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


     

Published in Agron J 100:819-829 (2008)
DOI: 10.2134/agronj2007.0209
© 2008 American Society of Agronomy
677 S. Segoe Rd., Madison, WI 53711 USA
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 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 Google Scholar
Google Scholar
Right arrow Articles by Francisco, E. A. B.
Right arrow Articles by Taylor, R. W.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Francisco, E. A. B.
Right arrow Articles by Taylor, R. W.
Agricola
Right arrow Articles by Francisco, E. A. B.
Right arrow Articles by Taylor, R. W.
Related Collections
Right arrow Rice
Right arrow Nutrient Cycling
Right arrow Nutrient Management
Right arrow Soil Fertility and Productivity

PHOSPHORUS MANAGEMENT

Characterization and Greenhouse Evaluation of Brazilian Calcined Nonapatite Phosphate Rocks for Rice

E. A. B. Franciscoa, S. H. Chienc, L. I. Prochnowb,*, E. R. Austinc, M. C. M. Toledod and R. W. Taylore

a Dep. of Biological Sci., Federal Univ. of Mato Grosso, Rondonópolis, Brazil
b IPNI Brazil and Dep. of Soil Sci., Univ. of São Paulo/ESALQ, C.P. 9, 13418-900, Piracicaba, Brazil
c Research and Market Development Div., IFDC-An International Center for Soil Fertility and Agricultural Development, P. O. Box 2040, Muscle Shoals, AL 35662
d School of Humanities and Geosciences Inst., Univ. of São Paulo, São Paulo, Brazil
e Dep. of Plant and Soil Sci., Alabama A & M Univ., Normal, AL 35762

* Corresponding author (lprochnow{at}ipni.net).


   ABSTRACT
TOP
 NOTES
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Little information is available on the agronomic effectiveness of calcined nonapatite phosphate rock (PR) sources containing crandallite minerals in the form of Ca–Fe–Al–P for flooded and upland rice (Oryza sativa L.). We conducted laboratory and greenhouse studies to (i) characterize the mineralogical composition, (ii) investigate the solubility and dissolution behavior, and (iii) evaluate the agronomic effectiveness of two nonapatite PR sources (Juquia and Sapucaia) from Brazil and compared them with (i) a highly reactive Gafsa PR (Tunisia) containing apatite in the form of Ca–P and (ii) a reference water-soluble triple superphosphate (TSP) for flooded and upland rice. After calcination at 500°C for 4 h, the solubility of Juquia PR and Sapucaia PR in neutral ammonium citrate (NAC) significantly increased from almost nil to a maximum of 39.3 and 114 g P kg–1, respectively. X-ray diffraction showed that crystalline crandallite mineral was transformed to an amophorus form after calcination. The solubility behavior of the two calcined PR sources followed the same trend as Gafsa PR, that is, P release decreased with increasing equilibrium pH in the 0.01 M KCl solution (pH 3.0–8.0). At pH 3, the solubility followed: Gafsa PR > calcined Sapucaia PR > calcined Juquia PR. No P release was detected from any of the PR sources at pH ≥ 5.0 in the solution, indicating the Ca–P characteristic of the Ca–Fe–Al–P mineral controlled P dissolution of the calcined PR. Without calcination, both Juquia PR and Sapucaia PR were totally ineffective for upland rice grown on a Hiwassee clay loam (fine, kaolinitic, thermic Rhodic Kanhapludult) with pH 5.4 whereas a significant P response was observed with the calcined PR samples. For flooded rice grown on Hiwassee soil, the calcined Juquia PR and Sapucaia PR were 66 and 72%, respectively, as effective as TSP in increasing rice grain yield whereas Gafsa PR was ineffective. For upland rice grown on the unlimed soil, Gafsa PR was as effective as TSP in increasing rice grain yield whereas calcined Juquia PR and Sapucaia PR were 89 and 83% of TSP. The effectiveness of Gafsa PR was reduced to 0% after the soil was limed to pH 7.0 whereas the two calcined PR sources were reduced to 49% of TSP. Soil available P extracted by iron oxide impregnated filter paper (Pi test) or anion-exchange resin after rice harvest correlated well with P uptake by rice grain for flooded and upland rice.

Abbreviations: CA, citric acid • NAC, neutral ammonium citrate • PR, phosphate rock • PDF, powder diffraction file • RAE, relative agronomic effectiveness • TSP, triple superphosphate • XRD, X-ray diffraction


   NOTES
TOP
 NOTES
ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher.

Received for publication June 17, 2007.
   INTRODUCTION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
THE MAJOR RAW phosphate material for acidulation to produce commonly used water-soluble P fertilizers is apatite PR that contains Ca–P minerals. Under certain conditions, direct application of apatite PR can be an economically attractive alternative to the use of more expensive water-soluble P, especially the use of indigenous PR for tropical acid soils in developing countries (Kuyvenhoven et al., 2004). Numerous studies on the agronomic use of apatite PR are well documented in literature and several comprehensive review articles have been published (Khasawneh and Doll, 1978; Hammond et al., 1986; Chien and Menon, 1995; Rajan et al., 1996). Also, a new computer-based PR decision support system that integrates various factors affecting the agronomic effectiveness of PR with respect to water-soluble P has been recently published (Smalberger et al., 2006).

