METABOLISM AND NUTRITION
Laboratory Evaluations of Feed-Grade and Agricultural-Grade Phosphates1
F. R. LIMA,*,2 J.I.M. FERNANDES,*,3 E. OLIVEIRA,† G. C. FRONZAGLIA,‡ and H. KAHN‡
*Departamento de Nutrição e Produção Animal, Faculdade de Medicina Veterinária e Zootecnia, Universidade de São Paulo,
Pirassununga, São Paulo, 13630-000, Brazil; †Instituto de Quı́mica, Universidade de São Paulo, São Paulo, São Paulo,
05508-900, Brazil; and ‡Laboratório de Caracterização Tecnológica, Departamento de Engenharia de Minas, Escola
Politécnica, Universidade de São Paulo, São Paulo, São Paulo, 05508-900, Brazil
ABSTRACT Nine samples of pure, feed-grade (FP) and
agricultural-grade (AP) phosphates were evaluated at
seven laboratories (six in Brazil and one in the U.S.) for
physical and chemical characteristics. Phosphates were
one “standard” pure dicalcium phosphate; four FP, two
dicalcium phosphates (FP-1 and FP-2) made in Brazil, one
di-monocalcium phosphate (FP-3), and one defluorinated
phosphate (FP-4) made in the U.S.; and four AP made in
Brazil [single superphosphate (AP-1), triple superphosphate (AP-2) and monoammonium (AP-3), and thermomagnesium (AP-4) phosphates]. Average analytical values for FP and AP, respectively, were 3.3 and 6.3% moisture, 1.0 and 2.5% insoluble residue, 16.2 and 28.4% loss
on ignition, 6.8 and 4.7 (pH), 1,028 and 1,023 g/L apparent
density, 9.6 and 55.0% P solubility in water, 83.6 and
88.4% P solubility in 2% citric acid, and 85.2 and 97.0% P
solubility in neutral ammonium citrate. Based on particle
size, six products were classified as “fine,” and three were
classified as “irregular.” Atomic absorption and plasma
spectrometry determinations were performed for 31 essential and potentially harmful or radioactive minerals.
The Na level was high in FP-4 (6.03%). Mineral concentrations were safe for all FP as compared with NRC standards. Levels in AP were toxic, exceeding the tolerance
limits for F, Fe, Mg, and Ba, and were particularly high
as compared with FP for S, Ti, and radioactive Th. The
AP-1 was high in F, Ba, S, and Th; AP-2 and AP-3 were
high in F and S; and AP-4 was high in F, Ba, Fe, Mg, Ti,
and Th. X-ray diffraction assays detected impurities for all
commercial samples and identified as major components
CaHPO4ⴢ2H2O (standard phosphate), CaCO3 and
CaHPO4 (FP-1, FP-2, and FP-3), Ca(H2PO4)2ⴢH2O (FP-3),
Na2Ca3Al2(PO4)2(SiO4)2 and Ca3(PO4)2 (FP-4), CaSO4ⴢnH2O
and (NH4)Fe3P6O20ⴢ(PO4)2 (AP-1), Ca(H2PO4)2ⴢH2O and
KFe3P6O20ⴢ10H2O (AP-2), (NH4)H2PO4 and CaSO4ⴢnH2O
(AP-3), and no definite molecular structure for AP-4, an
amorphous product. The biological consequences of feeding animals a mineral source with no definite molecular
structure, an amorphous product, is not known. A biological evaluation of all phosphates included in this article
is being published as a separate report (Fernandes et
al., 1999).
(Key words: feed phosphates, agricultural phosphates, chemical analysis, minerals, X-ray diffraction)
1999 Poultry Science 78:1717–1728
INTRODUCTION
Phosphates are salts of phosphoric acid, and a wide
variety of such products is commercially available.
Among the various types of inorganic phosphates, biological value varies widely because of differences in chemical
structure, crystallinity, particle size, pH, and concentration of contaminating elements. Phosphorus is not supplied as a pure chemical and many phosphates are ores
that may have had little processing or purification,
whereas others are by-products of some industrial process. These facts must be considered in estimating the
Received for publication November 19, 1998.
Accepted for publication July 1, 1999.
1
Sponsored by FAPESP. Project Number 1995/2006-6.
2
To whom correspondence should be addressed: frlima@usp.br
3
Present address: UFPR - Campus Palotina. R.24 de Junho, 698.
Palotina, Parana, 85950-000, Brazil.
quality of a mineral source (Ammerman et al., 1977). Physical and chemical evaluations of sources of supplemental
P, such as dicalcium and monocalcium phosphates, have
been published by Gillis et al. (1962), Griffith and Schexnailder (1970), NRC (1980, 1994), Burnell et al. (1990),
McDowell (1992), Sullivan et al. (1992, 1994), and Lima et
al. (1995).
Dicalcium phosphate is commonly used as a source of
supplemental P and is not a chemically defined entity. It
is, in fact, a mixture of varying amounts of dicalcium
and monocalcium phosphates, phosphoric acid, calcium
carbonate, and impurities, depending on the origin of the
raw material and procedures employed in its industrial
production (Lima et al., 1995).
Abbreviation Key: AP = agricultural-grade phosphate; FP = feedgrade phosphate.
1717
1718
LIMA ET AL.
Defluorinated phosphate is a feed-grade P supplement
thermochemically produced by the reaction of phosphate
rock concentrate and phosphoric acid in the presence of
soda ash. The raw material is processed through a mixer
and a kiln (1,300 to 1,500 C) so that fluorine is removed
as hydrofluoric acid, a volatile compound. After cooling
and screening, the final product is tricalcium phosphate
[Ca3(PO4)2] containing low F levels (0.17%) and a relatively high Na concentration (4.8 to 5.5%).
According to Sheve and Brink (1977), phosphate rock
concentrate and sulfuric acid react to produce single superphosphate, a mixture of nearly equal amounts of monocalcium phosphate [Ca(H2PO4)2] and calcium sulfate
[CaSO4]. The finely ground material is cured in large piles
so that reaction may go to completion, a process that
takes several weeks. Commercial superphosphate contains considerable amounts of unreacted phosphate rock
and sulfuric acid. Triple superphosphate results from the
reaction, at room temperature, of phosphate rock concentrate and phosphoric acid. Commercial triple superphosphate is a mixture of monocalcium phosphate
[Ca(H2PO4)2] and appreciable amounts of unreacted
phosphate rock and phosphoric acid. Monoammonium
and diammonium phosphates are obtained by reacting
phosphoric acid and anhydrous ammonia (Cardoso,
1991). If purified phosphoric acid is used in the process,
the products may be used safely to feed food-producing
animals. Otherwise, toxicity and potencially toxic mineral
residue in edible tissues may become a problem. Thermomagnesium phosphates are produced by very high temperature (1,600 C) fusion of rock phosphate concentrate
blends in the presence of silica (SiO2) and magnesium
(MgO) sources (Cardoso, 1991).
When agricultural-grade phosphates (AP) are priced
lower than feed-grade phosphates (FP), animal nutritionists become very much aware of the need for alternate
dietary P sources. Safety, however, is a matter of concern.
Agricultural-grade phosphates are not intended for use
in animal diets because they are not routinely produced
according to the strict manufacturing procedures necessary to guarantee the adequate degree of purity for feeding food-producing animals. Phosphorus in certain AP
is present in chemical forms that are easily used by plants,
but bioavailability for animals may be low. In addition,
high levels of certain contaminating mineral elements are
perfectly safe for soil and plants, but they may be toxic
to the animal.
When rock phosphate is solubilized in the process to
obtain phosphate fertilizers, little loss of contaminants
4
Laboratory 1: Departamento de Nutriçãoe Produção Animal, VNPFMVZ-USP, Pirassununga, SP, Brazil; Laboratory 2: Multimix Ltda.,
Campinas, SP, Brazil; Laboratory 3: M. Cassab Ltda., S. Paulo, SP, Brazil;
Laboratory 4: Socil S.A., Descalvado, SP, Brazil; Laboratory 5: Departamento de Quimica Analı́tica, IQ-USP, S. Paulo, SP, Brazil, Laboratory 6: IMC-Agrico, New Wales, Florida; Laboratory 7: LCT - Laboratório
de Caracterização Tecnológica, Departamento de Engenharia de Minas,
EP-USP, São Paulo, SP, Brazil.
