Bioresource Technology 161 (2014) 69–77
Contents lists available at ScienceDirect
Bioresource Technology journal homepage: www.elsevier.com/locate/biortech
Relationship between anaerobic digestion of biodegradable solid waste and spectral characteristics of the derived liquid digestate Wei Zheng a,b, Fan Lü a,b,⇑, Khamphe Phoungthong a,b, Pinjing He b,c a
State Key Laboratory of Pollution Control & Resource Reuse, Tongji University, Shanghai 200092, PR China Institute of Waste Treatment and Reclamation, Tongji University, Shanghai 200092, PR China c Centre for the Technology Research and Training on Household Waste in Small Towns & Rural Area, Ministry of Housing and Urban–Rural Development (MOHURD), PR China b
h i g h l i g h t s Fluorescence peaks of tyrosine were appropriate for protein evaluation during AD. SUVA254 and Ex/Em of 230/436 nm were suitable for evaluating humidification degree. E4/E6 and Ex/Em of 350/436 nm were unsuitable for evaluating humidification degree. Fluorescence peaks of tryptophan were inconsistent with dissolved protein content.
a r t i c l e
i n f o
Article history: Received 2 January 2014 Received in revised form 24 February 2014 Accepted 4 March 2014 Available online 15 March 2014 Keywords: Food waste Lignocellulose waste Fabric Biplot Humification
a b s t r a c t The evolution of spectral properties during anaerobic digestion (AD) of 29 types of biodegradable solid waste was investigated to determine if spectral characteristics could be used for assessment of biological stabilization during AD. Biochemical methane potential tests were conducted and spectral indicators (including the ratio of ultraviolet–visible absorbance at 254 nm to dissolved organic carbon concentration (SUVA254), the ratio of ultraviolet–visible absorbance measured at 465 nm and 665 nm (E4/E6), and the abundance of fluorescence peaks) were measured at different AD phases. Inter-relationship between organic degradation and spectral indicators were analyzed by principle component analysis. The results shows that from methane production phase to the end of methane production phase, SUVA254 increased by 0.16–10.93 times, the abundance of fulvic acid-like compounds fluorescence peak increased by 0.01– 0.54 times, the abundance of tyrosine fluorescence peak decreased by 0.03–0.64 times. Therefore, these indicators were useful to judge the course of mixed waste digestion. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction Global municipal solid wastes (MSWs) generation reached an estimated 1.3 billion tonnes per year in 2010 (Hoornweg and Bhada-Tat, 2012). MSWs contains many biodegradable components, but these components vary by region. For example, MSWs in the United States comprised 28% of paper, 14% of garden waste, and 15% of food waste in 2011 (US EPA, 2013). However, in Shanghai, the paper, garden waste, and food waste accounted for 11%, 1%, and 64% of the MSWs respectively in 2008–2009 (Shanghai Environmental Engineering Design and Science Academy, 2009). These components of MSWs are important organic sources of methane production through anaerobic digestion. ⇑ Corresponding author at: Institute of Waste Treatment and Reclamation, Tongji University, Shanghai 200092, PR China. Tel./fax: +86 21 65986104. E-mail address: [email protected] (F. Lü). http://dx.doi.org/10.1016/j.biortech.2014.03.016 0960-8524/Ó 2014 Elsevier Ltd. All rights reserved.
The residue of anaerobic digestion, called digestate, is usually used as soil amendments, and its biological stability is an important indicator of the environmental risk posed by the digestate (Senesi and Plaza, 2007). Several indicators have been proposed for assessment of the biological stability of MSWs and its derived materials, including the respirometric index, biomethane potential production, and the ratio of 5-day biological oxygen demand to chemical oxygen demand (BOD5/COD) (Cossu et al., 2001). However, these parameters are limited by their high cost, long testing time, and low representativity in the presence of inhibiting substances. Comparatively, ultraviolet–visible (UV–Vis) absorption (Vieyra et al., 2009; Hunger and Weitkamp, 2001) and fluorescence spectra (Antunes and Da Silva, 2005) as assessment methods to represent biological stability and stabilization of dissolved organic materials (DOM) are rapid, highly sensitive and nondestructive. Nevertheless, few studies have been conducted to investigate the use of spectral indicators to evaluate biological stability and
70
W. Zheng et al. / Bioresource Technology 161 (2014) 69–77
stabilization of anaerobic digestion, which were only limited to the leachate generated from anaerobic landfill for mixed MSWs (He et al., 2011; Lü et al., 2009). The ratio of the UV–Vis absorbance measured at k = 254 nm to the dissolved organic carbon (DOC) concentration (SUVA254) and the ratio of the UV–Vis absorbance measured at k = 465 nm to that measured at k = 665 nm (E4/E6) are widely used to represent biological stability in UV–Vis absorption spectral indicators. SUVA254 and E4/E6 have been reported to be particularly useful indicators for investigation of environmental samples (Weishaar et al., 2003; Chen et al., 1977), DOM of compost (Shao et al., 2009; Vieyra et al., 2009), and landfill leachate (He et al., 2011). The excitation–emission matrix (EEM) spectra technology is increasingly being applied to represent biological stability and stabilization of environmental samples (Ishii and Boyer, 2012), DOM of compost (Yu et al., 2010), and investigation of anaerobic digestion of sludge (Luo et al., 2013), livestock manure (Wan et al., 2012), and mixed municipal solid waste (Ghita et al., 2013). Nevertheless, since the biochemical characteristics of various MSWs are quite different (Zheng et al., 2013), the DOM of various MSWs has different fluorescence spectra characteristics. For example, lignin is fluorescent, but polysaccharide is not; therefore, the fluorescence spectrum of lignocellulose waste is quite different from that of low lignocellulose content waste such as food waste. Accordingly, it is necessary to investigate whether EEM fluorescence spectra have consistent variation during anaerobic digestion processes of diverse biodegradable solid waste. Most studies that have employed EEM spectra to probe the anaerobic digestion process have only used the ‘‘peak picking’’ technique (Ghita et al., 2013; Luo et al., 2013; Wan et al., 2012), which makes minimal use of the large quantity of data available in each EEM spectrum (Chen et al., 2003b). Comparatively, fluorescence regional integration (FRI) and parallel factor (PARAFAC) techniques (Shao et al., 2012) allow a more complete data mining than the traditional peak picking technique. Accordingly, combination of EEM with FRI and PARAFAC may provide deeper insight into the composition and behaviour of the DOM. This study was conducted to investigate the relationship between spectral indicators and biological stabilization properties during anaerobic digestion of various biodegradable solid wastes with different biochemical compositions. To accomplish this, a biochemical methane potential (BMP) assay was utilized to simulate anaerobic digestion, which was then evaluated during the lag phase, methane production phase, end of methane production phase and excessive acidification phase. A combination of the absorption and fluorescence spectral properties, EEM–FRI/PARAFAC analysis and multivariate analysis was then used to compare the applicability of spectroscopy indicators to represent biological stability and stabilization during biodegradable solid waste anaerobic digestion.