In contrast to apatite PR, the marginal nonapatite PR materials containing crandallite minerals cannot be used for acidulation because of their high Fe and Al content. In general, crandallite minerals are in the form of Ca–Fe–Al–P as Ca(Fe,Al)3(PO4)2(OH)5·H2O in which the Fe/Al mole ratio varies with PR deposits. The natural, uncalcined crandallite minerals are very low in reactivity as measured by NAC and are not suitable for agronomic use (Hoare, 1980). However, calcination of crandallite at 450 to 550°C drives off the water of hydration, the crandallite becomes amorphous, and its citrate solubility increases considerably (Doak et al., 1965; Gilkes and Palmer, 1979; Hoare, 1980). Bolland and Gilkes (1987) used canonical variate analysis for the data from 194 published pot and field experiments to assess the relative effectiveness of calcined crandallites from Christmas Island (Australia) and Thies (Senegal) with respect to water-soluble P. They concluded that the relative effectiveness of both calcined P sources were low (<10–50%) on highly P responsive soils when several levels of P were applied to provide complete response curves. However, most of the published reports on the agronomic effectiveness of calcined crandallite minerals have been conducted in Western Australia for upland crops, mainly pastures (Bolland and Bowden, 1984; Bolland et al., 1984, 1988; Bolland and Gilkes, 1991).

Little information is available in literature on the potential use of calcined crandallite minerals for upland and flooded rice, two important crops in developing countries. To our knowledge, there is only one recent preliminary report on the effectiveness of three calcined crandallite minerals in Brazil for upland and flooded rice; however, no detailed characterization of minerals was studied and the rice crops were grown only for 65 d instead of to maturity (Francisco et al., 2008). In this study, the relative effectiveness of calcined crandallite minerals for dry-matter yield with respect to a water-soluble P source was 48 to 57% for upland rice and 50 to 69% for flooded rice.

It is known that the solubility of Ca–P minerals decreases with increasing soil pH whereas the solubility of Fe–P and Al–P minerals increases with pH (Lindsay, 1979). However, little is known regarding the effect of soil pH on the solubility of Ca–Fe–Al–P minerals. Furthermore, it is known that the solubility of soil Fe–P and Al–P increases under flooded soil conditions due to the reduction of Fe+3 to Fe+2 and hydrolysis of Al+3 (Kirk et al., 1990). It is unknown whether the same effect may also be observed with Ca–Fe–Al–P in flooded soil. Thus the specific objectives of the present study on the calcined nonapatite crandallite minerals were to further (i) understand their chemical, mineralogical, and solubility characteristics, (ii) evaluate their agronomic potential for upland and flooded rice grown to maturity, (iii) investigate their available P as influenced by liming and soil moisture regime in an acid soil, and (iv) correlate soil available P and P uptake by upland and flooded rice.


   MATERIALS AND METHODS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Preparation and Characterization of Phosphate Rock Sources
Uncalcined PR samples containing crandallite minerals were collected from Juquia and Sapucaia deposits in Brazil. The samples were ground to <150 mm and calcined for 4 h in a muffle furnace at temperatures from 300 to 800°C with 50°C increments to select the calcined samples that were produced at the lowest temperature with the highest solubility for laboratory and greenhouse studies. The calcined samples were analyzed for total P and soluble P in NAC as described by Chien and Hammond (1978).

The two selected calcined samples of Juquia and Sapucaia that were used for laboratory and greenhouse studies and uncalcined samples of Juquia and Sapucaia were scanned using a Siemens D-500 X-ray powder diffractometer (Bruker AXS, Inc., Madison, WI). The data were then analyzed with Jade 5.0 software (XRD pattern processing) (Materials Data Inc., Livermore, CA) and the ICDD (International Center for Diffraction Data, 1999) powder diffraction files (PDF) database to which X-ray patterns of Fe–Al–P compounds published by Frazier et al. (1991) were added. The compounds present in the samples were then identified based on the XRD spectral lines of pure compounds in this database.

Phophate Rock Dissolution at Different Solution pH
In addition to the two selected calcined Juquia PR and Sapucaia PR samples, a commercial-grade highly reactive apatite Gafsa PR (Tunisia) was included for comparison. A 100-mg quantity of each finely ground ( <0.15 mm) PR sample was placed in a plastic tube (50 mL volumetric capacity) containing 30 mL of a 0.01 M KCl solution. Solution pH was adjusted to 3, 4, 5, 6, 7, and 8 by adding 0.01 or 0.1 M HCl or KOH solutions. The samples were shaken continuously until the solution pH reached equilibrium near the target pH. The pH was adjusted every 24 h by adding drops of 0.01 or 0.1 M HCl or KOH solutions until the pH was stabilized for 72 h. The amounts of OH– or H+, in mmol, added to the solutions and the final pH values at equilibrium are presented in Table 1 . After reaching equilibrium, the samples were filtered through Whatman No. 1 filter paper and the P concentration was measured by the method of Murphy and Riley (1962). Aluminum, Fe, and Ca concentrations were determined using a Thermo Elemental IRIS Intrepid inductively coupled plasma atomic emission spectrometer.


View this table:
[in this window]
[in a new window]

  		Table 1. Amount of OH– or H+ added as KOH or HCl to the target pH solution and final pH at equilibrium in the 0.01 M KCl solution.

