*2.2. Methods*

The methods used are acid digestion, thermochemical treatment of SSA and MLA investigation. SSA was subjected to digestion. Digestion was carried out by adding inorganic acid (HCl (5 N)) for 5 min at ambient temperature. The amount of solids was 5 g sample and 20 g acid. This was done in a beaker while stirring permanently. Following digestion, cake washing with distilled water was carried out using 2 × 50 mL. The resulting residue was dried at 110 ◦C for 2 h in a drying cabinet.

Then, acid digestion was undertaken after thermochemical pre-treatment: SSA was mixed with NH4Cl at a ratio of 1:1 and heated in a kiln. A first temperature rest was kept for 2 h at 250 ◦C, then another rest was kept for 2 h at 350 ◦C. Finally, the material was heated to 1000 ◦C in order to remove the remaining heavy metal volatiles. A continuous inert gas flow of N2 was applied at 15 L/h. Afterwards, the material was processed using inorganic acid (HCl (5 N)) at a digestion time of 5 min without additional heating and a liquid/solid ratio of 4:1. Both processes were accompanied by sampling for SEM-MLA investigation.

Three powder samples were collected: (i) untreated SSA; (ii) digested, rinsed and dried SSA residue; (iii) thermochemically treated, digested, rinsed and dried SSA residue. These samples were prepared and subjected to the grain mount preparation procedure. Preparation included mixing the dried sample powder 1:1 with graphite (for particle separation) and stirring this mix into epoxy in Teflon containers with 30 mm in diameter. Epoxy blocks were generated, whetted, polished and carbon coated for measurement. The samples were introduced into the vacuum chamber of the SEM and beam acceleration was set at 25 kV with a current of 10 nA; horizontal frame width (HWF) set to 373 pixels. For the acquisition of the grey-scale BSE image, Cu was selected as the BSE grey-scale standard (materials with a Z value of Cu or higher appear as 256 (very bright), all other material turn out darker, respectively). Spectral data were obtained using the GXMAP procedure, that is, particles are automatically detected and covered with a grid of measurement points [12]. Particles smaller than 20 pixels were identified by a single measurement point. Fractions of the sample are given in area-% in this work, since the density of SSA material is di fficult to assess. Keeping proportions representative, densities (even of known minerals) were set to one, so "% of the sample" refers to area-% (data were gathered from a polished surface, the actual measurement was that of an area). Proportions of the elemental contents in spectra are weight-%, thus, shares of phosphorus within the material are wt%.

In this study, sewage sludge ashes are analyzed with respect to their P-content and treated SSA checked for residual P-content. For data evaluation, a new approach to sorting and handling EDX spectra is applied. Similarly, workers have previously adapted data evaluation strategies, such as Ayling et al., 2012 [17] devising a new species identification protocol. A similar target-element focused evaluation using QEMSCAN to investigate eudialyte residue has been reported by Ma et al., 2018 [18]. The target component labelling looks at the P-content according to each spectrum and allows for a target material grouping. Conventionally, a mineral grain or particle is classified by comparing the collected EDX spectra to a set of reference spectra labelled by mineral names. EDX spectra of SSA are complex, with a multitude of peaks for Si, Al, Ca, Fe, S and P at various counts (97.6 area-% of the material is comprised of these and oxygen). Other elements found are Ba, Cl, Cr, Cu, K, Mg, Mn, Na and Ti. Due to small grain sizes and transitions between grains of di fferent compositions, there are also

some "mixed" spectra which display a combination of peaks from elements in adjacent grains. These conditions, in addition to amorphous, artificially produced material phases, render an interpretation of the spectra as a mineral impossible and do not allow for spectra labelling with mineral names. Using EDX-elemental analysis for quantifying the Si, Al, Ca, Fe, S and P-contents, 53 spectra were collected from the samples, which were then used as reference spectra for classifying the samples. For such reference spectra, generic labelling [16] is a viable approach.

Subsequently, these reference spectra are sorted by their P-content. Spectra groups summarizing the spectra within a range of P-contents (in wt%) are then named descriptively, "<1% P" or "5–10% P", for example (Table 1). This step is called the target component grouping. There is an interesting tendency of elemental complexity (number of elements detected) decreasing as P-content rises. In phosphate recycling, recovering a few other elements in addition to P is desirable as it reduces the need for product purification. The procedure di ffers from the other ways of grouping by sorts of minerals [16], as it focuses on a single elemental component in amorphous particles. All software operations used MLA datasuite 3.1.4.686.



Non-P-containing material components are quartz and other silica-rich species, calcium sulphate spectra, iron oxides and residuals/erroneous identification. The GXMAP measurements were classified against the reference EDX-ray spectra list using the MLA 3.1.4 software version. All spectra were collected with 11,000 counts minimum, with an error level of 0.5% or less. Thereby, this method tries to meet the general accuracy challenges of EDX with statistical significance. Less than 0.6% of the sample area remained classified as unknown, with no X-ray counts, or as unspecified due to a low number of counts. Not all phases have been detected in amounts of relevance. These were grouped as others. The arbitrary cut-o ff value (below—others, above—relevant compositional phase) is 0.4 area-%. For "BaCaO", the percentage is <0.3 area-% of the sample, represented by <1400 grains. All other spectra are below these values, and thus, only secondary in relevance. As an area was measured, values of percentages refer to area-% unless otherwise noted.