5
Perkin-Elmer, Norwalk, CT, 06859.
takes place, and the final product undergoes no further
purification (Cardoso, 1991).
The purpose of this study was to evaluate chemical
and physical characteristics of one pure, four FP (two
made in Brazil and two made in the U.S.), and four samples of AP made in Brazil.
MATERIALS AND METHODS
Nine samples of phosphate sources were evaluated at
seven different laboratories, six in Brazil and one in the
U.S., for physical and chemical characteristics. Phosphate
sources were four samples of FP [two dicalcium phosphates made in Brazil (FP-1 and FP-2) one di-monocalcium phosphate (FP-3), and one defluorinated phosphate
(FP-4) made in the U.S.] and four samples of AP made
in Brazil (AP-1, AP-2, AP-3, and AP-4). Analytical values
estimated for all experimental phosphates are presented
in Table 1.
Standard analytical methods recommended by the Association of Official Analytical Chemists (1984) were followed. All phosphates were analyzed for moisture, insoluble residue, loss on ignition, particle size, pH, P solubility in water, P solubility in 2% citric acid, P solubility in
neutral ammonium citrate. Mineral analysis included the
essencial Ca, P, Mg, Na, K, Co, Cu, Fe, Mn, Mo, S, Se,
Zn, and the potencially toxic elements Al, F, As, B, Ba,
Bi, Cd, Cr, Hg, Ni, Pb, Sb, Sn, Ti, V, W, and the radioactive
U and Th. Analyses were performed at seven different
laboratories in Brazil and in the U.S.4 Phosphorus and Ca
analyses were averaged from the results obtained at six
laboratories; fluorine, pH, apparent density, and particle
size analyses were based on the results from two labs.
The IMC-Agrico laboratory employed a segmented-flux
analytical system for Ca and P determinations and an ion
selective electrode for flourine analysis. Atomic absorption (Perkin-Elmer model 51005) was used for Al, Fe, and
Mg. Trace elements were analyzed by plasma spectrometry (Perkin-Elmer Soiax Elan 5005), and moisture determinations were performed using a vacuum oven. At the
Laboratory of Atomic Emission Spectrometry of the Institute of Chemistry (IQ-USP), plasma spectrometry (ICPAES) was used for analysis of 21 mineral elements. Values
of pH were obtained from 5-g phosphate samples in 250mL solutions (water pH = 7) using a potentiometer. Moisture was determined at 80 C, basically to quantify hygroscopicity of the products because water of crystalization
losses may occur with temperatures as low as 109 C. Loss
on ignition was determined at 1,000 C in dry samples
and represents losses in water of crystalization, carbon
dioxide from carbonates, and volatile mineral elements
such as As, Hg, and halogens. Insoluble residue in HCl
followed by HNO3 basically quantifies silica content. All
X-ray diffraction assays were performed at the Laboratório de Caracterização Tecnológica (LCT) laboratory using
the methodology described by Lima et al. (1995) for identification of the chemical species present in the samples.
Analytical procedures employed an X-ray diffractometer
1719
COMPOSITION OF FEED- AND AGRICULTURAL-GRADE PHOSPHATES
TABLE 1. Analytical values guaranteed by manufacturers (FP; feed-grade phosphate) or estimated from
the literature (AP; agricultural-grade phosphate) for the P sources studied
P source
Origin
Ca
(maximum)
SP1
FP-1
FP-2
FP-3
FP-4
AP-1
AP-2
AP-3
AP-4
Dicalcium, Brazil
Dicalcium, Brazil
Dicalcium, Brazil
Dicalcium, U.S.
Defluorinated, U.S.
Super single, Brazil
Super triple, Brazil
Monoammonium, Brazil
Thermomagnesium, Brazil
24.0
27.0
26.0
24.0
33.0
20.0
14.0
1.0
20.0
P
(minimum)
F
(maximum)
18.0
19.0
18.0
18.5
18.0
7.8
17.9
21.0
7.5
0.19
0.18
0.19
0.18
0.90
0.50
0.30
(%)
1
Standard phosphate.
PHILLIPS-PW18806 with a PW1710 controller and PCAPD (automated powder diffraction) software.
RESULTS AND DISCUSSION
The X-ray diffractogram interpretation is summarized
in Table 2. Moisture content, insoluble residue, loss on
ignition, pH, apparent density, P solubility in water, P
solubility in 2% citric acid, and P solubility in neutral
ammonium citrate values are presented in Table 3. Particle size data are shown in Table 4. Levels of 13 essential
mineral elements are shown in Tables 5 and 6. Levels of
18 potentially harmful or toxic mineral and radioactive
elements are presented in Tables 7 and 8.
Moisture, Insoluble Residue,
Loss on Ignition, and pH
Moisture levels (Table 3) were variable and averaged
3.3 and 6.3% for FP and AP, respectively. Particularly
high moisture content for AP-1 (13.3%) and AP-2 (8.7%)
seem to be primarily related to the industrial wet process
used to obtain superphosphates, in which sulphuric acid
is diluted in water to control product moisture conditioning the chemical reaction (Sheve and Brink, 1977). The
low moisture (0.1%) detected for the thermochemically
produced phosphate sources (FP-4 and AP-4) is in
agreement with Sullivan et al. (1994), who reported that
low moisture in such products may be explained by high
temperatures (1,250 to 1,550 C) employed in industrial
processing. Moisture was determined at a low temperature of 80 C, to quantify hygroscopicity of the products,
as some hydration water losses may occur at temperatures as low as 109 C. Moisture values for FP in our study
compare with those published by Potter (1988), Sullivan
et al. (1994), and Lima et al. (1995). Insoluble residue levels
(Table 3) were low, averaging 1.0% for FP and 2.5% for
AP. Considerable variation, however, was observed
among AP samples, ranging from 0% (AP-4) to a maximum of 9.2% (AP-1).
6
PHILLIPS, 7602 EA, Almelo, Holland.
Insoluble residue in HCl followed by HNO3 quantifies
the silica content of the sample (Lima et al., 1995). Single
superphosphate is the product of reacting sulphuric acid
and concentrate rock phosphate. Because no purification
processes are applied whatsoever, all impurities present
in rock phosphate, including silica, remain in the final
product, which may very well explain the high insoluble
residue value found for AP-1.
Loss on ignition levels (Table 3) were variable, with a
minimum of 0% (FP-4 and AP-4) and a maximum of
53.5% (AP-3). Loss on ignition was determined at 1,000
C in dry samples and represents losses in water of crystalization, carbon dioxide from carbonates, and volatile mineral elements such as As, Hg, and halogens. Defluorinated
rock phosphate (FP-4) and thermomagnesium phosphate
(AP-4) are the products of high temperature (1,250 to
1,550 C) industrial processing that eliminates calcium carbonate, volatile mineral elements, as well as water of
crystalization, which may explain the lack of detectable
losses at 1,000 C observed for these samples. Loss on
ignition values for commercial dicalcium phosphates (FP1, 2, and 3) in this study are in agreement with data
published by Lima et al. (1995) that reported, for seven
TABLE 2. X-ray diffraction presence of chemical species
in the P sources studied
Chemical species present
P source
Major
SP
FP-1
FP-2
CaHPO4ⴢ2H2O
CaHPO4
CaHPO4
FP-3
FP-4
AP-1
CaHPO4
Ca(H2PO4)2ⴢH2O
Na2Ca3Al2(PO4)2(SiO4)2
CaSO4ⴢnH2O2
AP-2
AP-3
AP-4
Ca(H2PO4)2ⴢH2O
(NH4)H2PO4
amorphous
1
Minor
CaCO3
CaCO3
CaSO4ⴢ2H2O
CaCO3
Ca3(PO4)2
Ca(H2PO4)2ⴢH2O
(NH4)Fe3P6O20ⴢ(PO4)2
KFe3P6O20ⴢ10H2O
CaSO4ⴢnH2O2
amorphous
1
SP = Standard phosphate; FP = feed-grade phosphate; AP = agricultural-grade phosphate.
