The results of these measurements indicated that the amount of settling chamber ash is several times greater than the amount of cyclone ash in each experiment. It is noted that the largest share of these products occurred in experiment E1, as the sum of the product from the settling chamber and the cyclone was, in this case, more than 59% wt., while the lowest share was in experiment E3, in which these products accounted for only 1.5% wt.
3.2.1. Chemical Composition of Ash and Grain Morphology
The combustion resulted in six ash samples, in which the dominant chemical components, occurring in amounts of several percent each, are CaO, P
2O
5, K
2O and loss on ignition (LOI) amounts to 36.04%–62.35% wt. An exception in this respect are samples from experiment E3, in which CaO, P
2O
5, K
2O contents are much higher (in the case of P
2O
5 almost twice), and LOI contents several times lower (6.42%–9.16% wt.) than in other samples (
Table 5). Components such as SiO
2, MgO, Na
2O, SO
3, Cl and for most samples, Fe
2O
3 are present in the amount of a few percent. Other chemical components do not exceed 1% wt. (TiO
2, Al
2O
3, MnO, SrO and BaO).
However, the content of chemical components varies between samples. It is noted that the composition of samples from pellet combustion differs in this respect from the composition of ash from the combustion of loose manure.
It is also noticeable that there is little variation in the chemical composition of ashes from the combustion of manure in loose form on different beds (E1 and E2). This variation is evidenced by slightly higher contents of: (i) MnO, K2O and SO3 in the E1 experiment conducted on a bed of ash from the combustion of poultry manure, (ii) Fe2O3 and Cl in the E2 experiment conducted on a bed of quartz sand. The reasons for the large variation of LOI in these samples is difficult to identify conclusively.
On the basis of the research carried out, it was found that the contents of some elements show a high, positive, significant correlation among themselves. In particular, this concerns the main components CaO, P
2O
5 and K
2O with other chemical components. This is clearly indicated by the high values of the coefficient R
2, which are, for most of the trend lines shown in
Figure 2,
Figure 3 and
Figure 4, above 0.90 for
p < 0.05. These correlations apply: (i) CaO with MnO, MgO, Na
2O, K
2O, SrO; (ii) P
2O
5 with SiO
2, MgO, CaO, Na
2O, SrO; (iii) K
2O with MnO, CaO, SO
3. A high, negative, significant correlation with CaO and P
2O
5 is only shown by LOI.
Due to the use of quartz sand beds, the correlation coefficient between the content of individual chemical components and the content of SiO2 was calculated. In this case, it was also found that the contents of chemical components such as Fe2O3, CaO and MgO also show a high, positive, significant correlation with SiO2, while LOI correlations are high, negative and significant.
Despite detailed correlation analyses and apparent differences in chemical composition between samples from the same experiment but taken from different locations in the reactor (cyclone, settling chamber), no regularities justifying these differences were observed.
In view of, (i) the sum of the four basic chemical components (CaO, P
2O
5, K
2O and LOI), fluctuating in the range 72%–89% wt.; (ii) geochemical affinity CaO with MgO and K
2O with Na
2O; (iii) the mutual correlations between these components, the results for the tested samples are presented in a triangular diagram in the system CaO + MgO—K
2O + Na
2O—P
2O
5, considering LOI as the fourth component in the form of the proportional size of symbols (bubbles) of the individual projection points of the samples (
Figure 5). The diagram shows line a, which clearly separates samples coming from the cyclone from those coming from the settling chamber.
Among the trace elements determined, Zn and Cu are clearly dominant, with contents ranging from 3280 ppm to 3790 ppm and from 583 ppm to 901 ppm, respectively (
Table 6). The contents of other elements rarely exceed 100 ppm, i.e., Ni (in sample E2/KO, E2/C), Pb (in sample E2/C), Rb (in samples E1/KO, E1/C, E2/KO, E2/C) and Zr (in samples E3/KO, E3/C). Concentrations of other elements (As, Cr, Mo, Sn, V and Y) usually did not exceed 50 ppm and, in some cases, were below the detection limit.
Some variation in the content of certain elements is apparent, (i) in experiment E1, lower contents of Ni, Pb, Sn and Zr are observed than in E2 and E3; (ii) in experiment E2, higher contents are present for As, Ni and Rb compared to experiments E1 and E3; (iii) in experiment E3 significantly higher contents of Cr and Zr and significantly lower Cu are observed compared to experiments E1 and E2.
