Waste Management 34 (2014) 817–824

Contents lists available at ScienceDirect

Waste Management journal homepage: www.elsevier.com/locate/wasman

On the understanding and control of the spontaneous heating of dried tannery wastewater sludge A. Biasin a, M. Della Zassa a, M. Zerlottin b, D. Refosco b, R. Bertani a, P. Canu a,⇑ a b

Dept. of Industrial Engineering, University of Padua, Via Gradenigo, 5/a, 35131 Padova, Italy Acque del Chiampo, SpA, Via Ferraretta, 20, 36071 Arzignano (Vi), Italy

a r t i c l e

i n f o

Article history: Received 8 September 2013 Accepted 31 December 2013 Available online 29 January 2014 Keywords: Spontaneous heating Dried sludge Self-heating Autoignition Iron/sulphur Tannery waste

a b s t r a c t We studied the spontaneous heating of dried sludge produced by treating wastewater mainly originating from tanneries. Heating up to burning has been observed in the presence of air and moisture, starting at ambient temperature. To understand and prevent the process we combined chemical and morphological analyses (ESEM) with thermal activity monitoring in insulated vessels. Selective additions of chemicals, either to amplify or depress the reactivity, have been used to investigate and identify both the chemical mechanism causing the sludge self-heating, and a prevention or a mitigation strategy. FeS additions accelerate the onset of reactivity, while S sustains it over time. On the contrary, Ca(OH)2, Na2CO3, NaHCO3, FeCl2, EDTA, NaClO can limit, up to completely preventing, the exothermic activity. All the experimental evidences show that the reactions supporting the dried sludge self-heating involve the Fe/S/O system. The total suppression of the reactivity requires amounts of additives that are industrially incompatible with waste reduction and economics. The best prevention requires reduction or removal of S and Fe from the dried solid matrix. Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction After the occurrence of unusual temperature in a dedicated landfill for exclusive disposal of dried sludge from tannery wastewater treatment, an extensive investigation to evaluate the hazard, the self-heating causes and the prevention strategies has been carried out. We already reported (Zerlottin et al., 2013) a large scale experimental study, based on 1 m3 big-bags, where we deterministically reproduced the self-heating processes, observing two distinct behaviors, depending on the drying operations. Either the sludge mass increases its temperature up to a maximum of approx. 90 °C before cooling or an unbounded temperature rise leads to self-combustion, without any flame. The second case is extremely critical for safety and pollution (Nammari et al., 2004), but we clearly proved that it occurs only after an uncontrolled drying operation. Yet, some heating is always observed, but in most cases it is weak enough to independently extinguish in a few days, depending on the heat exchange rate of the sludge mass and its aeration. The sequence of a faint heating that may trigger total combustion has been reported by many authors, on wastewater sludge (Poffet et al., 2008), as wells as other solid wastes (Moqbel et al., 2010; Hogland and Marques, 2007, Hogland et al., 2009; Buggeln and Rynk, 2002; Shimizu et al., 2009), and solid fuels, typ-

⇑ Corresponding author. Tel.: +39 049 8275463; fax: +39 049 8275461. E-mail address: [email protected] (P. Canu). 0956-053X/$ - see front matter Ó 2014 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.wasman.2013.12.023

ically coal (Carras and Young, 2004; Nelson and Chen, 2007; Ribeiro et al., 2010; Phillips et al., 2011; Day, 2000). We devised a procedure (Della Zassa et al., 2013) to reproduce the initial self-heating process at the laboratory scale, to manipulate either the solids or its environment, aiming at investigating the mechanism. So far we reported the enhancing (or damping) effect of aeration, moisture content (also by addition of water) particle size, bed and particle porosity and biological activity (Della Zassa et al., 2013). Apparently, fermentation is rather marginal in the heating process (Della Zassa et al., 2013; Li et al., 2006); a chemical route is clearly prevailing, as confirmed by spontaneous heating and combustion of solids that have little or no putrescible components, such as coals (Carras and Young, 2004; Nelson and Chen, 2007; Ribeiro et al., 2010; Phillips et al., 2011; Day, 2000), refuse derived fuel (RDF) (Yasuhara et al., 2010; Fu et al., 2005; Li et al., 2008) and refuse plastic/paper fuel (RPF) (Li et al., 2009). In most of these cases, the inorganic components are deemed responsible of the self-heating potential (Sujanti and Zhang, 1999), in addition to moisture that is always a key factor. Here we aim at better understanding the chemical mechanism that supports the self-heating process, beyond simple thermal monitoring, to include chemical and morphological analysis of the solids, mainly performed by ESEM–EDS technique, as well as discussing the effect of chemical additives expected to inhibit or delay the self-heating process. Finally, the process understanding’s goal is the identification of prevention policies, including

