Introduction

Ammonium nitrate (AN) is one of the most common commercially available nitrogen fertilizers, the content of nitrogen in which amounts up to 35% by mass. The main agrochemical advantage of AN compared to other simple nitrogen fertilizers is to present nitrogen both in ammonia and nitrate forms. Herewith, the high content of this component enables to mix it with other types of fertilizers and obtain complex fertilizer with the high content of basic nutrients—nitrogen, phosphorus and potassium. The main disadvantages of such types of fertilizers are their high hygroscopicity, caking [1] and the increased requirements for fire and explosion safety [2]. All the above-mentioned factors, and in particular the last, are the main disadvantages limiting the production of complex AN-based fertilizers.

Cases of explosion of AN and complex AN-based fertilizers are well known: in 1921 in the warehouse in Oppau (Germany), in 1947 in the warehouse in the bay in Texas City (USA), in 2001 in the warehouse in Toulouse (France), in 2013 in the warehouse in West (USA). The largest explosion of technological installations was recorded in 1952 in Nagoya (Japan), in 1978—in Chirchik (Uzbekistan) in 1981—in Cherepovets (Russia), in 1994—in Port Neil (USA), in 2009—in Kirovo-Chepetsk (Russia).

Ammonium phosphates NH4H2PO4 and (NH4)2HPO4, ammonium sulfate and potassium chloride are also used in the production of complex AN-based NPK fertilizers. Herewith, the following reactions take place:

$${\text{NH}}_{ 4} {\text{H}}_{ 2} {\text{PO}}_{ 4} + {\text{ KCl}} \rightleftarrows {\text{KH}}_{ 2} {\text{PO}}_{ 4} + {\text{ NH}}_{ 4} {\text{Cl}},$$
(1)
$${\text{NH}}_{ 4} {\text{NO}}_{ 3} + {\text{ KCl}} \rightleftarrows {\text{KNO}}_{ 3} + {\text{ NH}}_{ 4} {\text{Cl}},$$
(2)
$$({\text{NH}}_{ 4} )_{ 2} {\text{SO}}_{ 4} + {\text{ 2KCl}} \rightleftarrows {\text{K}}_{ 2} {\text{SO}}_{ 4} + {\text{ 2NH}}_{ 4} {\text{Cl}}.$$
(3)

KH2PO4, KNO3 and K2SO4 in combination with unreacted NH4H2PO4, NH4NO3 and (NH4)2SO4 (accordingly) form solid solutions—compounds of isomorphic-substituted type.

The composition of the solid solutions is determined by the extent of the conversion of the reactions (1–3). (NH4)2HPO4 does not react with KCl. Moreover, AN can form various double salts: NH4NO3·2KNO3, (NH4)2SO4·2NH4NO3, (NH4)2SO4·3NH4NO3. Formation of NH4NO3·2KNO3 depends on the extent of the conversion of the reaction (2) [3]. The double salts (NH4)2SO4·2NH4NO3 and (NH4)2SO4·3NH4NO3 in the presence of KCl can decompose with the formation of solid solutions [4].

Thus, complex AN-based fertilizers are complex salt systems, whose composition is defined by the ratio of initial components.

The presence of all the above-mentioned compounds can variously affect the decomposition of complex AN-based fertilizers and their propensity for detonation. The presence of NH4H2PO4, (NH4)2HPO4 and (NH4)2SO4 reduces the rate of AN decomposition [5, 6], and chloride-anions Cl−, on the contrary, act as catalysts for AN decomposition [7–9].

Despite this, increasing demands of the agrochemical sector leads to the necessity to develop new grades of the fertilizers, the production of which is possible only when using concentrated nitrogen fertilizers, especially ammonium nitrate and urea. However, considerable difficulties emerge in case of urea used, which consist in high hygroscopicity and caking, reduction of the amide nitrogen proportion in the product due to decomposition of urea at relatively low temperatures during granulation and drying, and complexity of the technological process because of heavy clogging of equipment [10, 11].

One of the ways to improve the quality of complex AN-based fertilizers and reduce the risk of explosion is to increase the ammoniation degree of wet-process phosphoric acid, which reduces the AN portion in the product. Such way can improve the properties of the final product (decrease hygroscopicity and caking), increase its thermal stability, decrease the amount of different compounds in exhaust gases (nitrous gases, chlorine and fluorine compounds) during thermal decomposition, increase fire and explosion safety, and also decrease viscosity of ammonium phosphate–nitrate slurries produced during the production of fertilizer that can decrease energy cost for their transportation. However, information on influence of the degree of phosphoric acid ammoniation on the above-mentioned properties of complex AN-based fertilizers and their intermediates is absent in the literature.

