Effects of initial particle sizes of Triarrhena lutarioriparia on processing performance, material properties, and heavy metal speciation in sewage sludge composting

The purpose of this study was to investigate the effect of initial particle size (IPS) on the environmental parameters and heavy metal speciation during sludge composting. Three piles were conducted: fine material (FM, screen underflow), coarse material (CM, oversize product), and mixed material (MM, mix FM and CM in 1:1). Results showed that the temperature trends of the three piles in different layers were highly repeatable during the thermophilic period. With the decrease of IPS, the heating rate and the highest temperature of the pile increased, the thermophilic period was prolonged, and the highest temperature area in the pile shifted to a lower layer. It also promoted the organic matter degradation, compost maturation, and nitrogen fixation effect. Composting had a good effect on the passivation of heavy metals, especially Cd, Cu, and Pb. The passivation effect on Cd and Cu was FM > CM > MM, and on Pb was CM > FM > MM. Fourier transform infrared spectroscopy, excitation-emission matrix, and thermogravimetric thermal analysis indicated that FM had the highest content of aromatic structure and humic-like substance on D40. The redundancy analysis revealed that MM was beneficial to improve the internal uniformity during composting.


Introduction
Being produced in wastewater treatment, dewatered sludge contains plenty of organic matter (OM) and is rich in nutrients. It is an ideal material for preparing soil amendments (Miranda-Carrazco et al. 2021). Moreover, the source of dewatered sludge is complicated and may contain high levels of heavy metals. Application of these dewatered sludge would aggravate the heavy metal pollutions in soils and water, which would be bound to pose a long-term menace to human health (Liu et al. 2017;Hua et al. 2009). Therefore, dewatered sludge must be properly treated before being applied to the land as biological fertilizer.
Composting, a typical biological treatment technology for dewatered sludge (Jia et al. 2021), has been widely proven to effectively passivate the heavy metals in the materials (Liu et al. 2017;Chen et al. 2019;Zhao et al. 2022). In general, these passivation effects are related to the complexation of humus formed during composting progress, the precipitation of heavy metals under alkaline conditions caused by material mineralization, the adsorption and oxidation of microorganisms, and other factors Xu et al. 2020).
Numerous further studies focused on the exogenous additives for strengthening the heavy metal passivation in the composting process. Some studies used alkaline materials, such as lime, bentonite, and activated alumina balls, to neutralize the organic acids released during composting and to reduce the solubility of metal carbonates and hydroxides, and finally to decrease the toxicity of most heavy metals (Jiang et al. 2014;Liu et al. 2017;Singh and Kalamdhad 2013). There were also some studies on adding porous Responsible Editor: Philippe Garrigues * Jing Huang gavinhj@163.com materials (including clinical stone, ceramsite, and biochar) (Wang et al. 2016Liu et al. 2017). On one hand, porous materials usually have a large specific surface area, a high cation exchange capacity, and abundant oxygen-containing functional groups, which can absorb various types of heavy metals. In addition, they could improve the activity of microorganisms in compost, then promote the mineralization and humification of OM, which indirectly affects the heavy metal passivation. Other studies found that some microbial agents not only directly passivate heavy metals by intracellular absorption or chelation of microorganisms, but also promote compost humification by promoting the degradation of lignocellulose (Zhang et al. 2017b). Therefore, no matter what kinds of exogenous additives have been used in the composting process, in addition to their direct effects on heavy metals, the changes of heavy metal speciation caused by the different organic matter conversion behaviors have been widely confirmed. Given that the characteristics of high viscosity, moisture content (MC), and the low carbon nitrogen (C/N) ratio of sludge, people widely employed the bulking agents (BA), such as corncobs and agricultural wastes to regulate the structure and nutrient of the composting material (Hu et al. 2020), which affects the fermentation effect of compost to a large extent. Hence, the characteristics of BA in sludge compost are bound to have a potential impact on heavy metal speciation. Numerous studies demonstrated that the initial particle size (IPS) of BA can greatly affect the composting process, and change the contact area between materials and microorganisms (Manpreet et al. 2005). Meanwhile, adjusting IPS can change the free air space (FAS) and void dispersion in the piles (Zhou et al. 2013), so water and gas can be fully exchanged between the gas phase and the solid phase. In this way, over-compaction of composting materials can be prevented. Once FAS is too low, the poor pore structure of pile would lead to anaerobic conditions, self-heating failure, and poor treatment (Zhou et al. 2013). Therefore, adjusting the IPS will inevitably affect the OM conversion and the the final compost product's physicochemical characteristics, thereby indirectly affect the heavy metal bioavailability. However, there were few studies on the relationship between sludge composting and heavy metal speciation under the influence of different IPS.
The wetland plants harvested in the process of constructed wetland management are ideal BA for sludge composting (Hu et al. 2020). From 2017 to now, the total construction area of the conversion of farmland to wetlands in Hunan Province, China, is 1136.6 ha, with Triarrhena lutarioriparia (Tl), Phragmites australis, and Typha orientalis as the main plants. This project had remarkable accomplishments in improving water quality, building a strong ecological economy, and enriching biodiversity. Meanwhile, these wetlands plants can be reaped twice a year, and the annual yield of the biomass is about 110,000 t (fresh weight). However, dealing with the large number of wetland plants produced every year has become a new challenge. As the BA and carbon source, T. lutarioriparia composted with the dewatered sludge in different IPS, which provided a feasible way for the resource utilization of wetland plant.
The purposes of our study were as follows: (1) confirm the IPS's influence on the composting process performance and material properties; (2) investigate the changes of heavy metal speciation in sludge composting, and the passivation mechanism; (3) analyses the relationship of the factors in the composting process performance, material properties, and heavy metal speciation with different IPS. The results of this study can help determine the appropriate IPS in sludge composting for actual production, so as to simultaneously promote the process productivity, lower the cost, reduce the heavy metal bioavailability, and realize the resource utilization of dewatered sludge and constructed wetland plants.

