Manufacturing of phosphoric acid from natural phosphate by sulfuric acid attack inevitably produces impurities such as sulphates, fluorine, organic matter and heavy metals. Sulphate ions, in particular, have about 2% weight, and affect undoubtedly the efficiency of 29% P₂O₅ phosphoric acid concentration unit, due to the solid formation which causes premature clogging of the phosphoric acid concentration unit. Consequently, there is a 20% or even 30% if one takes into account the washing downtime and the degradation of the heat exchangers, thus increasing the operating cost. The physical-chemical characterizations allowed its identification. In order to minimize the formation of the solid during the phosphoric acid concentration step, we propose to desulphate 29% P₂O₅ phosphoric acid by barium carbonate using a simple experimental protocol that can be inserted into the industrial process. Design of experiments methodology is used for modeling the process and to determine the optimum conditions depending on the quality of the acid to be pretreated.

Characterization of sediment were carried out by X-ray diffraction (XRD) on the powder (Bruker advance D8 eco diffractometer with CuKα radiation λ = 1.5418 Å). Sulphate was determined by titration with barium chloride and sulfonazo-3 (C₂₂H₁₂N₄Na₄O₁₄S₄) as an indicator. The determination of phosphoric acid impurities was carried out by inductively coupled plasma-emission spectroscopy ICP-OES (Jobin Yvon ULTIMA 2).

Phosphoric acid was produced from Moroccan fluoroapatite by wet process according to the following equation [1] [2] :

Ca 1 0 ( PO 4 ) 6 F 2 + 10 H 2 SO 4 + 10 x H 2 O → 6H 3 PO 4 + 10 CaSO 4 ⋅ x H 2 O ↓ +   2HF ↑

Dependent on the process conditions, such as the temperature, either calcium sulphate hemihydrates (x = 1/2, HH) or dehydrate (x = 2, DH or phospho-gypsum) is formed. In the dihydrate process (DH), the temperature of the reactor is maintained less than 80˚C. The obtained phosphoric acid contains 26% to 29% P₂O₅ with the main impurities shown in Table 1.

The profile of the concentrated acid (exiting the concentration step), obtained by water evaporation under vacuum, is shown in Table 2.

The deposited solid during the concentration of 29% P₂O₅ phosphoric acid is identified using ICP-OES, X-ray diffraction and by monitoring the variation of major impurity mass as a function of the %P₂O₅ evolution during the phosphoric acid concentration. Table 3 and Figure 1 show that the solid formed is mainly gypsum [3] , sodium fluosilicate [4] and other minor constituents.

Table 4 shows the variation of volume, density, P₂O₅, solid and major impurities during the concentration of phosphoric acid.

Figure 2 shows the profile variation of phosphoric acid during the concentration step. The masses of SO 4 2 − , CaO, F⁻, SiO₂, K₂O and Na₂O are decreased, while the masses of MgO, Al₂O₃ and Fe₂O₃ remain virtually constant during the concentration. This explains the formation of gypsum and (Na, K)₂SiF₆. From 47% of P₂O₅, dissolution of the solid and especially the sodium or potassium fluosilicate due to the evaporation of fluorine and silicon in the form of HF and SiF₄ [5] is observed, which leads to an increase in the solubility of (Na, K)₂SiF₆ (Figure 2(a)) and consequently the decrease of its content in the formed solid.

Solid deposit accumulation in the heat exchangers reduces the phosphoric acid cross-section and causes additional thermal resistance which subsequently increases the steam specific consumption and the pressure inside the heat-exchangers (Figure 3) [6] . As a result, significant decrease in productivity can be estimated to 30% less per production cycle (7 days).

In the light of these results, it is recommended to avoid the formation of this deposit by desulphation of phosphoric acid using barium carbonate before the concentration step.