2
Several degrees of hydration molecules were detected for the chemical species.
1720
LIMA ET AL.
TABLE 3. Moisture, insoluble residue, loss on ignition, pH, apparent density, and P solubility
of feed-grade phosphates (FP) and agricultural-grade phosphates (AP)
P source
Moisture
Insoluble
residue
SP1
FP-1
FP-2
FP-3
FP-4
Mean
SE
AP-1
AP-2
AP-3
AP-4
Mean
SE
1.6
3.4
5.8
3.8
0.1
3.3
2.4
13.3
8.7
3.1
0.1
6.3
5.9
0.0
0.2
0.1
1.8
1.9
1.0
1.0
9.2
0.6
0.2
0
2.5
4.5
Loss on
ignition
pH
24.4
12.5
16.5
19.5
0
16.2
3.5
30.5
29.6
53.5
0
28.4
21.9
6.8
7.0
7.5
4.4
8.2
6.8
1.7
2.9
3.1
4.2
8.7
4.7
2.7
(%)
Apparent
density
Water
solubility
Citric acid
solubility
2.0
0.9
11.2
25.6
0.8
9.6
11.7
35.9
87.6
95.3
1.0
55.0
44.6
93.5
85.3
93.9
86.8
68.4
83.6
10.8
76.0
92.0
98.2
87.4
88.4
9.4
(g/L)
580
828
888
966
1,429
1,028
273
795
997
798
1,504
1,023
334
NAC
solubility
(%)
99.8
93.2
na2
97.0
65.3
85.2
17.3
99.7
96.9
99.8
91.4
97.0
3.9
SP = Standard phosphate.
na = Not available.
1
2
commercial dicalcium phosphates, loss on ignition values
ranging from 13.7 to 23.9%.
Considering the X-ray diffraction data (Table 2) and
comparing loss on ignition values presented in Table 3,
the predominance of the dihydrate salt of calcium phosphate dibasic in SP may very well explain the relatively
high loss at 1,000 C (24.1%) observed for that product.
On the other hand, losses on ignition for FP-1 (12.5%)
and FP-2 (16.5%) seem to be primarily related to the presence of relatively high amounts of calcium carbonate (Table 2) in both samples. Loss at 1,000 C for FP-3 (19.5%)
is probably the result of losses in water of crystalization
from calcium phosphate monobasic monohydrate and
carbon dioxide from calcium carbonate (Table 2). Regarding feed-grade dicalcium phosphates, higher values of
loss on ignition may indicate a higher degree of hydration
on the calcium phosphate molecule, suggesting superior
P availability potential. Lower values for loss on ignition,
on the other hand, may be due to variations in processing
that influence the degree of hydration and the proportion
of mono to dibasic calcium phosphate. A higher degree
of hydration and higher proportions of the monocalcium
salt are expected to contribute to increase P availability
potential in commercial dicalcium phosphates (Sullivan
TABLE 4. Particle size of feed-grade phosphates (FP) and
agricultural-grade phosphates (AP)
P source
>1.00 mm
>0.30 mm
<0.30 mm
SP1
FP-1
FP-2
FP-3
FP-4
AP-1
AP-2
AP-3
AP-4
0
2.6
5.8
5.5
21.5
12.8
30.5
6.6
0
(%)
0
3.2
6.8
10.7
27.5
14.4
33.7
8.4
0
100.0
96.8
93.2
89.3
72.5
85.6
66.4
91.6
100.0
SP = Standard phosphate.
1
et al., 1994, Lima et al., 1995). Considering the AP, the
predominance of several hydrated forms of calcium sulphate and the monohydrate salt of calcium phosphate
monobasic in AP-1 (Table 2) may explain the high loss
at 1,000 C (30.5%) observed for that sample of single
superphosphate. Loss at 1,000 C for AP-2 (29.6%) is probably the result of losses in water of crystalization from
calcium phosphate monobasic monohydrate and other
hydrated phosphate salts (Table 2). On the other hand,
loss on ignition for AP-3 (53.5%) seems to be primarily
related to volatile ammonia loss from the ammonium
phosphate molecule (Table 2) at 1,000 C.
Values of pH (Table 3) were variable and averaged 6.8
and 4.7 for FP and AP, respectively, ranging from 2.9 (AP1) to 8.7 (AP-4). Phosphate products containing higher
proportions of monocalcium phosphate are expected to
show lower pH values (Lima et al., 1995). Our study has
shown that, for FP, the lowest pH value (4.4) was obtained
for the U.S.-made di-monocalcium phosphate FP-3 (Table
2), intermediate values (7.0 and 7.5) were assigned to
the exclusively dicalcium phosphate (Table 2) containing
products (FP-1 and FP-2), and the highest pH (8.2) was
recorded for the tricalcium phosphate containing (Table
2) U.S.-made defluorinated rock phosphate (FP-4). Our
data are in agreement with Potter (1988), who published
pH values of 3.4 for monocalcium phosphate and values
ranging from 4.4 to 7.4 for commercial dicalcium phosphate samples, and by Lima et al. (1995), who reported
pH values ranging from 3.2 to 4.8 for four commercial
dicalcium phosphates containing different proportions of
monocalcium phosphate and pH = 5.7 to 6.1 for three
products containing exclusively dicalcium phosphate. Regarding the AP, our interpretation of the pH data (Table
3) was reconfirmed by the X-ray diffraction assays (Table
2) that have detected levels of monocalcium phosphate
inversely proportional to pH values in each product. The
lowest pH values (2.9 and 3.1) were recorded for the
monocalcium phosphate containing single (AP-1) and triple (AP-2) superphosphates and were slightly higher (4.2)
1721
COMPOSITION OF FEED- AND AGRICULTURAL-GRADE PHOSPHATES
TABLE 5. Essential mineral element composition of feed-grade phosphates (FP)
and agricultural-grade phosphates (AP): Ca, P, Mg, Na, and K
P source
Ca1
P
Mg
SP2
FP-1
FP-2
FP-3
FP-4
Mean
SE
AP-1
AP-2
AP-3
AP-4
Mean
SE
24.5
26.2
25.4
21.1
31.9
26.2
4.4
18.3
14.7
2.6
19.5
13.8
7.7
18.0
19.2
18.1
18.1
17.8
18.3
0.6
8.4
19.6
22.8
7.3
14.5
7.8
0.02
0.45
0.91
0.78
0.35
0.62
0.27
0.11
0.70
0.80
9.89
2.88
4.68
Na
K
0.13
0.21
0.31
0.31
6.03
1.72
2.90
0.21
0.21
0.15
0.23
0.20
0.03
0.008
0.03
0.05
0.09
0.09
0.07
0.03
0.02
0.04
0.05
0.05
0.04
0.01
(%)
1
High variability on Ca results for phosphate AP-3 (1.2 and 4.1%).
SP = Standard phosphate.
2
for the monoammonium phosphate product. A high pH
value of 8.7 was recorded for the amorphous magnesium
thermophosphate sample (AP-4).
Apparent Density, Particle Size, and
Phosphorus Solubility
Particle size patterns (Table 4) were variable, allowing
for a classification of the products as “fine” (SP, FP-1, FP2, and AP-4), “coarse” (FP-3 and AP-3), and “irregular”
(FP-4, AP-1, and AP-2).
Apparent density (Table 3) values were higher (1,429
and 1,504 g/L) for thermochemically produced phosphates sources FP-4 and AP-4, respectively. For the remaining FP, density ranged from 828 and 888 g/L for the
Brazilian “fine” products FP-1 and FP-2, respectively, to
966 g/L for the U.S.-made “coarse,” or granular, di-monocalcium phosphate (FP-3). These results are in agreement
with Potter (1988) and Lima et al. (1995), who reported
higher density values for dicalcium phosphate samples
classified as coarse according to their particle size pattern,
as compared with finer products.
