| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Dipartimento di Produzione Vegetale, Università degli Studi di Milano, Via Celoria 2, 20133, Milan, Italy
* Corresponding author (fabrizio.adani{at}unimi.it).
Received for publication August 7, 2003.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: DRI, dynamic respiration index • DRIcum, cumulative dynamic respiration index value for 96 h • DRIcumadj, cumulative dynamic respiration index value for 96 h minus the lag phase • DRIDiProVe, average instantaneous dynamic respiration index value taken during the 24 h of the most intense biological activity • DRIhcumadj, cumulative dynamic respiration index value for 96 h minus the lag phase and standardized with respect to the number of hours • DRIimax, maximum instantaneous dynamic respiration index value measured during the entire test • MSW, municipal solid waste • OUR, oxygen uptake rate • VS, volatile solids
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Biological stability determines the extent to which readily biodegradable organic matter has decomposed (Lasaridi and Stentiford, 1996). It identifies the actual point reached in the decomposition process and represents a gradation on a recognized scale of values, which thus enable comparison of the process of decomposition (Lasaridi and Stentiford, 1996). Knowing the degree of biological stability possessed by the organic matter, not only during the aerobic biological processing but also to be found in the final products, is important for the process to be controlled effectively, for the products to be used beneficially, and in optimizing the design of the processing plant (Lasaridi and Stentiford, 1998). In fact, stability affects the potential for odor generation, biomass reheating, residual biogas production, regrowth of pathogens, phytotoxicity, plant disease suppression ability, and process parameters such as airflow rate and retention time (Iannotti et al., 1993; Müller et al., 1998).
Many analytical methods have been proposed for the determination of biological stability (Chanyasak and Kubota, 1981; Iannotti et al., 1992; Adani et al., 1995; Avnimelech et al., 1996; United States Composting Council, 1997a). Among these, methods determining the respiratory activity of the biomass have received more attention from researchers in considering both CO2 production (Naganawa et al., 1990; Willson and Dalmat, 1986) and O2 uptake (Iannotti et al., 1992; Paletski and Young, 1995; Lasaridi and Stentiford, 1998). For respirometric purposes oxygen uptake is preferred (Lasaridi and Stentiford, 1996) and has been proposed for adoption as the standard method (American Public Health Association, 1985; American Society for Testing and Materials, 1992; United States Composting Council, 1997b).
A respiration test by oxygen uptake can be subdivided into static and dynamic methods (Scaglia et al., 2000) since oxygen uptake measurement is performed in the absence (static) or presence (dynamic) of continuous aeration of the biomass. Static methods, such as the widely used Sapromat (Binner and Zach, 1999), SOLVITA (Changa et al., 2003), or United States Composting Council (1997b) methods, suffer from the disadvantage that they do not allow the oxygen to be dispersed throughout the biomass, thus limiting diffusion and mass transfer (Paletski and Young, 1995). Minimization of diffusion limits is important because limited oxygen transfer through the biomass layers and into the bacterial cell wall is typically considered to be the rate-limiting step in fixed-film biological reactions that exist in organic matrices (Paletski and Young, 1995). Therefore, when static methods are used, underestimation of oxygen uptake is possible, especially when fresh organic matter is present. These problems can be solved by continuous aeration of biomass (Adani et al., 2001) and/or by continuous biomass stirring combined with intermittent aeration, obtained under a liquid condition (SOUR method) (Lasaridi and Stentiford, 1998). Nevertheless, the SOUR method falls well short of real conditions for three reasons: the use of solid biomass in a liquid medium, the use of very low particle size (i.e., <1 mm), and the dependence of SOUR on the water-soluble fraction (Adani et al., 2003a). Consequently, a dynamic respiration index (DRI) reflecting the solid-state aerobic process has been developed (Adani et al., 2001).
The solid state aerobic process is a complex bioprocess involving many coupled physical and biological mechanisms (Chad and Walker, 2001). Over the years, investigators have proposed various mathematical approaches to describe this complex process (Keener et al., 1993; Kaiser, 1996; Hamoda et al., 1998; Chad and Walker, 2001). Recently, Hamelers (2001)(2002) proposed a deductive model containing the basic parameters representing the theoretical basis of the process (Hamelers, 2002; Liwarska-Bizukojc et al., 2002; Weber et al., 2002). It describes the aerobic biological process by using oxygen uptake rate (OUR) as well as microbiological, physical, and chemical aspects (Schumpe and Quicker, 1979; Sonnleitner, 1983; Luong and Volesky, 1983). The value of this model lies in the fact that it indicates the importance of oxygen uptake as a result of biological processes. As the DRI is a measure of the real-time OUR (Adani et al., 2001), the proposed model can also be expected to describe the respirometric test accurately. This will result in additional information on the basic principles and validity of the method.
