Systematic Evaluation of the Fate of Phosphorus in Fluidized Bed Combustion of Biomass and Sewage Sludge

http://www.diva-portal.org

This is the published version of a paper published in Energy & Fuels.

Citation for the original published paper (version of record):

Falk, J., Skoglund, N., Grimm, A., Ohman, M. (2020)

Systematic Evaluation of the Fate of Phosphorus in Fluidized Bed Combustion of Biomass and Sewage Sludge

Energy & Fuels, 34(4): 3984-3995

https://doi.org/10.1021/acs.energyfuels.9b03975

Access to the published version may require subscription.

N.B. When citing this work, cite the original published paper.

Permanent link to this version:

http://urn.kb.se/resolve?urn=urn:nbn:se:umu:diva-170545

Systematic Evaluation of the Fate of Phosphorus in Fluidized Bed Combustion of Biomass and Sewage Sludge

Joel Falk,* Nils Skoglund, Alejandro Grimm, and Marcus Öhman

Cite This:Energy Fuels 2020, 34, 3984−3995 Read Online

ACCESS

Metrics & More Article Recommendations

ABSTRACT: Comprehensive knowledge concerning the behavior of phosphorus (P) during combustion is necessary to enable more e fficient recovery of P from combustion ashes for agricultural purposes. To this end, parameters that influence the distribution and speciation of P in combustion ashes are important because they may in fluence which ash fractions are suitable for P recovery.

This study aims to determine the fate of P as a result of fuel ash composition and chemical association in the fuel during fluidized bed combustion by a systemic review of previous work. The synthesis was performed by comparing scanning electron microscopy − energy-dispersive X-ray spectroscopy and X-ray di ffraction chemical analyses of bed ash, fly ash particles, and deposits from fluidized bed combustion of di fferent blends of P-poor (logging residues or wheat straw) and P-rich (sewage sludge, dried distiller’s grain with solubles, or phosphoric acid) fuels and additives. The blends were produced to have a similar ash composition but with a di fferent P source. The distribution of P among ash fractions indicated that P is mainly found in the coarse ash fractions (bed and cyclone ash), irrespective of fuel ash composition or chemical association in the fuel. The chemical speciation of P in coarse ash fractions di ffered between biomass blends containing sewage sludge compared to blends with phosphoric acid or dried distiller ’s grain with solubles.

Phosphates in the ash from the two sewage sludge blends included predominantly Ca with minor inclusion of other cations. In contrast, ashes from the blends with phosphoric acid or dried distiller ’s grain with solubles contained phosphates with a significant amount of K, Ca, and Mg. The di fference in phosphate speciation could not solely be explained by the combustion conditions and the elemental composition of the ash fractions. These results show that it is necessary to consider the chemical association of P in the fuel to predict the type of phosphates that will form in fluidized bed combustion ashes.

1. INTRODUCTION

Mineral phosphorus (P) is an essential resource in society, primarily for its usage in fertilizers. It is considered as one of the key biogeochemical flows that must be considered for human society to thrive in the future because the use of P in agriculture has a severe impact on biodiversity as a result of freshwater eutri fication.

¹

Furthermore, on the basis of the current trends of the usage of mineral P in agriculture and the state of global phosphate reserves, a global P-scarcity crisis might occur in the immediate future.

²

A measure that would alleviate these problems is to increase the recycling of P in manure, human excreta, and food residues. The recovery of P from human waste was covered in detail by Desmidt et al.,

³

including a review of the di fferent P-recovery techniques from municipal wastewater treatment plants. A byproduct of this process with high potential for P recovery is sewage sludge. In the European Union (EU) alone, the annual production of sewage sludge exceeds 6 million tons on a dry basis (db) with a typical P content in the range of 0.5 −5 wt % given as P

2

O

₅

.

⁴

Sewage sludge can be applied to farmland directly; however, this requires that concentrations of phytotoxic heavy metals are within legislated limits.

⁵

Additionally, there is the potential of distributing hazardous organic compounds contained within the sludge.

A possible solution is to process the sewage sludge through thermochemical conversion. This treatment destroys organic compounds in the sludge

⁶

and may enable a signi ficant

separation between nutrients and heavy metals.

^7,8

Mono- incineration has been implemented extensively in countries, such as Germany,

⁹

yet only a small fraction of the resulting ash is used as fertilizer. A challenge for P recovery from mono- incinerated sewage sludge ashes is the often low P-plant availability assessed by its solubility in a neutral ammonium citrate solution (P

_NAC

).

¹⁰

The average P

_NAC

of ashes from 24 sewage sludge incinerators was 25.6% but with signi ficant di fferences between incinerators (9.6−82.6%). As a result, sewage sludge ash from mono-combustion may require further treatment, such as wet chemical methods

¹¹⁻¹³

or thermal treatment with di fferent additives,

^6,14

before it can be e ffectively used as a fertilizer.

Inorganic elements, such as P, undergo ash transformation reactions during fuel conversion as a result of the physical environment of the fuel particle during the thermal conversion process. Furthermore, the total and relative concentration of ash-forming elements in the fuel has a major in fluence on the overall ash transformation process.

¹⁵

This dependency implies

Received: November 15, 2019 Revised: March 20, 2020 Published: March 22, 2020

Review pubs.acs.org/EF

License, which permits unrestricted use, distribution and reproduction in any medium, provided the author and source are cited.

Downloaded via UMEA UNIV on May 27, 2020 at 12:52:29 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

that it may be possible to a ffect the chemical speciation of P in the ashes toward more plant-available phosphates and, thus, reduce the need for post-processing by adjusting the overall fuel composition through co-combustion or by use of additives. Sewage sludge co-combustion might also enable a more extensive utilization of alkali-rich biomasses, such as wheat straw, because it reduces problems with agglomer- ation,

^16,17

sintering,

¹⁸

deposition,

^16,19−21

and corrosion.

^16,20−22

The association of inorganic elements in the fuel may also in fluence the ash transformation reactions during combus- tion.

²³

In plant-based biomass, as a result of its many biological functions, P is found in inorganically and organically associated compounds and serves as an important linkage unit or binding site for alkali and alkaline metal ions.

²⁴

The concentration and association of P vary between the stem, bark, shoots, needles, or leaves in woody biomass.

²⁵

In sewage sludge, a high share of P is inorganic but also includes lower amounts of organically associated P.

²⁶

The association of the inorganic phosphates depends upon the coagulant(s) used in the coagulation − flocculation process of the wastewater, which includes Fe salts, Al salts, or lime.