The defluorinated phosphate sample differed in appearance among the FP, showing wide variations in particle size (very coarse), in shape (from fine powder to scalelike particles), and in color (predominantely dark brown
with scattered yellowish particles). Among AP, the thermomagnesium product markedly differed in appearance,
showing homogeneity in particle size (very fine with tendency of particles to stay together but not to lump), and
in color (dark brown). Considering the AP, the monoammonium product (AP-3) was a regular “coarse,” or granular, product (8.4% > 0.30 mm) with density equal to 798
g/L, whereas AP-1 and AP-2 were “lumpy” materials
that tended to form objects 14.4 to 33.7% > 0.30 mm, with
density ranging from 795 (AP-1) to 997 g/L (AP-2). These
differences seem to be primarily related to the type of
manufacturing procedures employed in the AP industry.
Particle size and density are characteristics related to the
miscibility of the product in conventional poultry diets
TABLE 6. Essential mineral element composition of feed-grade phosphates (FP)
and agricultural-grade phosphates (AP): Co, Cu, Fe, Mn, Mo, S, Se, and Zn
P Source
Co
Cu
Fe1
Mn
SP2
FP-1
FP-2
FP-3
FP-4
Mean
SE
AP-1
AP-2
AP-3
AP-4
Mean
SE
4
16
1
3
4
6
6.8
4
6
6
26
10
10.4
2
97
18
4
12
33
43.2
10
16
11
55
23
21.5
210
19,654
1,679
10,562
10,282
10,544
7,340
5,848
8,533
10,154
67,147
22,920
29,538
3
726
174
224
347
368
250
370
284
338
2,010
750
840
Mo
S
Se
Zn
<1
3
3
9
7
6
3.0
1
3
1
6
3
2.4
579
2,935
2,711
39
29
1,428
1,613
122,054
18,090
15,825
168
39,034
55,917
<5
<5
<5
<5
<5
<5
...
25
<5
<5
35
16
16
51
168
546
83
134
233
212
172
66
70
473
195
192
(ppm)
1
High variability on Fe results for phosphate AP-1 (155 and 11,541 ppm).
SP = Standard phosphate.
2
1722
LIMA ET AL.
TABLE 7. Potentially harmful mineral element levels of feed-grade phosphates (FP)
and agricultural-grade phosphates (AP): Al, F, As, B, Ba, Bi, Cd, Cr, and Hg
P source
Al
SP3
FP-1
FP-2
FP-3
FP-4
Mean
SE
AP-1
AP-2
AP-3
AP-4
Mean
SE
0.01
0.11
0.15
0.82
0.73
0.45
0.37
0.16
0.14
0.15
0.61
0.27
0.23
F
As
B1
Ba2
0.02
0.03
0.19
0.17
0.15
0.14
0.07
1.11
0.66
0.38
0.95
0.78
0.32
7
7
8
11
7
8
2
9
10
8
11
10
1.3
<0.4
<0.4
16
39
4
20
18
62
12
7
9
22
26
9
270
21
9
306
151
158
18,494
643
333
32,206
12,919
15,407
(%)
Bi
Cd
Cr
Hg
<1
1
3
5
<1
3
2
1
<1
<1
1
0.8
0.3
1
21
53
72
51
49
21
11
2
3
661
169
328
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
...
<0.5
<0.5
<0.5
<0.5
<0.5
...
(ppm)
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
...
<0.1
<0.1
<0.1
<0.1
<0.1
...
1
High variability for phosphate AP-1 (non-detectable and 124 ppm).
High variability for phosphate FP-1 (29 and 512 ppm), FP-4 (70 and 542 ppm), AP-1 (5,848 and 31,141 ppm),
and AP-3 (11 and 655 ppm).
3
SP = Standard phosphate.
2
based on ground cereal grain and oilseed meals. Larger
particle size in phosphates was shown to improve P utilization, probably by increasing retention time of material
in the gizzard of poultry (Griffith and Schexnailder, 1970;
Potter, 1988; Burnell et al., 1990), exposing food longer to
the more acidic conditions of the upper digestive tract,
which may solubilize P more completely (Burnell et al.,
1990). Very fine products are also difficult to handle at
the feed mill, exposing mill workers to uncomfortable
environmental conditions and increasing losses in exhaustion fans.
Phosphorus solubility in water, in 2% citric acid, and
in neutral ammonium citrate (Table 3) was variable and
averaged, respectively, 9.6, 83.6, and 85.2% for FP and
55.0, 88.4, and 97.0% for AP. Water solubility of P was
high for AP-2 (87.6%) and AP-3 (95.3%), which is in
agreement with the work of Sheve and Brink (1977), who
reported P from monocalcium and monoammonium
phosphate sources to be highly soluble in water. Compared with triple superphosphate (AP-2), single super-
phosphate (AP-1) was found to contain higher levels of
Ca, which is present in the product largely as nonwater
soluble compounds, and may explain the low water solubility of P in AP-1 (35.9%). Phosphorus solubility of defluorinated phosphate (FP-4) was low in water (0.8%), in
2% citric acid (68.4%), and in neutral ammonium citrate
(65.3%). Phosphorus was found to be highly soluble (all
values in excess of 85%) in 2% citric acid and in neutral
ammonium citrate for all phosphate samples studied, except for the defluorinated sample and for the citric acid
solubility for AP-1 (76%). Our results are in general accordance with Potter (1988), Sullivan et al. (1992), and Lima
et al. (1995), who reported P solubility in 2% citric acid
greater than 90% for FP samples. The work of Sullivan
et al. (1992) has shown that P solubility of feed phosphates
in HCl or neutral ammonium citrate was positively correlated with the bioavailability of P, and this is of great
interest to the phosphate industry and to nutritionists.
However, other researchers (Gillis et al., 1948; Day et al.,
1973) found little or no agreement between bioavailability
TABLE 8. Potentially harmful and radioactive mineral element levels of feed-grade phosphates (FP)
and agricultural-grade phosphates (AP): Ni, Pb, Sb, Sn, Ti, V, W, Th, and U
P source
Ni
Pb
Sb
Sn
Ti
V
W
Th
U
SP1
FP-1
FP-2
FP-3
FP-4
Mean
SE
AP-1
AP-2
AP-3
AP-4
Mean
SE
2
44
15
16
21
24
14
22
9
7
258
74
123
<1
21
16
24
26
22
4
39
18
4
95
39
40
<1
<1
1
3
1
1
1
<1
<1
<1
<1
<1
...
5
11
8
12
13
11
2
12
9
52
27
25
20
(ppm)
<0.6
519
37
122
398
269
227
339
146
225
2,010
680
890
1
32
81
139
129
95
49
33
101
122
111
92
40
<5
<5
<5
<5
<5
<5
...
<5
<5
<5
<5
<5
...
5
9
6
10
13
10
3
139
46
43
237
116
92
1
142
114
149
126
133
16
9
6
8
14
9
3
SP = Standard phosphate.
1
COMPOSITION OF FEED- AND AGRICULTURAL-GRADE PHOSPHATES
of P of the FP and their solubility in dilute acid solutions
and suggested that the in vitro tests could be useful only
to identify and eliminate insoluble compounds.
Essential Mineral Composition
Levels of Ca and P were determined at six laboratories;
Mg at four laboratories; Na, K, Co, Cu, Fe, Mn, Mo, Se,
and Zn at two laboratories; and S at one laboratory (Tables
5 and 6).
Calcium. Levels of Ca (Table 5) were variable, ranging
from 21.1 (FP-3) to 31.9% (FP-4), with an average of 26.2%
in FP. The average Ca content in AP was lower than that
in FP, averaging 13.8%, ranging from 2.6 (AP-3) to 19.5%
(AP-4). All samples were in compliance with the limits
of label guarantees (Table 1), not exceeding the maximum
Ca allowed, except for a deviation detected in sample
AP-3 (2.6 vs 1.0%).
Phosphorus. Levels of P (Table 5) were not very variable in FP, ranging from 17.8 (FP-4) to 19.2% (FP-1), with
an average of 18.3%. The P content in AP was lower than
that in FP, averaging 14.5%, ranging from 7.3 (AP-4) to
22.8% (AP-3). All samples were in compliance with the
limits of label guarantees (Table 1) for each product, not
showing values below the minimum P allowed.