 
Greenhouse Study
A Hiwassee clay loam soil was collected at the top 20-cm layer from an uncultivated land in Alabama. The soil had pH of 5.4 (1:2 soil to water ratio), 1.3 mg extractable P kg–1 by iron oxide-impregnated filter paper (Pi strip) (Menon et al., 1989), 2.5 cmolc kg–1 of effective cation- exchange capacity as measured by 1 M NH4Cl (Bertsch and Bloom, 1996), 1.8 g kg–1 of organic matter (Nelson and Sommers, 1996), and 320 g kg–1 of clay content. In addition to the two selected calcined PR sources and Gafsa PR, a commercial-grade, water-soluble granular TSP was also included as standard P source for comparison. Total and soluble P contents of these P sources are presented in Table 2 .


View this table:
[in this window]
[in a new window]

  		Table 2. Total and soluble P for the P sources used in the study.

 
The greenhouse study consisted of three experiments. In the first experiment, calcined Juquia PR and Sapucaia PR and their original, uncalcined PR samples were thoroughly mixed with 4 kg soil at total 100 mg P kg–1. A check with no P applied was included. Nitrogen and K were added at rates of 200 mg kg–1 N as urea and 200 mg kg–1 K as KCl and a solution containing secondary and micronutrients at rates of 96 mg kg–1 of Mg, 135 mg kg–1 of S, 5 mg kg–1 of Cu, 11 mg kg–1 of Zn, 8 mg kg–1 of Mn, 2 mg kg–1 of B, and 0.2 mg kg–1 of Mo was also added (Prochnow et al., 2003a, 2003b). The pots were arranged in a randomized complete block design with three replications. Six seeds of upland rice (cv. IR-47686) were planted at a depth of about 1 cm and subsequently thinned to two plants per pot 10 d after germination. The pots were watered using deionized water to maintain approximately 75% of field moisture capacity during the experiment. The aboveground plants were cut 45 d after germination followed by drying at 60°C for 2 wk and weighing. The concentration of P in the plant samples was determined after digestion of subsamples with H2SO4–H2O2 by the ammonium molybdate–ascorbic acid method (Murphy and Riley, 1962).

In the second experiment, 3.5-kg soil pots were flooded under 0.5 cm of water for 2 wk. After that period, calcined Juquia PR and Sapucaia PR, Gafsa PR, and TSP at total P rates of 0, 10, 25, 50, and 100 mg P kg–1 were mixed into the soil at a 10 cm depth with the help of a spatula. Secondary and micronutrients were applied at the same rates as described in the first experiment. Flooded rice seedlings (cv. IR-36) were then transplanted to the soil pots with four seedlings per pot. The rice plants were grown to maturity and harvested. The P concentration of P in the rice grain was determined after digestion as previously described.

To determine P availability of the flooded soil in situ in the second experiment, the same procedure used by Prochnow et al. (2003a) was employed. Briefly, after rice harvest a spatula was used to insert 10 Pi strips at different sites in each pot leaving one of the waxed ends above the flooded soil surface. After 24 h, the Pi strips from each pot were carefully pulled out from the soil and rinsed with distilled water. The Pi strips were then combined and extracted by shaking for 1 h with 40 mL of 0.1 M H2SO4 and filtering (Menon et al., 1989), followed by P determination (Murphy and Riley, 1962). Available P as measured by the Pi test was expressed in µg P per 10 strips.

In the third experiment, 4-kg soil pots were incubated with or without the addition of CaCO3. The pH of the limed soil was 7.0 after 3-wk incubation. Then the same four P sources were mixed with the unlimed and the limed soils at rates of 0, 10, 25, and 100 mg P kg–1. The planting procedures were the same as described in the first upland rice experiment except the plants were grown to maturity. The concentration of P in the rice grain was determined as previously described. After harvest, 50-g soil samples were collected, dried, and ground to 2-mm size, and available P was determined by the Pi strip (Menon et al., 1989) and the anion exchange resin method (Kuo, 1996).

Data Analysis
In the first greenhouse experiment, least significant difference (LSD) at the 5% level was used to detect any significant differences in dry matter yield of upland rice obtained with P sources. In the second and third experiments, the relationships between parameters, that is, rice grain yield, P uptake by rice grain, extractable P by Pi or resin and rate of P applied were evaluated using regression procedures (SAS Institute, 1985). Except for the relationships between rice grain yield or P uptake and extractable P by Pi or resin, a combined multiple-regression analysis using a dummy variable as described by Chien et al. (1988) was performed for all P sources in each experiment. This resulted in a common intercept and a single value of Sy^ (standard deviation of yield response) and R2 for all the regression equations (one for each P source) in each experiment. The response functions (linear, semi-log, and square root) were tested to describe the relationships between the parameters studied and the one presenting the highest R2 with the lowest Sy^ values was chosen. The models tested were as follows:

Formula 1[1]

Formula 2[2]

Formula 3[3]
where Yi is the rice grain yield, P uptake, or extractable P by Pi or resin obtained with P source i, X is the rate of P applied, βi is the slope of the response function for P source i, βo is the common intercept, and {varepsilon}i is the error term of the fitted model.

The relative agronomic effectiveness (RAE) was calculated for each P source with respect to TSP. It was defined as the ratio of the two slopes:

Formula 4[4]

where βi is the slope of the response function of the P source i and βTSP is the slope of the standard source of P as TSP. This expression ranks different P sources with respect to TSP according to their agronomic potential to produce a yield response (Chien et al., 1990). Use of RAE made it possible to compare the P effectiveness among the experiments with different P sources. Furthermore, one advantage of using the calculated RAE value according to Eq. [4] (vertical comparison) over the so-called substitution rate method (horizontal comparison) is that the RAE value based on the former is constant across all P rates whereas the RAE values based on the latter varies with the P rates selected (Chien et al., 1990). In the present study, a t test was used to determine if there was a statistically significant difference in the slopes of the response function between two P sources in the range of P rates applied (Steel and Torrie, 1960).