²⁶⁻²⁹

Chemical fractionation of sewage sludge and plant-based biomasses suggests a significant difference in P association. By dissolution of the fuel samples in progressively more aggressive aqueous solutions (water, NH

₄

Ac, and HCl), it is possible to determine the fraction of water-soluble, ion-exchangeable, and acid-soluble elements in the fuel. The major fraction of P in P- poor woody and agricultural biomass is water-soluble and ion- exchangeable ( ∼40−75 and ∼90 wt %, respectively). By comparison, P in sewage sludge is mainly acid-soluble (∼90−

95 wt %).

²³

The thermal conversion process may also have an impact on the type of phosphates that are found in the ash. For instance, combustion of P-rich rapeseed meal or dried distiller ’s grain with solubles in fluidized bed, fixed bed, and powder combustion results in some variation in the type of phosphate phases that are identi fied in the ash.

³⁰⁻³²

Similarly, thermal conversion of sewage sludge from a single wastewater treatment plant in di fferent combustion, pyrolysis, or gas- i fication technologies had a significant effect on the fertilizer value of char or ash,

³³

which implies a di fference in the chemical association of P.

The speciation of phosphorus in the ash produced during the combustion of biomass varies depending upon the overall fuel composition.

^18,34

In biomasses containing a signi ficant share of P (P-rich), commonly identi fied phosphate phases by X-ray di ffraction (XRD) are CaKPO

4

, KMgPO

₄

, and CaK

₂

P

₂

O

₇

.

^34,35

In ashes from P-poor fuels, such as woody biomass, the Ca-rich apatite [Ca

₅

(PO

₄

)

₃

OH] and whitlockite phase [Ca

₃

(PO

₄

)

₂

] are usually identi fied.

¹⁵

However, it seems that phosphates identi fied in the ashes from sewage sludge do not follow the same trend. Fluidized bed combustion of P-rich sewage sludge at 800 and 950 °C with a Ca and K additive did not cause any detectable changes to the identi fied whitlockite phase.

³⁶

These di fferences could potentially be caused by di fferences in the chemical association of P between plant- based biomass and sewage sludge. Understanding such di fferences would facilitate predictions of which phosphates are likely to form a wide range of biomass fuels in di fferent combustion conditions. Thus, to enable a more e fficient P recovery from combustion ashes, a more detailed under- standing of the reaction pathways of P during combustion is necessary.

To further this goal, the objective of this study is to determine the e ffect of fuel ash composition and chemical association of P in the fuel on the distribution and speciation of P in fluidized bed co-combustion ashes. The study was performed as a systematic review of data and experience gathered from mono- and co-combustion of P-poor and P-rich biomass fuels in a bench-scale bubbling fluidized bed.

16,30,34,37

From the original sets of data, six combustion experiments using two P-poor and four P-rich fuel blends of similar ash composition but with a di fferent P source (sewage sludge, dried distiller ’s grain with solubles, and phosphoric acid) were considered. The bed ash, fly ash particles, and deposits gathered from each experiment were analyzed utilizing scanning electron microscopy −energy-dispersive X-ray spec- troscopy (SEM −EDS) and X-ray diffraction (XRD).

2. MATERIALS AND METHODS

2.1. Fuels and Additives. The experimental data used in this study originated from six combustion experiments performed in separate studies16,30,34,37 (see Table 1). No additional experiments were performed beyond what was included in the original studies.

Because the studies were performed with a systematic approach, it

Table 1. Fuel Ash Content (wt %, db) and Content of Main Ash-Forming Elements (mmol kg

⁻¹

, db) in Raw Materials (Base Fuels), Fuel Blends, and Previous Literature

ash K Na Ca Mg Fe Al Si P S Cl reference

base fuels or additive

logging residues (LR) 2.4 43 6 127 25 4 13 103 15 13 0 34and37

wheat straw (WS) 5.7 320 13 100 41 1 2 285 42 59 73 34and37

dried distiller’s grain with solubles (DG) 4.4 271 43 27 114 2 0 36 266 321 62 30

sewage sludge (SS) 41.7 95 74 621 209 1379 867 1321 1369 421 14 16

P-Poor Blends^a

LR, 99.8%; PA, 0.2% 2.5 43 6 127 25 4 13 103 39 13 0 34and37

LR, 97%; SS, 3% 3.6 45 8 142 31 46 39 140 55 25 0

P-Rich Blends^a

LR, 60%; DG, 40% 3.2 135 21 87 61 3 8 76 115 136 25 30and34

WS, 50%; DG, 50% 5.1 295 28 64 78 1 1 160 154 190 68 30and34

WS, 99%; PA, 1% 6.4 316 13 99 41 1 2 282 158 59 73 34and37

WS, 91%; SS, 9% 8.9 300 19 147 56 125 80 378 161 92 68 16

aPercentages denote the share of base fuel(s) and additive (wt %, db). The ash compositions of blends were calculated from the composition of initial fuels and additives. Abbreviation: PA, phosphoric acid.

allowed for a cross comparison of the data, at which point a difference in the chemical association of P between fuel mixes was observed. The scope of the present study is how P speciation is affected by the chemical association of P present in the fuel, whereas previous studies investigated combustion characteristics of P-rich fuels and their fuel blends with respect to ash-related problems, such as agglomeration, slagging, and particulate matter emissions.

The data set included two experiments with P-poor fuel blends with a low K/Ca ratio and four P-rich blends with an intermediate ratio of K/(Ca + Mg) (Table 1). The blends were produced by mixing a P-poor fuel [logging residue (LR) or wheat straw (WS)]

with a P-rich fuel [digested sewage sludge (SS) or dried distiller’s grain with solubles (DG)] or with phosphoric acid (PA) in two approximate compositions with respect to K, Ca, Mg, and P. The mixing ratios between SS and LR or WS were based on molar ratios between K over P and S. As a result of the large relative difference in ash content between SS and LR or WS, this resulted in a low share of SS in the two blends. SS was produced at a local wastewater treatment plant, which employs iron(II) sulfate to remove P from the effluent by coagulation and subsequent flocculation. DG was produced by an ethanol producer in northern Europe and is a waste residue based on wheat. PA was used as a reactive P additive to increase the P content without causing changes to the other ash elements. For all blends, the P content is dominated by the share contributed by the P-rich fuel or the P additive (>62 mol %).

All fuel blends were mixed on a dry basis and completed in 20 kg batches using a ribbon mixer to ensure homogeneous blends. PA was

added as an aqueous solution (85% H3PO4) during mixing. The blends were subsequently pelletized in a pilot-scale pellet press (Ø, 8 mm; moisture, 10−12 wt %).

2.2. Combustion Experiments. The combustion experiments were performed in a bench-scale bubblingfluidized bed reactor using sieved quartz sand as the bed material (200−250 μm, >98% SiO2).