Magnesium. Magnesium levels (Table 5) were variable, averaging 0.62%, and ranging from 0.35 (FP-4) to
0.91% (FP-2) in FP sources. Magnesium in AP was higher
than that in FP, averaging 2.88% and ranging from 0.11
(AP-1) to 9.89% (AP-4). High content of Mg in AP-4 was
due to large amounts of this element added in the manufacturing procedure used to obtain magnesium thermophosphates. This high level of Mg (9.89%), per se, would
not be toxic for poultry because the maximum tolerable
Mg level (NRC, 1980) for poultry is 0.3% in the complete
diet or 15% in a P source used at a dietary level of 2%,
as proposed by Lima et al. (1995). However, many other
factors must be considered in Mg toxicity, such as chemical form in which Mg is present in the product and the
metabolic relationship of Mg and other mineral elements.
High levels of Mg are known to interfere with the metabolism of several mineral elements and to increase the incidence of wet litter. In addition, to provide an equal
amount of P as obtained from FP (18% P), 2.4 times as
much AP-4 is needed so that 8.89% Mg in that product
would be equivalent to 21.3% Mg, a value that exceeds
by 42% the maximum tolerable level of Mg (NRC, 1980)
for poultry. It should be emphasized that, according to
our X-ray diffraction assays (Table 2), AP-4 is an amorphous product, showing no regular molecular structure.
Consequences of feeding animals high levels (or any
level) of Mg in a nonregular molecular structure form is
unknown, as no references can be found in the literature
regarding such a situation. A biological evaluation for all
samples studied in this experiment was performed at our
laboratory (Fernandes et al., 1999), and signs of toxicity
were evident for the AP-4 sample, with great depression
in growth, low bone ash, and low bone strength.
1723
Sodium. Levels of Na (Table 5) were very variable,
ranging from 0.21 (FP-1) to 6.03% (FP-4), with an average
of 1.72% in FP sources. The Na content in AP was lower
than that in FP, averaging 0.20% and ranging from 0.15
(AP-3) to 0.23% (AP-4). High content of Na in FP-4 is due
to the addition of Na2CO3 in the industrial processing
methods used to obtain defluorinated phosphates (Association of Florida Phosphate Chemists, 1991). Levels of
Na in FP may be of little or no consequence in poultry feed
formulation; nevertheless, Sullivan et al. (1992) reported
difficulty in managing poultry litter and facilities because
of increased water consumption when excess Na diets
were fed. The use of defluorinated phosphates as a P
supplement for poultry may require adjustments in the
level of salt in the diet. The maximum tolerable Na level
(NRC, 1980) for poultry is 2% in the complete diet.
Potassium. Levels of K (Table 5) were low, averaging
0.07% and ranging from 0.03 (FP-1) to 0.09% (FP-3 and
FP-4) in FP sources. The K content in AP averaged 0.04%,
ranging from 0.02 (AP-1) to 0.05% (AP-3 and AP-4). These
levels may be of little or no consequence in poultry feed
formulation because of the abundance of this element in
feedstuffs. The maximum tolerable K level (NRC, 1980)
for poultry is 3% of the complete diet.
Cobalt. Levels of Co (Table 6) were low, ranging from
1 (FP-2) to 16 ppm (FP-1), with an average of 6 ppm in
FP sources. The Co content in AP averaged 10 ppm, ranging from 4 (AP-1) to 26 ppm (AP-4). These levels may be
of little or no consequence in poultry feed formulation
because the maximum tolerable Co level (NRC, 1980) for
poultry is 10 ppm in the complete diet or 500 ppm in a
P source used at a dietary level of 2%.
Copper. Levels of Cu (Table 6) were low, ranging from
4 (FP-3) to 97 ppm (FP-1), with an average of 33 ppm in
FP sources. The Cu content in AP averaged 23 ppm, ranging from 10 (AP-1) to 55 ppm (AP-4). These levels may
be of little or no consequence in poultry feed formulation
because the maximum tolerable Cu level (NRC, 1980) for
poultry is 300 ppm in the complete diet or 15,000 ppm
in a P source used at a dietary level of 2%. Considering
the poultry requirement for Cu (8 ppm in a complete
diet) as estimated by the NRC (1994), any possible contribution to Cu nutrition by the FP or AP tested would
be negligible.
Iron. Levels of Fe (Table 6) were variable, ranging from
1,679 (FP-2) to 19,654 ppm (FP-1), with an average of
10,544 ppm in FP sources. The Fe content in AP was
higher than that in FP, averaging 22,920 ppm and ranging
from 5,848 (AP-1) to 67,147 ppm (AP-4). The highest Fe
level was observed for AP-4 (67,147 ppm), and this level
may be toxic for poultry because the maximum tolerable
Fe level (NRC, 1980) for poultry is 1,000 ppm in the complete diet or 50,000 ppm in a P source used at a dietary
level of 2%. A biological evaluation for all samples studied
in this experiment was performed at our laboratory (Fernandes et al., 1999), and signs of toxicity were evident for
the AP-4 sample, with a strong depression in growth,
low bone ash, and low bone strength. Our observations
are in agreement with Sooncharerying and Edwards Jr.
1724
LIMA ET AL.
(1990), who reported low body weights, low bone ash,
and signs of rickets in birds fed diets containing 1,000
ppm Fe. Rock phosphates are frequently high in Fe and
Al, forming insoluble iron-aluminum-phosphate complexes (Ammerman et al. 1977).
Manganese. Levels of Mn (Table 6) were variable,
ranging from 174 (FP-2) to 726 ppm (FP-1), with an average of 368 ppm in FP sources. The Mn content in AP was
higher than that in FP, averaging 750 ppm, ranging from
284 (AP-2) to 2,010 ppm (AP-4). These levels of Mn may
be of little or no consequence in poultry feed formulation
because the maximum tolerable Mn level (NRC, 1980) for
poultry is 2,000 ppm in the complete diet or 100,000 ppm
in a P source used at a dietary level of 2%. Considering
the poultry requirement for Mn (60 ppm in a complete
diet), as estimated by the NRC (1994), any possible contribution to Mn nutrition by the commercial phosphate
products tested would be neglegible.
Molybdenum. Levels of Mo (Table 6) were low, averaging 6 and 3 ppm for FP and AP sources, respectively.
These levels may be of little or no consequence in poultry
feed formulation because the maximum tolerable Mo
level (NRC, 1980) for poultry is 2,000 ppm in the complete
diet or 100,000 ppm in a P source used at a dietary level
of 2%.
Sulfur. Levels of S (Table 6) were variable, ranging
from 29 (FP-4) to 2,935 ppm (FP-1), with an average of
1,428 ppm in FP sources. The S content in AP was higher
than that in FP, averaging 39,034 ppm and ranging from
168 (AP-4) to 122,054 ppm (AP-1). High content of S in
AP-1 is due to the presence of large amounts of calcium
sulphate in the product as detected by X-ray diffraction
(Table 2). Consequences of these levels for poultry feed
formulation are unknown because the maximum tolerable S level has not yet been established (NRC, 1980) for
any animal species.
Selenium. Levels of Se (Table 6) were undetectable
(<5 ppm) in all FP sources and in two AP sources. The
maximum tolerable Se level (NRC, 1980) for poultry is 2
ppm in the complete diet or 100 ppm in a P source used
at a dietary level of 2%. The poultry requirement for this
mineral is 150 ppb (NRC, 1994).
Zinc. Levels of Zn (Table 6) were variable, ranging
from 83 (FP-3) to 546 ppm (FP-2), with an average of 233
ppm in FP sources. The Zn content in AP averaged 195
ppm, ranging from 66 (AP-2) to 473 ppm (AP-4). These
levels of Zn may be of little or no consequence in poultry
feed formulation because the maximum tolerable Zn level
(NRC, 1980) for poultry is 1,000 ppm in the complete diet
or 50,000 ppm in a P source used at a dietary level of 2%.