   RESULTS AND DISCUSSION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Preparation and Characterization of Phosphate Rock Sources
Figure 1 shows the NAC solubility of Juquia PR and Sapucaia PR samples as a function of the calcination temperature. After the thermal treatment, the NAC solubility increased from about 5 g P kg–1 at 300°C for both samples to 39 and 114 g P kg–1 at 500°C for Juquia PR and Sapucaia PR, respectively. When the crandallite minerals are heated, the hydrated water in the structure is lost by evaporation, the mineral structure collapses, and they become amorphous. This is seen by comparing the XRD patterns of the uncalcined and calcined (at 500°C) samples of Sapucaia PR (Fig. 2 ). After calcination all of the crandallite peaks either disappeared or significantly decreased in intensity, indicating an amorphous structure. This disorganized structure after calcination subsequently results in an increase in the NAC solubility of crandallite minerals. The maximum solubility for calcined Juquia PR and Sapucaia PR occurred between 500 and 700°C. Beyond 700°C, the solubility began to decline with temperature (Fig. 1). The decline in solubility is probably due to the reorganization of the crystalline structure of Fe–Al–P or the formation of new mineral structures with low solubility such as whitlockite [β-Ca3(PO4)2] and phospho-cristobalite (AlPO4) as indicated by Gilkes and Palmer (1979).


Figure 1
View larger version (14K):
[in this window]
[in a new window]

  		Fig. 1. Effect of thermal treatment on neutral ammonium citrate (NAC) solubility of crandallite minerals from Juquia PR and Sapucaia PR deposits, Brazil.

 

Figure 2
View larger version (21K):
[in this window]
[in a new window]

  		Fig. 2. X-ray diffractograms of (A) the original sample from Sapucaia PR and (B) after the thermal treatment (4 h, 500°C), and (C) pure crandallite mineral.

 
Based on the results shown in Fig. 1, the two calcined Juquia PR and Sapucaia PR samples at 500°C were selected along with Gafsa PR and TSP for the laboratory and greenhouse studies. Table 2 shows their total P, NAC-soluble P, and 2% CA-soluble P contents. After calcination, total P content of all crandallite PR samples moderately increased due to the loss of hydrated water from the crandallite structure. As expected, the NAC solubility of all crandallite PR samples significantly increased after calcination compared with that of the uncalcined samples due to the transformation of crystalline to amorphous crandallite as shown in Fig. 2. Calcination had only a relatively small effect on increasing the solubility in 2% CA compared with that in NAC (Table 2). It is known that the chelating power of citrate with Al+3 and Fe+3 ions to form stable complexes increases with increasing pH. This favors the dissolution of Fe–Al–P minerals by reducing the presence of free Al+3 and Fe+3 ions in the NAC solution and results in an increase in NAC solubility of Fe–Al–P minerals (Chien, 1995). Furthermore, the solubility of Fe–Al–P increases with increasing pH (Lindsay, 1979). These two mechanisms explain the higher solubility in NAC compared to 2% CA for the calcined crandallite samples because of the lower pH of 2% CA (pH 2.2) than that of NAC (pH 7.0). In contrast, the NAC solubility of apatite Gafsa PR containing Ca–P mineral was lower than the 2% CA solubility. This is due to the fact that the solubility of Ca–P increases with lowering pH and the chelation of citrate with Ca+2 ion decreases with a decrease of pH. Thus the mechanism for the dissolution of apatite in 2% CA is mainly controlled by the acidity rather than by the chelation. This explains the differences in the solubility behavior of crandallite and apatite PR sources by NAC and 2% CA, that is, the NAC solubility of crandallite is higher than that of apatite whereas the 2% CA solubility of apatite is higher than that of crandallite.

Phosphate Rock Dissolution at Different Solution pH
The titration data of Table 1 indicate that the calcined nonapatite Juquia PR and Sapucaia PR containing Ca–Fe–Al–P minerals behaved similarly to the apatite Gafsa PR containing Ca–P mineral in the solution; namely, HCl was required to obtain equilibrium solution pH values ≤6.0. To reach pH 7.0, however, HCl was still required for Gafsa PR whereas KOH was required for calcined PRs. Based on the titration data, the alkalinity of P minerals follows the order of Gafsa PR > Sapucaia PR > Juquia PR. It is important to point out that the titration data of the calcined Ca–Fe–Al–P minerals in the present study differ from Prochnow et al. (2003b) who synthesized H8 [Fe3KH8(PO4)6·6H2O] and H14 [Fe3KH14(PO4)8·4H2O] compounds, which are Fe–P type minerals. These Fe–P compounds do not contain Ca and were acidic since they required KOH, instead of HCl, to maintain equilibrium solution pH ≤ 6.0. Thus, the alkalinity of P minerals follows the order of Ca–P (apatite) > Ca–Fe–Al–P (crandallite) > Fe–P (H8 and H14). This finding suggests that the solubility behavior of Ca–Fe–Al–P compounds may have a mixed solubility characteristic of Ca–P and Fe–Al–P compounds. However, this does not mean that Ca–Fe–Al–P compounds are a mixed discrete Ca–P and Fe–Al–P compounds in a form of solid solution such as adduct formation, for example, urea-phosphoric acid or urea-monocalcium phosphate complex.