818

A. Biasin et al. / Waste Management 34 (2014) 817–824

modifications of the sludge production process, as well as remediation actions, in case of uncontrolled heating in large storages. 2. Materials and methods 2.1. Materials All tests used dried sludge as produced by the same industrial drying plant (Zerlottin et al., 2013), which processes sludge resulting from the treatment of wastewater of a large tannery district and a minor amount (approx. 10%) of municipal sewage wastewater. Dried sludge appears as a coarse powder, made of brittle granules with an average particle size of 2 mm and a residual moisture content varying in the 8–15 wt% range. It is a complex solid matrix that is affected both by the large quantity (and variety) of chemicals used in the tanning industry and by the additives introduced during the water treatment process. The heterogeneity (and seasonality) of the original wastewater contributes to this complexity as well. The dried sludge is routinely analyzed to comply with legal requirements in order to be landfilled, but the analyses typically used are not specific enough to identify the chemical route leading to self-heating. Proximate analysis reports, beside moisture (by drying at 105 °C for 24 h), a large fraction of volatiles, leaving residual solids for approx. 25–30% of the initial weight (after incineration at 600 °C for 4 h). The elemental analysis has been already reported (Della Zassa et al., 2013). Milli-Q water (18.0 MX cm 1) generated from a NANOpure Diamond purifying system was used for all experiments. Ca(OH)2, CaCO3, Na2CO3, NaHCO3, S, FeS, FeCl2, NaClO and EDTA (Ethylenediaminetetraacetic acid) were used as additive in dedicated experiments. All were used as purchased (Sigma–Aldrich). 2.2. Batch experiments The dried sludge spontaneous heating always occurs immediately downstream of the drying process, as previously described (Zerlottin et al., 2013) but usually the heating rate is rather weak. Factors like the heat exchange rate of the sludge mass and its aeration affect the rate and intensity of self-heating. Simple techniques to bring the dried sludge to the laboratory preserving its reactivity have been adopted, by keeping the sludge samples under vacuum (Della Zassa et al., 2013). At the laboratory scale, the reactivity of the sludge has been determined by batch experiments in thermally insulated vessels (adiabatic tests), where temperature is continuously monitored over long periods of time (Della Zassa et al., 2013). Typically, a reference amount of dried sludge (250 g) was loaded inside an insulated flask (sometimes called Dewar) and the core temperature of the sample mass was monitored using K type thermocouples, approx. at its center. Data were continuously logged by a computer, through commercial data acquisition boards (Measurement Computing) The dried sludge has been studied either alone or mixed with additives, always using the same reference amount of dried sludge and a water quantity of 50% of the initial sludge sample’s mass, approaching a final moisture content of approx. 40 wt% (unless specified differently). Water addition and aeration (the upper surfaces of the insulated vessels were always kept open to air) are requirements to trigger the spontaneous heating, although their role in the mechanism was not clear, so far. Additives that have been tested include Ca(OH)2, CaCO3, Na2CO3, NaHCO3, S, FeS, FeCl2, NaClO and EDTA (Ethylenediaminetetraacetic acid). Because the process of self-heating is triggered by water, we tried to formulate all the mixtures of sludge and additives as powders, although the contact is less effective. The water content was then determined by our intentional addition. Additive amounts vary between 5% and 30% by weight of the initial

sludge. That was the case of all the additives, independently of the solubility, except for NaClO and EDTA that have been added as water solutions by spraying, over the sludge sample spread on a horizontal surface as a thin layer. In the case of solids mixtures, water was always added by spraying right before loading the mixture in the insulated vessel. It is noteworthy that other addition procedures can be envisaged (Sujanti and Zhang, 1999), including deionized water or acidwashing to remove most of the inorganic components, soaking the solids in the solution carrying the additives, followed by filtering and drying of the treated solids. We avoided these alternative procedures to prevent affecting the sample structure, both chemically and physically, as well as its reactivity. A large number of tests have been carried out, lasting between 2 weeks and a few months, spanning a period of almost 1 year, notwithstanding some parallelization of the tests. The drying plant yields a product with features varying at high- (days) and low(seasonal) frequency. Accordingly, the properties of collected samples of dried sludge gradually vary in time, so that new batches must be periodically collected. For these reasons, (i) comparative tests have been carried out simultaneously, on the same batch of sludge and (ii) the untreated sludge behavior reported in different figures may slightly vary because it originates from different sampling batches. 2.3. Analytical methods The main criteria to quantify the sludge thermal activity is always a time record of temperature. In addition, proximate analysis of the sludge has been carried out with conventional ovens and analytical balances. Solids residues have been analyzed for elemental composition by ICP (Perkin Elmer OES OPTIMA 7300 DV). Sludge microstructure and composition before and after the self-heating process have been studied by ESEM (environmental scanning electron microscope, Philips XL 30-EDS embedded) at 20–30 kV. 3. Results and discussion 3.1. ESEM analyses Scanning electron microscopy turned out to be quite useful both for the local elemental composition and the morphology, although limited to a shallow surface layer. Elemental analysis is obtained on superficial spots of variable size by X-ray fluorescence emissions using an Energy Dispersive Spectrometer (EDS). The composition is averaged on the spot area; a careful selection is required. Significant differences in the sludge samples, before and after the heat generation, have been observed. The elemental analyses averaged over a number of local surface spots of about 1 mm2 are reported in Table 1. Results are fairly consistent with ICP measurements already reported (Della Zassa et al., 2013). Carbon and oxygen prevail, while nitrogen, chromium, sulfur, calcium and iron are the most relevant minor species. The analysis of post-heating sludge reported in Table 1 was representative of an average result; slightly different composition might be observed depending on the intensity and extent of the self-heating phase. Still, we observe that the self-heating activity causes a decrease of carbon (approx. 10 wt%) and sulfur (approx. 40 wt%) while oxygen increases (approx. +10 wt%, it reflects the addition of water and the aeration). A very peculiar feature is the presence in all the samples of dried sludge, before the heating activity, of regular orthorhombic bi-pyramidal crystals (Fu and Manthiram, 2012; Meyer, 1976), as