This work was designed to evaluate the DRI as a descriptor of the biological stability of organic wastes possessing different characteristics and originating from different sources. In addition, a study was conducted to elucidate the most appropriate method of displaying results and propose possible limits to biological stability.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
Respirometer Apparatus
Respiration test were performed with a 148-L-capacity adiabatic respirometric reactor (Costech International, Cernusco S.N., Italy; DiProVe, Milan, Italy). The respirometer consisted of an insulated container (the reactor), a control cabinet, an air supply system, a PC unit, and a biofilter. A Clark-type temperature compensation electrode and differential-pressure electronic transmitter assured both oxygen and airflow measurements every 10 s. These instantaneous data where then used by the software for DRI calculation. An extensive description of the scientific apparatus was reported by Adani et al. (2001).
Dynamic Respiration Index Determination
The dynamic respiration index was determined according to Adani et al. (2001). This method was proposed for different types of compost derived from wastes, such as MSW and derived products, yard waste, source-separated organic waste, and other types of organic wastes, that do not have toxicity levels that are inhibitory to the microorganism (American Society for Testing and Materials, 1996).
The oxygen uptake was determined by measuring the difference in oxygen concentration (mL L–1) between the inlet and outlet air flow and the air having passed through the biomass, as well as by using knowledge of the absolute content of starting VS (kg) in the biomass, the flow rate (L h–1), and the time (h) during which oxygen consumption was measured. The instantaneous DRI was calculated as:
 | [1] |
is the acquisition time (2 h), 
O2 (mL L–1) is the difference in oxygen concentration in the inlet and outlet air flow of the reactor, Vg (L mol–1) is the volume occupied by one mole of gas at inlet air temperature, 31.98 (g mol–1) is the molecular weight of O2, and VS (kg) is the total volatile solids present at the time of measurement (starting VS). The trials lasted 96 h each and throughout this time oxygen concentration (v/v %) and DRI (mg O2 kg–1 VS h–1) measurements were recorded hourly. Each trial was made in one replicate (Adani et al., 2001). With regard to the proposed method, a normal distribution of DRI (Shapiro–Wilk test) and good repeatability and reproducibility indices (International Organization for Standardization, 1994) (variation coefficient ranged from 3.63 to 11.80% and 4.33 to 9.64%, respectively) have previously been described (Adani et al., 2003b).
Biological Stability Expression by Dynamic Respiration Index
The degree of biological stability was calculated using five methods representing different ways of expressing DRI (Table 2): (i) the average value of the instantaneous respiration index (DRIi) taken during the 24 h of the most intense biological activity (DRIDiProVe), (ii) the maximum value (DRIimax) measured during the entire test, (iii) the cumulative value for 96 h (DRIcum), (iv) the cumulative value for 96 h minus the lag phase (DRIcumadj), and (v) the cumulative value for 96 h minus the lag phase and standardized with respect to the number of hours (DRIhcumadj).
|
In Phase 1 the aerobic process is not limit by soluble substrate. If O2 and moisture are not limiting (e.g., O2 content in the free air space is >140 mL L–1), oxygen uptake increases with microbial growth rate (µm), and the maximum conversion rate will be maintained until soluble substrate concentration decreases below the threshold at which this growth rate is maintained. As soluble substrate is depleted, oxygen uptake decreases dramatically and microbial growth activity shifts to the hydrolysis of the insoluble substrate (curing phase).
The model describes the aerobic process at the single particle level (Hamelers, 2001, 2002). Thus, as the organic matrix is formed by particles differing in size (Lc), and as the water and air distribution through the composting bed determines the presence of secondary particles (solid particles linked by water) (Hamelers, 2001), oxygen uptake will be a measure of average oxygen uptake determined as the integral of particle size (from zero to infinity) of the DRI of a specific particle size distribution (scaled particle size, 
) and the probability density (characteristic shape factor, 
c) of this specific particle size (Hamelers, 2001, 2002).