Both P and Ca, especially P, released from nonapatite Sapucaia PR and Juquia PR followed the same trend as apatite Gafsa PR (Fig. 3 ), that is, a decrease in concentration with increasing equilibrium pH. On the other hand, Prochnow et al. (2003b) showed that P released from H8 and H14 increased with increasing equilibrium pH. The dissolution data (Fig. 3) thus agreed with the titration data (Table 1) for the three PR sources used in the present study. The data of P and Ca dissolved suggest that dissolution behavior of nonapatite Sapucaia PR and Juquia PR containing Ca–Al–Fe–P minerals in the equilibrium solution was more controlled by Ca–P than Fe–Al–P in the mixed characteristics of Ca–P and Fe–Al–P. For Al and Fe (Fig. 4 ), the trend was the same as P and Ca, that is, a decrease in concentrations with increasing equilibrium solution pH. However, the Al concentration was much greater than the Fe concentration at pH 3 (Fig. 4A vs. Fig. 4B) suggesting possible precipitation of Fe–P. Prochnow et al. (2003b) also found that little Fe was present in the equilibrium solution from H8 and H14 presumably due to the formation of water-insoluble FePO4·nH2O through hydrolysis of H8 and H14.


Figure 3
View larger version (16K):
[in this window]
[in a new window]

  		Fig. 3. (A) Phosphorus and (B) calcium released at equilibrium from calcined Juquia PR, calcined Sapucaia PR, and Gafsa as a function of the pH of 0.01 M KCl solution.

 

Figure 4
View larger version (15K):
[in this window]
[in a new window]

  		Fig. 4. (A) Aluminum and (B) iron released at equilibrium from calcined Juquia PR, calcined Sapucaia PR, and Gafsa as a function of the pH of 0.01 M KCl solution.

 
The amounts of P released from calcined Juquia PR and Sapucaia PR at equilibrium pH 7.0 were very low (Fig. 3A) whereas their soluble P by NAC extraction which has the same pH value was much higher (Table 2). Furthermore, the P solubility of calcined Juquia PR and Sapucaia PR in the equilibrium solution was much lower than that of Gafsa PR at pH 3 (Fig. 3A), similar to the trend of the solubility data of 2% CA but opposite to the NAC solubility (Table 2). The P solubility in the equilibrium solution at pH 3 follows the order of Gafsa PR > calcined Sapucaia PR > calcined Juquia PR while the P solubility was very low at higher pH for all PR sources (Fig. 3A).

Greenhouse Study
In the first experiment, dry matter yields of upland rice obtained with both uncalcined Juquia PR and Sapucaia PR were very low, as low as the control (Table 3 ) indicating low P availability due to their very low P solubility (Table 2). After calcination, both PR sources significantly increased dry matter yield of upland rice (Table 3). This confirms the premise that the uncalcined crandallite minerals cannot be used for direct application. There was no significant difference in agronomic effectiveness between the two calcined PR sources (Table 3), despite the fact that P solubility in NAC (Table 2) and the equilibrium solution (Fig. 3A) was higher for the calcined Sapucaia PR than that of the calcined Juquia PR. This demonstrates that the P availability of the calcined PR in the soil system may not correlate well with the P solubility in the pure solution system and this will be discussed later.


View this table:
[in this window]
[in a new window]

  		Table 3. Average dry matter yield of upland rice grown for 45 d in unlimed soil treated with calcined and uncalcined samples of Juquia PR and Sapucaia PR.

 
In the second experiment, the semi-log model as described by Eq. [2] was found to best describe the relationship between grain yield and P rate for flooded rice (Fig. 5A ). The effectiveness of P sources compared to TSP follows the order TSP > Sapucaia PR = Juquia PR > Gafsa PR for grain yield. The relationship between P uptake by rice grain and P rate as described by Eq. [1] in the linear model also follows the same trend among PR sources (Fig. 5B). The calculated values of regression coefficients and RAE of all P sources are shown in Tables 4 and 5 and 7 , respectively. The RAE values of calcined Juquia PR and Sapucaia PR compared to TSP are 66 and 72% for rice grain yield and 71 and 75% for P uptake, respectively (Table 5). There is no significant difference in agronomic effectiveness between the two calcined PR sources whereas Gafsa PR was totally ineffective compared to TSP. Although Gafsa PR is highly reactive, it is not suitable for flooded rice due to the well-known fact that soil pH increases on flooding. Ponnamperuma (1985) showed that soil pH drastically increased from 5.0 to 6.0 in only 2 wks after flooding and to 6.5 afterward. Since Gafsa PR was applied 2 wks after flooding the soil, this explains the very poor performance of the reactive Gafsa PR. Hellums (1991) showed that a highly reactive apatite Sechura PR (Peru) was as effective as TSP in increasing rice grain yield when the PR was applied 2 wks before flooding but the effectiveness was reduced to 50% of TSP when the PR was applied 2 wks after flooding of an acid soil with pH 4.8. Thus the reduction in the effectiveness of apatite Gafsa PR in the present study was attributed to an increase in soil pH on flooding before PR application. Further evidence to demonstrate the lack of P release from Gafsa PR can be seen in Fig. 3A that shows P release was essentially stopped at pH 5 in the equilibrium solution. The higher RAE value of reactive Sechura PR for flooded rice reported by Hellums (1991) than the RAE value of reactive Gafsa PR obtained in the present study was probably due to the initial soil pH used (4.8 vs. 5.3) rather than PR source since both PR samples had about the same reactivity (Chien, 1995).