Raw materials and composting tests
The dewatered sludge was taken from a wastewater treatment plant in Changsha, China, and the Triarrhena lutarioriparia was harvested from the conversion of farmland to wetlands project in late July. After being air-dried for 2 weeks and crushed twice with a crusher, the material can be divided into two equal parts through a 5.5-mm mesh screen. The dewatered sludge was mechanically mixed with T. lutarioriparia at a 3:2 weight ratio, then added water to about 60% moisture content and thoroughly mixed to obtain three piles: fine material (FM, screen underflow), coarse material (CM, oversize product), and mixed material (MM, mix FM and CM in 1:1). The physical and chemical properties of the dewatered sludge and T. lutarioriparia were presented in Table S1.

Experimental design
Composting tests were conducted on Xinyuan Biological Fertilizer CO., LTD in the local city. Then, we constructed three piles as co-composting reactors, and the actual dimensions were 3.5 × 1.7 × 0.8 m. These piles were installed on a concrete floor with a plastic roof to prevent rainwater from entering the compost. Two groups of temperature sensors (K thermocouples, China) were installed at the height of 20 cm and 60 cm in these piles, and an additional sensor temperature was placed outside piles to obtain environmental temperature. This study recorded these numerical values every 15 min.
The composting process continued for 40 days. It was manually overturned on days 2, 4, 6, 8, 10, 12, and 14 (D2, D4, D6, D8, D10, D12, and D14) to maintain aerobic conditions and high-temperature stages as long as possible after the average temperature of the pile showed a downward trend. Furthermore, the piles were turned once on day 25 (D25) and then remained undisturbed until the end of composting. During the whole composting process, additional tap water had been added into the piles on D4, D8, and D12 to maintain the moisture content at 60%.
We collected a set of samples on D0, D2, D8, D14, D25, and D40. The collected solid samples were divided into two parts. One part was promptly used to measure OM, MC, and germination index (GI). The other part was freeze-dried at − 40℃ for 48 h. The freeze-dried sample was crushed and sieved (100 mesh and 200 mesh, respectively) to determine physical and chemical properties. Gas samples were taken from the height of 0.03 m in the center of piles by a microporous aeration tube connected with an air pump. We collected gas samples by using 1-L sampling bags, which were made from E-Switch®. The filling time was 5 min. One gas sample was taken from each compost pile (collected at 0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 25, 32, and 40 days) to monitor the changes in O 2 and CO 2 during the whole composting process.