819

A. Biasin et al. / Waste Management 34 (2014) 817–824 Table 1 Elemental analyses of sludge samples surface by ESEM–EDS, before (sample C) and after (sample E) a self-heating cycle. Values in wt% and atomic % in parentheses. Type

C

N

O

Na

Mg

Al

Si

P

S

Cl

K

Ca

Cr

Fe

Ti

Dried

46.2 (59.0) 41.7 (53.2)

3.8 (4.0) 4.0 (4.5)

30.6 (29.2) 34.8 (34.2)

0.6 (0.4) 0.5 (0.3)

0.3 (0.2) 0.5 (0.2)

0.7 (0.4) 0.8 (0.4)

0.8 (0.2) 1.1 (0.5)

0.7 (0.3) 0.9 (0.4)

3.3 (1.6) 1.9 (0.9)

0.4 (0.1) 0.7 (0.3)

0.2 (0.07) 0.2 (0.08)

5.5 (2.0) 6.3 (2.4)

4.6 (1.3) 5.7 (1.4)

2.2 (0.6) 2.9 (0.8)

0.5 (0.1) 0.9 (0.2)

Post-heating

shown in Fig. 1 (a) and (b), quite large in size (approx. 20 lm). They have been unambiguously identified as pure sulfur by the EDS microanalysis, Fig. 1 (c). The crystals are quite abundant in the dried sludge as received, but they disappear after spontaneous heating, also in its most common moderate form that did not evolve into combustion. This is shown in Fig. 3 where the sludge surface, before and after selfheating, are compared. Also, Table 1 shows that sulfur is significantly consumed during the low temperature spontaneous heating. The remaining sulfur, not detected as crystals anymore, is at least partially converted to sulfate. Indeed, aggregates of hydrated calcium sulfate (gypsum) were observed on the surface of postheating sludge, Fig. 2. Quite interestingly, in our previous work (Della Zassa et al., 2013) we compared the spontaneous heating tendency of dried sludge from several production plants; the only one that did not show any significant temperature increase was also found lacking S in its matrix. We believe that the chemistry related to S in these materials is the key to understand the self-heating behavior. The ESEM images (not reported) also show the formation of microorganisms after long periods of spontaneous heating at mild temperature. That is quite likely, because of the relatively large amount of moisture (50% of the initial sludge weight), temperature (an average of approx. 60 °C) and time frame (months). 3.2. Effects of chemical additives The characterization of the sludge by means of analytical techniques suggested that self-heating process originated from some

chemical reactions that involved sulfur, oxygen and possibly iron. The complexity of the solid matrix limited the potential of such chemical analyses. Consequently, we arranged a set of macroscopic chemical tests, by selectively adding chemicals to the sludge, monitoring the effect on the self-heating behavior. Mixing of additives with sludge powder depends on their phase, either liquid or solid, as explained in the Materials and Methods section. 3.2.1. Sulfur related additives The possibility of sulfur-containing species to be involved in exothermic processes, leading to self-heating phenomena, has already been suggested in the literature referring to sulfide minerals (Poffet et al., 2008; Iliyas et al., 2010), mainly iron sulfides (Zhao et al., 2006; Soundararajan et al., 1996). ESEM analyses revealed the existence of elemental sulfur in the dried sludge, which disappeared after spontaneous heating. Total sulfur amount has been measured by ICP being about 2% of the moisture free dried sludge and it originated from the influent waste. However, in the sludge before drying it was expected as H2S and sulfides, which were precipitated as FeS with the intentional addition of FeCl2 in the plant (Zerlottin et al., 2013) before dewatering. Elemental sulfur in the dried sludge is quite surprising, but clearly proved by ESEM analysis (Figs. 1 and 3) and mentioned in the literature (Ros et al., 2006) as well. To spot the potential role of S and FeS, we forced it by adding extra amounts to the sludge and monitoring its spontaneous heating in a standard test in insulated vessels, as described in a previous work (Della Zassa et al., 2013). We added pure sulfur to standard dried sludge, 15 wt% of the initial sludge mass, then wetted with 50% water as usual. The