2. Methods 2.1. Degradable material and inoculum Based on the waste biochemical characteristics and production quantity, 29 types of precursor materials with the potential to be discarded as biodegradable solid waste were selected, including papers (newspaper, office paper, toilet paper), kitchen waste (fish bone, pork bone, lean pork, fat pork, soybean, potato, celery, lettuce), fruit waste (sugarcane residue, banana peel, orange peel, apple core and peel, watermelon peel, grapefruit peel), ligocellulosic materials (peanut shell, reed, the grass Cynodon dactylon, tea residue, bamboo leaf, bamboo branch, camphor tree leaf, camphor
tree branch, metasequoia leaf, metasequoia branch, and cotton), and fabric. All collected materials were shredded to a size of less than 1 mm and then prepared for biochemical composition as described by Zheng et al. (2013). The anaerobic inoculum was the digestate collected from an anaerobic digestion plant used to treat a mixture of sewage sludge and food waste. The inoculum had a total solids (TS) content of 23.8 ± 1.7 wt% and the volatile solids (VS) accounted for 81.9 ± 2.5 wt% of the TS. 2.2. Biochemical methane potential assay Each 1 L glass bottle contained 100 g (wet weight) of inoculum, 400 g of nutrient medium, and 10 g (dry weight) of the tested biodegradable material, except for the reactors that contained toilet paper, orange peel and grapefruit peel. Only 5 g (dry weight) of the latter three materials were added because preliminary tests indicated that a higher inoculum-to-substrate ratio was required to avoid acidification. A blank reactor that contained only inoculum was also incubated to measure the background methane production. All experiments were carried out in duplicate. The nutrient medium, cultivation condition and gas measurement method described by Zheng et al. (2013) was employed. The suspended mixtures in the bottles were obtained using a 5 ml syringe and then centrifuged for 10 min at 2000g. Next, the supernatant was then passed through a 0.45-lm, microfiber filter, after which was referred to as the liquid digestate. Based on the methane production curve (Zheng et al., 2013), the lag phase of all materials except bamboo leaf was longer than 2 days; therefore, the liquid digestate of the lag phase was collected on the first day. When the methane concentration was higher than the carbon dioxide concentration, the liquid digestate of methane production phase was collected. When the methane production curve reached a plateau, the liquid digestate was collected at the end of the methane production phase. When anaerobic digestion reactors were failed due to excessive acidification, the liquid digestate of acidification phase was collected. 2.3. Analytical methods 2.3.1. UV–Vis spectra UV–Vis spectroscopic measurements of the absorbance of liquid samples at 665, 465 and 254 nm were conducted using a UV-1800 UV–Vis spectrophotometer (Shimadzu, Kyoto, Japan) and 1-cm quartz cells. The unit of absorbance is m 1. And DOC whose unit is mg L 1 was measured using a total organic carbon (TOC) analyzer (TOC-V CPN, Shimadzu, Japan). SUVA254 values were determined by dividing the UV–Vis absorbance measured at k = 254 nm by the DOC concentration and then reported in L mg 1 m 1 (Shao et al., 2009). E4/E6 values were determined by dividing the UV–Vis absorbance measured at k = 465 nm by the UV–Vis absorbance measured at k = 665 nm. 2.3.2. Fluorescence spectra Prior to fluorescence spectroscopy analysis, the liquid digestate was diluted with 0.1 mol L 1 phosphate buffer to DOC < 10 mg L 1 to ensure that the maximum fluorescence signal was below the upper detection limit of the spectrometer and the disruption of inter- and intramolecular hydrogen bonds was eliminated (Ghita et al., 2013). The fluorescence EEM spectra of the liquid samples were recorded using a fluorescence spectrophotometer (Cary Eclipse, Varian, USA) in scan mode. To obtain fluorescence EEM, scanning emission (Em) spectral analysis was conducted from 250 to 600 nm at 2 nm increments by varying the excitation (Ex)
W. Zheng et al. / Bioresource Technology 161 (2014) 69–77
wavelength from 200 to 550 in 10 nm increments. The EEM of a control (Milli-Q water) was subtracted from each sample EEM. 2.4. Data analysis 2.4.1. Fluorescence regional integration (FRI) The Rayleigh and Raman scattering of EEM spectra were removed according to the protocol described by Bahram et al. (2006). EEMs were then normalized by dividing the spectrum by the corresponding DOC concentration, after which the corrected EEMs were plotted using SigmaPlot 11.0 (SPSS Inc., USA) with a contour plot graph. The FRI technique was employed to analyze the five excitation–emission regions of EEM spectroscopy according to Chen et al. (2003b). 2.4.2. PARAFAC modeling When compared with the FRI technique, PARAFAC can more precisely identify fluorophors with multiple excitation wavelengths at the same emission wavelength. PARAFAC analysis of EEM spectra was conducted as described by Stedmon and Bro (2008) using MATLAB 7.0 (Mathworks, Natick, MA) with the DOMFluor toolbox (www.models.life.ku.dk). The PARAFAC models with two to six components were computed for the EEMs. Determination of the correct number of components was primarily based on core consistency. 2.4.3. Statistical analysis To evaluate the relationship between spectroscopy indicators and anaerobic digestion status, PCA was employed using the CANOCO 4.0 (Biometris, Wageningen, Netherlands) software. 3. Results and discussion
71
the SUVA254 values were generally consistent with biological stability and stabilization of a routine anaerobic digestion process. 3.1.2. Evolution of E4/E6 E4/E6 is widely used to evaluate the degree of humification in aerobic environment because the E4/E6 value decreases as the polymerization degree, aromatization degree and molecular weight increase (Huang et al., 2006). The evolution of E4/E6 of liquid digestate during the anaerobic digestion of biodegradable solid waste is shown in Table 1, while Table 2 lists the E4/E6 values in the acidification phase. It could be found that the E4/E6 variation along AD process of different materials has not regular trend. E4/E6 is also influenced by pH, free radicals, total acidity (Chen et al., 1977), protein, polysaccharide (Vieyra et al., 2009) and microorganisms (Marschner et al., 2005). Though E4/E6 was widely used to investigate the biological stability of soil samples (Marschner et al., 2005; Chen et al., 1977) and compost liquid samples (Vieyra et al., 2009; Huang et al., 2006). But those samples were produced from aerobic environment. The reference to utilize E4/E6 as biological stability indicator in anaerobic environment was limited. And according to Table 1, from methane production phase to the end of methane production phase, E4/E6 values of metasequoia branch, reed, lettuce, peanut shell, watermelon peel, apple core and peel, tea residue, sugarcane residue, cotton, fabric fish bone, toilet paper and newspaper increased. Therefore, the E4/E6 variation of different materials was inconsistent with AD process. 3.2. Evolution of fluorescence characteristics during anaerobic digestion The evolution of EEM fluorescence spectra of liquid digestate is shown in Figs. A.1 and A.2 lists the EEM fluorescence spectra in the acidification phase. The EEM peaks overlapped; therefore, FRI and PARAFAC techniques were utilized to minimize mutual interference of fluorophores.