Thus, the purpose of this work is to study the properties of complex AN-based fertilizers and intermediates in their production depending on the degree of phosphoric acid ammoniation.

Experimental section

Preparation of the samples

To produce complex fertilizers, concentrated hemihydrate phosphoric acid, nitric acid, ammonium sulfate and potassium chloride (mineral concentrate “Silvin”) were used. Wet-process phosphoric acid was obtained from the Khibiny apatite concentrate (the Cola Peninsula, Russia) of composition: P2O5—51.72, CaO—0.67, MgO—0.23, F—1.33, SO3—4.53, Fe2O3—0.55, Al2O3—0.90, SiO2—0.43% by mass by sulfuric acid attack. Phosphoric and nitric acids were mixed in a certain ratio and ammoniated in a reactor equipped with the agitator device, the reflux condenser and the water jacket, which allowed ammoniation to be carried out under near-isothermal conditions at 70 ± 2 °C.

The degree of ammoniation of phosphoric acid NH3:H3PO4 (M) was determined by pH value of the 1% by mass aqueous solution of the slurry obtained and using the reference source [12]. Ammonium sulfate and potassium chloride were introduced into the slurry in an amount necessary to obtain the desired grade, mixed thoroughly and dried at 65 °C. Then, the charge mixture was crushed and put in a pan granulator with diameter of 300 mm and length of 150 mm. Granules of 2–4 mm were finally dried at 65 °C to reach the required humidity. The product obtained was analyzed for content of basic elements.

X-ray diffraction analysis

X-ray diffraction analysis of the investigated samples was performed when used powder diffractometer «STADI-MP» (STOE, Germany) with curved Ge (111) monochromator and radiation of CuKα (λ = 1.54056 Å). The data acquisition was carried out in stepwise overlapping of scanning area mode by means of position-sensitive linear detector, the capture angle of which amounted 5° over 2θ with channel width of 0.02°. The reliability and accuracy of compounds in X-ray patterns obtained were established by means of database of 2013 International Centre for Diffraction Data.

Derivatographic analysis

Derivatographic analysis was carried out when used Paulik–Erdei derivatograph (MOM, Hungary) of Q-1500 series while heating in the air at atmospheric pressure in open quartz crucibles with heating rate of 2.5°/min. Al2O3 pre-calcinated at 1000 °C was used as a reference. The sample weight amounted 0.2 g. The thermocouple was Pt/Pt–Pd. The interpretations of the dependencies obtained were carried out in compliance with the literature data [13–16].

Hygroscopicity

Hygroscopicity (K) of the samples obtained was determined by means of climatic chamber BINDER KBF 115 (BINDER, Germany) with internal circulation. The value of K was determined by means of conditioning of granule samples with the diameter of 3–4 mm with the mass of 3.500 ± 0.006 g in the chamber at 25 °C and the relative air humidity (φ) of 80% for 1 h. Granules were uniformly distributed in a cup with the diameter of 50 mm and height of 10 mm in a single layer. The value of K was determined as the amount of water absorbed with a sample of unit mass for 1 h.

Caking

Determination of caking (σ) of samples obtained was conducted by means of climatic chamber with internal circulation BINDER KBF 115 (BINDER, Germany) at temperature of 45 °C, φ = 40%, and special presses equipped with calibrated spring. The spring load for each sample was 340 kPa. The samples detention time in the chamber was 6 h. Caking was determined as averaged maximum force required for breaking of formed cylindrical pellet divided by its cross-section area (pellet size: diameter 33 mm, height 40 mm).

Static strength

Determining the static strength, P was conducted by means of IPG-1M (Urals Scientific Research Institute of Chemistry with Experiment Plant, Russia) according to the formula:

$$P\text{ = }\frac{{\sum\nolimits_{{i\text{ = 1}}}^{N} {F_{i} } }}{{\frac{{\pi d_{m}^{\text{2}} }}{4}N}},$$
(4)

where F is the mean force required for breaking of one granule, d m is the mean diameter of one granule equal to 3.5 mm, and N is the number of measured granules.

Microcalorimetry

The microcalorimetric studies of the thermal decomposition kinetics were conducted by measuring the heat release rate in the samples under study with differential automatic calorimeter DAC-1-2 [17]. Tests were carried out in the vacuum-sealed glass ampoules with inner volume of about 2 cm3, a mass of each tested mixture sample was 1 g. The free inner volume after putting each sample and sealing an ampoule was in the range 0.7–1.2 cm3 per 1 g of the mixture tested. These ampoules were entirely put into the calorimeter and had no cold surfaces, and reaction products could not leave the boundaries of the reaction space.