Figure 5
View larger version (25K):
[in this window]
[in a new window]

  		Fig. 5. (A) Grain yield and (B) P uptake by rice grain from several P sources and rates of P applied. Models followed by the same letter are statistically not different from each other in the slope (P ≤ 0.05).

 

View this table:
[in this window]
[in a new window]

  		Table 4. Regression estimates for the models describing the relationship between grain yield and P uptake by grain as a function of rate of P applied for flooded rice.

 

View this table:
[in this window]
[in a new window]

  		Table 5. Calculated values of relative agronomic effectiveness (RAE) of P sources for flooded rice.

 

View this table:
[in this window]
[in a new window]

  		Table 7. Calculated values of relative agronomic effectiveness (RAE) of P sources for upland rice grown on unlimed and limed soils.

 

View this table:
[in this window]
[in a new window]

  		Table 6. Regression estimates for the models describing the relationship between grain yield and P uptake by grain as a function of rate of P applied to unlimed soil for upland rice.

 
In the equilibrium dissolution study, it was shown that P released from calcined Juquia PR and Sapucaia PR was less than that from Gafsa PR at pH 3 and little P was released from the calcined PR sources at pH ≥ 4 (Fig. 3A). However, a significant rice response to these two calcined PR sources was observed under flooded conditions whereas there was no response for Gafsa PR (Fig. 5). Therefore, although the P dissolution of the Ca–P characteristic of the Ca–Fe–Al–P minerals in the calcined PR sources was similar to the Ca–P characteristic of apatite mineral in Gafsa PR in the equilibrium solution, there may be other factors to account for the P dissolution of the calcined Juquia PR and Sapucaia PR at high pH in the flooded soil. One explanation is that after the soil was flooded, it resulted in a reduction of Fe+3 to Fe+2 along with Al-hydrolysis that increased the solubility of Fe–P and Al–P (Kirk et al., 1990). This explains why the calcined Juquia PR and Sapucaia PR were much more effective than Gafsa PR at high pH in the flooded soil (Fig. 5) despite the fact that little P release was observed from all PR sources at high pH in the equilibrium solution (Fig. 3A).

A highly significant relationship between in situ Pi-extractable P and applied P rate for flooded rice was best described by Eq. [1] (linear model) as shown in Fig. 6 . The order of Pi-extractable P from P sources follows the order of TSP > calcined Juquia PR > calcined Sapucaia PR > Gafsa PR. Little P release from Gafsa PR as shown by the Pi test thus confirms the lack of rice response to Gafsa PR in the flooded soil (Fig. 5). It should be pointed out, however, that there were no significant differences in rice grain yield and P uptake between calcined Juquia PR and calcined Sapucaia PR (Fig. 5) although the solubility of calcined Sapucaia PR was higher than that of calcined Juquia PR (Table 2). Apparently the P solubility and Pi-extractable P were more sensitive than rice response to available P in distinguishing the two calcined PR sources in the flooded soil. However, we could not explain why Pi-extractable P of calcined Juquia PR was higher than that of calcined Sapucaia PR at 100 mg P kg–1 rate (Fig. 6). There is also a significant relationship as described by Eq. [2] (semi-log model) between P uptake and Pi-extractable P for flooded rice (Fig. 7 ). It is known that soil P tests for flooded rice soils are difficult due the fact that soil available P changes after collecting the wet soil samples followed by air-drying before chemical analysis. Direct P analysis of the collected wet soil samples also often encounters numerous handling problems. The application of the Pi strips to measure available P in situ for flooded rice soil was first demonstrated by Prochnow et al. (2003a) and the present results support their suggestion that the Pi test is indeed a potential in situ method to monitor available P in flooded soil. Future field trials are required to test this proposition.


Figure 6
View larger version (20K):
[in this window]
[in a new window]

  		Fig. 6. Phosphorus extracted by iron oxide-impregnated filter paper strips (Pi) inserted into the flooded soil after rice harvest as related to P rate applied. Models followed by the same letter are statistically not different from each other in the slope (P ≤ 0.05). Sy^ is the standard deviation of yield response.

 

Figure 7
View larger version (15K):
[in this window]
[in a new window]

  		Fig. 7. Relationship between P uptake by rice grain and P extracted by iron-impregnated filter paper (Pi strip) inserted into the flooded soil treated with P sources after rice harvest. Sy^ is the standard deviation of yield response.