3.1. Evolution of UV–Vis absorption spectra during anaerobic digestion 3.1.1. Evolution of SUVA254 SUVA254 can reflect the aromatic carbon content in dissolved organic matter (DOM) (Weishaar et al., 2003) and indirectly reflect the degree of humification and biological degradability of the sample. Evolution of SUVA254 of liquid digestate during the anaerobic digestion of biodegradable solid waste is shown in Table 1, while Table 2 lists the SUVA254 values in the acidification phase. As shown in Table 1, SUVA254 values of most liquid digestates increased constantly as anaerobic digestion proceeded. However, unsaturated carbon bond compounds (such as aromatic protein) existed in the liquid digestate of food waste (such as grapefruit peel, celery, lettuce, etc.) and lignin-derived aromatics existed in the liquid digestate (Kataeva et al., 2013) of lignocellulose material (such as camphor tree branch, bamboo branch, newspaper, toilet paper, etc.); therefore, some SUVA254 values in the lag phase were higher than in the methane production phase. The SUVA254 values at the end of the methane production phase were highest (>1.0 L mg 1 m 1). The SUVA254 values were increased by 0.16– 10.93 times from the methane production phase to the end of the methane production phase. Shao et al. (2009) observed that the SUVA254 values of water extracts of MSWs after the hydrolysis-aerobic process were around 3.0 L mg 1 m 1, suggesting that the anaerobic digestion product in the present study was still unstable. The lowest SUVA254 values were observed when anaerobic digestion of food waste failed due to excessive acidification. Comparatively, although the digestion of lignocellulose waste maybe failed as a result of excessive acidification as well, the SUVA254 value did not decrease obviously owing to the lignin-derived aromatics dissolved in the liquid digestate. Nevertheless,
3.2.1. Evolution of EEM–FRI Chen et al. (2003b) proposed an FRI technique to quantify fluorescence spectra and divided the EEM fluorescence spectra into five emission–excitation regions. The percentage of fluorescence response (Pi,n) can quantitatively reflect the abundance of a specific region that represents a specific structural compound. The evolution of Pi,n during anaerobic digestion of biodegradable materials is shown in Table 1, while Table 2 lists the Pi,n values in the acidification phase. The PI,n values (except newspaper and fat pork) decreased by 0.05–0.54 times from the methane production phase to the end of methane production phase. In addition, the abundance of tyrosine-like compounds decreased, suggesting that protein was converted during anaerobic digestion. Region II and IV were adjacent to the fluorescent region of humus-like compounds. As a result, PII,n and PIV,n incorrectly reflect the abundance of tryptophan-like compounds and soluble microbial by-product-like compounds due to the interference of humus-like compounds; therefore, the variation regulation of PII,n and PIV,n was not unified among different types of biodegradable materials. PIII,n increased by 0.01–0.54 times and PV,n (except soybean) increased by 0.02–0.78 times. However, Nicotinamide Adenine Diuncleotide Hydrogen (NADH) (Li et al., 2011) and lignin (Muller et al., 2011) showed fluorescence peaks at Ex/ Em wavelengths of 350/440 nm and (270) 380/440 nm, respectively. PV,n was determined by humic acid-like compounds as well as by NADH and lignin. While anaerobic digestion failed due to excessive acidification, the fluorescence properties of liquid digestate were influenced by the biochemical characteristics of materials. Since amino acids accumulated during the acidification phase of food waste, PI,n increased (PI,n > 0.40) and PIII,n decreased
72
W. Zheng et al. / Bioresource Technology 161 (2014) 69–77 Table 1 Evolution of spectra characteristics during anaerobic digestion.
W. Zheng et al. / Bioresource Technology 161 (2014) 69–77
73
Table 1 (continued)
(continued on next page)
74
W. Zheng et al. / Bioresource Technology 161 (2014) 69–77 Table 1 (continued)
a b c
L: lag phase. M: methane production phase. E: end of methane production phase.
Table 2 Spectra characteristics of liquid samples from failed digesters. SUVA254 (L mg Grapefruit peel Orange peel Toilet paper a
0.146 0.152 0.214
1
m
1
)
E4/E6
PI,n
PII,n
PIII,n
PIV,n
PV,n
C1/C2
C2/C3
2.63 3.09 3.38
0.420 0.682 0.245
0.219 0.125 0.299
0.135 0.046 0.268
0.149 0.144 0.114
0.077 0.003 0.075
1.176 naa 0.653
0.314 0 1.025
na = not applicable.
(PIII,n < 0.15). However, lignin-derived aromatics dissolved in liquid digestate and the production of amino acid was limited in the acidification phase; therefore, PI,n and PIII,n did not change significantly. In general, PI,n was consistent with the changes in protein content during anaerobic digestion and PIII,n was consistent with the stabilization degree. 3.2.2. Evolution of EEM–PARAFAC Since the core consistency of three and four components were 92.6% and 87.9% respectively, the three-components PARAFAC model was used to quantitatively analyze the EEM dataset. The three components identified by the PARAFAC model are shown in Fig. 1. Component 1 (C1) has fluorescence peaks at an Ex/Em wavelength of 230(260, 350)/436 nm (Fig. 1(a)). C1 in this study was similar to the humic-like compounds identified by Coble et al. (1996, 1993), which had an Ex/Em wavelength of 230(260, 350)/ 420–450 nm. Additionally, NADH could accumulate during anaerobic digestion (Farabegoli et al., 2003) and lignocellulose waste contains lignin. NADH has an Ex/Em wavelength of 350/440 nm (Li et al., 2011) and lignin has an Ex/Em wavelength of 270(380)/ 440 nm (Muller et al., 2011). Therefore, the fluorescence intensity of C1 can be influenced by NADH and lignin. Component 2 (C2) has fluorescence peaks at an Ex/Em wavelength of 230(290)/ 388 nm (Fig. 1(b)). Baddi et al. (2013) indicated that peaks of Em at 390 nm were attributed to fulvic acid-like compounds; accordingly, C2 was similar to fulvic acid-like compounds. Component 3 (C3) generated fluorescence peaks at an Ex/Em wavelength of 220(270)/304 nm (Fig. 1(c)). C3 was similar to tyrosine-like compounds identified by Wolfbeis (1985), which had Ex/Em wavelengths of 220–275/300–305 nm. Ternary scatter plots of liquid digestate during anaerobic digestion of different biodegradable solid wastes are shown in Fig. 2. The C1 and C2 of most materials increased from the methane production phase to the end phase, while some C1 values (such as soybean and banana) and C2 values (such as fat pork, grapefruit peel, watermelon peel, newspaper and office paper) decreased. C3 decreased by 0.03–0.64 times from the methane production
phase to the end phase. While anaerobic digestion of food waste failed due to excessive acidification, C3 values increased to more than 0.55. Therefore, C3 was generally consistent with protein content during anaerobic digestion. Chen et al. (2003a) reported that a red shift of emission wavelength to longer wavelengths implied an increase of molecular size, aromatic polycondensation, and humification degree of humic-like compounds. Therefore, C1/C2 can be used to indicate the degree of polymerization of humic-like compounds. To avoid the mutual effects of NADH, lignin and humic-like compounds, C2/C3 was selected to evaluate the degree of humification. The evolution of C1/C2 and C2/C3 during biodegradable solid waste anaerobic digestion is shown in Table 1, while Table 2 lists the C1/C2 and C2/C3 values in the acidification phase. Since polymerization degree may be decreased and lignin can be transformed (Kataeva et al., 2013) to guaiacylglycerol, syringylglycerol, and phenolic acids, C1/C2 of most lignocellulose materials (except reed, office paper and newspaper) decreased from the methane production phase to the end phase (Table 1). Although the C2/C3 values of fat pork, grapefruit peel, office paper and newspaper decreased, those of most materials increased. And C2/C3 value of watermelon peel was increased, however, C2 value of watermelon peel was decreased. Therefore, C2/C3 values of more materials have consistent variation during anaerobic digestion. While anaerobic digestion of food waste failed due to excessive acidification, the C2/C3 values decreased to no more than 0.50. 3.3. PCA analysis of absorption and fluorescence properties PCA was used to evaluate the similarities and differences between anaerobic digestion stages and spectroscopy indicators of liquid digestate. The first four principle components, PC1, PC2, PC3, and PC4, explained 44%, 28%, 13%, and 7% of the variation, respectively, which accounted for 92% of the total variance. A biplot of PCA analysis was used to investigate the inter-relationships between the cases (liquid digestate of different degradable solid wastes at different anaerobic digestion phases) and the variables (spectroscopy indicators). The biplot of the first two components
75
W. Zheng et al. / Bioresource Technology 161 (2014) 69–77
(a) Component 1
.4
550
(d)
.3 450
Loadings
Excitation wavelength (nm)
500
400 350 300
Excitation Emission .2
.1
250 200 250
300
350
400
450
500
550
0.0 200
600
300
550
.6
500
.5
450
.4 Loadings
Excitation wavelength (nm)
(b) Component 2
400 350
500
600
(e)
Excitation Emission
.3 .2
300
.1
250 200 250
300
350
400
450
500
550
0.0 200
600
300
(c) Component 3
.7
500
.6
450
.5 Loadings
550
400 350
250
.1
350
400
450
500
600
550
600
Emission wavelength (nm)
Excitation Emission
.3 .2
300
500
(f)
.4
300
200 250
400 Wavelength (nm)
Emission wavelength (nm)
Excitation wavelength (nm)
400 Wavelength (nm)
Emission wavelength (nm)
0.0 200
300
400
500
600
Wavelength (nm)
Fig. 1. Three PARAFAC components identified from EEMs dataset. Contour plots (a–c); Loadings of excitation and emission (d–f).