Gravimetric study of the thermal decomposition

Studies of mass loss in the thermal decomposition were conducted by maintaining granulated samples with mass of 20.00 ± 0.05 g in the electric oven without forced convection at the given temperature for a given period of time. The content of ammonium and nitrate nitrogen, chlorine, fluorine and sulfur was determined in products of the thermal decomposition. The fraction of these elements that have been released into the gas phase was calculated according to the formula:

$$X_{A} \text{ = }\frac{{\upomega_{\text{0}} \text{(}A\text{)}m_{\text{0}} -\upomega_{t} \text{(}A\text{)}m_{t} }}{{m_{\text{0}} }},$$
(5)

where X A is the fraction of A (A = Namm, Nnitr, Cl, F) released into the gas phase per the unit mass of the initial sample; ω0(A) is the mass fraction of A in the initial sample; m 0 is the mass of the initial sample; ω t (A) is the mass fraction of A in the sample after the decomposition for time t; m t is the mass of the sample after the decomposition for a time t.

Dynamic viscosity

The dynamic viscosity of slurries was determined by means of rotation viscometer HAAKE VT 74 Plus (Thermo Scientific, USA). In order to do that, the slurry obtained was placed in the cylindrical vessel provided with a thermostatic jacket and connected to circulation bath in which a polysilicon oil was circulated. After viscosity measurements, the slurry humidity was measured.

Processing experimental data obtained and the determination of confidence intervals for 95% confidence probability were conducted with the mathematical statistics methods by means of software application of origin.

Results and discussion

The composition of the fertilizer samples and X-ray diffraction analysis

Table 1 shows the results of analyses of fertilizer samples.

Table 1 The composition of the fertilizer samples (%mass.)
Full size table

Figure 1 shows X-ray patterns for samples 1 and 2 (grade 26:13:0), 3 and 4 (grade 22:11:11), 5 and 6 (grade 16:16:16).

Fig. 1
figure 1

X-ray patterns of the fertilizer samples: a—1, b—2, c—3, d—4, e—5, f—6; 1 (NH4)2HPO4, 2 NH4H2PO4, 3 NH4NO3, 4 (NH4)2SO4, 5 2NH4NO3·(NH4)2SO4, 6 3NH4NO3·(NH4)2SO4, 7 (NH4,K)H2PO4, 8 (NH4,K)NO3, 9 (NH4,K)2SO4, 10 KCl, 11 NH4Cl, 2θ Bragg angle (degree)

Full size image

X-ray patterns for the samples of grades 16:16:16 and 22:11:11 demonstrate the presence of solid solutions (NH4,K)NO3, (NH4,K)H2PO4 and (NH4,K)2SO4, as well as of NH4Cl and KCl. For samples 3 and 5, the presence of (NH4)2HPO4 was established.

Comparing X-ray patterns for the samples 3 and 4 of grade 16:16:16 and 5 and 6 of grade 22:11:11 shows that the intensity of the main diffraction peak of NH4Cl decreases with a higher degree of ammoniation. This is associated with a reduction of the original content of AN in the composition of samples that results in reducing the amount of NH4Cl produced in reaction (2).

X-ray patterns for sample 1 of 26:13:0 grade demonstrate the presence of (NH4)2HPO4, NH4H2PO4, (NH4)2SO4, 2NH4NO3·(NH4)2SO4 and 3NH4NO3·(NH4)2SO4 and for the sample 2 the presence of NH4NO3, NH4H2PO4, 2NH4NO3·(NH4)2SO4 and 3NH4NO3·(NH4)2SO4.

Comparing X-ray patterns for samples 1 and 2 of grade 26:13:0 demonstrates that the composition of sample 2 has the unbound AN, which could not fully converted to 2NH4NO3·(NH4)2SO4 and 3NH4NO3·(NH4)2SO4 due to a high content of AN and a low content of (NH4)2SO4 in the composition of the fertilizer. This may lead to significant deterioration of the properties of sample 2 compared with sample 1.

All these compounds are typical for complex AN-based fertilizers that is noted in [1, 3, 4, 12].

Derivatographic analysis

Figures 2, 3, 4 and 5 show the results of the derivatographic analysis for samples 1, 2, 3 and 4. Analysis of curves of the differential thermal analysis (DTA) and of the differential thermogravimetric analysis (DTG) confirms the data of X-ray diffraction analysis.