The model assumes the following form (Hamelers, 2001):
 | [2] |
c is the characteristic shape factor, is the scale particle size, and Ah is the hydrolytic activity (mg O2 kg–1 VS h–1).
By solving Eq. [2] for c = 2 (Hamelers, 2001), and considering the terms on the right in Eq. [2] as constant and equal to Ah, the oxygen uptake can be modeled by the equation:
 | [3] |
This assumes that for t 
ts (ts is the time in which DRIimax is reached, i.e., switch time), the DRI(t) only depends on the microbial growth rate; at the end of this phase the soluble fraction of the smallest particles is completely depleted. For t 
ts, we assume that the process continues through the depletion of the residual soluble fraction of the larger particles. In this case the substrate becomes the limiting factor and DRI becomes a function of the scaled particle size (). For t ts1 (ts1 is the curing time), it is assumed that the process continues through the respiration of the fraction that became soluble after hydrolysis. As the hydrolysis rate is much lower than the oxidation rate, the hydrolytic activity can be directly measured by DRI(t)tts1 
DRIc (Hamelers, 2001), assuming DRIc = Ah. For a short period it can be assumed that the hydrolysis rate (Ah) is constant.
Assuming an initial (DRI0) and a maximum (DRIimax) oxygen uptake as direct measurements of the minimum and maximum biomass density (Adani, 2002), the ß parameters can be calculated as:
 | [4] |
 | [5] |
 | [6] |
These equations introduce the DRI as a direct measure of biomass density, which is itself a function of time. In this case, the DRI trend would be represented more correctly by a no-root-logistic approach (Eq. [7]) (Hamelers, 2001), and so the model would become an inductive model:
 | [7] |
Both model equations (i.e., no-root and root logistic approaches), were used in our study, using parameters calculated by Eq. [4] to [6].
Model calculations were made by ad hoc software prepared using Microsoft Visual Basic (Microsoft, 1997). This software calculated simulated values by solving the equations proposed by the model (Eq. [3]–[7]). The model was applied to each trial by starting from the end of the lag phase. The lag phase indicates the time during which there is no respiration activity; its duration varied with the tested sample as it depends on the sample history (e.g., process temperature, moisture, and conservation) (Table 3).
|
/+, optimum = 1), which, if positive, indicates that the model is a better predictor than the average of the measured values; and (iv) the coefficient of residual mass (range = 0–1, optimum = 0), which assumes a positive sign in cases of model underestimation. Moreover, the parameters of the regression equation between observed and predicted values were calculated. |
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
|
|
|
The trends for biodried wastes were similar. The DRI profiles (Fig. 2) were very similar to those of biostabilized samples (Samples 4–6). This was due to the fact that the biodried process caused a partial degradation of easily degradable organic matter giving, more properly, a biodried-biostabilized waste (Adani et al., 2002). Once again, under this condition the test accurately described the respirometric activity.
Composted wastes showed a similar trend to fresh material (Samples 11–13; Fig. 3). Long lag phases characterizing these samples did not permit the respirometric activity to be completed within the 96 h. Above all, this was true of Samples 11 and 13.
The different amounts of organic fraction used in the mixtures (Table 1) produced differing respirometric activity. In particular, Sample 11, characterized by a lower content of organic fraction, gave lower DRI values than Sample 12 (Table 4). Conversely, the higher content of the organic fraction for Sample 13 did not produce any increase in the respirometric activity with respect to Sample 12. In this case, the long lag phase that occurred in Sample 13 probably did not allow the respirometric activity to be completed within 96 h. High-rate composting processes led to the stabilization of the composted organic wastes (Table 4; Fig. 3), with the exception of Sample 12, which showed DRI values similar to those of the initial sample (Fig. 3). This result can be explained by the fact that mixtures characterized by a high organic fraction can fail for optimal bulk density and porosity. This led to semi-anaerobic conditions with organic acid production and low pH (pH of Sample 16 was 5.6). This condition determined a slowing of the degradation activity and consequently an accumulation of easily degradable soluble fraction coming from hydrolysis of polymers. Therefore, the test gave high DRI values and longer lag phase.
