased the ionic strength in the deionized water; the increase was higher in the Betsele soil than in the Flakastugan soil (Fig. 6). The increase in ionic strength in the Betsele soil correlated (r = 0.98, P < 0.001) to an increase in the divalent cations Mg and Ca (Fig. 3). In contrast, phosphate additions decreased the ionic strength in both soils using ionic medium (Fig. 3), and the decrease was correlated to a decrease in Ca and Mg ions in the solutions (r = 0.96 and 0.92, respectively, for the Betsele and Flakastugan soil; P < 0.001 in both cases). There was an increase in Al in the Flakastugan soil, with increasing P additions independent of medium (Fig. 3). In the Betsele soil, there was a smaller increase using 0.1 mol dm–3 NaCl at the lower phosphate additions, but no change at higher additions.


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
CONCLUSIONS
 REFERENCES
 
Phosphate Sorption
Many processes may interact in a complex matrix such as humus soils and affect phosphate sorption. Competition between phosphate and anions, especially organic anions, are likely to be high in this matrix where the organic C content is about 30% or more. The high organic C content will also determine the cation exchange capacity and will have a large impact on the pH buffering capacity in these soils. The composition of the cation exchange complex affects the composition of cations in the solution and can indirectly affect phosphate sorption since di- and trivalent cations can affect the solubility of DOC (Skyllberg and Magnusson, 1995; Reemtsma et al., 1999; Oste et al., 2002).

Aluminum and Fe within this matrix are probably present in several different forms. For instance, the operationally defined extractions of Al and Fe suggest that organically bound Al and Fe dominate in the Flakastugan soil. In this soil, almost all of the Al and Fe were pyrophosphate extractable, with only small differences compared with the other extractions (Table 1). In the Betsele soil, pyrophosphate extracted about 84% Al and 44% Fe compared with CD. This indicates that amorphous and crystalline forms especially of Fe are present in the Betsele soil, but not in the Flakastugan soil.

The difference spectra from the FTIR measurements indicate that the phosphate speciation varies with phosphate concentration for the Betsele samples, but it is more or less constant for the Flakastugan samples (Fig. 5). The Flakastugan spectra are consistent with precipitation of some kind hydrous metal phosphate (e.g., Al or Fe phosphate), which typically display broad and poorly resolved P–O stretching bands (Ross, 1974). However, we cannot exclude the possibility that ternary complexes may form as proposed by Bloom (1981) and Gerke (1993). To investigate this issue further, experimental data using molecular level techniques are needed. Other forms, such as Ca phosphates, are not likely owing to both the relatively low pH (5.0) and the low availability of Ca. The total Ca content in the Flakastugan soil was about 60 mmol kg–1 (Giesler et al., 2002) and could only support the formation of 20 mmol kg–1 of Ca2(PO4)3 consuming 40 mmol P kg–1. This is well below the total P sorption of about 210 mmol kg–1. The same value for the Betsele soil was about 155 mmol kg–1, reflecting the higher Ca content in this soil. We can thus not exclude some formation of Ca phosphate in this soil. The variation in the spectra of the Betsele samples is tentatively explained by a combination of sorption mechanisms. At high phosphate concentration, the difference spectrum is similar to those obtained for the Flakastugan samples, and thus indicates a precipitation mechanism. At low phosphate concentration, the additional bands might be caused by a complexation mechanism where phosphate interacts with metal ions present in the organic matrix, either as mononuclear species or at the surfaces of small metal (hydr)oxide clusters/particles. This is in agreement with the wet-chemical extractions, suggesting that significant amounts of amorphous Fe (hydr)oxides are present in the Betsele soil, while these are much less abundant in the Flakastugan soil.

The higher phosphate sorption obtained for the Flakastugan soil as compared with the Betsele soil may at least partly be related to the higher total amount of extractable Al and Fe (AlCD + FeCD; 548 and 729 mmol kg–1 for the Betsele and Flakastugan soils, respectively). Earlier studies by Cuttle (1983) and Richardson (1985) have shown that the phosphate sorption is positively related to the amounts of extractable Al and Fe in peat soils and supports the above assumption. Another factor that affects the sorption capacity is pH of the soil solution. A pH decrease of 1 pH unit for the Betsele soil, giving a pH more similar to the Flakastugan soil, would give a phosphate sorption increase of about 28% in deionized water (Fig. 4), and is thus sufficient to explain the difference between the two soils. This explanation is, however, not applicable to the situation in ionic medium because the pH dependent sorption there was minor.