Fig. 2
figure 2

Curves of the differential thermal analysis (DTA) and of the differential thermogravimetric analysis (DTG) of sample 1: t time (min)

Full size image
Fig. 3
figure 3

Curves of the differential thermal analysis (DTA) and of the differential thermogravimetric analysis (DTG) of sample 2: t time (min)

Full size image
Fig. 4
figure 4

Curves of the differential thermal analysis (DTA) and of the differential thermogravimetric analysis (DTG) of sample 3: t time (min)

Full size image
Fig. 5
figure 5

Curves of the differential thermal analysis (DTA) and of the differential thermogravimetric analysis (DTG) of sample 4: t time (min)

Full size image

Curves of the differential thermal analysis (DTA) for 22:11:11 samples are characterized by the following peaks: the reverse phase transition of (NH4,K)NO3 in NH4NO3·2KNO3 (113.1 and 129.9 °C, respectively) [15]; melting (132.8 and 145.3 °C) [15]; the exothermal decomposition of the product including the decomposition of NH4NO3 [15], the polycondensation of (NH4,K)H2PO4 and the decomposition of (NH4)2HPO4 for samples 3 and 4 (197.2 and 221.5 °C [13].

It can be concluded by comparing the DTG and DTA curves that sample 3 has the higher thermal stability as compared to sample 4, which may be related to a lower content of AN and a higher content of (NH4)2HPO4. It is also worth noting that there is no peak characteristic for the (NH4)2HPO4 decomposition in the DTA and DTG curves of sample 3, which would be in the range of 120–200 °C. It may be assumed that its absence is due to the interaction between (NH4)2HPO4 and HNO3, which is formed as the result of the partial dissociation of NH4NO3, according to the reaction:

$$( {\text{NH}}_{ 4} )_{ 2} {\text{HPO}}_{ 4} + {\text{ HNO}}_{ 3} \to {\text{ NH}}_{ 4} {\text{H}}_{ 2} {\text{PO}}_{ 4} + {\text{ NH}}_{ 4} {\text{NO}}_{ 3} .$$
(6)

The decomposition of (NH4)2HPO4 is apparently to occur at higher temperatures due to the course of reaction (6). In the case of sample 3, this process takes place in the intensive exothermal decomposition of the product.

At heating sample 2, the peaks are observed on the DTA curve, which is related to the following phenomena: the reverse phase transition of AN IV → III (39.9 °C) [13]; the reverse phase transition of AN III → II (85.4 °C) [13]; the reverse phase transition of AN II → I (116.3 °C) [13]; melting and partial decomposition of adducts 2NH4NO3·(NH4)2SO4 and 3NH4NO3·(NH4)2SO4 (162.9 °C) [16]; the polycondensation of NH4H2PO4 (209.8 °C) [13]; the AN decomposition (220.6 °C) [13].

There are no peaks, which are characteristic to AN in the DTA curve of sample 1. The thermal decomposition of this sample is characterized by the following processes: the decomposition of (NH4)2HPO4 (138.6 °C) [13]; the melting and partial decomposition of adducts 2NH4NO3·(NH4)2SO4 and 3NH4NO3·(NH4)2SO4 (152.2 °C) [16]; the polycondensation of NH4H2PO4 (210.8 °C) [13]; the AN decomposition (219.5 °C) [13]; the (NH4)2SO4 decomposition (244.5 °C) [14].

It can be concluded when compared the DTG and DTA curves that the presence of (NH4)2HPO4 as a part of sample 1 leads to the fact that at temperatures over 100 °C (NH4)2HPO4 decomposes to NH4H2PO4 to release NH3 into a gas phase. However, sample 2 exhibits the higher thermo-stability than sample 1 when further heated.

It should also be noted that the decomposition of samples 1 and 2 takes place endo-thermally as opposed to samples 3 and 4, whose decomposition proceeds with the release of the large amount of heat. This is due to the absence of chlorine compounds in the composition of samples 1 and 2, which are capable to accelerate the exothermal AN and complex AN-based fertilizers decomposition [7–9, 15].

Hygroscopicity, caking and static strength

Table 2 presents the results of studying hygroscopicity, caking and static strength of the fertilizer samples obtained.