describing most of the system variability was represented as orthogonal axes, and samples were projected in the bi-dimensional space of the biplot (Fig. 3). As shown in Fig. 3, SUVA254, PIII,n and C2/C3 increased from the methane production phase to the end phase; however, PI,n and C3 decreased. These changes illustrate that the degree of stabilization increased during the anaerobic digestion of biodegradable materials. While the anaerobic digestion of food waste failed due to excessive acidification, SUVA254, PIII,n and C2/C3 decreased, and PI,n and C3 increased due to the accumulation of protein hydrolysate and the inhibition of humification processes. Comparatively, other parameters such as E4/E6, PII,n, PIV,n, PV,n, C1, C2, and C1/C2 were not significantly related to the degree of stabilization. The
PCA results were consistent with the analyses described above, and PCA analysis was more intuitive for evaluation. 3.4. Selection of suitable indicators for evaluation of anaerobic stabilization The results of this study indicate that the increase of SUVA254 always corresponded to the increase of humification degree during anaerobic reaction for different waste, although it may fluctuate at the beginning of anaerobic digestion owing to aromatic amino acids produced from food waste and lignin-derived aromatics dissolved from lignocellulose materials. However, SUVA254 values were highest at the end phase (>1.0 L mg 1 m 1). Weishaar et al.
76
W. Zheng et al. / Bioresource Technology 161 (2014) 69–77
0.00 0.9
2 nt one mp Co
Co mp one nt 3
0.25 0.6
0.50
0.3
0.75
1.00 0.0
0.3
0.6
0.9
0.0
Component 1 Fig. 2. Ternary scatter plot of liquid digestate during different biodegradable solid waste anaerobic digestion based on EEM–PARAFAC model. , lag phase of food waste; , methane production phase of food waste; , the end of methane production phase of food waste; , acidification phase of food waste; , lag phase of lignocellulose waste; , methane production phase of lignocellulose waste; , the end of methane production phase of lignocellulose waste; , acidification phase of lignocellulose waste; , lag phase of fabric; , methane production phase of fabric; , the end of methane production phase of fabric; , lag phase of blank; , methane production phase of blank; , the end of methane production phase of blank.
waste interfere with the fluorescence peak of humic acid-like compounds, and NADH that accumulates during anaerobic digestion of food waste, lignocellulose waste, and fabric interferes with the fluorescence peak of humic acid-like compounds. Therefore, humic acid-like compounds with an Ex/Em wavelength of 260(350)/ 436 nm are unsuitable as a humification degree indicator. Owing to interference of fulvic acid-like compounds, the fluorescence peaks of tryptophan were inconsistent with the dissolved protein content. In contrast, undisturbed fluorescence peaks such as the fluorescence peak of fulvic acid-like compounds that had an Ex/ Em wavelength of 230/436 nm and the fluorescence peaks of tyrosine like compounds that had an Ex/Em wavelength of 220(270)/ 304 nm can be used as monitoring indicators. From the methane production phase to the end phase, PI,n decreased by 0.05–0.54 times, PIII,n increased by 0.01–0.54 times, and C3 decreased by 0.03–0.64 times. When anaerobic digestion of food waste failed due to excessive acidification, PI,n was higher than 0.40, C3 was higher than 0.55 and PIII,n was lower than 0.15. By now the application of spectroscopy indicators to anaerobic environment was mainly limited to the field of landfill leachate. He et al. (2011) indicated that SUVA254 of landfill leachate was increased with the landfill age increasing. Shao et al. (2012) indicated that PI,n decreased and PIII,n increased with the landfill age increasing. Those results were consistent with this study. However, the landfill waste contains several fluorescent chemical compounds that may release into leachate, and the huge heterogeneity of the landfilled waste may lead to that the collected leachate is generated from different waste zones with diverse degradation degree. Therefore, in the future works, there is a demand on more humification data collected from the anaerobic digesters of mixed waste with known biochemical components and known process status, in order to further verify the effectiveness of the distinct indicators identified in the present study. And compared with a study conducted Ghita et al. (2013), we identified fluorescence peaks that can be used as indicators during anaerobic digestion as well as a single fluorescence peak that can be selected as a monitoring indicator during the actual anaerobic digestion process as needed.
4. Conclusion
Fig. 3. Biplot illustrating the inter-relationships between anaerobic digestion phase and spectra parameters. , lag phase of food waste; , methane production phase of food waste; , the end of methane production phase of food waste; , acidification phase of food waste; , lag phase of lignocellulose waste; , methane production phase of lignocellulose waste; , the end of methane production phase of lignocellulose waste; , acidification phase of lignocellulose waste; , lag phase of fabric; , methane production phase of fabric; , the end of methane production phase of fabric; , lag phase of blank; , methane production phase of blank; , the end of methane production phase of blank.
