Desorption of phosphate from three Finnish mineral soil samples during adsorption of vanadate , molybdate and tungstate

Adsorption of V(V) and Mo(VI) from 10 4 M and 10 ' M solutions and W(VI) from a 10" M solution (in 0.02 M KCI) by three Finnish mineral soils was studied in two series of experiments. In the first experiment, the adsorption of V, Mo and W by soil and the desorption of P were measured at the soils’ natural pH after an equilibration time of 3,5, 22, 29, 46 and 70 h. Adsorption of molybdate occurred mainly within the three first hours, whereas adsorption of vanadate and tungstate were slower processes. During the first few hours, the presence of molybdate seemed to increase the desorption of phosphate most effectively, but after a longer equilibration period, the differences between additions of V, Mo, and W became smaller. In the second experiment, the adsorption process was followed as a function of the acidity of the suspension (pH 2.3-7.5; for W pH 2.8-7.5). Adsorption of V(V), Mo(VI) or W(VI) resulted in a statistically significant increase in the amounts of P desorbed from all three soils over the pH range studied. The aqueous chemistry of V(V), Mo(VI) and W(VI) is briefly discussed.


Introduction
In studying adsorption of anions by different materials, soils, without doubt, are the most com- plicated adsorbents.Even in simplified adsorption and desorption experiments carried out in laboratory conditions, processes such as cation and anion exchange, dissolution of fertilizer particles or soil constituents, and precipitation can be expected (Lindsay 1979,Barber 1984).When different anions are present in soil solution or in the solution in contact with adsorbent, competitive adsorption may occur (Murali and Aylmore 1983, Roy et al. 1986a,b, Goldberg and Traina   1987, Roy et al. 1989, Barrow 1989, Mikkonen   and Tummavuori 1993 a).Addition of a specifi- cally adsorbed ion may affect desorption of other ions already adsorbed (Gorlach et al. 1969, Bar-  row 1974).
When studying retention of V(V), Mo(VI) and W(VI) by three Finnish soils from sodium oxy- salt solutions (Mikkonen and Tummavuori 1993  b,c,d), we followed how quickly adsorption of V, Mo and W occurs at the natural soil pH and how much P is released to the aqueous phase.In addi- tion, we measured the amounts of P present in the aqueous phase at the end of the 72-hour equilibration period to see if the desorption of phos-phate is affected by the addition of rather high amounts of these anions.
The reasons for using 10"* M and KT 5 M solu- tions of added anions was that batch analyses were comfortable to perform at these concentra- tions and there were no polyanionic species present in molybdate solutions.Molybdate is the most important of these three analytes.
In addition to the isopolyanions mentioned above, it is possible that V(V), Mo(VI), and W(VI) form heteropolyanions with each other or with many other elements present in aqueous solu- tions.Examples of such heteropolyanions that might also occur in soil suspensions are molybdophosphates (H +   )   p (Mo0 4 2 ) S (HP0 4 2 ) 2 , p = 8,9,10 (Pettersson et al. 1985) and tetra-decavanadophosphate HPV |4 0 42 8- (Harrison and Howarth  1985).Both of these have been shown to exist in laboratory conditions, using high ionic strength and 10 -1 -10" 2 M solutions.At present, it is still uncertain if these kinds of species exist in very dilute solutions.

Material and methods
Test soils were a clay (soil 1) and two finesand samples (soils 2 and 3).Soil 2 had the coarsest texture.Selected characteristics of the soils are given in Table 1.The ammonium oxalate-oxalic acid (pH 3.3) extractable amounts of Mo and W, and HCIO, + HNO-extractable amounts of V are 4 3 given in our previous papers (Mikkonen and Tum- mavuori 1993b,c,d).
The soils were air-dried and hand-crushed to pass through a 2-mm sieve.When measuring adsorption of V, Mo and W as a function of time, aliquots of 100.0 ml of 10" 4 M NaVO v Na.MoO., or Na,WO., all in 0.02 M KCI, were 2 4 7 2 4' ' mixed to the 1.00-g samples of soils (two repli- cates) in 250-ml beakers, covered with Parafilm and allowed to equilibrate for 3,5, 22, 29, 46, and 70 h at room temperature.Before filtration, the pH values of the suspensions were measured.From these filtrates, V, Mo or W as well as P concentrations were measured by Perkin-Elmer  ICP 2000.
For experiments, where adsorption was stud- ied as a function of pH, subsamples of 1.00 g were placed into 250-beakers and aliquots of 100.0 ml of 0.02 M KCI, 10"* or KF 5 M Na,Mo0 4 , I O' 4 M Na 2 W0 4 , or or 10" 5 M NaVO v all in 0.02 M KCI were added.KCI was used to keep the ionic strength constant.pH adjustments to obtain final pH 2.3-7.5 were made by adding dilute HCI or NaOH.The beakers were covered with Parafilm, shaken manually for 2-3 min, and left to equilibrate at room tempera- ture for 72 h.The final pH values of the sus- pensions were recorded prior to the filtration using a standard combination electrode.During the pH measurements, the samples were stirred using a magnetic rod.The samples were then filtered through Whatman 40 filter paper, and the concentrations of the analytes in the filtrates were measured by a Perkin-Elmer ICP 5000 spectrophotometer.The concentrations were vol- ume-corrected because of the H + /OH addition.From each soil, arbitrarily-selected filtrates were also analysed using the standard addition method, so that the accuracy of the ICP measurements was checked both for the blank sam- ples and for the samples into which V, Mo or W was added.The added 100.0-ml aliquots of I0" 4 M V, Mo or W solution contained 509.4 pg of V, 959.4 pg of Mo and 1838.5 pg of W, respec- tively.