The bed material isfluidized by a primary airflow of ∼80 NL min⁻¹ that was injected from below through a perforated plate.Figure 1 shows a schematic overview of the experimental setup. The height of the reactor is 2.4 m and includes a primary air preheater, a propane burner, and electrical heaters to regulate the temperature of the bed and freeboard sections (100 and 220 mm Ø). Type N thermocouples are distributed along the reactor. Pressures are measured in the upper and lower bed sections.

Sampling points are placed throughout the reactor to enable sampling (see the numbered positions inFigure 1). Bed samples were gathered from a sampling port above the bed (1), andfly ash deposits were gathered from an air-cooled deposition probe (steel) that was inserted at the top of the freeboard (2). Fly ash particles were gathered from a cyclone (cut of size 10 μm) located after the freeboard section (3) and from a 13 stage low-pressure impactor connected to theflue gas channel after the cyclone (4). The impactor (Dekati, Ltd.) fractionates the particulate matter according to aerodynamic diameter in the range of 0.03−10 μm. For more details on the experimental setup, thefluidized bed reactor was evaluated in depth by Öhman and Nordin.³⁸

Figure 1.Schematic overview of the bench-scale bubblingfluidized bed (not to scale), adapted with permission from ref30. Copyright 2012 Elsevier.

Before each experiment, the reactor was slowly preheated (3°C/

min) until it reached the operational temperature, at which point fuel feeding was initiated. The feeding rate was maintained at∼0.6 kg h⁻¹. The experiments lasted for 8 h, with the exception of the blend of WS and DG, which suffered total defluidization after 2 h. For all combustion experiments, the dry oxygen content in theflue gas was 8−10 vol %. The bed temperature was maintained at 730 °C during the combustion of WS blends and 800°C for LR blends. The lower bed temperature for the experiments with WS blends was intended to reduce the risk of early defluidization during the experiments as a result of the known tendency of WS ash to form low-temperature melting alkali silicates. The air-cooled deposit probe was kept at a surface temperature of 450°C, while the surrounding flue gas had a temperature of 800−850 °C. The flue gas temperature at the impactor sampling point was 100−150 °C, and the impactor was preheated to a similar temperature to prevent the condensation of water vapor.

Aluminum foil (non-greased) was used as substrates in the impactor.

The deposition probe was mounted once stable combustion conditions were achieved. A bed sample was gathered from the bed directly after fuel feeding was stopped, and the deposition probe was removed (sampling time of∼6−7 h) before initializing a controlled defluidization test.³⁸ The early defluidization event for the blend between WS and DG prevented the gathering of bed samples before defluidization and significantly shortened the exposure time of the deposition probe. Cyclone ash and an additional bed sample were gathered once the reactor had cooled to room temperature. More details on the experimental procedure for each fuel or fuel blend can be found in the original references.16,30,34,37

2.3. Chemical Characterization of Ash Fractions. Bed ash particles were separated from the bed samples by passing the material through a sieve with a mesh size significantly smaller (100 μm) than the initial mean particle size of the quartz bed material (200−250 μm). Except for the experiment with the blend of LR and SS, a sufficient amount of bed ash particles could be separated from the bed samples to enable further XRD analysis. The deposits on the deposition probe ring were manually removed and separated into a coarse wind-side fraction and afine lee-side fraction. Representative subsamples of the bed ash particles, wind-side deposits, and cyclone ash were gathered and thoroughly homogenized before chemical characterization. The measured weights of the impactor substrates 1−

7 and 8−13 were grouped as fine particulate matter (<1 μm, PM1) and coarse particulate matter (1−10 μm), respectively, and saved for chemical analysis.

The morphology and elemental composition of the ash fractions were analyzed using Philips model XL30 SEM coupled with EDAX EDS. Bed samples were encased in epoxy resin, followed by dry polishing to give a smooth cross section of the quartz grains.

Representative quartz bed grains and bed ash particles were identified in the bed samples using a backscatter electron detector. The ash layers on the quartz grains and bed ash particles were subsequently analyzed at several points, evenly spread out over the grain/particles by SEM−EDS spot analysis. The wind-side deposits, cyclone ash, and fine/coarse fly ash subsamples were mounted on carbon tape, and three 100× 100 μm SEM−EDS area scans were performed.

The crystalline phase composition of ash fractions was charac- terized by XRD using a Bruker D8 ADVANCE instrument in θ−θ mode and Cu Kα X-ray tube. Initial qualitative identification was made using the PDF-2 database³⁹ together with Diffrac^plus EVA.

Subsequently, semi-quantitative Rietveld analysis was performed using structures from the inorganic crystal structure database (ICSD)⁴⁰and the Diffrac^plussoftware (TOPAS). Continuous scans were applied to each sample repeatedly for a total sampling time of at least 6 h. During the analysis, a scan speed of 1−4°/min was used with a data range of 2θ 10−70°.

Previous studies indicated two whitlockite phases [Ca₃(PO₄)₂and Ca₉KMg(PO₄)₇] in the ash.^16,34,37Whitlockite phases have a structure that allows for limited solid solutions with cations, such as K, Mg, and Fe. Such substitutions are complicated to accurately estimate because it does not cause any changes to the overall crystal structure of the phase but may cause minor shifts in the height and position of the

peaks in the diffractogram. Therefore, no exact stoichiometric composition should be assigned to the identified whitlockite phases.

More likely, they exist as a limited solution series between Ca3(PO4)2

and Mg₃(PO₄)₂ or between Ca₉KMg(PO₄)₇ and Ca₉Fe(PO₄)₇. However, to what extent the cations can be substituted before it causes changes to the overall crystal structure is not currently known.

As a result of the interaction between ash and bed material and the varying presence of bed material in the different coarse ash fractions, i.e., bed ash, cyclone ash, wind-side and lee-side deposits, and PM₁₋₁₀, it was not possible to establish a reliable mass balance of P between the gathered ash fractions. Instead, the share of P in PM₁of total P was estimated by comparing the mass of P in PM₁with the amount of P fed with the fuel. The share of P in coarse ash fractions was estimated by difference. The total mass of P in PM1 during the experiment was estimated by multiplying the mass concentration of P in PM1 (calculated from impactor data and SEM−EDS analysis of PM₁) with the totalflue gas flow for the duration of the experiment.

The amount of P that was introduced with the fuel was estimated from the total amount of fuel used during the experiment, the P content (g/kg, db), and the moisture content of the fuel blend.

3. RESULTS

The total mass concentration of PM

₁

was in a similar range for the P-poor fuel blends and the P-rich fuel blends, respectively (Figure 2). The total mass concentration of coarse PM (1 −10

μm) was in a similar range for all fuel blends. The concentration of P in PM

₁

was low for all fuel blends but tended to be higher in the PM

₁

fraction of the P-poor blends.