Physicochemical analysis
The determination method of O 2 and CO 2 content was shown in the literature (Hu et al. 2020). Samples were extracted at the solid-to-water ratio of 1:10 (w/v, dry weight basis) for the subsequent determinations of pH, EC, and excitationemission matrix (EEM). The values of pH and EC were measured according to the method of Hu et al. (2020). The determination methods of MC, OM, total organic carbon (TOC), total nitrogen (TN), ammonia nitrogen (NH 4 + -N), nitrate nitrogen (NO 3 − -N), humic acid (HA), and fulvic acid (FA) were described by Zhang et al. (2017a), Brittain et al. (2012), and Bohacz (2019). The C/N ratio was the quotient of TOC to TN. The GI was assessed using cabbage seeds (Zucconi et al., 1982). The total heavy metal was digested using HNO 3 -HCl-H 2 O 2 (v/v/v, 2:6:1) on an electric hot plate, and the heavy metal speciation was determined according to a modified BCR method extraction procedure (Lu et al. 2014). Their contents were measured by ICP-OES.

Spectroscopic and thermal analysis
The functional groups and humic acid structure in the samples collected on D0 and D40 were identified by EEM fluorescence spectra, FTIR spectra, and thermogravimetric analysis (TGA). As for the EEM spectra and parallel factor analysis, we referred to that of Yu et al. (2010a). The FTIR spectra and TGA were detected as per Ouaqoudi et al. (2014).

Statistical analysis
The means of at least three replicates were reported in data. Canoco 5 (Microcomputer Power, USA) was used for redundancy analysis (RDA). This research mainly used SPSS software version 11.0 for Windows (SPSS, USA), OriginPro 8.5 (OriginLab Corp, USA), and Matlab R2016a (Mathworks Inc., USA) to statistically analyze and plot the collected data.

Changes in composting temperature
Temperature reflects the microbial activity and is often used to monitor the composting process. Figure 1a shows the average temperature evolution during the whole composting process, where the environmental temperature varied from 22 to 39℃. The temperature evolution of the three piles displayed a curve, which includes: the mesophilic stage, thermophilic stage, cooling stage, and maturation stage. It coincided with the typical temperature curve (Muhammad et al. 2021). The temperature of the three piles raised rapidly and reached the thermophilic stage on the first day. In the thermophilic stage, the maximum temperatures of the three piles were, respectively, FM (69.7℃) > MM (64.6℃) > CM (57.3℃), and the thermophilic stage lasted longer in FM (14 days) than MM (13 days) and CM (3 days). In the cooling phase, the temperature of FM and MM were higher than that of CM. Then, with the degradation of OM and multiple turning, the material was gradually compacting and the temperature was gradually close to the ambient temperature in the maturation stage. Furthermore, the temperature trends of FM and MM were rising fast and falling slowly after turning; on the contrary, CM was rising slowly and falling fast. These phenomena showed that the pile with smaller IPS presented better heat preservation during sludge composting, with FM having the highest temperature and longest thermophilic stage. An increased temperature indicated that OM, such as carbohydrates, decomposed in large quantities and generated a lot of heat, which leads to an increase in composting temperature (Hu et al. 2020). Meanwhile, the distribution of IPS influenced the microorganisms' accessibility to materials (Manpreet et al. 2005). It also changed FAS and voided dispersion in the pile (Iqbal et al. 2010), which allowed sufficient water and gas exchange between the gas-solid phases. Therefore, adjusting IPS can improve the heat preservation effect, and then promote compost maturity.
The temperatures of three piles in both upper (60 cm) and lower (20 cm) layers were recorded from Fig. 1 b-d, which were helpful to further analyze the temperature distribution in the piles. There was a temperature gradient during composting, and in the mesophilic and thermophilic phase, the temperature trends in the "unit" between each turning were highly repeatable, both in rising and decreasing stage. In each unit, the upper temperature was higher than the lower temperature except for FM. Zhou et al. (2013) found that the temperature gradient in the piles was usually like this: bottom < upper < middle. Meanwhile, the cooling rate of the upper layer in FM was slower than that of the lower layer, while the two layers respectively in MM and CM had similar cooling rates. That was probably because the void structures of MM and CM were larger than FM. It would benefit the heat and mass transfer in the pile, thus led to the inapparent temperature difference of the two layers. On the other hand, the upper layer of FM may contain higher MC and the structure was compact, which formed an energy storage mode and achieved a better heat preservation effect. This phenomenon was consistent with some researches (Gea et al. 2007;Zhang and Sun 2014), which the heat preservation effect was enhanced with the increase of MC. This indicated that the temperature difference between the two layers of the pile may result from the difference in MC. Therefore, with the increase of IPS, the highest temperature area in the pile shifted to the upper layer, and the time to reach the highest temperature was prolonged.