From the above considerations it appears that DRI profiles were not similar for all of the organic wastes tested. Lower respirometric activity (more stable wastes) gave less pronounced DRI profiles (Fig. 1–3) that became practically flat for samples characterized by low DRI values (higher biological stability) (Samples 14 and 15). Both the length of the lag phase and the degree of the biological stability affected the extent of the respirometric activity, which, as a result, cannot be completed within 96 h (Samples 2, 11, 12, 13, and 16).
Agreement between the Dynamic Respiration Index and the Model
Figures 1 to 3 show the agreement between the observed and the simulated DRI values. The statistical indices reflecting agreement are detailed in Table 5. Both the no-root and the root logistic approaches gave good results for Samples 1 and 2 and 4 to 13.
|
The model did not perform well for Samples 3, 14, 15, and 16. Samples 3 and 16 (Fig. 1 and 3, respectively) showed a similar trend characterized by two peaks, probably due to experimental error during the recording of the oxygen uptake rate. Nevertheless, the model was able to predict the general DRI trend, although with less precision than for the other samples. The poor precision displayed by the models for Samples 14 and 15 was of a different nature (Fig. 3). In this case, the samples showed high biological stability, as indicated by the low DRI values (Table 4) (Scaglia et al., 2000). High biological stability means a low content of readily available degradable substrate (soluble fraction) (Hamelers, 2001; Adani et al., 2003a). Therefore, if a low content of soluble fraction is assumed for Samples 14 and 15, the logistic approach cannot describe these processes precisely. In this case, the oxygen uptake rate only depended on the soluble fraction that became available after processes of hydrolysis took place. Thus, hydrolysis kinetics should be used to describe the processes [i.e., DRI(t) = Ah].
A deductive model was proposed to describe the oxygen uptake rate. Nevertheless, the earlier assumption and the method used to determine the model parameters µn, ß, and , resulted in an inductive approach. The aim of this study was to use the model to improve understanding of oxygen uptake from a theoretical point of view.
The use of parameters calculated by Eq. [4], [5], and [6] also provided a good fit with the root-logistic equation (deductive approach) (Fig. 1–3), except for the first stages of the trials when DRIs were higher than those of the experimental data. It was found that as the DRI approached its maximum value (DRIimax), the experimental and calculated values became similar.
The root logistic model is able to consider the fact that as DRI(t) approached maximal biomass density, the oxygen penetration depth (
p,min) decreased, leading to a decrease in aerobic biomass activity (Hamelers, 2001). Oxygen penetration varied with biomass density, reaching its minimum when the biomass concentration reached its maximum (Hamelers, 2001). If p,min influenced the process (Oostra et al., 2001; Bishop et al., 1995), then an effective growth rate constant (µeff) (Eq. [8]) (Hamelers, 2001) and an effective dimensionless biomass concentration (ßeff) (Eq. [9]) (Hamelers, 2001) should be reconsidered in the root logistic approach, as reported in the original model (Eq. [3]):
 | [8] |
 | [9] |
If particle thickness (Lc) is greater than minimal oxygen diffusion depth (Lp,min), the dimensionless minimal penetration depth, defined as p,min = Lp,min/Lc, is close to zero. In this case µeff equals µn (Eq. [8]) and ßeff equals ß (Eq. [9]) (Hamelers, 2001), and so the no-root logistic function can be used to describe OUR.
Hamelers (2001) reported that if data were generated by a root logistic rate equation with a growth rate µeff, a no-root logistic relationship works well, with a net growth rate constant µn of 72% µeff. In the present work the two models were shown to differ little for the same microbial growth rate (Table 4). Hamelers (2001) indicated that Lc of >4 mm led to no change in µeff. Thus, as a first approximation, this particle size could be used as the limit up to which oxygen penetration depth does not influence the process (i.e., p,min << 1). The respirometric test proposed in the present work used larger masses (>10 kg) than those generally proposed in other tests (maximum 1 kg) (e.g., American Society for Testing and Materials, 1996). This allowed the use of samples of larger particle size than those generally used (<10 mm). In this way, if conducted on a biomass with a particle size greater than the usual 4 mm, the respirometric test was not dependent on the depth of oxygen penetration. This was especially true under the experimental conditions adopted (i.e., O2 = 140 mL L–1) and the biomass' free air space (FAS > 30%) (Adani et al., 2001; Rudrum et al., 2002). Therefore, the assumption that particle size (Lc) is greater than minimal diffusion depth (Lp), giving dimensionless minimal penetration depth (p,min) << 1, is correct.