The increase in phosphate sorption with increasing ionic strength at midrange additions is similar to findings using mineral soils (cf. Pardo et al., 1992). They also found that differences in sorption due to pH diminished using 0.1 mol dm–3 NaCl. An increased ionic strength can affect sorption of organic acids. Low molecular weight organic acids have shown to be mainly adsorbed by forming outer-sphere surface complexes, in which the organic anions are electrostatically bound to positively charged surface sites (Nordin et al., 1997). Such interactions are strongly sensitive to high salt concentrations, leading to suppressed adsorption at high ionic strength. This would in turn favor phosphate in its competition with the organic matter for surface sites. The humus soils we used were complex, including several forms of Al and Fe as indicated by the wet-chemical extractions and the FTIR measurements. More controlled experiments are needed with model substrates to better understand the mechanisms behind sorption for these types of soils. Obviously, several factors can influence the sorption of phosphate in these soils, either directly or indirectly.

The phosphate sorption capacities obtained in our study were higher than those reported by Richardson (1985) on peat soils with a high Al content (highest amounts about 370 mmol kg–1 pyrophosphate extractable Al). A comparison with studies where humus material or peat has been manipulated by additions of Al or Fe to create phosphate sorption sites gave higher or about equal sorption capacities compared with our results (Haynes and Swift, 1989; Gerke and Hermann, 1992). The larger sorption capacities in the studies are most likely explained by higher amounts of Al and Fe in the samples used. Recent studies (Norrström 1993, 1995; Giesler et al., 2002) on humus soils from ground water discharge areas have found comparable or higher amounts of Al and Fe than those in our humus soils. For instance, Norrström (1993)( 1995) found total Fe concentration in humus soils ranging from 24 to 625 mmol kg–1 (average 231 mmol kg–1) and oxalate-extractable Al concentrations ranging between 80 and 441 mmol kg–1 (average 234 mmol kg–1). The organically bound fraction ranged between 40 to 80% for Fe (pyrophosphate extractable Fe versus total Fe) and about 30 to 90% for Al (CuCl2 extractable Al versus oxalate extractable Al). Compared with phosphate sorption capacities of spodic mineral B horizons, the sorption capacities in the humus soil from our discharge areas are almost 10 times higher (Zhou et al., 1997; Nair et al., 1999). The concentrations of B horizon Al and Fe in the two above studies were also between six to 10 times lower compared with the Flakastugan humus soil. The molar ratios between phosphate sorbed and the sum of Al and Fe in our humus soils (about 0.4) are less than earlier findings using manipulated humus soils (Gerke and Hermann, 1992) but much higher than, for instance, for microcrystalline goethite (0.01; Nilsson et al., 1992). In these organic C-rich soils, more sorption sites related to Al and Fe are probably available since the presence of organic matter can prevent crystallization (Schwertmann et al., 1986).

Release of Dissolved Organic Carbon
The release of DOC in ionic medium suggests that phosphate competes for surface sites occupied by organic compounds. Previous studies on mineral soils have also shown that phosphate can displace DOC from sorption sites in soils (Kaiser and Zech, 1996, 1997; Bhatti et al., 1998; Beck et al., 1999; Arbestain et al., 2002). With ionic medium, competition between phosphate ions and DOC seems to be the major mechanism explaining the increase in DOC and contrasts the fact that the ionic strength actually decreases in the solutions. The decreased ionic strength may decrease the DOC concentration (Kalbitz et al., 2000), but this effect is obviously minor. The decrease in ionic strength is explained by the decrease in the divalent cations Ca2+ and Mg2+ with increasing phosphate sorption. Wiklander (1978) showed similarly that phosphate sorption resulted in a decrease in the added cations and related this to co-sorption of the cations and anions. Bloom (1981) suggested that the sorption of phosphate on Al–peat results in an excess negative charge and concluded that precipitated phosphate probably is not an important factor in the system studied.

The initial decrease in DOC concentrations in deionized water (Fig. 6) may be explained by a decreased DOC solubility caused by the increasing ionic strength (Kalbitz et al., 2000), resulting from the addition of phosphate. However, additional amounts of phosphate seem to counteract the decreased DOC solubility, and the competition between DOC and phosphate becomes more important. The effect of ionic strength on DOC solubility seems to be much less for the Betsele soil compared with the Flakastugan soil. The difference in Al-saturation between the two soils may explain the difference in DOC solubility (Table 1). The addition of 0.1 mol dm–3 NaCl resulted in a cation exchange and gave a larger release of Al into the soil solution as indicated by Fig. 3 in the Flakastugan soil compared with the Betsele soil. A cation exchange between Na and Al may increase the flocculation of DOC in solution and decrease the ratio between DOC and soil organic matter. The Al remaining in solution is most likely organically bound, and the increase is related to the increase in DOC. In the Betsele soil, the increase in Na concentration mainly affected the release of Ca and Mg, reflecting the difference in the cation exchange properties (Table 2).