(2003) observed that the data range of fulvic acid SUVA254 from environmental samples was between 0.6 and 3.9 L mg 1 m 1 and the humic acid SUVA254 from environmental samples was >5.0 L mg 1 m 1. Therefore, fulvic acid-like compounds were the main products of humus-like compounds during biodegradable solid waste anaerobic digestion and the digestate was still unstable. While anaerobic digestion of food waste failed due to excessive acidification, the SUVA254 value was
Contents lists available at ScienceDirect
Bioresource Technology journal homepage: www.elsevier.com/locate/biortech
Relationship between anaerobic digestion of biodegradable solid waste and spectral characteristics of the derived liquid digestate Wei Zheng a,b, Fan Lü a,b,⇑, Khamphe Phoungthong a,b, Pinjing He b,c a
State Key Laboratory of Pollution Control & Resource Reuse, Tongji University, Shanghai 200092, PR China Institute of Waste Treatment and Reclamation, Tongji University, Shanghai 200092, PR China c Centre for the Technology Research and Training on Household Waste in Small Towns & Rural Area, Ministry of Housing and Urban–Rural Development (MOHURD), PR China b
h i g h l i g h t s Fluorescence peaks of tyrosine were appropriate for protein evaluation during AD. SUVA254 and Ex/Em of 230/436 nm were suitable for evaluating humidification degree. E4/E6 and Ex/Em of 350/436 nm were unsuitable for evaluating humidification degree. Fluorescence peaks of tryptophan were inconsistent with dissolved protein content.
a r t i c l e
i n f o
Article history: Received 2 January 2014 Received in revised form 24 February 2014 Accepted 4 March 2014 Available online 15 March 2014 Keywords: Food waste Lignocellulose waste Fabric Biplot Humification
a b s t r a c t The evolution of spectral properties during anaerobic digestion (AD) of 29 types of biodegradable solid waste was investigated to determine if spectral characteristics could be used for assessment of biological stabilization during AD. Biochemical methane potential tests were conducted and spectral indicators (including the ratio of ultraviolet–visible absorbance at 254 nm to dissolved organic carbon concentration (SUVA254), the ratio of ultraviolet–visible absorbance measured at 465 nm and 665 nm (E4/E6), and the abundance of fluorescence peaks) were measured at different AD phases. Inter-relationship between organic degradation and spectral indicators were analyzed by principle component analysis. The results shows that from methane production phase to the end of methane production phase, SUVA254 increased by 0.16–10.93 times, the abundance of fulvic acid-like compounds fluorescence peak increased by 0.01– 0.54 times, the abundance of tyrosine fluorescence peak decreased by 0.03–0.64 times. Therefore, these indicators were useful to judge the course of mixed waste digestion. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction Global municipal solid wastes (MSWs) generation reached an estimated 1.3 billion tonnes per year in 2010 (Hoornweg and Bhada-Tat, 2012). MSWs contains many biodegradable components, but these components vary by region. For example, MSWs in the United States comprised 28% of paper, 14% of garden waste, and 15% of food waste in 2011 (US EPA, 2013). However, in Shanghai, the paper, garden waste, and food waste accounted for 11%, 1%, and 64% of the MSWs respectively in 2008–2009 (Shanghai Environmental Engineering Design and Science Academy, 2009). These components of MSWs are important organic sources of methane production through anaerobic digestion. ⇑ Corresponding author at: Institute of Waste Treatment and Reclamation, Tongji University, Shanghai 200092, PR China. Tel./fax: +86 21 65986104. E-mail address: [email protected] (F. Lü). http://dx.doi.org/10.1016/j.biortech.2014.03.016 0960-8524/Ó 2014 Elsevier Ltd. All rights reserved.
The residue of anaerobic digestion, called digestate, is usually used as soil amendments, and its biological stability is an important indicator of the environmental risk posed by the digestate (Senesi and Plaza, 2007). Several indicators have been proposed for assessment of the biological stability of MSWs and its derived materials, including the respirometric index, biomethane potential production, and the ratio of 5-day biological oxygen demand to chemical oxygen demand (BOD5/COD) (Cossu et al., 2001). However, these parameters are limited by their high cost, long testing time, and low representativity in the presence of inhibiting substances. Comparatively, ultraviolet–visible (UV–Vis) absorption (Vieyra et al., 2009; Hunger and Weitkamp, 2001) and fluorescence spectra (Antunes and Da Silva, 2005) as assessment methods to represent biological stability and stabilization of dissolved organic materials (DOM) are rapid, highly sensitive and nondestructive. Nevertheless, few studies have been conducted to investigate the use of spectral indicators to evaluate biological stability and
70
W. Zheng et al. / Bioresource Technology 161 (2014) 69–77
stabilization of anaerobic digestion, which were only limited to the leachate generated from anaerobic landfill for mixed MSWs (He et al., 2011; Lü et al., 2009). The ratio of the UV–Vis absorbance measured at k = 254 nm to the dissolved organic carbon (DOC) concentration (SUVA254) and the ratio of the UV–Vis absorbance measured at k = 465 nm to that measured at k = 665 nm (E4/E6) are widely used to represent biological stability in UV–Vis absorption spectral indicators. SUVA254 and E4/E6 have been reported to be particularly useful indicators for investigation of environmental samples (Weishaar et al., 2003; Chen et al., 1977), DOM of compost (Shao et al., 2009; Vieyra et al., 2009), and landfill leachate (He et al., 2011). The excitation–emission matrix (EEM) spectra technology is increasingly being applied to represent biological stability and stabilization of environmental samples (Ishii and Boyer, 2012), DOM of compost (Yu et al., 2010), and investigation of anaerobic digestion of sludge (Luo et al., 2013), livestock manure (Wan et al., 2012), and mixed municipal solid waste (Ghita et al., 2013). Nevertheless, since the biochemical characteristics of various MSWs are quite different (Zheng et al., 2013), the DOM of various MSWs has different fluorescence spectra characteristics. For example, lignin is fluorescent, but polysaccharide is not; therefore, the fluorescence spectrum of lignocellulose waste is quite different from that of low lignocellulose content waste such as food waste. Accordingly, it is necessary to investigate whether EEM fluorescence spectra have consistent variation during anaerobic digestion processes of diverse biodegradable solid waste. Most studies that have employed EEM spectra to probe the anaerobic digestion process have only used the ‘‘peak picking’’ technique (Ghita et al., 2013; Luo et al., 2013; Wan et al., 2012), which makes minimal use of the large quantity of data available in each EEM spectrum (Chen et al., 2003b). Comparatively, fluorescence regional integration (FRI) and parallel factor (PARAFAC) techniques (Shao et al., 2012) allow a more complete data mining than the traditional peak picking technique. Accordingly, combination of EEM with FRI and PARAFAC may provide deeper insight into the composition and behaviour of the DOM. This study was conducted to investigate the relationship between spectral indicators and biological stabilization properties during anaerobic digestion of various biodegradable solid wastes with different biochemical compositions. To accomplish this, a biochemical methane potential (BMP) assay was utilized to simulate anaerobic digestion, which was then evaluated during the lag phase, methane production phase, end of methane production phase and excessive acidification phase. A combination of the absorption and fluorescence spectral properties, EEM–FRI/PARAFAC analysis and multivariate analysis was then used to compare the applicability of spectroscopy indicators to represent biological stability and stabilization during biodegradable solid waste anaerobic digestion.