Fig. 1. ESEM images from sample of dried sludge. Evidence of crystalline sulfur. (a) Secondary electron analysis, (b) backscattered electron analysis and (c) EDS spectrum of the crystals.

820

A. Biasin et al. / Waste Management 34 (2014) 817–824

Fig. 2. Aggregates of hydrated calcium sulfate (gypsum) after self-heating and their EDS spectra allowing for chemical identification.

Fig. 3. Sludge surface appearance before (a) and after (b) self-heating.

test has been replicated on a new batch of sludge, with 30 wt% of S, and again with 30% FeS. All the results of the self-heating process are compared in Fig. 4 together with those of the untreated sludge (only 50% water was added). It can be observed that elemental sulfur had an effect mainly on the duration of the self-heating process, not affecting its intensity neither its initial rate. It could suggest that sulfur is a limiting reagent for the exothermic reactions: the greater the amount, the longer the time over which the sludge continues to generate heat, in a single hydration step. However, sulfur did not increase the average temperature during heating, which was lower compared to standard tests carried out with addition of water only. These features suggest that sulfur supports some reactions in the process, while limiting others. As for the temperature and the duration of self-heating process it must be considered the increase in total mass amount, due to the substantial sulfur addition (15–30 wt%),

Fig. 4. Self-heating processes carried out in insulated vessels of standard dried sludge after the addition of 50% water and of S and FeS.

which can affect the thermal inertia and the heat dissipation rate of the sample mass. On the other hand, it is striking the effect of adding FeS, which was able to immediately trigger the temperature rise, totally removing the initial latency interval (ranging between 1 and 2 days), while not significantly affecting the total amount of heat generated (see the maximum temperature achieved). It could be argued that the addition of FeS was quantitatively large, compared to the expected amount in the sludge. In addition, FeS produced directly within the sludge, during its treatment in the wastewater plant, was intimately dispersed in the solid, in contrast with the powder mixed with difficulty to the sludge, the latter being less reactive. Still, the behavior observed in Fig. 4 suggests a relevant role of FeS in the thermal activation process of the sludge which has been confirmed by SEM images of the surface of a reacted sludge after the addition of FeS. The formation of iron sulfate and iron oxide species has been observed with the disappearance of sulfur crystals. In Fig. 5 the typical result of an oxidation process on a sludge containing iron sulfides is shown: iron oxides (Fe2O3, hematite and/or Fe3O4, magnetite, depending on the temperature and oxygen concentration) together with sulfates (mainly ferrous, FeSO4 and ferric sulfate, Fe2(SO4)3) can be clearly spotted, as already reported for similar materials (Hu et al., 2006). 3.2.2. Addition of FeCl2 To investigate the role of iron in the exothermic process, independently from sulfur, the addition of iron species has been considered. Ferrous chloride, FeCl2, which was already used in previous stages of the wastewater purification process, has been added further. Because of the ferrous chloride solubility in water, a more effective contact with the solid matrix was expected. Two tests,

A. Biasin et al. / Waste Management 34 (2014) 817–824

821

itates, (FeO(OH)1 xClx), where four double chains of FeO3(OH)3 octahedral form a tunnel partly occupied by 1–9% of Cl , essential for the structural stability, which are known to act as flocculant and absorbent for removal of metal or metalloid elements (Xiong et al., 2008).

Fig. 5. ESEM image of the surface of the sludge mixed with 30% of FeS, after selfheating activity.