The mass fraction of total fuel P found in PM

₁

was estimated to be well below 1 wt % for all fuel blends.

Ash in the bed sample was found as individual bed ash particles or as layers formed on the bed material grains, distributed between inner reaction layers and outer coating layers. Further details on the morphology and composition of the characteristic layers and agglomerate necks are out of scope for this study but have previously been described in detail.

^16,34

The spatial correlation of elements in outer coating layers and bed ash particles was determined by SEM −EDS spot analysis and indicated a signi ficant concentration of P. The outer coating layers and bed ash particles were heterogeneous, with considerable variation in composition between spot measure- ments but with an average composition similar to the overall fuel ash composition.

Figure 2.Average total mass concentration offine (PM1) and coarse (PM₁₋₁₀) particle emissions (mg/Nm₃at 10% O₂) as determined by impactor measurements are shown with bars (left y axis). The right axis shows the concentration of P in the PM₁fraction (wt %, C-free basis, gray dots) and the mass fraction of P introduced with the fuel that was found in PM₁(orange dots).

Most bed ash particles could be separated into two groups:

one group with a very low P concentration and one group with high concentrations. Particles with a low P concentration were mainly found in bed samples from the combustion of LR blends and had a composition that closely resembled that of either albite or microcline (NaAlSi

₃

O

₈

or KAlSi

₃

O

₈

). To more accurately re flect the spatial correlation of P in the particles with high concentrations of P, only spot analyses with over 5 mol % P on an O- and C-free basis was considered. Because the number of bed ash particles and the share of P-rich bed ash particles in the bed ash samples varied, the number of included data points in Figure 3 varies between experiments.

The particles with a high P concentration were scarce in the bed samples from P-poor fuel blends, with only one point having a signi ficant P concentration in the blend between LR and PA (Figure 3A). Except for a few bed ash particles from the blend of WS and SS, the bed ash particles from bed samples of the P-rich fuel blends all had P concentrations over 5 mol % (Figure 3B). To simplify the comparison, the ash composition of the P-rich fuel or the fuel blend is superimposed on the composition of the bed ash particles as empty and outlined columns. The composition of these particles, excluding PA blends, seems to more closely resemble the P-rich fuel in the blend rather than the composition of the overall mixture but with higher concentrations of K and Na and lower concentrations of S and Cl.

The average elemental composition of wind-side deposits and cyclone ash were determined by SEM −EDS area analysis (Figure 4, Figure 5). The wind-side deposits of the P-poor

fuels had a composition that closely resembled the fuel ash composition but increased concentrations of S and Cl (Figure 4A). Similarly, the composition of the wind-side deposits of the

Figure 3.Average elemental composition (mol % on an O- and C-free

basis) from SEM−EDS spot analysis of bed ash particles from experiments with (A) P-poor blends and (B) P-rich blends. Only spot analyses with over 5 mol % P were included in the results, where n is the included number of data points of total data points (error bar of

±1σ). The empty, outlined columns indicate the calculated fuel ash composition (mol %) of the P-rich fuel in the blend (DG and SS) or the fuel ash composition of the mixture (PA).

Figure 4.Average elemental composition (mol % on an O- and C-free basis) from SEM−EDS area analysis of wind-side deposits from the (A) P-poor experiments and (B) P-rich experiments (error bar of

±1σ; n = 3). The empty, outlined columns indicate the calculated fuel ash composition (mol %) of the fuel blend.

Figure 5.Average elemental composition (mol % on an O- and C-free basis) from SEM−EDS area analysis of cyclone ash from the (A) P- poor experiments and (B) P-rich experiments (error bar of±1σ; n = 3). The empty, outlined columns indicate the calculated fuel ash composition (mol %) of the fuel blend.

Table 2. Crystalline Phases (wt %) Identi fied in Bed Ash Particles Sieved from the Bed Sample

P-poorP-rich phasephasecompositionLR,99.8%;PA,0.2%LR,97%;SS,3%LR,60%;DG,40%WS,50%;DG,50%WS,99%;PA,1%WS,91%;SS,9% phosphatehydroxyapatiteCa5(PO4)3OHNAa6 whitlockiteCa3(PO4)23636 whitlockiteCa9MgK(PO4)718 CaKPO43 KMgPO4142714 CaMgP2O719 CaK2P2O76181625 silicatemerwiniteCa3Mg(SiO4)2141511 åkermaniteCa2MgSi2O766 albiteNaAlSi3O87 microclineKAlSi3O83 leuciteKAl(SiO3)220 quartzSiO2651 cristobaliteSiO231128 sulfateanhydriteCaSO42211 aphthitaliteK3Na(SO4)2 arcaniteK2SO430106 oxidepericlaseMgO2 hematiteFe2O318 a NA=notanalyzed.

Table 3. Crystalline Phases (wt %) Identi fied in the Wind-Side Deposit of the Deposition Probe

P-poorP-rich phasephasecompositionLR,99.8%;PA,0.2%LR,97%;SS,3%LR,60%;DG,40%WS,50%;DG,50%WS,99%;PA,1%WS,91%;SS,9% phosphatehydroxyapatiteCa5(PO4)3OH673 whitlockiteCa3(PO4)29 CaKPO44 KMgPO415 CaK2P2O761925 silicatemerwiniteCa3Mg(SiO4)226 albiteNaAlSi3O8157610 microclineKAlSi3O81210 quartzSiO225166515 cristobaliteSiO28 sulfateanhydriteCaSO4152924324 aphthitaliteK3Na(SO4)23103 arcaniteK2SO4822172952 othersylviteKCl7433028 calciteCaCO319113 limeCaO22 periclaseMgO322 hematiteFe2O322

Table 4. Crystalline Phases (wt %) Identi fied in Cyclone Ash

P-poorP-rich phasephasecompositionLR,99.8%;PA,0.2%LR,97%;SS,3%LR,60%;DG,40%WS,50%;DG,50%WS,99%;PA,1%WS,91%;SS,9% phosphatehydroxyapatiteCa5(PO4)3OH713515 whitlockiteCa3(PO4)261616819 whitlockiteCa9MgK(PO4)7 CaKPO418 CaK2P2O74132632 silicatemerwiniteCa3Mg(SiO4)274 albiteNaAlSi3O81671016 microclineKAlSi3O8148109 quartzSiO2163853 cristobaliteSiO211 sulfateanhydriteCaSO4462245 aphthitaliteK3Na(SO4)29 arcaniteK2SO428292235 othersylviteKCl216121717 calciteCaCO3212 limeCaO1 periclaseMgO31 hematiteFe2O336

P-rich fuels resembled the fuel ash composition (Figure 4B).