Competition with Organic Phosphorus
The absence of an increase in DOP in the presence of phosphate may indicate that (i) organic P compounds are not associated with Al and Fe, (ii) that they are more strongly bonded to Al and Fe surface sites and thus are not easily outcompeted by phosphate, or (iii) that the surface sites are in excess of the sorbed P and DOC in this experiment. The first suggestion contrasts with earlier findings, which show that organic P accumulates in humus soils with increasing Al and Fe content (Giesler et al., 2002, 2004). Figure 7 illustrates that organic P increases with increasing Al + Fe content in humus and that the increase is independent of the C content. Several studies have shown that many organophosphates and phosphonates sorb to surface sites of Al and Fe precipitates and some of them equally efficient as phosphate (Shang et al., 1990; Ognalaga et al., 1994; Nowack and Stone, 1999; Celi et al., 2001, 2003; Sheals et al., 2002). Anderson et al. (1974) also found that inositol hexaphosphate suppressed sorption of phosphate, particularly with large additions. Leytem et al. (2002) showed that sorption of organic P was positively correlated with organic matter and Al and Fe contents in coastal plain soils. The sorption experiments with organic P compounds, the relationship between both the content and sorption of organic P in soils and the Al and Fe content, and the absence of relationship with C content in our soils supports the contention that organic P most likely is associated with Al and Fe in our humus soils.



View larger version (14K):
[in this window]
[in a new window]
  		Fig. 7. The relationship from between (a) acid digestible Al + Fe and organic P (filled circles) and inorganic P (unfilled squares) and (b) the relationship between C content and organic P in humus soils from five groundwater discharge areas (Giesler et al., 2002) including both within-site and between-site variation.

 
The amount of Al and Fe surface sites present in our two soils is most likely sufficient to support both the amount of sorbed phosphate in the experiments and an inherent organic P pool. For instance, the organic P pool in the Betsele humus soil was about 60 mmol kg–1 dry wt. and the phosphate sorption maximum was about 216 mmol kg–1 in ionic medium, this gave a total of 276 mmol kg–1 dry wt. and the Al + Fe pool in the same humus soil was 548 mmol kg–1 dry wt. (see above). This yields a molar ratio between sorbed P and Al + Fe of 0.50, which is lower than the molar ratio of about 1 that Gerke and Hermann (1992) found in their sorption experiments using Fe-organic matter complexes. The same molar ratio for the Flakastugan humus soil was 0.49. The absence of DOP release may thus relate to insufficient phosphate additions, and further experiments are needed to address questions on organic P sorption and accumulation in these soils.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
REFERENCES
 
The humus soil samples from two groundwater discharge areas studied herein showed a high phosphate sorption capacity. Despite differences in Al and Fe content and forms, cation exchange properties, and DOC, the sorption behavior of the soils was relatively similar. An increased sorption of phosphate and a simultaneous desorption of organic matter in the presence of ionic medium indicated competition between phosphate and DOC for surface sites. The DOP release was low and not related to phosphate additions. We hypothesize that groundwater discharge areas may be important sinks for organic phosphorus compounds. Our results suggest that accumulation of Al and Fe in humus soils has a large impact on the phosphate sorption capacity despite the presence of large amounts of soil organic matter. Humus soils with a high P sorption capacity are likely to occur in local groundwater discharge areas, and may thus affect the transport of both phosphate and DOP from terrestrial to aquatic ecosystems.


   ACKNOWLEDGMENTS
 
This study was funded by the Center of Environmental Research, CMF, at Umeå University. R.G. received financial support from the Swedish Research Council for Environment, Agricultural Sciences and Spatial planning (FORMAS). We thank three anonymous reviewers for valuable comments.

Received for publication March 7, 2003.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 




This article has been cited by other articles:


Home page
 J. Environ. Qual.Home page
K. W. Goyne, H.-J. Jun, S. H. Anderson, and P. P. Motavalli
Phosphorus and Nitrogen Sorption to Soils in the Presence of Poultry Litter-Derived Dissolved Organic Matter
J. Environ. Qual., January 4, 2008; 37(1): 154 - 163.
[Abstract] [Full Text] [PDF]

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 Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (11)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Giesler, R.
Right arrow Articles by Persson, P.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Giesler, R.
Right arrow Articles by Persson, P.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Giesler, R.
Right arrow Articles by Persson, P.
Related Collections
Right arrow Forest Soils
Right arrow Biogeochemical Processes
Right arrow Nutrient Cycling
Right arrow Phosphorus
Right arrow Soil Chemistry

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		 Journal of
Environmental Quality