Table 2 Hygroscopicity, caking and static strength of granulated fertilizer samples
Full size table

The presented data show that for the same grade of the fertilizer the increase of M reduces the hygroscopicity and caking; however, the static strength of granules decreases also. The reduction of hygroscopicity can be associated with a reduced content of AN, which is highly hygroscopic. The reduction of caking can also be associated with a reduced ammonium chloride content with increasing M, which is apparent from intensity of peaks for NH4Cl in the presented X-ray patterns [1, 18]. The reduction of static strength of granules can be the result of lower strength of phase contacts between granules with increase of M in the granulation process [19].

The maximum difference in hygroscopicity and caking is observed for 26:13:0 grade, which can be due to the presence of AN in sample 2, whereas in sample 1 AN is connected in double salts (NH4)2SO4·2NH4NO3 and (NH4)2SO4·3NH4NO3. The minimum difference in hygroscopicity and caking is observed for 27:6:6 grade, which can be explained by the high content of nitrate nitrogen in both samples and the small difference in its content between them.

It should also be noted that the highest increase of the caking was observed for 16:16:16 grade (σ 6/σ 5 = 1.76), whereas for the other grades this ratio is much lower. This is possible due to the high ratio of the content of NH4Cl in two samples of 16:16:16 grade and almost twofold increase in the content of AN in sample 6 when M simultaneously reduced. The closest value to this one is σ 2/σ 1 = 1.49 for 26:13:0 grade. The high ratio σ 2/σ 1 for 26:13:0 grade is apparently due to the fact that in sample 2 the part of NA presents in the free form, while in sample 1 NA is fully bound in double salts.

Microcalorimetry

Figures 6 and 7 show the curves of the heat release rate dependence on time in the thermal decomposition of samples 3 and 4 in the temperature range of 183.5–245.9 °C.

Fig. 6
figure 6

Dependence of the heat release rate dQ/dt (mW g−1) on time t (min) in the thermal decomposition of sample 3

Full size image
Fig. 7
figure 7

Dependence of the heat release rate dQ/dt (mW g−1) on time t (min) in the thermal decomposition of sample 4

Full size image

As indicated above, chloride-anions Cl− contained in samples under study are catalysts of the AN decomposition, and their catalytic effect increases with the increase of the content of nitric acid in the system and virtually does not occur when its content is low. The accelerating action of Cl− in the AN decomposition is related to accumulation of nitryl chloride NO2Cl, nitrosyl chloride NOCl and chlorine Cl2 in the system, being more effective oxidizers of ammonium cation NH4 + and ammonia as compared to nitric acid. The presence of NH4H2PO4, (NH4)2HPO4 and (NH4)2SO4 together with Cl− reduces Cl− catalytic effect in AN decomposition.

The study of the heat release rate for sample 4 revealed its low thermal stability. In the decomposition of sample 4 Cl− accelerating action prevails over decreasing the AN decomposition rate in response to H2PO4 −, HPO4 2− and SO4 2− anions and, therefore, the decomposition of this sample occurs with the self-acceleration.

Sample 3 has a lower content of AN as compared to sample 4, herewith in its composition a large portion of H2PO4 − is substituted with HPO4 2−. Anion of HPO4 2− is capable to a higher degree to reduce the concentration of undissociated nitric acid, and so to increase the thermal stability of sample 3. Besides, the content of (NH4)2SO4 in sample 3 is also higher than in sample 4. All this contributes to the fact that the accelerating action of Cl− is not detected, and the decomposition occurs without self-acceleration. Thus, sample 3 has significantly higher thermal stability as compared to sample 4.

Besides the study of the fertilizer samples, the heat release rate was also measured as a function of time in the thermal decomposition of nitrate–phosphate–ammonium slurries at obtaining sample 3 with M = 1.0 (sample 3a) and M = 1.4 (sample 3b) with humidity of about 8% mass in the temperature range of 243.5–277.0 °C (Figs. 8, 9).

Fig. 8
figure 8

Dependence of the heat release rate dQ/dt (mW g−1) versus time t (min) in the thermal decomposition of sample 3a

Full size image
Fig. 9
figure 9

Dependence of the heat release rate dQ/dt (mW g−1) versus time t (min) in the thermal decomposition of sample 3b

Full size image

The study of the heat release rate for these samples revealed their high thermal stability, while sample 3b was more thermally stable than sample 3a, which can be explained by the higher content of (NH4)2HPO4 in it.

Figure 10 shows the temperature dependencies of the initial heat release rates (dQ/dt) t=0 in the thermal decomposition of samples 3, 4, 3a and 3b in Arrenius coordinates. For comparison, Fig. 5 also illustrates the temperature dependence of the initial heat release rates in the AN thermal decomposition studied previously [20].