The protocol of 29% P₂O₅ phosphoric acid desulphation is judiciously chosen to be easily inserted into the phosphoric acid manufacturing process (Figure 4). Volume of phosphoric acid containing an initial quantity of H₂SO₄ (X₁ = quantity of initial H₂SO₄ per volume of acid) is heated to the desired temperature (X₃) in a stirred reactor and then a quantity (X₂) of BaCO₃ is introduced. After a reaction time (X₄), the mixture is decanted to obtain 29% P₂O₅ acid intended to be concentrated with a low level of sulphate to prevent solid formation at the heat exchangers.

In order to determine the optimum conditions for the desulphation process, we proceed by modeling the amount of residual H₂SO₄ as a function of the inlet phosphoric acid quality and the operating conditions (q, T, t).

DOE methodology with two-levels full factorial design strategy, which assumes that the mean response can be adjusted correctly using a linear model, reduces the number of trials to determine with great precision the effects of the factors and their interactions on the response and avoids the alias phenomenon [7] . Four factors were studied to determine their effects on the desulphation process, namely: X₁: initial H₂SO₄ rate in g/l (reflecting the quality of phosphoric acid), X₂ the amount of barium carbonate in g/l, X₃ la Temperature in ˚C and X₄ the time in min. The response to be minimized is the residual rate of H₂SO₄ (Y in g/l). The amount of the phosphoric acid to be treated is fixed at 100 ml. The experimental field of variation of these factors (Table 5) takes into account the range of qualities that can be obtained for phosphoric acid and the operating conditions available on the industrial site.

For four factors and two levels, 16 trials (2⁴) were performed under the conditions established by the methodology of the two-levels full factorial design. The results obtained are summarized in Table 6.

Obs.: Observed.

The codified mathematical model is written in the form (Equation (1))

Y = b 0 + b 1 X 1 + b 2 X 2 + b 3 X 3 + b 4 X 4 + b 12 X 1 X 2 + b 13 X 1 X 3 + b 14 X 1 X 4     + b 23 X 2 X 3 + b 24 X 2 X 4 + b 34 X 3 X 4 + b 123 X 1 X 2 X 3 + b 124 X 1 X 2 X 4     + b 134 X 1 X 3 X 4 + b 234 X 2 X 3 X 4 + b 1234 X 1 X 2 X 3 X 4 + ε (1)

where Y is the amount of residual sulphate (response to be fitted and to be minimized), X₁, X₂, X₃ and X₄ (factors in codified units), b₀ is the global mean, b_i are the effects of the main and interaction factors and ε is the random error associated with the response. The results of Table 6 are treated by the JMP-SAS software to evaluate the effects of different factors with and without interaction on the residual rate of H₂SO₄ (Y in g/l) to analyze and validate the mathematical model obtained. Statistical analysis by STUDENT test for 5% risk made it possible to identify the factors having a significant influence on the observed response [8] . Table 7 present the effects, t ratio and p value of each factor.

Factors that have statistically significant effects on desulphation are: X₁, X₂ and the interaction between them X₁ × X₂ (p value < 0.05).

Taking into account the values of t ratios (Table 7) and the factor effects graph (Figure 5), it is found that the initial quality of the acid (X₁) and the amount of the barium carbonate (X₂) have a significant and antagonistic effect on the residual content of the sulphates (Y) while the temperature (X₃) and the time (X₄) have no effect on the desulphation process due to the high affinity between barium ions and sulphate ion.

The mathematical model is therefore written in the form (Equation (2)):

Y = 9.0074 + 5.0838 ∗ X 1 − 5.6044 ∗ X 2 − 1.6844 ∗ X 1 ∗ X 2 (2)

ANOVA (Analysis of variance) shown in Table 8 confirms the statistical validity of the mathematical model with a determination factor R² of 99.96%. The graphical fit of the model (Figure 6) shows that the estimated responses (Y_est) by the experimental data with a correlation factor R² = 99.9%.