2. Methods 2.1. Degradable material and inoculum Based on the waste biochemical characteristics and production quantity, 29 types of precursor materials with the potential to be discarded as biodegradable solid waste were selected, including papers (newspaper, office paper, toilet paper), kitchen waste (fish bone, pork bone, lean pork, fat pork, soybean, potato, celery, lettuce), fruit waste (sugarcane residue, banana peel, orange peel, apple core and peel, watermelon peel, grapefruit peel), ligocellulosic materials (peanut shell, reed, the grass Cynodon dactylon, tea residue, bamboo leaf, bamboo branch, camphor tree leaf, camphor
tree branch, metasequoia leaf, metasequoia branch, and cotton), and fabric. All collected materials were shredded to a size of less than 1 mm and then prepared for biochemical composition as described by Zheng et al. (2013). The anaerobic inoculum was the digestate collected from an anaerobic digestion plant used to treat a mixture of sewage sludge and food waste. The inoculum had a total solids (TS) content of 23.8 ± 1.7 wt% and the volatile solids (VS) accounted for 81.9 ± 2.5 wt% of the TS. 2.2. Biochemical methane potential assay Each 1 L glass bottle contained 100 g (wet weight) of inoculum, 400 g of nutrient medium, and 10 g (dry weight) of the tested biodegradable material, except for the reactors that contained toilet paper, orange peel and grapefruit peel. Only 5 g (dry weight) of the latter three materials were added because preliminary tests indicated that a higher inoculum-to-substrate ratio was required to avoid acidification. A blank reactor that contained only inoculum was also incubated to measure the background methane production. All experiments were carried out in duplicate. The nutrient medium, cultivation condition and gas measurement method described by Zheng et al. (2013) was employed. The suspended mixtures in the bottles were obtained using a 5 ml syringe and then centrifuged for 10 min at 2000g. Next, the supernatant was then passed through a 0.45-lm, microfiber filter, after which was referred to as the liquid digestate. Based on the methane production curve (Zheng et al., 2013), the lag phase of all materials except bamboo leaf was longer than 2 days; therefore, the liquid digestate of the lag phase was collected on the first day. When the methane concentration was higher than the carbon dioxide concentration, the liquid digestate of methane production phase was collected. When the methane production curve reached a plateau, the liquid digestate was collected at the end of the methane production phase. When anaerobic digestion reactors were failed due to excessive acidification, the liquid digestate of acidification phase was collected. 2.3. Analytical methods 2.3.1. UV–Vis spectra UV–Vis spectroscopic measurements of the absorbance of liquid samples at 665, 465 and 254 nm were conducted using a UV-1800 UV–Vis spectrophotometer (Shimadzu, Kyoto, Japan) and 1-cm quartz cells. The unit of absorbance is m 1. And DOC whose unit is mg L 1 was measured using a total organic carbon (TOC) analyzer (TOC-V CPN, Shimadzu, Japan). SUVA254 values were determined by dividing the UV–Vis absorbance measured at k = 254 nm by the DOC concentration and then reported in L mg 1 m 1 (Shao et al., 2009). E4/E6 values were determined by dividing the UV–Vis absorbance measured at k = 465 nm by the UV–Vis absorbance measured at k = 665 nm. 2.3.2. Fluorescence spectra Prior to fluorescence spectroscopy analysis, the liquid digestate was diluted with 0.1 mol L 1 phosphate buffer to DOC < 10 mg L 1 to ensure that the maximum fluorescence signal was below the upper detection limit of the spectrometer and the disruption of inter- and intramolecular hydrogen bonds was eliminated (Ghita et al., 2013). The fluorescence EEM spectra of the liquid samples were recorded using a fluorescence spectrophotometer (Cary Eclipse, Varian, USA) in scan mode. To obtain fluorescence EEM, scanning emission (Em) spectral analysis was conducted from 250 to 600 nm at 2 nm increments by varying the excitation (Ex)
W. Zheng et al. / Bioresource Technology 161 (2014) 69–77
wavelength from 200 to 550 in 10 nm increments. The EEM of a control (Milli-Q water) was subtracted from each sample EEM. 2.4. Data analysis 2.4.1. Fluorescence regional integration (FRI) The Rayleigh and Raman scattering of EEM spectra were removed according to the protocol described by Bahram et al. (2006). EEMs were then normalized by dividing the spectrum by the corresponding DOC concentration, after which the corrected EEMs were plotted using SigmaPlot 11.0 (SPSS Inc., USA) with a contour plot graph. The FRI technique was employed to analyze the five excitation–emission regions of EEM spectroscopy according to Chen et al. (2003b). 2.4.2. PARAFAC modeling When compared with the FRI technique, PARAFAC can more precisely identify fluorophors with multiple excitation wavelengths at the same emission wavelength. PARAFAC analysis of EEM spectra was conducted as described by Stedmon and Bro (2008) using MATLAB 7.0 (Mathworks, Natick, MA) with the DOMFluor toolbox (www.models.life.ku.dk). The PARAFAC models with two to six components were computed for the EEMs. Determination of the correct number of components was primarily based on core consistency. 2.4.3. Statistical analysis To evaluate the relationship between spectroscopy indicators and anaerobic digestion status, PCA was employed using the CANOCO 4.0 (Biometris, Wageningen, Netherlands) software. 3. Results and discussion
71
the SUVA254 values were generally consistent with biological stability and stabilization of a routine anaerobic digestion process. 3.1.2. Evolution of E4/E6 E4/E6 is widely used to evaluate the degree of humification in aerobic environment because the E4/E6 value decreases as the polymerization degree, aromatization degree and molecular weight increase (Huang et al., 2006). The evolution of E4/E6 of liquid digestate during the anaerobic digestion of biodegradable solid waste is shown in Table 1, while Table 2 lists the E4/E6 values in the acidification phase. It could be found that the E4/E6 variation along AD process of different materials has not regular trend. E4/E6 is also influenced by pH, free radicals, total acidity (Chen et al., 1977), protein, polysaccharide (Vieyra et al., 2009) and microorganisms (Marschner et al., 2005). Though E4/E6 was widely used to investigate the biological stability of soil samples (Marschner et al., 2005; Chen et al., 1977) and compost liquid samples (Vieyra et al., 2009; Huang et al., 2006). But those samples were produced from aerobic environment. The reference to utilize E4/E6 as biological stability indicator in anaerobic environment was limited. And according to Table 1, from methane production phase to the end of methane production phase, E4/E6 values of metasequoia branch, reed, lettuce, peanut shell, watermelon peel, apple core and peel, tea residue, sugarcane residue, cotton, fabric fish bone, toilet paper and newspaper increased. Therefore, the E4/E6 variation of different materials was inconsistent with AD process. 3.2. Evolution of fluorescence characteristics during anaerobic digestion The evolution of EEM fluorescence spectra of liquid digestate is shown in Figs. A.1 and A.2 lists the EEM fluorescence spectra in the acidification phase. The EEM peaks overlapped; therefore, FRI and PARAFAC techniques were utilized to minimize mutual interference of fluorophores.