The result from the experiment with the blend of WS and DG is presented on a Fe-free basis as a result of the presence of Fe oxides from the deposition probe ring. The cyclone ashes shared a composition and distribution similar to the wind-side deposits (panels A and B of Figure 5). The wind-side deposit and cyclone ash of the two SS blends had lower concentrations of Fe, Al, and P compared to the fuel ash composition.

Semi-quantitative analysis (wt %) of the crystalline phases of the bed ash particles (sieved from the bed sample), wind-side deposits, and cyclone ashes was determined by XRD coupled with Rietveld re finement ( Tables 2 − 4). The bed ash particles were dominated by crystalline phosphates and silicate phases (Table 2). A large variety of crystalline Ca-, Mg-, and K- containing phosphate phases were detected from the bed ash particles, except for the bed sample from the WS and SS blend, where only one whitlockite phase [Ca

₉

KMg(PO

₄

)

₇

] was detected. Furthermore, leucite (KAlSi

₂

O

₆

) and hematite (Fe

₂

O

₃

) were detected but were not observed in bed ash particles from the other blends.

The phase composition of wind-side deposits contained signi ficant shares of phosphates, silicates, sulfates, chlorides, carbonates, and oxides (Table 3). The two P-poor blends had a similar phase composition and were dominated by silicates, sulfates, chlorides, carbonates, and oxides and a small share of Ca phosphates. The phase composition of the P-rich blends containing PA or DG was similar and had high shares of mixed Ca, Mg, and K phosphates. No phosphates were detected in the wind-side deposit of the blend between WS and SS, which contained mainly silicates (SiO

₂

), sulfates (CaSO

₄

and K

₂

SO

₄

), and chlorides (KCl). The cyclone ashes had similar composition and distribution as the wind-side deposits but tended to have a higher share of phosphates (Table 4). Similar to the bed ash particles, the detected phosphate phases in cyclone ash from SS blends were Ca-rich, including hydroxyapatite [Ca

₅

(PO

₄

)

₃

OH] and a whitlockite phase [Ca

₃

(PO

₄

)

₂

].

4. DISCUSSION

P was predominately found in bed samples and cyclone ashes, i.e., the coarse ash fractions for all fuel blends, which is consistent with other studies on fluidized bed combustion of

SS.

^36,41

Therefore, for P recovery from fluidized bed

combustion ashes, e fforts should be focused on coarse ash fractions, such as bed ash and cyclone ash.

A signi ficant difference in the detected crystalline phosphate phases was observed for the coarse ash fractions between blends containing SS and blends containing either PA or DG.

For the latter, a wide range of crystalline phosphate phases was detected in the ash with a signi ficant share of Ca, Mg, and K, i.e., Ca

₅

(PO

₄

)

₃

OH, Ca

₃

(PO

₄

)

₂

, CaKPO

₄

, KMgPO

₄

, CaMg- P

₂

O

₇

, and CaK

₂

P

₂

O

₇

. In the ashes from SS blends, only Ca- rich orthophosphates (PO

₄³⁻

) were detected, including Ca

₅

(PO

₄

)

₃

OH, Ca

₉

KMg(PO

₄

)

₇

, and Ca

₃

(PO

₄

)

₂

. Furthermore, in comparison of the stoichiometric composition of the phosphates to the bulk composition of the bed ash particles, there is an insu fficient amount of Ca in the SS samples for all P to form Ca orthophosphates. On the basis of the result from a study assessing the molecular environment of P in fluidized bed filter ash,

⁴²

this is likely explained by amorphous Fe phosphates, which are not detectable by XRD.

There is a qualitative agreement between the stoichiometric composition of the identi fied phosphate phases and the bulk

composition of the sample, except for the bed ash particles from the SS blends and the LR and PA blend. The poor match between SEM−EDS and XRD results is significantly improved when comparing the XRD results to the composition of the bed ash particles with over 5 mol % P in the bed sample.

Furthermore, the elemental composition of the bed ash particles from the SS blends is similar to the composition of the original SS, and with a phase composition of phosphates that resembles those found in mono-combusted SS ashes.

⁴²

Considering the similarities in the fractionation of ash elements between ash fractions within the respective fuel group (P-rich and P-poor), the di fference in the detected crystalline phosphate phases between SS blends and blends with PA or DG is likely caused by di fferences in the chemical association of P in the base fuels.

Phosphorus is mainly found as orthophosphate (PO

₄³⁻

)

^24,43

in all of the base fuels, bonded to organic and inorganic elements. Phosphorus in WS and LR is mainly water-soluble or ion-exchangeable ( ∼70:20 and 40:15 wt %, respectively),

²³

while P in DG is mainly water-soluble, together with K, Na, and Mg.

⁴⁴

In contrast, P in SS is mainly acid-soluble ( ∼90−95 wt %).

²³

Blends with DG would likely contain a high share of inorganic and organic phosphate salts associated with K, Ca, Mg, C, and possibly H. A high share is in the form of phytates (32.4 −36.6 wt %, db),

⁴⁵

which is an organic orthophosphate salt formed from phytic acid. The similarities in XRD results from the ash fractions from blends with PA and DG would suggest that P in these blends had a similar chemical association and would follow a similar reaction pathway during initial fuel conversion. Even though PA only has covalent H bonds, that would change when added to the P-poor biomass.

Wheat straw and LR have a high share of water-soluble, ion- exchangeable, or acid-soluble P, K, Na, Ca, and Mg.

²³

With the addition during fuel mixing, PA would likely react with ash compounds (K, Na, Ca, and Mg) and organic structures in the biomass by substituting hydrogen, forming inorganic and organic phosphate salts associated with K, Ca, Mg, C, and H.

The high share of Ca, K, and Mg and the lack of C and H among the detected crystalline phosphates suggest that phosphates in the fuel retain bonds, including K, Ca, and Mg during the thermal conversion process, while bonds including C and H are lost. Furthermore, the coarse ash fractions from the blends with PA and DG contained pyrophosphates (P

₂

O

₇⁴⁻

), including CaMgP

₂

O

₇

and CaK

₂

P

₂

O

₇

. Phosphates associated with H in the fuel are likely the initial source for these phosphates because, upon heating, orthophosphate salts containing H are dehydrated, leading to the loss of H and O and subsequent formation of pyrophosphates and metaphosphates.

^46,47

The formation of metaphosphates cannot be veri fied in the present work because no metaphosphate was identi fied in the ash but has previously been identified in combustion ashes.

³⁵

Thus, for blends containing PA or DG, the initial stages of fuel conversion would likely result in the formation of ortho-, pyro-, and metaphosphates associated with K, Ca, and Mg. Furthermore, the low share of P in the fine fly ash particles of <1 μm would suggest that these initial phosphates are solid or liquid rather than gaseous at the process temperature.