with the addition of 5 and 30 wt% of FeCl2 have been carried out and the results compared to the untreated sludge. Results are shown in Fig. 6. The measured heating history was quite surprising. Instead of the expected catalytic enhancement by Fe(II), 30% of FeCl2, inhibited any reactivity within 2 weeks. With the smaller addition of 5% the sludge activation was significantly delayed (more than 1 week instead of the usual 1–2 days), and the exothermic process slightly damped. ESEM analysis, Fig. 7, revealed a dramatic transformation in the sludge surface. The sludge samples treated with 30% of ferrous chloride show a sort of compact crust eventually surrounding the solid matrix. Such a compact coating was made of Fe(OH)2, easily oxidized to Fe(OH)3 as already observed on pyrite (Zhang and Evangelou, 1996; Pérez-López et al., 2007), dramatically limiting the further oxidation rate. Below this skin, sulfur crystals remained visible and intact like those that characterize all the dried sludge samples before the spontaneous heating. It suggests that the water dissolved FeCl2 reacted with the solid matrix, yielding a uniform, impermeable layer of iron hydroxide which limited the permeation of air (i.e. O2), preventing or delaying the onset of the exothermic reactions, now known to consume the sulfur crystals. This feature is a further confirmation of the role of S in the self-heating reaction, as well as a demonstration that it is a surface reaction, extending into the internal surface of the porous matrix. Any limitation to mass transfer on the external surface of the sludge or modification of its porosity is expected to affect its reactivity. The presence of chlorine ions could also induce the formation of akaganéite precip-

3.2.3. Addition of Ca(OH)2 The addition of Ca(OH)2 has been taken into account because it is expected in the dry sludge, being continuously used in the drying plant to facilitate the aggregation of dust in the exhaust gas and to prevent both the formation of acid condensates within the bag filters and their clogging, thus protecting the dusty sludge transfer lines during big-bags filling. Some tests have been performed by adding to the sludge samples Ca(OH)2 in increasing amounts of 5, 15 and 30 wt%, in addition to the usual 50% of activating water. The results are shown in Fig. 8. Increasing amounts of Ca(OH)2 progressively delay the onset of self-heating; also the intensity of the reaction, measured by the extension over time, gradually reduces, up to disappear completely with the largest addition (30%). The expected delivery of heat after hydration (‘wetting heat’) could be clearly observed at the very beginning, and increases with the amount of Ca(OH)2 added. It was reproduced twice, at 30 wt%, as shown in Fig. 9 where the variability due to hydration effectiveness can be appreciated, together with a confirmation of the inert state of the resulting sludge. Interestingly, we get a clear quantitative perception of the wetting heat mechanism, frequently mentioned as a possible self-heating cause (Chen, 1998; Buggeln and Rynk, 2002; Fu et al., 2005; Novack et al., 2001). A large amount (30 wt%) of anhydrous Ca(OH)2 could lead to a relatively modest heating, which shows up immediately after hydration and do not trigger further reactivity. Accordingly, we conclude that the heat of hydration must be a marginal contribution to our dried sludge spontaneous heating. By extension, it must be the case also of other wastes, hardly approaching such a large fraction of inorganics to hydrate. Figs. 8 and 9 demonstrate the suppression of the sludge spontaneous heating by adding 30 wt% of Ca(OH)2. The deactivation was irreversible; further water additions and aeration did not trigger any reaction with measurable thermal effects, differently from what observed on plain dried sludge in the large scale (Zerlottin et al., 2013) and in the same laboratory vessels (Della Zassa et al., 2013). We formulated a few hypothesis about this effect by Ca(OH)2: (i) in its hydration, it could distribute in the sludge pores, blocking the inter-granular oxygen diffusion; (ii) it changed the pH of the reacting environment; and (iii) it contributed Ca2+ e OH ions to other reactions, competing with the exothermic ones. The latter hypothesis agrees with the known behavior of Ca(OH)2 saturated aqueous solutions. Lime water (0.5 wt% aqueous solution) or calcium hydroxide suspensions (lime milk) are widely used as strong bases, with the specific ability to bind several metals in

Fig. 6. Thermal profile of self-heating processes carried out in insulated vessels of standard dried sludge after the addition of 50% water and of FeCl2 (5%, and 30 wt%).

822

A. Biasin et al. / Waste Management 34 (2014) 817–824

Fig. 7. Sludge surface by ESEM, after 30% FeCl2 and 50% water additions and 14 days in insulated vessel. Details of the crust (a–c) and morphology of the underlying matrix (d).

Fig. 8. Self-heating in insulated vessels of dried sludge after the addition of 50% water. Comparison of selective additions of Ca(OH)2: 5%, 15% and 30 wt%.

3.2.4. Addition of different CO23 precursors To develop one of the hypotheses on the role of calcium hydroxide, we investigated the possibility that pH variations in the hydrated sludge could modify the heat generation process. We proceeded by alternatively adding the same conventional amount of 30 wt% of CaCO3, Na2CO3 (mono or deca-hydrate) or NaHCO3. All the additives were powders and mixed with the dry sludge before hydration. The pH measurements reported that the original value of the untreated sludge, 7.8, did not vary with these additions. On the contrary, the self-heating behavior was dramatically different, as shown in Fig. 10. While the addition of CaCO3 (characterized by a low water solubility at ambient temperature) did seem to not affect the typical self-heating behavior of untreated sludge, the samples treated with both type of Na2CO3 and NaHCO3 (both water soluble) appeared irreversibly deactivated. Note that the addition Na2CO3 deca-hydrate activates endothermic processes from the very beginning, that bring the sludge temperature well below the ambient (20 °C), approaching 12 °C. Having proved that pH was

Fig. 9. Reproducibility of the 30% addition of Ca(OH)2.

water and are known to react with sulfur giving rise to lime sulfur used in pest control. Hence, the idea of dissolving Ca(OH)2 in the hydrated sludge, to bind metal ions including the most suspected iron III, and sulfur, too, preventing their participation in exothermic reactions.