Results and discussion
Adsorption of V, Mo and W begins within the first three hours, but adsorption of vanadate oc- curs more slowly than adsorption of molybdate or tungstate (Fig. la-c).The pH remained rather constant in all samples.
If we take a 30-cm layer of soil and use a bulk density of 1.00 kg/1, the soils could retain at least 510-840 kg/ha of V, 90-300 kg/ha of Mo, and 1200-1800 kg/ha of W, respectively.These are, of course, only estimates, because the adsorption capacity was not determined using a series of more concentrated solutions.
In addition that phosphate can displace e.g.adsorbed molybdate, phosphate can be displaced by high amounts of other specifically adsorbed anions.Within the first few hours, the presence of molybdate seems to increase the desorption of phosphate most, but after a longer equilibration period, the differences in displacing ability of V, Mo, and W become smaller (Fig. 2a-c).
The desorption of P from the blank samples in 72 h and the changes in the desorption because of the addition of Mo (at the lO' 4 mol/1 level) as a function of pH are presented in Fig. 3.The differences in the shape of the curves of soils 2 and 3 at pH > 6 compared with that of soil 1 are  Table 2. P-release from untreated soils and differences in means of desorbed P from soil samples with a Mo, V, or W application compared to means of amounts desorbed from blank samples.
Set of samples the statistical significance of the differences soil 1 soil 2 soil 3 blank mean u,g/g 47.5 47. attributable to the addition of OH ions that displaced some of the adsorbed phosphate ions.Be- cause the natural pH of soil I was 7.3, no addi- tion of OH ions was necessary for obtaining final high pH values of 6-7.In the most acidic suspensions, phosphorus bound to oxide surfaces or present in apatitic form or in the organic mat- ter may also begin to dissolve, in addition to the other forms of P in soil.
In statistical analyses, the results at pH 2.3-7.5 (pH 2.8-7.5 for W) after every V/MoAV treatment were used as one group of test points.A paired t-test confirmed that the sieved soils were homogenous enough for accurate analytical work.When pairs of test points after each V/Mo/W application were divided into two subgroups and the means of the latter compared, the means were identical only when the pH of the test pairs was within 0.05 pH units.
ANOVA was used for statistically testing the effect of adding competing anions on the desorption of P (Table 2).The statistical tests showed that addition of a foreign heavy metal (V, Mo,  W) as sodium oxosalt increased the release of P from all soils.A correlation between adsorbed Mo, V, or W, and desorbed P, however, existed only in the case of W (Table 3).This correlation is linear at pH 3-6.Both at low solubility of P at pH 6-7 and at high solubility of P from soil 3 at pH < 3, the W-P correlation deviates from a straight line.
In creating adsorption isotherms and adsorption envelopes, solution chemistry, properties of the adsorbent, and the analytical conditions should all be taken into account.Today, scientists should carefully investigate especially the possibility of vanadium's forming heteropolyanions.SI V NMR studies, for example, are in progress (Mikkonen and Kolehmainen 1994).As soon as better re- sults are obtained in laboratory experiments, the behaviour of Mo, V, and W in soils could be more thoroughly discussed.Table 3. Correlation between retained W (in pg/g) and desorbed P (in pg/g) at pH 3-6.caused by the addition of 10 4 M molybdate.SI = soil 1, S 2 = soil 2, S 3 = soil 3; (*) denotes blank samples.

Fig. 3 .
Fig. 3. Mobility of P from the blank samples and changes

Table I .
Selected characteristics of the soil samples.determination made at the Agricultural Research Centre of Finland, Jokioinen.AO-OA = ammonium oxalate-oxalic acid extraction solution, pH 3.3 * = A