A recently developed method for quantifying various forms

of P in solid fuels indicates that, in biosolids (i.e., sewage

sludge), acid-soluble inorganic P dominates ( ∼80 wt %),

followed by acid-insoluble inorganic P ( ∼10 %) and a minor

share of organic P ( ∼5 wt %).

⁴⁸

The major share of P in the SS

blends is expected to be inorganically associated salts with a smaller share of organically associated phosphates.

²⁶⁻²⁹

Because SS was coagulated with iron(II) sulfate (FeSO

₄

), the inorganic fraction likely contains a high share of Fe orthophosphates, such as vivianite, but may also include orthophosphates associated with Ca, Mg, and Al.

23,29,43,49,50

On the basis of the previous reasoning, the initial stages of fuel conversion would mainly result in orthophosphates associated with Fe and possibly Ca, Mg, and Al.

Even though the fuels were thoroughly mixed and co- pelletized, the composition of bed ash particles would suggest that the initial fuel conversion resulted in ash particles that had a composition that more closely resembled the P-poor fuel or the P-rich fuel rather than the mixture between them. Some interaction between ash compounds in the P-poor and P-rich fuels can be observed considering the increase in the K concentration of bed ash particles with >5 mol % P in comparison to the P-rich fuel (Figure 2). However, considering the similarities in the phase composition of phosphates from the bed ash particles of the P-poor and P-rich blends, the interaction between ash compounds from the P-poor fuel and P-rich fuel is limited.

The phase composition of wind-side deposits and cyclone ashes would suggest a more significant ash interaction between the P-poor fuel and the P-rich fuel. With increasing concentrations of Ca to P in the fuel blends with PA or DG, the share of Ca orthophosphates [Ca

₅

(PO

₄

)

₃

OH and Ca

₃

(PO

₄

)

₂

] and mixed Ca and Mg −K orthophosphates (CaKPO

₄

and KMgPO

₄

) increased in the wind-side deposit and cyclone ash. These phases are likely the results of ash transformation reactions between the initial meta- and pyrophosphate compounds and Ca-containing compounds in the P-poor biomass. By comparison, only Ca-rich orthophos- phates [Ca

₅

(PO

₄

)

₃

OH and Ca

₃

(PO

₄

)

₂

] were detected in the wind-side deposits and cyclone ashes from the SS blends.

These results are consistent with the results from studies of thermochemical treatment of SS or combustion of SS with the addition of Ca additives.

^6,51

This suggests that the phosphates that formed during initial fuel conversion mainly interacted with Ca compounds in subsequent ash transformation reactions, yielding increasingly Ca-rich phosphates.

Thus, the di fferences in the detected crystalline phosphate phases between blends with SS and PA or DG can be explained by di fferences in the chemical association of inorganic elements in the fuel and kinetic limitations during fuel conversion. The initial phosphates that form during fuel conversion likely have a chemical association that is similar to the original fuel. If signi ficant kinetic barriers exist for the reaction between the ashes in the fuel blend, which was the case in the bed ash particles, the phosphates retained a phase composition that was similar to the chemical association of the P-rich biomass.

In the wind-side deposits and cyclone ashes, phosphate became increasingly Ca-rich with an increasing content of Ca in the fuel blend, suggesting a more signi ficant interaction between the inorganic elements of the P-poor and P-rich fuels.

This implies that it might be challenging to alter the chemical association of P in SS ashes toward K-containing phosphates by co-combustion or using fuel additives in typical fluidized bed combustion conditions. Because phosphates in the ashes from co-combustion retain a chemical association similar to SS ashes, the plant availability would likely be poor, similar to SS ashes.

¹⁰

Therefore, co-combusted SS ashes from a fluidized

bed would probably require further processing by wet chemical methods or thermochemical treatment with additives to be e ffectively used as a fertilizer.

5. CONCLUSION

The bed ash and cyclone ash contained the majority of fuel P for all fuel blends, suggesting that the P distribution between ash fractions is less dependent upon the fuel ash composition or the chemical association of P in the fuel. At similar combustion conditions and ash composition concerning K, Ca, Mg, and P, the speciation of P in the coarse ash fractions was di fferent in biomass blends containing sewage sludge compared to phosphoric acid or dried distiller ’s grain with solubles. In ash fractions from blends with phosphoric acid or only plant-based biomasses, a wide range of Ca, Mg, and K phosphates was identi fied in the ash, with significant variation depending upon the ash composition. By comparison, only Ca- rich phosphates were identi fied in the coarse ash fractions from sewage sludge blends, independent of the ash composition. On the basis of the elemental composition of the analyzed ash fractions, di fferences in the elemental composition of ash fractions were insu fficient to explain the difference in P speciation between the fuel blends.

The chemical characterization of ash fractions suggests that the ash transformation reactions of P are dependent upon the chemical association in the fuel. By comparison to the other fuel blends, P in the sewage sludge blends was less prone to form K-containing phosphates for a given fuel ash composition.

For the P-rich fuel blends, the di fference in P association in the fuel had a signi ficant effect on the types of crystalline phosphate phases that were detected in the coarse ash fractions. The di fference was smaller between the P-poor fuel blends because, in both cases, the formation of Ca orthophosphates was favored.

■ AUTHOR INFORMATION

Corresponding Author

Joel Falk − Energy Engineering, Department of Engineering Sciences and Mathematics, Luleå University of Technology, SE- 97187 Luleå, Sweden; orcid.org/0000-0003-3738-555X;

Phone: +46-70-6068331; Email: joel.falk@ltu.se

Authors

Nils Skoglund − Thermochemical Energy Conversion Laboratory, Department of Applied Physics and Electronics, Umeå University, SE-90187 Umeå, Sweden; orcid.org/

0000-0002-5777-9241

Alejandro Grimm − Department of Forest Biomaterials and Technology, Swedish University of Agricultural Sciences, SE- 90183 Umeå, Sweden; orcid.org/0000-0001-8502-8069 Marcus Öhman − Energy Engineering, Department of

Engineering Sciences and Mathematics, Luleå University of Technology, SE-97187 Luleå, Sweden

Complete contact information is available at:

https://pubs.acs.org/10.1021/acs.energyfuels.9b03975

Notes

The authors declare no competing financial interest.

■ ACKNOWLEDGMENTS

The financial support from the Swedish Research Council for

Environment, Agricultural Sciences and Spatial Planning

(FORMAS, Grant 942-2015-619) is gratefully acknowledged.

Further, Nils Skoglund gratefully acknowledges the financial support from the Swedish Research Council (Grant 2017- 05331).