The poultry requirement for Zn is 40 ppm (NRC, 1994),
so that any possible contribution to Zn nutrition by the
phosphate products tested would be neglegible.
Potentialy Harmful
and Radioactive Minerals
Levels of Al, F, As, B, Ba, Cd, Cr, Hg, Ni, Pb, Sn, V,
and W were determined at two laboratories, and Bi, Sb,
Ti, Th, and U were determined at one laboratory (Tables
7 and 8).
Aluminum. Levels of Al (Table 7) were variable, ranging from 0.11 (FP-1) to 0.82% (FP-3), with an average of
0.45% in FP sources. The Al content in AP averaged 0.27%,
ranging from 0.14 (AP-2) to 0.61% (AP-4). The highest Al
levels were observed for FP-3 and FP-4 (0.82 and 0.73%,
respectively) and were produced from sedimentary rock
phosphate. These results are in agreement with Lima et al.
(1995), who published higher values of Al for phosphates
produced from rocks of sedimentary origin. Sedimentary
phosphate deposits contain higher levels of Al than rock
phosphates from igneous origin. These levels may be of
little or no consequence in poultry feed formulation because the maximum tolerable Al level, from soluble salts
of high bioavailability (NRC, 1980) for poultry is 200 ppm
in the complete diet, or 10,000 ppm in a P source used
at a dietary level of 2%. Higher levels of less soluble forms
can be tolerated (NRC, 1980). According to Cakir et al.
(1978), poultry can tolerate 300 to 400 ppm. In our studies,
the X-ray diffraction assays detected the presence of Al
in the form of the highly unsoluble (low toxicity) silicates
and phosphates in FP-4, which is in agreement with previous work from our laboratory (Lima et al., 1995) that
reported the presence of low solubility Al salts in all seven
commercial FP studied.
Fluorine. Levels of F (Table 7) ranged from 0.03 (FP1) to 0.19% (FP-2), with an average of 0.14% in FP sources.
The F content in AP was higher, averaging 0.78%, ranging
from 0.38 (AP-3) to 1.11% (AP-1). All samples of FP were
in compliance with the limits of label garantees (Table 1)
for each product, not exceeding the maximum F allowed.
On the other hand, all samples of AP contained excessive
F, as compared with a recommendation of the AAFCO
(1973), that feed phosphate supplements should contain
no more than 1 part F to 100 parts P. The proportions of
F to 100 parts of P determined were 13.2, 3.4, 1.7, and
13.0 for AP-1, AP-2, AP-3, and AP-4, respectively, unacceptable under the AAFCO (1973) recommendation,
which was exceeded by 1,210, 240, 70, and 1,200%, respectively, in those products.
Arsenic. Levels of As (Table 7) were very low, averaging 8 and 10 ppm for FP and AP sources, respectively.
That level would be of little or no consequence in poultry
feed formulation because the maximum tolerable As level
(NRC, 1980) for poultry is 100 ppm in the complete diet
or 5,000 ppm in a P source used at a dietary level of 2%.
Boron. Levels of B (Table 7) were variable, ranging
from undetectable (<0.4 ppm) in FP-1 to 39 ppm (FP-3),
with an average of 20 ppm in FP sources. The B content
in AP was equally low, averaging 22 ppm and ranging
from 7 (AP-3) to 62 ppm (AP-1). That level would be
of little or no consequence in poultry feed formulation
because the maximum tolerable B level (NRC, 1980) for
poultry is 150 ppm in the complete diet or 7,500 ppm in
a P source used at a dietary level of 2%.
Barium. Levels of Ba (Table 7) were variable, ranging
from 9 (FP-3) to 306 ppm (FP-4), with an average of 151
ppm in FP sources. The Ba content in AP was markedly
COMPOSITION OF FEED- AND AGRICULTURAL-GRADE PHOSPHATES
higher than that in FP, averaging 12,919 ppm, ranging
from 333 ppm (AP-3) to 18,494 (AP-1) and 32,206 ppm
(AP-4). The higher Ba levels observed for AP-1 and AP4 may be toxic to poultry because the maximum tolerable
Ba level (NRC, 1980) for poultry is 20 ppm in the complete
diet or 1,000 ppm in a P source used at a dietary level of
2%. There is little information in the literature concerning
Ba toxicosis in food-producing animals. According to the
NRC (1980), Ba is a highly toxic mineral when absorbed.
Despite the information that Ba tends to accumulate in
bone tissue, there are no reports describing pathological
bone disorders that can be accounted for by the ingestion
of high Ba diets (NRC, 1980).
Bismuth. Levels of Bi (Table 7) were undetectable (<0.1
ppm) in all samples evaluated. The maximum tolerable
Bi level (NRC, 1980) for poultry is 400 ppm in the complete
diet or 20,000 ppm in a P source used at a dietary level
of 2%.
Cadmium. Levels of Cd (Table 7) were very low in all
samples, ranging from undetectable (<0.1 ppm) in three
experimental phosphates to 5 ppm (FP-3), with averages
of 2.4 ppm for FP sources and 0.8 ppm for AP. That
level would be of little or no consequence in poultry
feed formulation because the maximum tolerable Cd level
(NRC, 1980) for poultry is 0.5 ppm in the complete diet
or 25 ppm in a P source used at a dietary level of 2%.
Chromium. Levels of Cr (Table 7) were variable, ranging from 21 (FP-1) to 72 ppm (FP-3), with an average of
49 ppm in FP sources. The Cr content in AP was higher
than that in FP, averaging 169 ppm and ranging from 2
(AP-2) to 661 ppm (AP-4). That amount would be of little
or no consequence in poultry feed formulation because
the maximum tolerable Cr level (NRC, 1980) for poultry
is 3,000 ppm in the complete diet or 150,000 ppm in a P
source used at a dietary level of 2%.
Mercury. Levels of Hg (Table 7) were undetectable
(<0.5 ppm) in all samples evaluated. The level would be
inconsequential in poultry feed formulation because the
maximum tolerable Hg level (NRC, 1980) for poultry is
2 ppm in the complete diet or 100 ppm in a P source used
at a dietary level of 2%.
Nickel. Levels of Ni (Table 8) were low, ranging from
16 (FP-3) to 44 ppm (FP-1), with an average of 24 ppm
in FP sources. The Ni content in AP was higher than that
in FP, averaging 74 ppm and ranging from 7 (AP-3) to
258 ppm (AP-4). This level would be of little or no consequence in poultry feed formulation because the maximum
tolerable Ni level (NRC, 1980) for poultry is 300 ppm in
the complete diet, or 15,000 ppm in a P source used at a
dietary level of 2%.
Lead. Levels of Pb (Table 8) were very low in all samples, with averages of 22 ppm for FP sources and 39 ppm
for AP. This amount would be of little or no consequence
in poultry feed formulation because the maximum tolerable Pb level (NRC, 1980) for poultry is 30 ppm in the
complete diet or 1,500 ppm in a P source used at a dietary
level of 2%.
Antimony. Levels of Sb (Table 8) were very low, found
to be undetectable (<1 ppm) in most samples. Dietary
1725
levels of Sb may be of no consequence in poultry feed
formulation because the maximum tolerable Sb level has
not been established (NRC, 1980) for poultry.
Tin. Levels of Sn (Table 8) were very low in all samples,
with averages of 11 ppm for FP sources and 25 ppm for
AP. This level may be of no consequence in poultry feed
formulation because the maximum tolerable Sn level has
not been established (NRC, 1980) for poultry.
Titanium. Levels of Ti (Table 8) were variable, ranging
from 37 (FP-2) to 519 ppm (FP-1), with an average of 269
ppm in FP sources. The Ti content in AP was markedly
higher than that in FP, averaging 680 ppm and ranging
from 146 (AP-2) to 2,010 ppm (AP-4). Dietary levels of Ti
may be of no consequence in poultry feed formulation
because the maximum tolerable Ti level has not been
established (NRC, 1980) for poultry.