Fig. 10
figure 10

Dependence of lg[dQ/dt (mW g−1)] t=0 on 103/T (K−1) for samples 3 (1), 4 (2), 3a (3), 3b (4) and ammonium nitrate (5)

Full size image

The equations of the obtained dependence of (dQ/dt) t=0 (mW g−1) on temperature (K) are as follows:

for sample 3

$$\left( {\frac{{\text{d}Q}}{{\text{d}t}}} \right)_{{t\text{ = 0}}} \text{ = 10}^{{\text{11}{.7 \pm 0}\text{.7}}} \text{exp}\left( { - \frac{{(\text{17}{.2 \pm 0}\text{.8}) \times \text{10}^{\text{3}} }}{T}} \right),$$
(7)

for sample 4

$$\left( {\frac{{\text{d}Q}}{{\text{d}t}}} \right)_{{t\text{ = 0}}} \text{ = 10}^{{\text{18}{.1 \pm 0}\text{.7}}} \text{exp}\left( { - \frac{{(\text{22}{.3 \pm 0}\text{.9}) \times \text{10}^{\text{3}} }}{T}} \right),$$
(8)

for sample 3a

$$\left( {\frac{{\text{d}Q}}{{\text{d}t}}} \right)_{{t\text{ = 0}}} \text{ = 10}^{{\text{16}{.1 \pm 0}\text{.3}}} \text{exp}\left( { - \frac{{(\text{22}{.8 \pm 0}\text{.4}) \times \text{10}^{\text{3}} }}{T}} \right),$$
(9)

for sample 3b

$$\left( {\frac{{\text{d}Q}}{{\text{d}t}}} \right)_{{t\text{ = 0}}} \text{ = 10}^{{\text{7}{.3 \pm 0}\text{.8}}} \text{exp}\left( { - \frac{{(\text{12}{.7 \pm 0}\text{.9}) \times \text{10}^{\text{3}} }}{T}} \right).$$
(10)

The dependencies presented in Figs. 6, 7, 8, 9 and 10 show that the initial heat release rate of sample 4 is on average by 1–2 orders higher than that for sample 3. Herewith the initial heat release rate of sample 4 significantly exceeds that of AN, while for sample 3 the situation is inverse. Samples 3a and 3b have even higher thermal stability as compared to sample 3, which may be explained by lack of Cl− in their composition and the high water content.

In any real conditions of conducting the discussed reaction, the thermal explosion is only possible when the values of external parameters of the process exceed the critical ones for the thermal explosion, but calculation of the critical conditions for a real complex production process is a very time-consuming task, and the adiabatic induction period of thermal explosion τ ad is calculated simply. If the value τ ad is much greater than the real time of the production process at an appropriate temperature, then the thermal explosion will not occur, and in any real process conditions the induction period may only be greater than under adiabatic conditions. However, if the value τ ad and process real time are close enough or if τ ad is even less, it is necessary to calculate the critical conditions of the thermal explosion (the critical temperature for the actual size of the unit and the conditions of heat transfer from it). Only these calculations can give final decision on possibility of the thermal explosion in the process considered.

Calculation of the adiabatic induction period is the most simple and available method to assess the possibility of the thermal explosion for any particular composition. In the complete absence of heat removal (adiabatic conditions) and at a sufficiently high value of the process heat, the thermal explosion will always occur; besides, the degree of conversion in the reaction discussed during induction period will be very small, because all the heat is used for heating a substance. As far as there is no heat removal, the adiabatic induction period is independent of the sample mass and heat removal conditions and it is considered as a characteristic for a substance or mixture discussed. In the theory of thermal explosion because of the weak influence of the process acceleration, the exact quantitative equation for calculating the adiabatic induction period was obtained only for zero-order reaction, and the reaction rate change in the subsequent stages is assumed to have a very small action on the adiabatic induction period [21]:

$$\uptau_{{\text{ad}}} \text{ = }\frac{{c_{\text{p}} }}{{Q_{\text{0}} k_{\text{0}} }} \cdot \frac{{RT_{\text{0}}^{\text{2}} }}{E} \cdot \text{exp}\left( {\frac{{E_{\text{c}} }}{{RT_{\text{0}} }}} \right),$$
(11)

where c p is the heat capacity of the sample; Q 0 is the total process heat; k 0 and E c are the pre-exponential factors and the activation energy of the decomposition rate constant; T 0 is the absolute temperature of the decomposition; R = 8.314 J mole−1 K−1 is the universal gas constant.