Evolution of CO 2 and O 2 contents
The rate of O 2 consumption and CO 2 production can reflect the degree of composting reaction (Hu et al. 2020). At the beginning of composting, the O 2 and CO 2 content were Fig. 1 a Average temperature evolution during composting process; b, c, d temperature evolution in the upper (60 cm) and lower layers (20 cm) during composting process maintained within the range of 17-19% (Fig. 2a) and 2-4% (Fig. 2b), respectively. The CO 2 content of FM increased rapidly and reached the highest value (12.56%) on D4, which was faster than that of MM (7.46%) on D6 and CM (16.08%) on D12. Furthermore, the peak level of FM was generally higher than that of MM and CM during composting. This is probably because that pile with small IPS was conducive to microbial degradation of organic matter, resulting in high oxygen consumption and a sharp increase in the production of CO 2 . This also results in increases in temperature and the rapid entering into the thermophilic phase, as shown in Fig. 1a. The content of O 2 and CO 2 tended to be stable after day 25, indicating that these piles gradually reached stabilization and maturation. Generally, the O 2 content of 8-18% is a reasonable interval. As Fig. 2a shows, in the thermophilic stage, the O 2 content in CM was always at a high level (16.08-19.55%), while the O 2 content in FM fluctuated greatly (9.41-15.34%) and did not respond significantly to the turning operation. However, the MM fluctuated slightly (12.89-16.58%) with the turning, indicating that the pore structure characteristics of MM and the turning provided a relatively reasonable and stable O 2 content for the pile, which improve the uniformity of the composting process.

Organic matter degradation and moisture content evolution
Organic matter degradation is an important index for evaluating composting effectiveness. As it is shown in Fig. 3a, the OM decreased continuously during the composting process. The OM contents in FM, MM, and CM decreased respectively within 40 days, from the original 86.8 to 70.6%, 86.5 to 73.1%, and 87.3 to 78%. To quantify the mineralization of OM, the loss of organic matter (OML) during composting was calculated as 16.2%, 13.4%, and 9.3%, for FM, MM, and CM, respectively. The OML value of the FM was higher than that of MM and CM. A significant difference was observed between FM and CM (P < 0.05), but not between FM and MM (P > 0.05). These results indicated that reducing IPS could promote the OML, probably because smaller IPS had a larger specific surface area and accelerated the degradation of OM. This corresponded to FM's characteristics, such as the longest thermophilic phase and the highest gas peak.
As Fig. 3b shows, it was possible to observe a clear tendency towards spontaneous moisture reduction. It was reported that the optimum range of MC value for the composting process was between 40 and 60% (Diaz and Savage 2007), which resulted in a necessary addition of water. In this study, water was replenished on D4, D8, and D12 (60 L, 15 L, and 15 L, respectively), and we found that MC had a significant effect on the temperature difference between the upper and lower layers. Among them, FM had obvious reaction on replenishing 60 L and 15 L water, and the temperature difference between the two layers reduced after replenishing water; meanwhile, CM had obvious feedback on replenishing 60 L water, but after replenishing 15 L water, there was still a significant temperature difference between the two layers; the temperature difference between the two layers of MM was small no matter it replenished water or not. Furthermore, the MC of CM was reduced by 35.3% during the composting process, which was higher than that of FM (25.1%) and MM (25.4%). These results further confirmed that the void structure of CM was better than that of FM and MM (Iqbal et al. 2010).

Evaluation of C/N ratio and seed germination index
The C/N ratio and seed germination index can reflect the maturity and biological toxicity of compost (Huang et al. 2004;Zucconi et al. 1981). In Fig. 3c, the C/N ratio shows an upward trend in the initial stage, because the TOC and TN in the pile were consumed simultaneously by microbial activity, and the TN content decreased faster than the TOC (Jain et al. 2019). As the composting progressed, the C/N ratio gradually decreased. This was because the simple and easily degradable OM, such as glycogen and fats, was used by microorganisms, resulting in a significant reduction in TOC (Fuentes et al. 2007), and nitrogen remained through the cellular synthesis by microorganisms. The C/N ratio for FM, MM, and CM decreased, respectively, from the initial value of 26.87 on D0 to 15.20 on D40, 24.40 to 16.33, and 25.09 to 21.51. The change of C/N ratio was FM > MM > CM, and there was a significant difference between FM and CM, but not between FM and MM.
Meanwhile, as Table S2 shows, the GI of FM, MM, and CM on D40 were 117%, 111%, and 66%, respectively. The C/N ratio < 20 indicated that the compost met the maturity requirements (Heerden et al. 2002). Huang et al. (2004) suggested that when the GI > 80%, the compost material may be regarded as mature. These results showed that FM and MM had reached maturity and were conducive to plant growth, while CM required longer composting time.