■

(1) Steffen, W.; Richardson, K.; Rockstrom, J.; Cornell, S. E.; Fetzer,

^REFERENCES

I.; Bennett, E. M.; Biggs, R.; Carpenter, S. R.; de Vries, W.; de Wit, C.

A.; Folke, C.; Gerten, D.; Heinke, J.; Mace, G. M.; Persson, L. M.;

Ramanathan, V.; Reyers, B.; Sorlin, S. Planetary Boundaries: Guiding Human Development on a Changing Planet. Science 2015, 347 (6223), 1259855.

(2) Cordell, D.; Drangert, J.; White, S. The Story of Phosphorus:

Global Food Security and Food for Thought. Glob. Environ. Chang.

2009, 19, 292−305.

(3) Desmidt, E.; Ghyselbrecht, K.; Zhang, Y.; Pinoy, L.; Van Der Bruggen, B.; Verstraete, W.; Rabaey, K.; Meesschaert, B. Global Phosphorus Scarcity and Full-Scale P-Recovery Techniques: A Review. Crit. Rev. Environ. Sci. Technol. 2015, 45 (4), 336−384.

(4) Alabaster, G. P.; Leblanc, R. J.; Mattews, P.; Richard, R. P.

Global Atlas of Excreta, Wastewater Sludge, and Biosolids Management:

Moving Forward Wastewater Biosolids Sustainability Technical Mana- gerial and Public Synergy; Leblanc, R. J., Mattews, P., Richard, R. P., Eds.; United Nations Human Settlements Programme (UN- HABITAT): Nairobi, Kenya, 2008.

(5) Singh, R. P.; Agrawal, M. Potential Benefits and Risks of Land Application of Sewage Sludge. Waste Manage. 2008, 28, 347−358.

(6) Adam, C.; Peplinski, B.; Michaelis, M.; Kley, G.; Simon, F.-G.

Thermochemical Treatment of Sewage Sludge Ashes for Phosphorus Recovery. Waste Manage. 2009, 29 (3), 1122−1128.

(7) Narodoslawsky, M.; Obernberger, I. From Waste to Raw MaterialThe Route from Biomass to Wood Ash for Cadmium and Other Heavy Metals. J. Hazard. Mater. 1996, 50, 157−168.

(8) Møller, H. B.; Jensen, H. S.; Tobiasen, L.; Hansen, M. N. Heavy Metal and Phosphorus Content of Fractions from Manure Treatment and Incineration. Environ. Technol. 2007, 28 (12), 1403−1418.

(9) Krüger, O.; Grabner, A.; Adam, C. Complete Survey of German Sewage Sludge Ash. Environ. Sci. Technol. 2014, 48 (20), 11811− 11818.

(10) Krüger, O.; Adam, C. Recovery Potential of German Sewage Sludge Ash. Waste Manage. 2015, 45, 400−406.

(11) Petzet, S.; Peplinski, B.; Cornel, P. On Wet Chemical Phosphorus Recovery from Sewage Sludge Ash by Acidic or Alkaline Leaching and an Optimized Combination of Both. Water Res. 2012, 46 (12), 3769−3780.

(12) Kalmykova, Y.; Karlfeldt Fedje, K. Phosphorus Recovery from Municipal Solid Waste Incineration Fly Ash. Waste Manage. 2013, 33 (6), 1403−1410.

(13) Takahashi, M.; Kato, S.; Shima, H.; Sarai, E.; Ichioka, T.;

Hatyakawa, S.; Miyajiri, H. Technology for Recovering Phosphorus from Incinerated Wastewater Treatment Sludge. Chemosphere 2001, 44 (1), 23−29.

(14) Stemann, J.; Peplinski, B.; Adam, C. Thermochemical Treatment of Sewage Sludge Ash with Sodium Salt Additives for Phosphorus Fertilizer ProductionAnalysis of Underlying Chemical Reactions. Waste Manage. 2015, 45, 385−390.

(15) Boström, D.; Skoglund, N.; Grimm, A.; Boman, C.; Öhman, M.; Broström, M.; Backman, R. Ash Transformation Chemistry during Combustion of Biomass. Energy Fuels 2012, 26 (1), 85−93.

(16) Skoglund, N.; Grimm, A.; Öhman, M.; Boström, D. Effects on Ash Chemistry When Co-Firing Municipal Sewage Sludge and Wheat Straw in a Fluidized Bed: Influence on the Ash Chemistry by Fuel Mixing. Energy Fuels 2013, 27 (10), 5725−5732.

(17) Ren, Q.; Li, L. Co-Combustion of Agricultural Straw with Municipal Sewage Sludge in a Fluidized Bed: Role of Phosphorus in Potassium Behavior. Energy Fuels 2015, 29 (7), 4321−4327.

(18) Wang, L.; Skjevrak, G.; Hustad, J. E.; Skreiberg, Ø.

Investigation of Biomass Ash Sintering Characteristics and the Effect of Additives. Energy Fuels 2014, 28 (1), 208−218.

(19) Åmand, L. E.; Leckner, B.; Eskilsson, D.; Tullin, C. Deposits on Heat Transfer Tubes during Co-Combustion of Biofuels and Sewage Sludge. Fuel 2006, 85 (10−11), 1313−1322.

(20) Elled, A. L.; Davidsson, K. O.; Åmand, L. E. Sewage Sludge as a Deposit Inhibitor When Co-Fired with High Potassium Fuels.

Biomass Bioenergy 2010, 34 (11), 1546−1554.

(21) Aho, M.; Yrjas, P.; Taipale, R.; Hupa, M.; Silvennoinen, J.

Reduction of Superheater Corrosion by Co-Firing Risky Biomass with Sewage Sludge. Fuel 2010, 89 (9), 2376−2386.

(22) Karlsson, S.; Åmand, L.-E.; Liske, J. Reducing High-Temper- ature Corrosion on High-Alloyed Stainless Steel Superheaters by Co- Combustion of Municipal Sewage Sludge in a Fluidised Bed Boiler.

Fuel 2015, 139, 482−493.

(23) Zevenhoven, M.; Yrjas, P.; Skrifvars, B.-J.; Hupa, M.

Characterization of Ash-Forming Matter in Various Solid Fuels by Selective Leaching and Its Implications for Fluidized-Bed Combus- tion. Energy Fuels 2012, 26 (10), 6366−6386.

(24) Clarkson, D. T.; Hanson, J. B. The Mineral Nutrition of Higher Plants. Annu. Rev. Plant Physiol. 1980, 31, 239−298.

(25) Werkelin, J.; Skrifvars, B. J.; Zevenhoven, M.; Holmbom, B.;

Hupa, M. Chemical Forms of Ash-Forming Elements in Woody Biomass Fuels. Fuel 2010, 89 (2), 481−493.