When \(\left( {\frac{{\text{d}Q}}{{\text{d}t}}} \right)_{{t\text{ = 0}}} \text{ = }Q_{\text{0}} k_{\text{0}} \text{exp}\left( { - \frac{E}{{RT_{\text{0}} }}} \right)\) Eq. (11) takes the following form:

$$\uptau_{{\text{ad}}} \text{ = }\frac{{\text{RT}_{\text{0}}^{\text{2}} }}{\text{E}} \cdot \frac{{\text{c}_{\text{p}} }}{{\left( {\frac{{\text{dQ}}}{{\text{dt}}}} \right)_{{\text{t = 0}}} }}.$$
(12)

The results from paper [12] were used to determine the heat capacity of samples under study, provided that in a first approximation the heat capacities of samples 3, 4 and 3a, 3b are equal in pairs. The value of the AN heat capacity was taken according to the data in [2]. The τ ad values obtained are given in Table 3. The τ ad values for the same initial temperature may be considered as the characteristics of a relative explosion risk of a substance. The adiabatic induction periods of the thermal explosion for samples 3, 3a and 3b are greater than that for AN, and for sample 4 they are almost by an order less, which reveals the potential danger of thermal spontaneous ignition of the sample during production operations at high temperatures.

Table 3 Adiabatic induction period of the thermal explosion τad of samples 3, 4, 3a and 3b and AN depending on temperature T
Full size table

Gravimetric study of the thermal decomposition

The study of the mass loss in the thermal decomposition was carried out for samples 3 and 4 at temperatures of 170, 180, 190 and 200 °C. In addition to the study of the mass loss, the release of ammonium nitrogen, nitrate nitrogen, chlorine and fluorine to the gas phase was also evaluated. The research results are presented in Figs. 11, 12 and 13.

Fig. 11
figure 11

Dependence of thermal decomposition degree β = (m − m 0)/m 0·100 (%) versus time t (min) for samples 3 and 4 at constant temperature; sample 3: 1 170 °C, 2 180 °C, 3 190 °C, 4 200 °C; sample 4: 5 180 °C

Full size image
Fig. 12
figure 12

Release of chlorine XCl, ammonium XNamm and nitrate nitrogen XNnitr (g kg−1) into the gas phase in the thermal decomposition of samples 3 and 4 at temperature of 180 °C versus time t (min); sample 3: curve 1 Cl, 2 Nam, 3 Nnit; sample 4: curve 4 Cl, 5 Nam, 6 Nnit

Full size image
Fig. 13
figure 13

Release of fluorine into the gas phase XF (g kg−1) in the thermal decomposition of samples 3 (curve 1) and 4 (curve 2) at temperature 180 °C versus time t (min)

Full size image

The decomposition intensity for sample 4 is much higher than that for sample 3. The release of chlorine, fluorine, ammonium nitrogen, and nitrate nitrogen from sample 4 to the gas phase is also much more intensive than that from sample 3. Ammonium nitrogen in the initial decomposition stage is released from sample 3 in a greater quantity than that from sample 4. It is related to the higher content of (NH4)2HPO4, which starts to decompose in NH3 and NH4H2PO4 at low temperatures.

It is also worth mentioning that the maximum amount of fluorine released into the gas phase for both samples is almost the same. It is related to the fact that fluorine in both samples according to [22] is present in the form of compounds (NH4)2SiF6, NH4F, NH4NO3·(NH4)2SiF6, KNO3·K2SiF6, (NH4)2SiF6·NH4F, etc., the decomposition of which depends only on the process temperature. The higher fluorine release rate for sample 4 is related to the more intense exothermal decomposition of this sample.

For chlorine, the release into the gas phase depends on the content of AN, so for sample 4 a significantly greater amount of chlorine is released into the gas phase than for sample 3.

The release of chlorine, fluorine, nitrous gases and ammonium compounds into the gas phase leads to the essential complication and more expensive purification of exhaust gases from them, as well as to the more intense corrosion of equipment.

Dynamic viscosity

The study of dynamic viscosity was performed for ammonium–phosphate–nitrate slurries obtained at production of 22:11:11 grade when M = 1.7. To obtain such slurries, phosphoric and nitric acids were mixed in the ratio P2O5:HNO3 = 0.36:1 (by mass.) and ammoniated up to the specified value of M.

Figure 14 shows the dependence of dynamic viscosity of such slurry on M for different values of humidity at 110 °C. The slurry viscosity is apparent to reach the minimum value at M = 1.45 for all the values of humidity.