Variation in pH and electrical conductivity
As shown in Fig. 3d, the initial pH values of the three piles were 7.49, 7.46, and 7.41, respectively. They all increased in the beginning and then decreased during the composting process. The pH value increased rapidly in the early stage and peaked on D2 (7.65, 7.55, and 7.73, respectively), and then dropped to 6.66, 6.77, and 7.05 on D40, and the fluctuation ranges were 0.99, 0.78, and 0.68, respectively. As reported, the increase of pH was due to the ammonification of organic nitrogen with the release of hydroxyl ions and ammonium, and the disappearance of organic acids (Said-Pullicino et al. 2007). The decrease of pH may be attributed to the volatilization of ammonia, nitrification of microorganisms, and formation of acidic substances after the degradation of organic matter (Soobhany 2018). These results indicated that the final pH of the three piles was between 6.66 and 7.05, which met the mature standard (pH < 9) proposed by Pera (1983). Reducing IPS can change the pH of the materials in sludge composting from slightly alkaline to slightly acidic.
The EC indicates the salt content of the material, indicating its potential impact on plants as a fertilizer (Lin 2008). As the composting went on, soluble salts produced by the decomposition of OM can increase EC value, and the maximum values of EC would be found in the final composting product. It can be seen from Fig. 3e that the EC values of the three piles presented a gradually increasing trend in the sludge composting except the CM. The EC values for FM, MM, and CM increased from the initial value of 1.88 mS/cm on D0 to 2.32 mS/cm on D40, 1.88 mS/cm to 2.10 mS/cm, and 1.93 mS/cm to 1.96 mS/cm, respectively. Among the three piles, the maximum EC value was observed in the FM, indicating that the organic matter in FM was decomposed more easily, which was consistent with the result of OML in Fig. 3a. Furthermore, the EC values were all in the range of 1.96-2.32 mS/cm, which met the requirements of compost maturity and safety (Awasthi et al. 2014).

Changes of total nitrogen, ammonium nitrogen, and nitrate nitrogen
Nitrogen loss is common in the thermophilic phase, which is due to the volatilization of ammonia, the mineralization and denitrification of microorganisms. Figure 3f shows the changes of TN during composting. The TN of each pile followed a pattern of decreasing-then-increasing during the composting process. The result of Chan et al. (2016) also found that the TN content decreased in the early period was mainly due to a large amount of volatilization of ammonia in the process of microbial mineralization. During the composting process, due to the relative decrease in ammonia volatilization, the acceleration of organic matter decomposition and the rate of water decline resulted in the increase of TN (Li and Song 2020). Forty days after the composting began, the TN contents of the FM, MM, and CM increased respectively from 1.4 to 1.82%, 1.46 to 1.80%, and 1.51 to 1.5%. It indicated that small IPS was beneficial to enhance the nitrogen fixation effect of pile.
It can be seen from Fig. 3g that the NH 4 + -N contents of all piles increased at first, then decreased and stabilized, and they showed a significant peak during the composting time 8 ~ 14 days. This was similar to the results found by Zhang et al. (2017a). The NH 4 + -N contents of FM and MM peaked at D8 (2.55 and 1.4 mg/g, respectively), while CM continued to rise and reached the peak (1.9 mg/g) at D14. Having reached the peak value, the NH 4 + -N contents of FM, MM, and CM increased by 64.51%, 33.3%, and 26.7%, respectively, and there was extremely significant difference between FM and the other piles (P < 0.01). This reflected the ammonification of organic nitrogen caused by the degradation of organic matter, which was consistent with the result of the thermophilic period (Fig. 1a) and OML (Fig. 3a) from D0 to D4. Meanwhile, as the composting progressed, the NH 4 + -N contents of the three piles decreased due to the dual effects of microbial nitrification and NH 3 volatilization. Furthermore, compared with MM and CM, FM significantly increased the accumulation of NH 4 + -N and promoted the conversion of NH 4 + -N in sludge composting. In Fig. 3h, the NO 3 − -N content of each pile showed a gradually increasing trend during the composting progress. At the end of composting, the NO 3 − -N contents of these piles, respectively increased from 1.3, 1.25, and 1.45 mg/g to 2.2, 2.1, and 1.65 mg/g. The increase of NO 3 − -N contents in the three piles was FM > MM > CM, and the difference between FM and CM was extremely significant (P < 0.01). This was probably because the lower pH and higher ammonia absorption capacity of small IPS could promote the microbial nitrification and accumulation of NO 3 − -N in sludge composting.