 
In the third experiment, the relationship between grain yield of upland rice and P rate is best described by the semi-log model for both unlimed (pH 5.4) and limed (pH 7.0) soils (Fig. 8A and 9A ) whereas the linear model describes best for the relationship between P uptake and P rate (Fig. 8B and 9B). The calculated values of regression coefficients and RAE are shown in Tables 6, respectively. In the unlimed soil, the effectiveness of P sources follows the order of TSP = Gafsa PR ≥ calcined Juquia PR = calcined Sapucaia PR (Table 7). The good performance of highly reactive Gafsa PR compared to TSP for upland crops, especially for upland rice grown on acid soils, has been reported (Chien and Hammond, 1978; Hammond et al., 1986; Chien and Menon, 1995). It is interesting to note that effectiveness of calcined Juquia PR and calcined Sapucaia PR was also rather high, 89 and 83%, respectively, compared to TSP in increasing upland rice grain in the unlimed soil. This result differs from the reported low effectiveness (<10–50%) of calcined nonapatite Christmas Island PR and Senegal PR compared to superphosphate for other upland crops (Bolland and Gilkes, 1987). Therefore, crop species may play an important role to use calcined nonapatite PR. Upland rice is known to use apatite PR more effectively than other crops (Chien and Menon, 1995). This may also be the case for calcined nonapatite PR as observed in the present study. It also should be pointed out that the RAE values of calcined Juquia PR and Sapucaia PR were about the same yet the NAC solubility of calcined Sapucaia PR was higher than that of Juquia PR (Table 2). This suggests that the NAC solubility may not be a good indicator for predicting the agronomic effectiveness of calcined Ca–Fe–Al–P compounds as also reported by Prochnow et al. (2003b) for the synthesized Fe–P compounds. This is probably due to different mechanisms for P dissolution from Ca–Fe–Al–P in NAC extraction and soil system. In the NAC solution, the mechanism is mainly chelation of cations with citrate; in the soil system, acidity, chelation, Fe reduction, and Al hydrolysis provide the driving forces for P release from Ca–Fe–Al–P.


Figure 8
View larger version (26K):
[in this window]
[in a new window]

  		Fig. 8. (A) Grain yield and (B) P uptake by upland rice grain grown on unlimed soil related to P sources and rates of P applied. Models followed by the same letter are statistically not different from each other in the slope (P ≤ 0.05).

 

Figure 9
View larger version (24K):
[in this window]
[in a new window]

  		Fig. 9. (A) Grain yield and (B) P uptake by upland rice grain grown on limed soil related to P sources and rates of P applied. Models followed by the same letter are statistically not different from each other in the slope (P ≤ 0.05).

 
For the limed soil (pH 7.0), Gafsa PR was totally ineffective for upland rice (Table 7). This is consistent with Khasawneh and Doll (1978) who showed that increasing soil pH and Ca by liming significantly reduced the dissolution and agronomic effectiveness of a highly reactive apatite North Carolina PR. For calcined Juquia PR and calcined Sapucaia PR, liming also reduced their RAE from 83 to 89% without liming to 49% with liming (Table 7). This observation agrees with Bolland and Gilkes (1987) who also reported that the effectiveness of calcined Christmas Island PR decreased with increasing soil pH. However, unlike Gafsa PR with RAE = 0%, calcined Juquia PR and Sapucaia PR were 49% as effective as TSP in increasing upland rice grain yield at soil pH 7.0. The following mechanisms are proposed to explain the complex effects of soil pH and P mineralogical composition on the agronomic effectiveness of apatite PR vs. nonapatite PR. In the unlimed soil, the Ca–P characteristic of Ca–Fe–Al–P minerals of calcined nonapatite Juquia PR and Sapucaia PR dominates the P release, similar to Ca–P of apatite Gafsa PR as shown in the equilibrium study (Fig. 3A). Since the P dissolution of Gafsa PR was greater than that of calcined Juquia PR and Sapucaia PR at low pH, this explains that Gafsa PR was more effective than the calcined Juquia PR and Sapucaia PR in the unlimed soil.

In the limed soil, P dissolution from the calcined Ca–Fe–Al–P minerals, similar to apatite (Ca–P mineral), was reduced by increasing soil pH and Ca. However, liming did not completely stop P dissolution of Ca–Fe–Al–P of calcined crandallites like it did to Ca–P of apatite. Unlike Gafsa PR, which contains Ca–P but no Fe–Al–P, the solubility of Fe–Al–P of the Ca–Fe–Al–P may increase with increasing soil pH as indicated by Lindsay (1979) for Fe–P and Al–P. Although the equilibrium study showed no P dissolution from calcined PR at pH 7.0 (Fig. 3A), it should be pointed out that there are no sinks for Fe and Al ions in the equilibrium solution system. In the soil system Fe and Al release can be adsorbed by clay minerals and this may result in further P release. This explains why the calcined Juquia PR and Sapucaia PR were still able to provide some available P to upland rice at soil pH 7.0 (Fig. 9) whereas Gafsa PR totally failed. The proposed mechanisms also suggest that the driving force for P dissolution from calcined Ca–Fe–Al–P in the flooded soil at high pH may not be entirely attributed to Fe-reduction and Al-hydrolysis; additional effect of increasing P dissolution from Fe–Al–P characteristic of Ca–Fe–Al–P at high pH may also contribute to available P. Thus the effectiveness of calcined Juquia PR and Sapucaia PR in increasing rice grain yield for flooded rice (RAE = 66–72% in Table 5) was higher than that for upland rice in the limed soil (RAE = 49% in Table ). However, the effectiveness of calcined PR for flooded rice was lower than that for upland rice in the unlimed soil (RAE = 83–89% in Table 7). The latter result is opposite to the studies by Prochnow et al. (2003a,b) in which H8 and H14 containing only Fe–P but no Ca–P was more effective for flooded rice than upland rice grown on the unlimed Hiwassee soil. In their studies, H14 was as effective as TSP for flooded rice mainly due to the effects of Fe-reduction and increased pH on P release from Fe–P under flooded conditions.