(26) García-Albacete, M.; Martín, A.; Cartagena, M. C. Fractiona- tion of Phosphorus Biowastes: Characterisation and Environmental Risk. Waste Manage. 2012, 32, 1061−1068.

(27) Frossard, E.; Tekely, P.; Grimal, J. Y. Characterization of Phosphate Species in Urban Sewage Sludges by High-Resolution Solid-State³¹P NMR. Eur. J. Soil Sci. 1994, 45, 403−408.

(28) Pardo, P.; López-Sánchez, J. F.; Rauret, G. Relationships between Phosphorus Fractionation and Major Components in Sediments Using the SMT Harmonised Extraction Procedure. Anal.

Bioanal. Chem. 2003, 376, 248−254.

(29) Frossard, E.; Bauer, J. P.; Lothe, F. Evidence of Vivianite in FeSO₄-Flocculated Sludges. Water Res. 1997, 31 (10), 2449−2454.

(30) Eriksson, G.; Grimm, A.; Skoglund, N.; Boström, D.; Öhman, M. Combustion and Fuel Characterisation of Wheat Distillers Dried Grain with Solubles (DDGS) and Possible Combustion Applications.

Fuel 2012, 102, 208−220.

(31) Eriksson, G.; Hedman, H.; Boström, D.; Pettersson, E.;

Backman, R.; Öhman, M. Combustion Characterization of Rapeseed Meal and Possible Combustion Applications. Energy Fuels 2009, 23 (8), 3930−3939.

(32) Boström, D.; Eriksson, G.; Boman, C.; Öhman, M. Ash Transformations in Fluidized-Bed Combustion of Rapeseed Meal.

Energy Fuels 2009, 23 (5), 2700−2706.

(33) Thomsen, T. P.; Sárossy, Z.; Ahrenfeldt, J.; Henriksen, U. B.;

Frandsen, F. J.; Müller-Stöver, D. S. Changes Imposed by Pyrolysis, Thermal Gasification and Incineration on Composition and Phosphorus Fertilizer Quality of Municipal Sewage Sludge. J. Environ.

Manage. 2017, 198, 308−318.

(34) Grimm, A.; Skoglund, N.; Boström, D.; Öhman, M. Bed Agglomeration Characteristics in Fluidized Quartz Bed Combustion of Phosphorus-Rich Biomass Fuels. Energy Fuels 2011, 25 (3), 937− 947.(35) Lindström, E.; Sandström, M.; Boström, D.; Öhman, M.

Slagging Characteristics during Combustion of Cereal Grains Rich in Phosphorus. Energy Fuels 2007, 21 (2), 710−717.

(36) Skoglund, N.; Grimm, A.; Öhman, M.; Boström, D.

Combustion of Biosolids in a Bubbling Fluidized Bed, Part 1: Main Ash-Forming Elements and Ash Distribution with a Focus on Phosphorus. Energy Fuels 2014, 28 (2), 1183−1190.

(37) Grimm, A.; Skoglund, N.; Boström, D.; Boman, C.; Öhman, M.

Influence of Phosphorus on Alkali Distribution during Combustion of Logging Residues and Wheat Straw in a Bench-Scale Fluidized Bed.

Energy Fuels 2012, 26 (5), 3012−3023.

(38) Öhman, M.; Nordin, A. A New Method for Quantification of Fluidized Bed Agglomeration Tendencies: A Sensitivity Analysis.

Energy Fuels 1998, 12 (1), 90−94.

(39) International Centre for Diffraction Data (ICDD). Powder Diffraction File (PDF)-2 Database; ICDD: Newtown Square, PA, 2004; Release 52.

(40) Hellenbrandt, M. The Inorganic Crystal Structure Database (ICSD)Present and Future. Crystallogr. Rev. 2004, 10 (1), 17−22.

(41) Pettersson, A.; Åmand, L. E.; Steenari, B. M. Leaching of Ashes from Co-Combustion of Sewage Sludge and WoodPart 1: Recovery of Phosphorus. Biomass Bioenergy 2008, 32 (3), 236−244.

(42) Nanzer, S.; Oberson, A.; Huthwelker, T.; Eggenberger, U.;

Frossard, E. The Molecular Environment of Phosphorus in Sewage Sludge Ash: Implications for Bioavailability. J. Environ. Qual. 2014, 43 (3), 1050−1060.

(43) Jiang, J.; Graham, N. J. D. Pre-Polymerised Inorganic Coagulants and Phosphorus Removal by CoagulationA Review.

Water SA 1998, 24 (3), 237−244.

(44) Piotrowska, P.; Zevenhoven, M.; Hupa, M.; Giuntoli, J.; De Jong, W. Residues from the Production of Biofuels for Trans- portation: Characterization and Ash Sintering Tendency. Fuel Process.

Technol. 2013, 105, 37−45.

(45) Mjoun, K.; Kalscheur, K. F.; Hippen, A. R.; Schingoethe, D. J.

Ruminal Phosphorus Disappearance from Corn and Soybean Feedstuffs. J. Dairy Sci. 2008, 91 (10), 3938−3946.

(46) Hughes, D. M.; Grégoire, D. C.; Naka, H.; Chakrabarti, C. L.

The Vaporization of Phosphorus Compounds and the Use of Chemical Modifiers for the Determination of Phosphorus by Electrothermal Vaporization Inductively Coupled Plasma Mass Spectrometry. Spectrochim. Acta, Part B 1997, 52 (4), 517−529.

(47) Abdel-Kader, A.; Saleh, S. I. Thermal Analysis of Dihydrated Sodium. Thermochim. Acta 1991, 181, 203−214.

(48) Chen, X.; Liaw, S. B.; Wu, H. A New Method for Quantifying Phosphorus of Various Occurrence Forms in Solid Fuels. Energy Fuels 2019, 33 (4), 3311−3321.

(49) Wilfert, P.; Kumar, P. S.; Korving, L.; Witkamp, G. J.; Van Loosdrecht, M. C. M. The Relevance of Phosphorus and Iron Chemistry to the Recovery of Phosphorus from Wastewater: A Review. Environ. Sci. Technol. 2015, 49 (16), 9400−9414.

(50) Wilfert, P.; Mandalidis, A.; Dugulan, A. I.; Goubitz, K.; Korving, L.; Temmink, H.; Witkamp, G. J.; Van Loosdrecht, M. C. M. Vivianite as an Important Iron Phosphate Precipitate in Sewage Treatment Plants. Water Res. 2016, 104, 449−460.

(51) Li, R.; Zhang, Z.; Li, Y.; Teng, W.; Wang, W.; Yang, T.

Transformation of Apatite Phosphorus and Non-Apatite Inorganic Phosphorus during Incineration of Sewage Sludge. Chemosphere 2015, 141, 57−61.