EEM fluorescence spectra
The EEM fluorescence spectra of samples on D0 and D40 were shown in Fig. 4. According to the scheme of Chen et al. (2003), the EEM spectra can be delineated into five regions. At the beginning of composting (D0), the three piles had similar fluorescence distribution, indicating that soluble microbial by-products and simple aromatic proteins were the main components of the initial compost samples. After 40 days of composting, due to the degradation of OM caused by microbial activity, the peak A and B of the three piles had an obvious tendency to move to regions III and V. In the meantime, the fluorescence intensity of FM on D40 was higher than that of MM and CM, suggesting a higher degree of humification.
To obtain more details about EEM spectra, three fluorescent components were determined by PARAFAC analysis ( Fig. 4g-i). Compared with the identified components, it was obvious that component 1 (C1), component 2 (C2), and component 3 (C3) were respectively related to fulvic acid-like and humic acid-like substances, soluble microbial by-products and tryptophan, and humic acid-like substances (Yu et al. 2010b;Zhang et al. 2016a). The changes in the maximum fluorescence intensity (F max ) of these components were presented in Fig. 4j. The different components of the three piles showed the same pattern throughout the composting period. At the beginning of composting (D0), the F max value of C2 in FM was 5.31 a.u., higher than MM (4.49 a.u.) and CM (4.78 a.u.), indicating that the pile with small IPS was conducive to the dissolution of soluble organic matter in the initial stage. Within 40 days, the F max value of C2 in FM, MM, and CM respectively decreased from the original 5.31 to 2.21 a.u., 4.49 to 2.19 a.u., and 4.78 to 1.69 a.u., which could be related to the strong biological oxidation of protein-like substances by microorganisms (He et al. 2011). The increase of C1 and C3 suggested the increase of humification degree. Similar results were found in many previous studies (He et al. 2014;Zhang et al. 2016b), indicating that the reduction of protein-like substances promoted the formation of humic substances in the composting process. These results were consistent with the result in Table S2: the final HA concentration of FM was the highest among the three piles.

FTIR spectra
The functional groups presented in the three piles on D0 and D40 were determined by FTIR spectroscopy (Fig. 5). The same transmittance peaks were identified in the three piles, but the intensities were different. The peaks or bands of 1030-1110, 1260, 1600-1650, 2919, and 3410 cm −1 were dominant on D0. It indicated that there were abundant polysaccharides, aliphatic, phenolic, and aromatic structures in the three piles. Meanwhile, the presence of the typical lignin peak (1525 cm −1 ) in the pile was completely consistent with the lignocellulosic properties of Triarrhena lutarioriparia. According to the changes occurring in the structures of pile, the calculation of various aromatic indexes or transmittance ratios was made (Table S3).
For the three piles, a decrease was noted in 1643/2840, 1643/2919, and 1525/1919 ratios, which could be interpreted by the increase in aromatic structure, and the reduction of polysaccharide and carboxylate concentration (Baddi et al. 2003). Meanwhile, these ratios of FM on D40 were lower than those of the other piles. This stressed the aromatization of the three piles and indicated that the OM structure had become more complicated as the pile matured (Som et al. 2009). The increase of 1425/1050 ratio in FM and MM could be attributed to a reduction in carboxylate ions and carbohydrates (C-O groups), and the reduction in CM may be related to the microbial activities during the composting process. These differences observed in the three piles distinctly explained the differences in maturity. The reason for the lowest maturity of CM was attributed to its larger IPS and higher content of difficultly biodegradable OM. Therefore, a longer compost time would be required to reach the same maturity as FM.

Thermogravimetric thermal analysis
The results of thermal analysis for the raw materials and three piles were shown in Fig. S1. It can be seen from Fig. S1, there were three peaks respectively in the dewatered sludge and the three piles on D0 and D40. After burning, the residual amount of Tl was about 10%, which was lower Fig. 5 Evolution of the FTIR spectra during composting than dewatered sludge. It indicated that the ash content of dewatered sludge was higher than that of Tl. At D0, the peak intensities of the three piles were similar. After 40 days of composting, the intensities of peak A and B were as follows: CM > FM > MM. The increased intensity of peak A in FM and MM may be due to the release of the aliphatic structures bonded or captured in lignocellulose complexes after microorganisms attack lignified substrates. The enhancement of peak B was associated with the condensation of the aromatic structures released by the metamorphism of the lignified complexes (Ouaqoudi et al. 2014). However, the enhancement of peak A and B in CM did not match its phenomena in Fig. 1a and Fig. 3a. In the meantime, the position of peak A and B coincided with the first exothermic peak of Tl. This indicated that the proportion of Tl in CM had increased. The above results showed that the mature degree of the pile was like this: FM > MM > CM.