3.1. Evolution of UV–Vis absorption spectra during anaerobic digestion 3.1.1. Evolution of SUVA254 SUVA254 can reflect the aromatic carbon content in dissolved organic matter (DOM) (Weishaar et al., 2003) and indirectly reflect the degree of humification and biological degradability of the sample. Evolution of SUVA254 of liquid digestate during the anaerobic digestion of biodegradable solid waste is shown in Table 1, while Table 2 lists the SUVA254 values in the acidification phase. As shown in Table 1, SUVA254 values of most liquid digestates increased constantly as anaerobic digestion proceeded. However, unsaturated carbon bond compounds (such as aromatic protein) existed in the liquid digestate of food waste (such as grapefruit peel, celery, lettuce, etc.) and lignin-derived aromatics existed in the liquid digestate (Kataeva et al., 2013) of lignocellulose material (such as camphor tree branch, bamboo branch, newspaper, toilet paper, etc.); therefore, some SUVA254 values in the lag phase were higher than in the methane production phase. The SUVA254 values at the end of the methane production phase were highest (>1.0 L mg 1 m 1). The SUVA254 values were increased by 0.16– 10.93 times from the methane production phase to the end of the methane production phase. Shao et al. (2009) observed that the SUVA254 values of water extracts of MSWs after the hydrolysis-aerobic process were around 3.0 L mg 1 m 1, suggesting that the anaerobic digestion product in the present study was still unstable. The lowest SUVA254 values were observed when anaerobic digestion of food waste failed due to excessive acidification. Comparatively, although the digestion of lignocellulose waste maybe failed as a result of excessive acidification as well, the SUVA254 value did not decrease obviously owing to the lignin-derived aromatics dissolved in the liquid digestate. Nevertheless,
3.2.1. Evolution of EEM–FRI Chen et al. (2003b) proposed an FRI technique to quantify fluorescence spectra and divided the EEM fluorescence spectra into five emission–excitation regions. The percentage of fluorescence response (Pi,n) can quantitatively reflect the abundance of a specific region that represents a specific structural compound. The evolution of Pi,n during anaerobic digestion of biodegradable materials is shown in Table 1, while Table 2 lists the Pi,n values in the acidification phase. The PI,n values (except newspaper and fat pork) decreased by 0.05–0.54 times from the methane production phase to the end of methane production phase. In addition, the abundance of tyrosine-like compounds decreased, suggesting that protein was converted during anaerobic digestion. Region II and IV were adjacent to the fluorescent region of humus-like compounds. As a result, PII,n and PIV,n incorrectly reflect the abundance of tryptophan-like compounds and soluble microbial by-product-like compounds due to the interference of humus-like compounds; therefore, the variation regulation of PII,n and PIV,n was not unified among different types of biodegradable materials. PIII,n increased by 0.01–0.54 times and PV,n (except soybean) increased by 0.02–0.78 times. However, Nicotinamide Adenine Diuncleotide Hydrogen (NADH) (Li et al., 2011) and lignin (Muller et al., 2011) showed fluorescence peaks at Ex/ Em wavelengths of 350/440 nm and (270) 380/440 nm, respectively. PV,n was determined by humic acid-like compounds as well as by NADH and lignin. While anaerobic digestion failed due to excessive acidification, the fluorescence properties of liquid digestate were influenced by the biochemical characteristics of materials. Since amino acids accumulated during the acidification phase of food waste, PI,n increased (PI,n > 0.40) and PIII,n decreased
72
W. Zheng et al. / Bioresource Technology 161 (2014) 69–77 Table 1 Evolution of spectra characteristics during anaerobic digestion.
W. Zheng et al. / Bioresource Technology 161 (2014) 69–77
73
Table 1 (continued)
(continued on next page)
74
W. Zheng et al. / Bioresource Technology 161 (2014) 69–77 Table 1 (continued)
a b c
L: lag phase. M: methane production phase. E: end of methane production phase.
Table 2 Spectra characteristics of liquid samples from failed digesters. SUVA254 (L mg Grapefruit peel Orange peel Toilet paper a
0.146 0.152 0.214
1
m
1
)
E4/E6
PI,n
PII,n
PIII,n
PIV,n
PV,n
C1/C2
C2/C3
2.63 3.09 3.38
0.420 0.682 0.245
0.219 0.125 0.299
0.135 0.046 0.268
0.149 0.144 0.114
0.077 0.003 0.075
1.176 naa 0.653
0.314 0 1.025
na = not applicable.
(PIII,n < 0.15). However, lignin-derived aromatics dissolved in liquid digestate and the production of amino acid was limited in the acidification phase; therefore, PI,n and PIII,n did not change significantly. In general, PI,n was consistent with the changes in protein content during anaerobic digestion and PIII,n was consistent with the stabilization degree. 3.2.2. Evolution of EEM–PARAFAC Since the core consistency of three and four components were 92.6% and 87.9% respectively, the three-components PARAFAC model was used to quantitatively analyze the EEM dataset. The three components identified by the PARAFAC model are shown in Fig. 1. Component 1 (C1) has fluorescence peaks at an Ex/Em wavelength of 230(260, 350)/436 nm (Fig. 1(a)). C1 in this study was similar to the humic-like compounds identified by Coble et al. (1996, 1993), which had an Ex/Em wavelength of 230(260, 350)/ 420–450 nm. Additionally, NADH could accumulate during anaerobic digestion (Farabegoli et al., 2003) and lignocellulose waste contains lignin. NADH has an Ex/Em wavelength of 350/440 nm (Li et al., 2011) and lignin has an Ex/Em wavelength of 270(380)/ 440 nm (Muller et al., 2011). Therefore, the fluorescence intensity of C1 can be influenced by NADH and lignin. Component 2 (C2) has fluorescence peaks at an Ex/Em wavelength of 230(290)/ 388 nm (Fig. 1(b)). Baddi et al. (2013) indicated that peaks of Em at 390 nm were attributed to fulvic acid-like compounds; accordingly, C2 was similar to fulvic acid-like compounds. Component 3 (C3) generated fluorescence peaks at an Ex/Em wavelength of 220(270)/304 nm (Fig. 1(c)). C3 was similar to tyrosine-like compounds identified by Wolfbeis (1985), which had Ex/Em wavelengths of 220–275/300–305 nm. Ternary scatter plots of liquid digestate during anaerobic digestion of different biodegradable solid wastes are shown in Fig. 2. The C1 and C2 of most materials increased from the methane production phase to the end phase, while some C1 values (such as soybean and banana) and C2 values (such as fat pork, grapefruit peel, watermelon peel, newspaper and office paper) decreased. C3 decreased by 0.03–0.64 times from the methane production
phase to the end phase. While anaerobic digestion of food waste failed due to excessive acidification, C3 values increased to more than 0.55. Therefore, C3 was generally consistent with protein content during anaerobic digestion. Chen et al. (2003a) reported that a red shift of emission wavelength to longer wavelengths implied an increase of molecular size, aromatic polycondensation, and humification degree of humic-like compounds. Therefore, C1/C2 can be used to indicate the degree of polymerization of humic-like compounds. To avoid the mutual effects of NADH, lignin and humic-like compounds, C2/C3 was selected to evaluate the degree of humification. The evolution of C1/C2 and C2/C3 during biodegradable solid waste anaerobic digestion is shown in Table 1, while Table 2 lists the C1/C2 and C2/C3 values in the acidification phase. Since polymerization degree may be decreased and lignin can be transformed (Kataeva et al., 2013) to guaiacylglycerol, syringylglycerol, and phenolic acids, C1/C2 of most lignocellulose materials (except reed, office paper and newspaper) decreased from the methane production phase to the end phase (Table 1). Although the C2/C3 values of fat pork, grapefruit peel, office paper and newspaper decreased, those of most materials increased. And C2/C3 value of watermelon peel was increased, however, C2 value of watermelon peel was decreased. Therefore, C2/C3 values of more materials have consistent variation during anaerobic digestion. While anaerobic digestion of food waste failed due to excessive acidification, the C2/C3 values decreased to no more than 0.50. 3.3. PCA analysis of absorption and fluorescence properties PCA was used to evaluate the similarities and differences between anaerobic digestion stages and spectroscopy indicators of liquid digestate. The first four principle components, PC1, PC2, PC3, and PC4, explained 44%, 28%, 13%, and 7% of the variation, respectively, which accounted for 92% of the total variance. A biplot of PCA analysis was used to investigate the inter-relationships between the cases (liquid digestate of different degradable solid wastes at different anaerobic digestion phases) and the variables (spectroscopy indicators). The biplot of the first two components
75
W. Zheng et al. / Bioresource Technology 161 (2014) 69–77
(a) Component 1
.4
550
(d)
.3 450
Loadings
Excitation wavelength (nm)
500
400 350 300
Excitation Emission .2
.1
250 200 250
300
350
400
450
500
550
0.0 200
600
300
550
.6
500
.5
450
.4 Loadings
Excitation wavelength (nm)
(b) Component 2
400 350
500
600
(e)
Excitation Emission
.3 .2
300
.1
250 200 250
300
350
400
450
500
550
0.0 200
600
300
(c) Component 3
.7
500
.6
450
.5 Loadings
550
400 350
250
.1
350
400
450
500
600
550
600
Emission wavelength (nm)
Excitation Emission
.3 .2
300
500
(f)
.4
300
200 250
400 Wavelength (nm)
Emission wavelength (nm)
Excitation wavelength (nm)
400 Wavelength (nm)
Emission wavelength (nm)
0.0 200
300
400
500
600
Wavelength (nm)
Fig. 1. Three PARAFAC components identified from EEMs dataset. Contour plots (a–c); Loadings of excitation and emission (d–f).