Fig. 10. Self-heating in insulated vessels of dried sludge after the addition of 50% water. Comparison of selective additions of 30 wt% of CaCO3 or Na2CO3 or NaHCO3.

A. Biasin et al. / Waste Management 34 (2014) 817–824

not a variable, the different effect of CO23 brought by calcium or sodium, and the NaHCO3 effect, must be otherwise explained. Thanks to their water solubility, one hypothesis on the role of sodium carbonate and bicarbonate was the formation of complex surface structures. It has been reported that the addition of Na2CO3 water solutions to ferrous salts (hydroxide, carbonate, sulfate) promotes the formation of stoichiometric carbonate-containing green rust, determined to be [Fe4(II)Fe2(III)(OH)12][CO3
2H2O], which is stable at approx. pH = 8 (Drissi et al., 1995), yielding an impervious surface layer, similar to the one observed after FeCl2 addition. More interesting is the known property of both Na2CO3 and NaHCO3 to act as oxygen-derived free radical scavengers, also at low temperature. Radicals are thought (Poffet et al., 2008; Moqbel et al., 2010; Li et al., 2009), to be involved in low temperature oxidation. However, their concentration was deemed low enough to assume a marginal support to the exothermic. Still, radical chain reactions can be very productive even with a small radicals concentration, that regenerate in the chain sequence. The addition of radical scavengers could definitely turn off also these reactions (Li et al., 2006). A preliminary experiment carried out by adding a 30% water solution of 2,6-diterbutyl-4-methylphenol as radical trap did not show any difference with respect the thermal behavior exhibited by dried sludge treated only with 50 wt% of water.

3.2.5. Addition of EDTA So far, we believe that metals are the most likely species involved in the early exothermic reactions that eventually might lead to combustion of the organic fraction of the sludge. We tested this hypothesis using a chelating additive, able to sequester metal ions by forming complexes. EDTA was selected, which is known to strongly chelate several metal ions, and specifically Fe2+ and Fe3+; it was added as a 30% water solution. The results are shown in Fig. 11. The results confirmed that a complexation functionality could effectively delay the onset of self-heating (approx. 6 days later) and dampen its intensity (approx. 15–20 °C less). However, it is likely that, over time, the effect of EDTA faded away, reasonably due to the low stability of the iron/EDTA complexes at pH > 6. These may slowly get involved in a sequence of complexation equilibria with other metal ions in the sludge, releasing Fe3+ ions which support the self-heating process (Novack et al., 2001).

3.2.6. Addition of NaClO The effect of the addition of sodium hypochlorite on the thermal behavior of dried sludge has been investigated because it is an oxidant alternative to O2 and contains Cl , which typically inhibits the catalytic activity of transition metals, including iron. Each sludge sample has been added 50 wt% of NaClO aqueous solutions, instead of water. Concentrations of NaClO of 2, 4 and 8 wt% in water have been used. Results are shown in Fig. 12. The inhibition effect of NaClO was evident: it increased with the concentration of NaClO

Fig. 11. Self-heating in insulated vessels of dried sludge after the addition of 50% water. Comparison of additions of EDTA: 0% and 30 wt%.

823

Fig. 12. Self-heating in insulated vessels of dried sludge after the addition of 50 wt% water solution of NaClO. Comparison of different NaClO concentration in the water: 0, 2, 4, 8 wt%.

in the hydrating solution, up to preventing any reaction to occur in 3 weeks, with 8% NaClO in water. Inhibition as shown in Fig. 12 reveals as an increased delay of the reaction onset and attenuation of its rate and extent, corresponding to progressively lower heating rate and maximum temperature achieved. Although sodium hypochlorite solutions above 5 wt% are frequently used as a disinfectant (and this effect is certainly present in the moist sludge) we already proved (Della Zassa et al., 2013) that developing microorganisms and exothermic reactions were uncorrelated. Inhibition should be an effect of the formation of akaganéite precipitates, similarly to what observed in the case of FeCl2 addition.