Speciation of heavy metal in the sludge composting
Since composting land use is one of its most important purposes, it is essential to investigate the concentration and variation trend of heavy metals such as Cd, Cr, Cu, Ni, Pb, and Zn. The heavy metal concentrations of each pile were shown in Table S4. The concentration of Cr, Cu, Ni, Pb, and Zn in FM and MM increased during the composting process. Conversely, compared with before, the Cd concentration decreased after composting. However, all kinds of heavy metals in CM showed a downward trend during the composting process. On the one hand, the proportion of sludge in CM decreased at the end of composting, which led to the decrease of heavy metal concentration. On the other hand, the different trends of heavy metal concentration during sludge composting were caused by the varied speciation of heavy metal. Before composting, exchangeable Cd accounted for a large proportion in each pile. It could easily flow off through the leachate, resulting in a decrease in the concentration of heavy metal post compost (Li and Song 2020). The Cr, Cu, Ni, Pb, and Zn concentrations increased slightly, which might be attributed to the small proportion of exchangeable form and a self-concentration effect (Guerra-Rodriguez et al. 2006). The phenomenon of CM may ascribe to its large IPS, which resulted in the weak water storage capacity of the pile (Fig. 3b), so the metal loss after composting is accompanied by the loss of leachate. Furthermore, each pile in this study was an independent system, and heterogeneity of materials would influence the sludge composting process, which led to the difference of heavy metal concentration.
The distribution rate is an important index to evaluate heavy metal bioavailability. In the form of heavy metals, the activity in descending order were exchangeable state (F1), reductive state (F2), oxidation state (F3), and residual state (F4). The exchangeable and reductive states had high activity and were easy to be absorbed by plants. The oxidation and residual states were not easy to be released under natural condition, and can stably bind with sediment, thus not easily absorbed by plants (Lu et al. 2014). As shown in Fig. 6a-f, the changes of Cr, Ni, and Zn speciation in the three piles were roughly the same, while there were obvious differences in the speciation of Cd, Cu, and Pb. At the beginning of composting (D0), the speciation of heavy metal Cd and Pb were dominated by reductive state, and Cu was dominated by oxidation state. The residual state of the three metals was low. At the end of composting (D40), the exchangeable Cd decreased substantially with FM (20.91%) > CM (13.96%) > MM (6.51%). The distribution ratios of reductive Cu in FM, MM, and CM were respectively 26.44%, 16.75%, 18.34%. Before composting, the proportion of Pb in the oxidation state were 40.21%, 40.13%, and 40.99%. After composting, it increased to 69.74%, 63.05%, and 71.70%, respectively. There was no obvious change in the residual state of Cd, Cu, and Pd in each pile, and the oxidation Cu increased drastically. It can be seen from Fig. 6g that the passivation effect of the three piles on Cu was in the following order: FM > CM > MM. And the passivation effect of FM and CM was 9.67% and 1.49% for each, both higher than that of MM. Simultaneously, the passivation effect on Pb was in such order: CM (30.70%) > FM (29.53%) > MM (22.91%). These results showed that composting could effectively reduce the bioavailability of heavy metals Cd, Cu, and Pb, and the passivation effect of FM on both Cd and Cu was stronger than that of MM and CM. The reason may be that FM had the highest degree of humification, which facilitated the heavy metal passivation (Manpreet et al. 2005). Meanwhile, CM had the best passivation effect on Pb, probably due to the following two aspects: On the one hand, due to the large IPS of CM, the water storage capacity of CM was weak (Fig. 3b), and the speciation of Pb was dominated by reductive state at the beginning of composting (Fig. 6e), which may lead to the loss of leachate accompanied by the loss of reductive Pb in the process of composting, thus leading to the increase of the proportion of oxidized Pb. On the other hand, in the process of composting, the pH of CM was higher than that of the other two piles (Fig. 3d), which may also be the reason for its better passivation effect on Pb. Therefore, the high degree of humification effectively controlled the heavy metal activation caused by the decrease of pH in the composting process.