Variations in the content of the main chemical components and trace elements in the tested poultry manure ashes in individual experiments are most likely due to the conditions of their conduct—two types of starting deposits and two forms of preparation of the samples fed to the reactor (loose and pellets), which is related to, (i) residence time of combusted particles in the combustion chamber; (ii) the influence of the chemical components of the bed on the chemical composition of ashes from the combustion of manure.
It was therefore observed that the highest phosphorus concentration and the lowest unburned organic matter content (lowest LOI), i.e., the best efficiency, were obtained for experiment E3, in which the starting bed was quartz sand and the poultry manure was fed to the reactor in the form of pellets.
The morphology (
Figure S3) and chemical composition of 20 randomly selected grains in the sample of ash obtained in experiment E3, both taken from the cyclone (
Figure 6a) and the settling chamber (
Figure 6b) by EDS, were also examined. In the case of the main elements, i.e., calcium, phosphorus and potassium, the difference between the content of a given element determined by the XRF method and the average content calculated on the basis of EDS tests does not exceed 15%. The chemical composition of the ash grains was compared with the theoretical composition of the main phosphate minerals identified in the tested ash samples (
Figure 7). It can be observed that the grains selected for the analysis consisted mainly of amorphous substances.
3.2.2. Phase Composition of Ash
X-ray diffraction (XRD) identified the following phase components in all ash samples (
Table 7,
Figure S4): potassium magnesium phosphate (V), nonacalcium magnesium sodium heptakis (phosphate (V)), nagelschmidtite, periclase and sylvine. Apatite (except for sample E3/C), wopmayite, arcanite (except for sample E2/KO) and calcite (except for sample E2/C) were also present in most of the samples examined. Only some samples contained whitlockite (E1/KO, E1/C and E2/C), calcium iron magnesium hydrogen phosphate (E1/C and E2/C) and metathenardite (only in samples from experiment E1).
Diffractograms of all samples show a high background in the range of 2 theta 15–30°, which clearly indicates the presence of an amorphous substance in the ashes studied, associated with the presence of unburned organic matter, as well as the glass.
Among the identified phases, due to their chemical composition, two types of components can be distinguished, i.e., (i) those containing phosphorus, i.e., potassium magnesium phosphate (V), nonacalcium magnesium sodium heptakis (phosphate (V)), nagelschmidtite, wopmayite, apatite, whitlockite, calcium iron magnesium hydrogen phosphate, (ii) other components, i.e., arcanite, calcite, m etathenardite, periclase, sylvine.
Thus, phosphorus in crystalline form occurs in combination with the following metals, (i) calcium (apatyt, nagelschmidtite), (ii) calcium and magnesium (whithlockite) and sodium (nonacalcium magnesium sodium heptakis (phosphate (V)), (iii) potassium and magnesium (potassium magnesium phosphate (V)), (iv) magnesium and iron (calcium iron magnesium hydrogen phosphate) and sodium and manganese (wopmayite).
It was found that among the crystalline phases, the phases containing phosphorus—magnesium phosphate (V), nonacalcium magnesium sodium heptakis (phosphate (V)) and nagelschmidtite, whose amounts are usually several percent each, and in some samples even above 20% wt. (samples from experiment E2). Their total amount was in the range of 25.8% wt. to 59.9% wt. The sum of all crystalline phases containing phosphorus ranged from 37.7% wt. to 70.2% wt. It is noteworthy that the highest amounts of these phases were found in samples from experiment E2, while in experiments E1 and E3, these amounts were comparable (37.7%–42.3% wt.).
The proportion of other crystalline components varied considerably. The highest amounts were shown for experiment E1 (about 18.5% wt.), while the lowest was in experiment E3 (about 1% wt.).
The proportion of amorphous substances ranged from 21.3% wt. to 59.0% wt., with significantly higher contents in samples from experiment E3 (56.7%–59.0% wt.) than in E1 and E2 (21.3%–42.8% wt.).
Practically no relationships were shown between phase components among, as well as between, phase components and chemical components. The calculated values of correlation coefficients were mostly insignificant (p > 0.05), even at high values. This is most likely due to the sample population being too small.
On the basis of the quantitative XRD analysis of all crystalline phases, taking into account their stoichiometric formulations, the shares of the main chemical components in the ashes from experiment E3, which had the best process efficiency, were calculated. The results of these calculations were used to estimate the proportion of the main chemical components in the amorphous phase, taking into account the results of the chemical composition determined by the XRF method (
Table 8).
A higher proportion of phosphorus, calcium, potassium and sodium was observed in the amorphous phase than in the crystalline phase.