Vanadium. Levels of V (Table 8) were variable, ranging
from 32 (FP-1) to 139 ppm (FP-3), with an average of 95
ppm in FP sources. The V content in AP averaged 92
ppm, ranging from 33 (AP-1) to 122 ppm (AP-3). This
amount would be of little or no consequence in poultry
feed formulation because the maximum tolerable V level
(NRC, 1980) for poultry is 10 ppm in the complete diet
or 500 ppm in a P source used at a dietary level of 2%.
Tungsten. Levels of W (Table 8) were undetectable (<5
ppm) in all samples and that would be inconsequential
in poultry feed formulation because the maximum tolerable W level (NRC, 1980) for poultry is 20 ppm in the
complete diet or 1,000 ppm in a P source used at a dietary
level of 2%.
Thorium. Levels of Th (Table 8) were variable, ranging
from 6 (FP-2) to 13 ppm (FP-4), with an average of 10
ppm in FP sources. The Th content in AP was higher than
that in FP, averaging 116 ppm and ranging from 43 (AP3) to 237 ppm (AP-4). This level may be of no consequence
in poultry feed formulation because the maximum tolerable Th level has not been established (NRC, 1980) for
poultry.
Uranium. Levels of U (Table 8) were variable, ranging
from 114 (FP-2) to 149 ppm (FP-3) with an average of 133
ppm in FP sources. The U content in AP was lower than
that in FP, averaging 10 ppm, ranging from 6 ppm (AP2) to 14 ppm (AP 4). This amount may be of no consequence in poultry feed formulation because the maximum
tolerable U level has not been established (NRC, 1980)
for poultry.
X-Ray Diffraction Analysis
The diffractograms generated by X-ray diffraction analysis were interpreted, and the results are summarized in
Table 2.
The predominant chemical substances present in all
commercial FP were calcium phosphates and calcium carbonates. For the reference standard, calcium phosphate
dibasic dihydrated purified grade, two forms of calcium
phosphate dibasic dihydrated were found to be the most
predominant products present, followed by smaller proportion of an anhydrous form of the same salt.
1726
LIMA ET AL.
Among the FP, monocalcium phosphate was found
only in FP-3. The anhydrous salt of dicalcium phosphate
was identified as the most predominant substance present
in FP-1, FP-2, and FP-3. Calcitic limestone was found in
large proportions in all three comercial dicalcium phosphate products and may be the result of excess calcium
carbonate added to neutralize phosphoric acid during
industrial processing.
These results are in agreement with a previous publication from our laboratory (Lima et al., 1995) that reported
X-ray diffraction evaluations for one pure and seven commercial dicalcium phosphates and found that monocalcium phosphate was present in only four commercial
products. According to that report, the most predominant
chemical species present in all commercial dicalcium
phosphates were calcium carbonate and anhydrous dicalcium phosphate.
The U.S.-made defluorinated phosphate was found to
contain tricalcium phosphate, predominantely in the form
of a sodium calcium aluminum silica phosphate, and this
is in agreement with Na analysis of this product that has
shown comparatively higher Na values (6.03%) than other
FP samples (0.21 to 0.31%). These results are consistent
with the pH values obtained, showing the lowest value
for FP-3 (pH = 4.4), intermediate values for samples FP1 (pH = 7.0) and FP-2 (pH = 7.5), and the highest value
for FP-4 (pH = 8.2).
Considering the AP, the most predominant chemical
substances present in single superphosphate (AP-1) were
one anhydrous and two hydrated forms of calcium sulphate, followed by smaller amounts of monocalcium
phosphate, plus an ammonium and iron salt of P. The
triple superphosphate product (AP-2) was found to contain monocalcium phosphate as the major chemical species, and a potassium and iron salt of P as a minor component. Monoammonium phosphate was found to be the
major component in the commercial monoammonium
phosphate sample (AP-3). Surprisingly, no crystaline
structure was identified by X-ray diffraction in the thermomagnesium phosphate (AP-4). Apparently, the high
temperature used in industrial processing of rocks to obtain AP-4 produces an amorphous product with no definite crystaline structure. There is no information in the
literature regarding the biological consequences to animals of feeding a product in which P is not present in a
regular molecular form.
In routine feed formulation for poultry and swine,
phosphate supplements are largely used at levels not to
exceed the 2% limit in conventional diets. Higher nutrient
concentrations in feed ingredients are generally more desirable, so that high P and Ca phosphate sources will add
flexibility to the formulation of high performance animal
ration. Considering the poultry requirements for minerals
according to the NRC (1994), Cu, Mg, Mn, Se, and Zn
levels present in feed phosphates are inconsequential.
However, supplementing diets with defluorinated phosphate can make a meaningful contribution to the requirement of the animal for Na (Miles and Henry, 1997). Considering the defluorinated phosphate sample evaluated
in this study (FP-4), the relatively high Na level observed
(6.03%) may require adjustment on salt addition to the
diet under certain practical feed formulation conditions.
A comparison of the maximum tolerable level of several
mineral elements for domestic animals, as suggested by
the NRC (1980), with values observed in this study indicates no toxicity risks for the use of any of the feed phosphates studied in normal diets for poultry. According to
the findings of Sullivan et al. (1994), F, Cd, and V are
potentially toxic elements of greatest concern in feed
phosphates. A comparison of the average results for three
dicalcium phosphate samples obtained in this study with
the levels reported for 19 mineral elements by Sullivan
et al. (1994) in 13 di-monocalcium phosphates shows comparable values for F (0.14 vs 0.17%), Cd (3 vs 8 ppm), and
V (84 vs 96 ppm), indicating that no serious problems
should be expected when using any of the dicalcium
phosphates studied in animal feeds. Higher average values were obtained for Fe (10,631 vs 6,468 ppm) and Pb
(21 vs 4 ppm), and comparable levels were obtained for
Ca, P, Mg, Na, K, Cu, Mn, Se, Zn, Al, As, Cr, Hg, and
Ni. The paper published by Sullivan et al. (1994) also
includes analytical results for 14 defluorinated phosphates. Values obtained in this study for the defluorinated
sample were comparable with those reported by Sullivan
et al. (1994) for F (0.15 vs 0.14%), Cd (non detectable), and
V (129 vs 94 ppm), indicating that no serious problems
should be expected when using the defluorinated product
studied in animal feeds. Higher average values were obtained for Fe (10,282 vs 6,520 ppm) and Pb (26 ppm vs
nondetectable), lower value was observed for Cr (51 vs
110 ppm), and comparable levels were obtained for Ca,
P, Mg, Na, K, Cu, Mn, Se, Zn, Al, As, Hg, and Ni. Mineral
element analyses performed in this study that were not
included in the paper published by Sullivan et al. (1994)
are Co, Mo, B, Ba, Bi, W, S, Sb, Sn, Ti, Th, and U. As
far as the reference standard, calcium phosphate dibasic
dihydrate, is concerned, all of the analytical mineral levels
obtained in this study were comparable or lower than
the values reported for the same reference standard by
Sullivan et al. (1994).
Previous publication from our laboratory (Lima et al.,
1995) reported analytical values for 26 mineral elements
in one pure dicalcium phosphate and in seven commercial
dicalcium phosphates. A comparison of the average results for three dicalcium phosphate samples obtained in
this study with the levels reported by Lima et al. (1995)
shows comparable values for F (0.14 vs 0.13%), Cd (3 vs
6 ppm), and V (84 vs 134 ppm). Higher average values
were obtained for Fe (10,631 vs 7,515 ppm) and U (135
vs 51 ppm), lower value was observed for Mg (0.17 vs
1.13%), and comparable levels were obtained for Ca, P,
Co, Cu, Mn, Mo, Se, Zn, Al, As, B, Ba, Bi, Cr, Hg, Ni, Pb,
Sb, W, and Th. Mineral element analyses performed in
this study that were not included in the paper published
by Lima et al. (1995) are Na, K, S, Sn, and Ti. As far as
the reference standard, calcium phosphate dibasic dihydrate, is concerned, all of the analytical mineral levels
obtained in this study were comparable with the values
COMPOSITION OF FEED- AND AGRICULTURAL-GRADE PHOSPHATES
reported for the same reference standard by Lima et al.
(1995).
Nutritionally essential element concentrations in common macromineral supplements have been reported by
the NRC (1994). Levels of Fe, Cu, Mg, Mn, and Zn generally compare very well with values obtained in this study.