Relationship among the environmental parameters and BCR-extractable metals in compost
RDA was used to evaluate the relationship between the physicochemical properties and heavy metal speciation of the compost, in order to explain the impact of physicochemical properties on heavy metal speciation, and also to analyze the parameters with the greatest impact. Overall, Fig. 7 showed that RDA1 and RDA2 explained 95.18% of the total variation. In all physicochemical properties, OM (P = 0.002), pH (P = 0.002), MC (P = 0.002), C/N (P = 0.002), NO 3 − -N (P = 0.002), EC (P = 0.004), and temperature (P = 0.016) had a very significant correlation with the succession of heavy metal speciation, indicating that these properties had a great influence. The physicochemical properties, including temperature, OM and C/N ratio of the piles, were associated with OM degradation and compost maturity. The MC could affect the metabolic activities of microorganisms (Diaz and Savage 2007). The pH and EC in the pile could affect the maturity and security of compost. The change of NO 3 − -N reflected the compost maturation process. The composting process can be approximately divided into three parts Fig. 6 a, b, c, d, e, f Percentage of different fractions of Cd, Cr, Cu, Ni, Pb, Zn comprising total metals from the three piles before and after composting; g the passivation effect of Cd, Cr, Cu, Ni, Pb, Zn according to the scatter of sample points, namely, the initial stage of composting (part I), the middle stage of composting (part II), and the end stage of composting (part III). At the initial stage of composting, the RDA values of the three piles were similar, indicating that there was little difference among the three piles. As the composting progressed, the difference in the heavy metal speciation among the three piles gradually became obvious. Meanwhile, the three sample points of MM showed a good polymerization degree in the three stages, while that of FM and CM showed obvious dispersion degree in the middle and end stage of composting, respectively, and the dispersion of FM decreased at the end of composting. This phenomenon showed that the internal uniformity of MM was better than that of the other piles in the whole process of composting, which was confirmed in Fig. 1b-d. Meanwhile, according to positions of sample points and blue arrows, Cd-F1, Cu-F1, Cu-F2, and Pb-F2 were mainly distributed in part I (Fig. 7). As confirmed by the change of heavy metal speciation shown in Fig. 6, they had a positive correlation with temperature, pH, OM, MC, and C/N, while Cd-F2, Cu-F3, Cu-F4, and Pb-F3 were mainly distributed in part III, and were positively correlated with EC, and NO 3 − -N. Therefore, a pile with different IPS can change the environmental parameters by changing the structure of the pile, and it can provide suitable conditions for the complication of metal ions to reduce heavy metal bioavailability. Furthermore, although the passivation effect of MM on heavy metals was weaker than that of FM, the internal uniformity of MM was stronger than that of FM in the whole composting process. Therefore, MM would help us obtain a composting product with better uniformity in a larger scale composting process under conditions where the risk of heavy metal pollution in sludge was controllable.

Conclusions
In this study, Triarrhena lutarioriparia was added to dewatered sludge to study the effects of different IPS on sludge composting. The results showed that IPS had a significant effect on processing performance, material properties, and especially heavy metal speciation. With the decrease of IPS, the heating rate and the highest temperature of the pile increased and the high-temperature period was prolonged. Meanwhile, the highest temperature area in the pile shifted to the lower layer, which represented a better fermentation effect. FM had the highest level of OML, humification, nitrogen fixation, and heavy metal passivation. MM was beneficial to improve the internal uniformity in the sludge composting and was conducive to obtaining a composting product with better uniformity in a larger scale composting process under conditions where the risk of heavy metal pollution in sludge was controllable. The findings of this study can directly improve the efficiency and safety of sludge composting by adjusting IPS.
Funding The authors gratefully acknowledge financial supports from the National Natural Science Foundation of China (51808216, 51608052), Science and Technology International Cooperation Project of Changsha City (KQ1907082), Hunan Forestry Science and Technology Project (XLK202108-3, XLK201908), Natural Science Foundation of Hunan Province (2019JJ50665, 2021JJ40284, 2021JJ30713), Training Program for Excellent Young Innovators of Changsha (KQ2009085, KQ2009086), The Science and Technology Innovation Program of Hunan Province (2020RC5008).
Data availability All data and materials included in this study are available upon request by contact with the corresponding author (J. Huang).

Declarations
Ethical approval Approval was obtained from the ethics committee of Changsha University of Science and Technology. The procedures used in this study adhere to the tenets of the Declaration of Helsinki.
Consent to participate Informed consent was obtained from all individual participants included in the study.