Additional tests (Table 9) were carried out in the experimental field with experimental conditions not used in the experimental matrix (Table 6). The residual between the observed (Y_obs) and the estimated (Y_est) results is not significant. The desulphation protocol can therefore be represented by the mathematical model (Equation (2)) in order to determine the optimum experimental conditions for efficient reduction of sulphates, depending on the quality of the phosphoric acid to be treated (X₁).

For economic reasons, 29% P₂O₅ phosphoric acid can be pre-treated at its temperature upon its arrival in the desulphation reactor (74˚C) and for a time which does not affect the phosphoric acid manufacturing process (60 min). Depending on the phosphoric acid quality to be pre-treated, the amount of barium carbonate is determined using Equation (3) where Y Tolerated is the maximum residual amount of H₂SO₄ having no effect on the formation of deposit solid during the phosphoric acid concentration step.

X 2 = 9.0074 − Y tolerated + 5.0838 ∗ X 1 5.6044 + 1.6844 ∗ X 1 (3)

According to Table 2, phosphoric acid 54% P₂O₅, obtained from 29% P₂O₅ acid non-desulphated, is saturated with sulphate and calcium (2.95% SO 4 2 − and 0.68% CaO) which corresponds to a concentration of 49.59 g/l of sulphate and 8.16 g/l of calcium given a solubility product of 404.89 (g/l)². In the case of desulphation the mass of the calcium initially present in the acid 29% is preserved in the acid 54% with a concentration of 17.25 g/l. Therefore, the sulphate concentration in the 54% acid at saturation should be 23.46 g/l. The maximum residual mass of Tolerated Sulphate Y_Tolerated is 8.97 g/l.

For example, for an initial acid with quality of X₁ = 22 g H₂SO₄/l and Y_Tolerated = 8.97 g/l, the amount of BaCO₃ to be used for desulphation is 25.25 g/l. After concentration of the desulphated acid, the total amount of the formed solid is

Obs.: Observed; Est.: Estimated.

0.37% instead of 2.24% (Table 2), without taking into account the solid deposited at the heat exchanger for the second value. This represents a reduction rate of formed solid around 83.5% at least.

For industrial scale, desulphation by barium carbonate can be carried out directly at the level of the aging tank without addition of specific equipment to the phosphoric acid manufacturing unit (Figure 7).

Deposition of gypsum during the 29% P₂O₅ phosphoric acid concentration stage is a real technical-economic problem for producing phosphoric acid 54% P₂O₅. Physico-chemical characterization by ICP-OES and X-ray diffraction showed that this deposit is mainly composed by gypsum and low amount of fluosilicates. Desulphation of phosphoric acid 29% P₂O₅ by barium carbonate will reduce the

formation of gypsum in the concentrator. The proposed desulphation protocol can be easily inserted into the industrial process; it is limited to a simple attack of phosphoric acid 29% P₂O₅ by BaCO₃ under stirring. Regarding to DOE methodology based on the strategy of full factorial design, taking into account the initial amount of H₂SO₄ in phosphoric acid 29% P₂O₅, the amount of barium carbonate, temperature and reaction time, it has been shown that only the quality of the acid and the amount of BaCO₃ have a significant effect on desulphation. The model obtained is validated with R² = 99.96% and can be used to determine the effective amount of BaCO₃ as a function of the phosphoric acid quality and the maximum residual mass of H₂SO₄ to avoid the formation of gypsum. The amount of the solid formed can therefore be reduced to 83.5%. This improves the production cycle and reduces maintenance and operating costs.

“Research carried out within the framework of the University Center for Research Applied Chemistry and Sustainable Development CUR CA2D of Chouaïb Doukkali University”.

The authors declare no conflicts of interest regarding the publication of this paper.

Aboulhassane, A., El Idrissi, A.N., Bounou, Y. and Zakaria, D. (2019) 29% P₂O₅ Phosphoric Acid Desulphation: Improving the Performance of the Unit of Concentration. American Journal of Analytical Chemistry, 10, 65-75. https://doi.org/10.4236/ajac.2019.103006