In contrast to the safety demonstrated for FP in supplementing animal diets, this was not found to be true for
the AP tested. The AP may represent considerable risk
of toxicity for use in animal diets, considering their levels
of F, Fe, Mg, S, Ba, Ti, and Th. According to a recommendation of the AAFCO (1973), phosphate sources for use
in animal feeds should not contain more than 1 part F to
100 parts P, and, according to the NRC (1980), the maximum tolerable levels for poultry in phosphate sources
used at a dietary level of 2% are 1,000 ppm for Ba, 50,000
ppm for Fe, and 15% for Mg. The maximum tolerable
levels for S, Ti, and Th have not yet been established
(NRC, 1980) for any animal species. Single superphosphate (AP-1) was found to contain 1,210% excess F,
1,750% excess Ba, and (compared with the average content in FP) 8,550% excess S (122,054 vs 1,428 ppm). The
radioactive Th level found in the AP-1 source was particularly high as compared with the FP, exceeding by 1,300%
the average Th content detected for FP (139 vs 10 ppm).
Triple superphosphate (AP-2) level of F exceeded the
tolerance limit by 240%, and S content was in excess of
1,270% as compared with the average S content in FP
(18,090 vs 1,428 ppm). Monoammonium phosphate F level
was found to exceed by 70% the tolerance limit, and an
excess of 1,110% S was observed as compared with FP
(15,825 vs 1,428 ppm).
Concerning the thermomagnesium phosphate (AP-4),
a product identified by X-ray diffraction as amorphous
with no definite crystalline structure, there is no information in the literature on the consequences of feeding animals such products in which mineral elements are not
present in an organized, well-defined molecular fashion.
Mineral analysis revealed that AP-4 exceeded the tolerance levels by 1,200% for F, 3,120% for Ba, and 34% for
Fe. Magnesium level was particularly high (9.89%), exceeding by 1,500% the average Mg content for FP (0,62%).
The amorphous product was also found to contain an
excess of 650% Ti compared with an FP average Ti content
(2,010 vs 269 ppm). The radioactive Th level found in the
AP-4 source was particularly high compared with the FP,
exceeding by 2,300% the average Th content detected for
FP (237 vs 10 ppm).
Scarce information can be found in the literature concerning micromineral composition of AP. Ammerman et
al. (1977) reported that a level of 1,100 ppm V was found
in a sample of monoammonium phosphate originating
from a localized area.
Analytical values obtained for the four FP sources studied were all in compliance with the manufacturers’ levels
of guarantee, and no mineral element was found at levels
that would represent a toxicological problem considering
the use of the products in normal animal diets. X-ray
diffraction assays were found to be a useful analytical
1727
tool for application, combined with chemical and physical
analysis, in the evaluation of phosphates for use in animal
feeds. Considering that no purification procedures are
applied in industrial production of AP, all impurities
present in the precursory rocks remain in the product, and
levels of contaminating mineral elements will be variable,
according to the origin of the rock phosphate. Analytical
evaluations of the four AP sources studied revealed that
these products can be toxic to animals.
ACKNOWLEDGMENTS
The authors are grateful to IMC-Agrico, Mundelein, IL,
60060; Multimix Ltda., Campinas, SP, 13082-380, Brazil;
M. Cassab Ltda., São Paulo, SP, 04728-000, Brazil; and
Socil S.A., Descalvado, SP, 13690-000, Brazil for laboratory
analysis. The assistance of Mitsui S.A., São Paulo, SP,
01311-940, Brazil, and Fosfértil, Uberaba, MG, 38102-970
in obtaining some phosphate source samples is acknowledged. This material is based on work supported in part
by the Fundação de Amparo à Pesquisa do Estado de São
Paulo, FAPESP, under project number 1995/2005-0 and
1995/2006-6.
REFERENCES
Ammerman, C. B., S. M. Miller, K. R. Fick, and S. L. Hansard,
1977. Contaminating elements in mineral supplements and
their potential toxicity: A review. J. Anim. Sci. 44:485–508.
Association of American Feed Control Officials, 1973. Page 62 in:
Official Publication. Association of American Feed Control
Officials, Charleston, WV.
Association of Florida Phosphate Chemists, 1991. Page 131 in:
Methods Used and Adopted. 7th ed. Association of Florida
Phosphate Chemists, Bartow, FL.
Association of Official Analytical Chemists, 1984. Methods of
Analysis, 14th ed. Association of Official Analytical Chemists, Washington, DC.
Burnell, T. W., G. L. Cromwell, and T. S. Stahly, 1990. Effects
of particle size on the biological availability of calcium and
P in defluorinated phosphate for chicks. Poultry Sci.
69:1110–1117.
Cakir, A., T. W., Sullivan, and F. B. Mather, 1978. Alleviation
of fluorine toxicity in starting turkeys and chicks with aluminum. Poultry Sci. 57:498.
Cardoso, J.L.A., 1991. Produção, processamento e perspectivas
do fosfato na alimentação animal. Pages 35–52 in: Anais do
Mini Simpósio do Colégio Brasileiro de Nutrição Animal.
Colégio Brasileiro de Nutrição Animal, Campinas, SP, Brazil.
Day, E. J., J. McNaughton, and B. C. Dilworth, 1973. Chemical
versus chick bioassay for P availability of feed grade sources.
Poultry Sci. 52:393–395.
Fernandes, J.I.M., F. R. Lima, C. X. Mendonça, Jr., I. Mabe, R.
Albuquerque, and P. M. Leal, 1999. Relative bioavailability
of phosphorus in feed- and agricultural-grade phosphates
for poultry. Poultry Sci. 78:1729–1736.
Gillis, M. B., H. M. Edwards, Jr., and R. J. Young, 1962. Studies
on the availability of calcium orthophosphates for chickens
and turkeys. J. Nutr. 78:155–162.
Gillis, M. B., L. C. Norris, and F. G. Heuser, 1948. The utilization
by the chick of P from different sources. J. Nutr. 35:195–207.
Griffith, M., and R. Schexnailder, 1970. The relation of dietary
particle size to P availability in purified diets. Poultry Sci.
49:1271–1274.
Lima, F. R., C. X. Mendonça, Jr., J. C. Alvarez, G. Ratti, S.L.R.
Lenharo, H. Kahn, and J.M.F. Garzillo, 1995. Chemical and
1728
LIMA ET AL.
physical evaluations of commercial dicalcium phosphates as
sources of P in animal nutrition. Poultry Sci. 74:1659–1670.
McDowell, L. R., 1992. Minerals in Animal and Human Nutrition. 1st ed. Academic Press, Inc., New York, NY.
Miles, R. D., and P. R. Henry, 1997. Defluorinated phosphates
may provide advantages. Feedstuffs June 1:12–15.
National Research Council, 1980. Mineral Tolerance of Domestic
Animals. National Academy Press, Washington, DC.
National Research Council, 1994. Nutrient Requirements of
Poultry. 9th ed. National Academy Press, Washington, DC.
Potter, L. M., 1988. Bioavailability of P from various phosphates
based on body weight and toe ash measurements. Poultry
Sci. 67:96–102.
View publication stats
Sheve, R. N., and J. A. Brink, Jr., 1977. P Industries. 7th ed. R.
N. Sheve and J. A. Brink, Jr., ed. Chemical Process Industries,
Tokyo, Japan.
Sooncharernying, S., and H. M. Edwards, Jr., 1990. Aluminium
and iron in poultry rations. In: Georgia Nutrition Conference.
University of Georgia, Athens, GA.
Sullivan, T. W., J. H. Douglas, and N. J. Gonzales, 1994. Levels
of various elements of concern in feed phosphates of domestic and foreign origin. Poultry Sci. 73:520–528.
Sullivan, T. W., J. H. Douglas, N. J. Gonzales, and P. L. Bond,
Jr., 1992. Correlation of biological value of feed phosphates
with their solubility in water, dilute hydrogen chloride, dilute
citric acid, and neutral ammonium citrate. Poultry Sci.
71:2065–2074.