The RAE value (89% in Table 7) of calcined Juquia PR for grain yield of upland rice grown to maturity on the unlimed Ultisol (pH 5.4) obtained in the present study is higher than the RAE value (53%) of the same calcined Juquia PR for dry matter yield of upland rice grown only for 65 d on an unlimed Oxisol (pH 5.6) as reported by Francisco et al. (2008). However, the RAE value for grain yield of flooded rice grown to maturity obtained in the present study (66% in Table 5) is close to the RAE value (69%) as reported by Francisco et al. (2008) for dry matter yield when the flooded rice was harvested after 65 d. One possible explanation could be that upland rice treated with calcined PR samples took a longer time to reach maturity (about 150 d) than flooded rice (about 100 d) so that significant difference in RAE for rice grain yield was observed with upland rice whereas none was with flooded rice between the results of Francisco et al. (2008) and the present study. However, future research is needed to further investigate and explain these different results.

Both Pi and resin P show that soil available P after upland rice from all P sources increased with P rate in the unlimed (Fig. 10 ) and limed soils (Fig. 11 ). In the unlimed soil, the order of Pi-P and resin-P follows TSP = Gafsa PR > calcined Juquia PR = calcined Sapucaia PR (Fig. 10). In the limed soil, the order follows TSP > calcined Juquia PR = calcined Sapucaia PR > Gafsa PR (Fig. 11). The data for soil available P thus agree well with the data of rice grain yield and P uptake in ranking the agronomic effectiveness of P sources. Both Pi and resin results also show good correlation between soil available P and P uptake in the unlimed soil (Fig. 12 ) and the limed soil (Fig. 13 ). Several reports have shown that Pi and resin P are the only two soil P tests that are able to correlate available P with P uptake or crop yield in the soils treated with PR or water-soluble P while other commonly used soil tests such as Bray 1 and Olsen have failed (Saggar et al., 1992a, 1992b; Menon and Chien, 1995; Habib et al., 1998; Chien, 2004).


Figure 10
View larger version (25K):
[in this window]
[in a new window]

  		Fig. 10. Phosphate extracted by (A) iron oxide-impregnated filter paper or (B) anion exchangeable resin of soil samples (unlimed soil) collected after upland rice was harvested. Models followed by the same letter are statistically not different from each other in the slope (P ≤ 0.05).

 

Figure 11
View larger version (25K):
[in this window]
[in a new window]

  		Fig. 11. Phosphorus extracted by (A) iron oxide-impregnated filter paper (Pi strip) or (B) anion exchangeable resin of soil samples related to P rate applied to limed soil. Models followed by the same letter are statistically not different from each other in the slope (P ≤ 0.05).

 

Figure 12
View larger version (19K):
[in this window]
[in a new window]

  		Fig. 12. Relationship between P uptake by upland rice grain grown on unlimed soil and P extracted by (A) iron oxide-impregnated filter paper (Pi strip) or (B) anion exchangeable resin of soil samples after rice harvest.

 

Figure 13
View larger version (20K):
[in this window]
[in a new window]

  		Fig. 13. Relationship between P uptake by upland rice grown in limed soil and P extracted by (A) iron oxide- impregnated filter paper or (B) anion exchangeable resin of soil samples collected after plants were harvested.

 

   CONCLUSIONS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
The results of this study show that the nonapatite PR sources containing crandallite minerals in the form of Ca–Fe–Al–P cannot be used for direct application without calcination. The agronomic effectiveness of calcined nonapatite PR for flooded and upland rice is greatly influenced by the soil pH and moisture regime. In the unlimed soil with pH 5.4 for upland rice, the Ca–P characteristic of the Ca–Fe–Al–P dominated PR dissolution over the Fe–Al–P characteristic to provide available P similar to apatite PR containing only Ca–P. In the limed soil with pH 7.0 for upland rice, available P of the calcined nonapatite PR came mainly from the dissolution of the Fe–Al–P characteristic which increased with soil pH while the Ca–P characteristic played no role as evidenced by the lack of upland rice response to the highly reactive apatite PR in the limed soil. After the soil was flooded, increased soil pH, Fe reduction, and Al-hydrolysis were responsible for the dissolution of Fe–Al–P characteristic to provide available P for flooded rice while the Ca–P characteristic played no role as shown by the totally ineffectiveness of apatite PR for flooded rice. In summary, the results show that the conditions for the best agronomic use of calcined nonapatite PR in increasing rice grain yield follow the order of upland rice without liming > flooded rice > upland rice with liming. The agronomic effectiveness of the calcined nonapatite PR sources for flooded and upland rice was much higher than other upland crops that have been reported in literature. However, future field trials are needed to validate the present greenhouse results.


   ACKNOWLEDGMENTS
 
The senior author expresses his gratitude to CAPES (Coordenacao de Aperfeicoamento de Pessoal de Nivel Superior) for the research scholarship that made this study possible and to the staff of IFDC for their help during the study. He also wants to thank Prof. Dr. Marcondes Lima da Costa, Federal University of Para, Brazil, for his help to provide the Sapucaia PR samples used in the present study.

All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher.


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





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 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 Google Scholar
Google Scholar
Right arrow Articles by Francisco, E. A. B.
Right arrow Articles by Taylor, R. W.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Francisco, E. A. B.
Right arrow Articles by Taylor, R. W.
Agricola
Right arrow Articles by Francisco, E. A. B.
Right arrow Articles by Taylor, R. W.
Related Collections
Right arrow Rice
Right arrow Nutrient Cycling
Right arrow Nutrient Management
Right arrow Soil Fertility and Productivity

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Crop Science Vadose Zone Journal
Journal of Natural Resources
and Life Sciences Education		 Soil Science Society of America Journal
Journal of Plant Registrations Journal of
Environmental Quality		 The Plant Genome