describing most of the system variability was represented as orthogonal axes, and samples were projected in the bi-dimensional space of the biplot (Fig. 3). As shown in Fig. 3, SUVA254, PIII,n and C2/C3 increased from the methane production phase to the end phase; however, PI,n and C3 decreased. These changes illustrate that the degree of stabilization increased during the anaerobic digestion of biodegradable materials. While the anaerobic digestion of food waste failed due to excessive acidification, SUVA254, PIII,n and C2/C3 decreased, and PI,n and C3 increased due to the accumulation of protein hydrolysate and the inhibition of humification processes. Comparatively, other parameters such as E4/E6, PII,n, PIV,n, PV,n, C1, C2, and C1/C2 were not significantly related to the degree of stabilization. The
PCA results were consistent with the analyses described above, and PCA analysis was more intuitive for evaluation. 3.4. Selection of suitable indicators for evaluation of anaerobic stabilization The results of this study indicate that the increase of SUVA254 always corresponded to the increase of humification degree during anaerobic reaction for different waste, although it may fluctuate at the beginning of anaerobic digestion owing to aromatic amino acids produced from food waste and lignin-derived aromatics dissolved from lignocellulose materials. However, SUVA254 values were highest at the end phase (>1.0 L mg 1 m 1). Weishaar et al.
76
W. Zheng et al. / Bioresource Technology 161 (2014) 69–77
0.00 0.9
2 nt one mp Co
Co mp one nt 3
0.25 0.6
0.50
0.3
0.75
1.00 0.0
0.3
0.6
0.9
0.0
Component 1 Fig. 2. Ternary scatter plot of liquid digestate during different biodegradable solid waste anaerobic digestion based on EEM–PARAFAC model. , lag phase of food waste; , methane production phase of food waste; , the end of methane production phase of food waste; , acidification phase of food waste; , lag phase of lignocellulose waste; , methane production phase of lignocellulose waste; , the end of methane production phase of lignocellulose waste; , acidification phase of lignocellulose waste; , lag phase of fabric; , methane production phase of fabric; , the end of methane production phase of fabric; , lag phase of blank; , methane production phase of blank; , the end of methane production phase of blank.
waste interfere with the fluorescence peak of humic acid-like compounds, and NADH that accumulates during anaerobic digestion of food waste, lignocellulose waste, and fabric interferes with the fluorescence peak of humic acid-like compounds. Therefore, humic acid-like compounds with an Ex/Em wavelength of 260(350)/ 436 nm are unsuitable as a humification degree indicator. Owing to interference of fulvic acid-like compounds, the fluorescence peaks of tryptophan were inconsistent with the dissolved protein content. In contrast, undisturbed fluorescence peaks such as the fluorescence peak of fulvic acid-like compounds that had an Ex/ Em wavelength of 230/436 nm and the fluorescence peaks of tyrosine like compounds that had an Ex/Em wavelength of 220(270)/ 304 nm can be used as monitoring indicators. From the methane production phase to the end phase, PI,n decreased by 0.05–0.54 times, PIII,n increased by 0.01–0.54 times, and C3 decreased by 0.03–0.64 times. When anaerobic digestion of food waste failed due to excessive acidification, PI,n was higher than 0.40, C3 was higher than 0.55 and PIII,n was lower than 0.15. By now the application of spectroscopy indicators to anaerobic environment was mainly limited to the field of landfill leachate. He et al. (2011) indicated that SUVA254 of landfill leachate was increased with the landfill age increasing. Shao et al. (2012) indicated that PI,n decreased and PIII,n increased with the landfill age increasing. Those results were consistent with this study. However, the landfill waste contains several fluorescent chemical compounds that may release into leachate, and the huge heterogeneity of the landfilled waste may lead to that the collected leachate is generated from different waste zones with diverse degradation degree. Therefore, in the future works, there is a demand on more humification data collected from the anaerobic digesters of mixed waste with known biochemical components and known process status, in order to further verify the effectiveness of the distinct indicators identified in the present study. And compared with a study conducted Ghita et al. (2013), we identified fluorescence peaks that can be used as indicators during anaerobic digestion as well as a single fluorescence peak that can be selected as a monitoring indicator during the actual anaerobic digestion process as needed.
4. Conclusion
Fig. 3. Biplot illustrating the inter-relationships between anaerobic digestion phase and spectra parameters. , lag phase of food waste; , methane production phase of food waste; , the end of methane production phase of food waste; , acidification phase of food waste; , lag phase of lignocellulose waste; , methane production phase of lignocellulose waste; , the end of methane production phase of lignocellulose waste; , acidification phase of lignocellulose waste; , lag phase of fabric; , methane production phase of fabric; , the end of methane production phase of fabric; , lag phase of blank; , methane production phase of blank; , the end of methane production phase of blank.
(2003) observed that the data range of fulvic acid SUVA254 from environmental samples was between 0.6 and 3.9 L mg 1 m 1 and the humic acid SUVA254 from environmental samples was >5.0 L mg 1 m 1. Therefore, fulvic acid-like compounds were the main products of humus-like compounds during biodegradable solid waste anaerobic digestion and the digestate was still unstable. While anaerobic digestion of food waste failed due to excessive acidification, the SUVA254 value was