Fig. 14
figure 14

The dependence of dynamic viscosity η (mPa s) of ammonium–phosphate–nitrate slurry on M at temperature 110 °C and for different values of humidity: 1 5% mass., 2 6% mass., 3 7% mass., 4 8% mass., 5 10% mass

Full size image

It should be mentioned that the same behavior of viscosity was observed for phosphate ammonia slurries obtained from various types of a phosphate raw [3, 23]. The presence of minimum in the viscosity curve is probably due to the high solubility of ammonium phosphates at M = 1.4–1.5. The presence in the slurry of impurities of iron, aluminum, magnesium, fluorine, silicon, etc. leads to increasing viscosity due to the formation of poorly soluble compounds [3, 24, 25].

Figures 15 and 16 show the dependences of the slurry dynamic viscosity (for M = 1.05 and M = 1.45) on humidity at different temperatures. The figures show that the slurry viscosity increases with decreasing humidity. Herewith for the slurries at M = 1.05, a more rapid increase of viscosity with decreasing humidity is observed.

Fig. 15
figure 15

Dependence of dynamic viscosity η (mPa s) of ammonium–phosphate–nitrate slurry for M = 1.05 on humidity (%mass.) for different values of temperature: 1 100 °C, 2 105 °C, 3 110 °C, 4 115 °C, 5 120 °C

Full size image
Fig. 16
figure 16

Dependence of dynamic viscosity η (mPa s) of ammonium–phosphate–nitrate slurry for M = 1.45 on humidity (%mass.) for different values of temperature: 1 100 °C, 2 105 °C, 3 110 °C, 4 115 °C, 5 120 °C

Full size image

Temperature influence on dynamic viscosity of the slurries obeys the law of Arrhenius–Andrade [26]:

$${\eta = }A_{{\text{exp}}} \left( {\frac{{E_{\text{v}} }}{RT}} \right),$$
(13)

where A is the pre-exponential factor and E v is the activation energy for viscous flow.

Table 4 presents the equations of dynamic viscosity dependence on temperature at the different values of humidity for slurries studied. As can be seen from the equations presented, the activation energy of a viscous flow, the values of pre-exponential factor and the value of dynamic viscosity of the slurry for M = 1.45 are substantially less than for M = 1.05.

Table 4 Equations of dynamic viscosity (mPa·s) dependence for ammonium phosphate nitrate slurries for M = 1.05 and M = 1.45 for the various humidity values W (% mass.)
Full size table

Using the slurries having higher mobility and flowability during their processing in the granular product can significantly reduce the energy costs for the removal of moisture from the granules and reduce the amount of the recirculated product obtainable by a recycle method.

Conclusions

On the basis of the studies performed, it has been found that increasing the degree of phosphoric acid ammoniation with M = 1.0–1.1 to M = 1.6–1.7 influences on the properties of the complex AN-based NP and NPK fertilizers.

It has been shown by X-ray diffraction and derivatographic analysis that the composition of NPK fertilizer (16:16:16 and 22:11:11) contains (NH4,K)H2PO4, (NH4,K)2SO4, (NH4,K)NO3, KCl and NH4Cl. When M = 1.6–1.7, (NH4)2HPO4 presents also additionally. The composition of NP fertilizer (26:13:0) contains NH4H2PO4, 2NH4NO3·(NH4)2SO4 and 3NH4NO3·(NH4)2SO4. When M = 1.0–1.1, NH4NO3 presents additionally in the system, when M = 1.6–1.7, (NH4)2HPO4 and (NH4)2SO4 present additionally.

It is found that the decomposition of NPK fertilizers occurs with the strong exothermal effect and NP fertilizers decomposition occurs with the endothermal effect. The strong exothermal effect of the thermal NPK fertilizer decomposition is associated with the presence of chlorine-contained compounds.

It has been shown that hygroscopicity and caking for 26:13:0, 22:11:11, 16:16:16, 20:10:10, 19:9:19 and 27:6:6 grades decrease by increasing M from 1.0–1.1 to 1.6–1.7.

The study of the thermal decomposition by the example of 22:11:11 grade has demonstrated that increasing the degree of ammonization up to the specified values increases the thermal stability and reduces the intensity of the release of compounds of chlorine, fluorine and nitrous gases into the gas phase.

The study of thermal and rheological properties of ammonium–phosphate–nitrate slurries has allowed to set their high thermal stability, which increases with the increase of the phosphoric acid ammoniation degree. The viscosity of the slurries changes extremely having the minimum value at M = 1.4–1.5 and the maximum value at M = 1.0. The viscosity of the slurries increases with decreasing moisture content and decreases with increasing temperature according to the law of Arrhenius–Andrade.