4. Conclusions This work summarizes an experimental investigation aiming at clarifying the chemical mechanism responsible of the spontaneous heating of dried wastewater treatment sludge (originating mainly from a tannery district) and at providing useful suggestions to control it. We already reported in about plant scale experiments that dramatically revealed the hazard potential, but were constrained by the difficulty of manipulating large quantities, to investigate the mechanism. The process has then been successfully reproduced at the laboratory scale, still with a significant amount of sludge, allowing us to investigate the role of moisture, air, particle size, particle and bed porosity, and biological activity. In this work we investigated the causes of the exothermic phenomena and the opportunity of preventing or mitigating the self-heating process by means of analytical techniques and selective additions of chemicals to the unreacted dried sludge. Chemical and morphological analyses provide evidences of the presence of elemental sulfur crystals before the any reacting activity that consumed after heating, increasing the amount of sulfates. The addition of FeS anticipated the onset of self-heating, but it did not affect significantly the maximum temperature achieved. Additions of pure S extended the heat generation phase proportionally, but they did not modify the typical delay in its onset, differently from FeS. On the contrary, Ca(OH)2, Na2CO3, NaHCO3, FeCl2, EDTA, NaClO could limit, up to completely preventing, the exothermic activity. The increasing additions of FeCl2 depressed the spontaneous heating, until the complete inhibition. This effect has been correlated to the formation of a uniform coating around sludge grains which limited the O2 diffusion to the underlying reacting solid, where unconverted sulfur was found. Ca(OH)2 had an inhibiting effect proportional to the amount added, again reaching the total inhibition of any heating reaction. The most likely reason was its triggering reactions involving Ca2+ and (OH) ions, competitive with possible exothermic

824

A. Biasin et al. / Waste Management 34 (2014) 817–824

mechanisms. CaCO3 was ineffective to modify the reactivity, differently from Na2CO3 and NaHCO3. We did not measure any significant pH changes for these additions; but Na2CO3 and NaHCO3 could scavenge radicals. Again, more effective was the formation of a coating limiting the O2 permeation. EDTA proved effective in delaying and reducing the reactivity (maximum temperature is reduced by about 20 °C), by chelating metals (mainly Fe). However, the chelating effect is fading with time. With an aqueous solution 8 wt% of sodium hypochlorite, the heat generation was completely prevented, reasonably by precipitating iron. All the experimental evidences showed that the reactions supporting the dried sludge self-heating involved the Fe/S/O system. However, the chemical complexity of the dried sludge causes chemical interactions which is worth a systematic investigation. Concerning the prevention and mitigation actions, we demonstrated that the total suppression of the reactivity is possible, but requires amounts of additives that are industrially incompatible with waste reduction and economics. So far, the best prevention practice identified requires the reduction or removal of S and Fe from the dried solid matrix. Unfortunately, all the S and most of Fe are in the incoming wastewater and requires modifications of the tanning processes, which is behind the scope of this work. Alternatives within the wastewater process to remove S from the sludge, before drying, have been evaluated. All the viable options cannot achieve quantitative and economic S removal. Acknowledgement Authors acknowledge Dr. C. Furlan (CUGAS Padova) for ESEM measurements. References Buggeln, R., Rynk, R., 2002. Self-heating in yard trimmings: conditions leading to spontaneous combustion. Compost Sci. Utilization 10, 162–182. Carras, J.N., Young, B.C., 2004. Self-heating of coal and related materials: models, application and test methods. Prog. Energy Combust. Sci. 20, 1–15. Chen, X.D., 1998. On the fundamentals of diffusive self-heating in water containing combustible materials. Chem. Eng. Process. 37, 367–378. Day, R., 2000. Notice to shipmasters loading coal. TP10944E. Minister of Public Works and Government Services Canada. Della Zassa, M., Biasin, A., Zerlottin, M., Refosco, D., Canu, P., 2013. Self-heating of dried industrial wastewater sludge: lab-scale investigation of supporting conditions. Waste Manage. (Oxford) 33, 1469–1477. Drissi, H., Refait, P., Abdelmoula, M., Genin, J.M.R., 1995. The preparation and thermodynamic properties of Fe(II)-Fe(III) hydroxycarbonate (green rust one): pourbaix diagram of iron in carbonate-containing aqueous media. Corros. Sci. 37, 2025–2041. Fu, Y., Manthiram, A., 2012. Orthorhombic bipyramidal sulfur coated with polypyrrole nanolayers as a cathode material for lithium–sulfur batteries. J. Phys. Chem. C116, 8910–8915.