Expression of the Biological Stability by the Dynamic Respiration Index
Different methods of calculating DRI (Table 2) yielded similar results for the degree of biological stability (Table 4). Correlation analysis confirmed the agreement among the different methodologies used (Table 6). Thus, DRIimax described the biological stability as well as the cumulative data (DRIcum or DRIhcum). This was in agreement with the theoretical model that describes the DRI trend over time as a function of the maximum respiration activity (DRIimax) (Fig. 1–3; Eq. [3]).
|
The best method for determining biological stability will be the one requiring minimal analytical time. Respiration peak could well describe biological stability, but the use of a single value could lead to inaccuracy due to possible errors and particle size effects (Hamelers, 2001). On the other hand, cumulative data require a larger amount of analytical time and, sometimes, the presence of a long lag phase does not allow the test to be completed within 96 h. The best solution is to take the average DRI over the 24 h spanning the periods of most intense biological activity (DRIDiProVe) (Adani et al., 2001). Nevertheless, it should be considered in the future to extend the incubation time (t > 96 h), if oxygen consumption in the last 24 h is higher than during the previous 24 h (American Society for Testing and Materials, 1996).
Dynamic Respiration Index: Proposed Biological Stability Limits
Scaglia et al. (2000) indicate biological stability for fresh and mature compost for DRIDiProVe of 1000 and 500 mg O2 kg–1 VS h–1, respectively. Values in this range are not yet reported in the literature, probably because the method and the results-expression format are relatively new.
In 1996, the American Society for Testing and Materials (ASTM) proposed a dynamic respirometric method to test compost stability. The ASTM method differs from that proposed here because it is performed at a preset temperature (58°C) and not under adiabatic conditions. Moreover, the results are reported as the cumulative for 96 h (Table 7). Nevertheless, using the results from the present work, a cumulative DRI (DRIcum) value corresponding to the above-mentioned stability values, expressed as DRIDiProVe, was determined. Regression calculation for DRIcum versus DRIDiProVe (R2 = 0.96, p < 0.01) was used to carry this out. Results were then compared with the cumulative values reported by ASTM and the corresponding stability classes of the self-heating test (American Society for Testing and Materials, 1996) (Table 7). A stability class for DRIDiProVe of 1000 mg O2 kg–1 VS h–1 and DRIcum of 57190 mg O2 kg–1 VS 96 h–1 indicated medium biological stability (i.e., in line with the ASTM compost Classes II and III and the self-heating stability Classes III and IV). On the other hand, the DRIDiProVe of 500 mg O2 kg–1 VS h–1 and the DRIcum of 28950 mg O2 kg–1 VS 96 h–1 indicated high biological stability, corresponding to compost Class 4 of the ASTM classification and the self-heating test.
|
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The methodology proposed in the present work represents a good tool compared with other methods proposed: it is able to fully describe the aerobic process, providing a realistic description of the biological stability. Moreover, this method works on large waste particle sizes, allowing the use of full-scale particle size and reducing the particle-size effect on the oxygen uptake rate (Hamelers, 2001). All these advantages mean that additional information can be obtained from the test such as the specific airflow rate (m3 Mg–1 dry matter h–1) required by the waste to degrade organic matter under optimal conditions (Adani et al., 2003c).
The theoretical model proposed by Hamelers (2001) works well, and can be used to predict behavior during respiration index determination. Furthermore, if it is considered as a mechanistic model, it will be a useful tool for further studies (e.g., the effect of the particle size on OUR), reducing the needs for a huge experimental effort.
In conclusion, the DRI can be used for stability measurement. Dynamic respiration index values, calculated as a mean of 24 h of the highest microbial activity (DRIDiProVe), of 1000 and 500 mg O2 kg–1 VS h–1 could be used to indicate medium (e.g., fresh compost) and high (e.g., mature compost) biological stabilities, respectively.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Related articles in JEQ:
This article has been cited by other articles:
|
|
 |
|
  F. Adani and M. Spagnol Humic Acid Formation in Artificial Soils Amended with Compost at Different Stages of Organic Matter Evolution J. Environ. Qual., June 23, 2008; 37(4): 1608 - 1616. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Vadose Zone Journal | Journal of Plant Registrations | ||||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal |
used in the dynamic respiration index (DRI) models.