Fu, Z.M., Li, X.R., Koseki, H., 2005. Heat generation of refuse derived fuel with water. J. Loss Prev. Process Ind. 18, 27–33. Hogland, W., Marques, M., 2007. Fires in storage areas for organic waste. In: Proceedings of the International Conference on Sustainable Solid Waste Management, Chennai, India, pp. 189–196. Hogland, W., Lönnermark, A., Marques, M., Persson, H., 2009. Storage of organic materials, solid waste and biofuesl – risks for fires and fire fighting. In: Margherita di Pula, S., (Ed.), Proceedings Sardian 2009, the 12th International Waste Management and Landfill Symposium, Cagliari, Italy. Hu, G., Johansen, K.D., Wedel, S., Hansen, J.P., 2006. Decomposition and oxidation of pyrite. Prog. Energy Combust. Sci. 32, 295–314. Iliyas, A., Hawboldt, K., Khan, F., 2010. Thermal stability investigation of sulfide minerals in DSC. J. Hazard. Mater. 178, 814–822. Li, X.R., Koseki, H., Momota, M., 2006. Evaluation of danger from fermentationinduced spontaneous ignition of wood chips. J. Hazard. Mater. A135, 15–20. Li, X.R., Lim, W.S., Iwata, Y., Koseki, H., 2008. Thermal behavior of sewage sludge derived fuels. Therm. Sci. 12, 137–148. Li, X.R., Lim, W.S., Iwata, Y., Koseki, H., 2009. Thermal characteristics and their relevance to spontaneous ignition of refuse plastics/paper fuel. J. Loss Prev. Process Ind. 22, 1–6. Meyer, B., 1976. Elemental sulfur. Chem. Rev. 76, 367–388. Moqbel, S., Reinhart, D., Chen, R.H., 2010. Factors influencing spontaneous combustion of solid waste. Waste Manage. (Oxford) 30, 1600–1607. Nammari, D.R., Hogland, W., Marques, M., Nimmermark, S., Moutavtchi, V., 2004. Emissions from a controlled fire in municipal solid waste bales. Waste Manage. (Oxford) 24, 9–18. Nelson, M.I., Chen, X.D., 2007. Survey of experimental work on the self-heating and spontaneous combustion of coal. Rev Eng. Geol. 18, 31–83. Novack, B., Kari, F.G., Kruger, H.G., 2001. The remobilization of metals from iron oxides and sediments by metal-EDTA complexes. Water Air Soil Pollut. 125, 243–257. Pérez-López, R., Cama, J., Nieto, J.M., Ayora, C., 2007. The iron-coating role on the oxidation kinetics of a pyritic sludge doped with fly ash. Geochim. Cosmochim. Acta 71, 1921–1934. Phillips, H., Chabedi, K., Uludag, S., 2011. Spontaneous Combustion Prevention and Control. Coaltech Research Association. Poffet, M., Kaeser, K., Jenni, T.A., 2008. Thermal runaway of dried sewage sludge granules in storage tanks. CHIMIA Int. J. Chem. 62, 29–34. Ribeiro, J., Ferreira da Silva, E., Flores, D., 2010. Burning of coal waste piles from Douro coalfield (Portugal): petrological, geochemical and mineralogical characterization. Int. J. Coal Geol. 81, 359–372. Ros, A., Montes-Moran, M.A., Fuente, E., Nevskaia, D.M., Martin, M.J., 2006. Dried sludges and sludge-based chars for H2S removal at low temperature: influence of sewage sludge characteristics. Environ. Sci. Technol. 40, 302–309. Shimizu, Y., Wakakura, M., Arai, M., 2009. Heat accumulations and fire accidents of waste piles. J. Loss Prev. Process Ind. 22, 86–90. Soundararajan, R., Amyotte, P.R., Pegg, M.J., 1996. Explosibility hazard of iron sulphide dusts as a function of particle size. J. Hazard. Mater. 51, 225–239. Sujanti, W., Zhang, D.K., 1999. A laboratory study of spontaneous combustion of coal: the influence of inorganic matter and reactor size. Fuel 78, 549–556. Xiong, H.X., Liao, Y.H., Zhou, L.X., 2008. Influence of chloride and sulfate on formation of akaganeite and schwertmannite through ferrous biooxidation by acidithiobacillus ferrooxidans cells. Environ. Sci. Technol. 42, 8681–8686. Yasuhara, A., Amano, Y., Shibamoto, T., 2010. Investigation of the self-heating and spontaneous ignition of refuse derived fuel (RDF) during storage. Waste Manage. (Oxford) 30, 1161–1164. Zerlottin, M., Refosco, D., Della Zassa, M., Biasin, A., Canu, P., 2013. Self-heating of dried wastewater sludge. Waste Manage. (Oxford) 33 (1), 129–137. Zhang, Y.L., Evangelou, V.P., 1996. Influence of iron oxide forming conditions on pyrite oxidation. Soil Sci. 161, 852–864. Zhao, X., Jiang, J., Meng, Y., 2006. Mechanism and influencing Factors of spontaneous combustion of oil tank containing sulfur. In: International Symposium on Safety Science and Technology, Changsha, China, pp. 1399– 1403.