A thermophilic aerobic membrane reactor (TAMR) treating high-strength COD liquid wastes was submitted to an integrated investigation, with the aim of characterizing the biomass and its rheological behaviour. These processes are still scarcely adopted, also because the knowledge of their biology as well as of the physical-chemical properties of the sludge needs to be improved. In this paper, samples of mixed liquor were taken from a TAMR and submitted to fluorescent in situ hybridization for the identification and quantification of main bacterial groups. Measurements were also targeted at flocs features, filamentous bacteria, and microfauna, in order to characterize the sludge. The studied rheological properties were selected as they influence significantly the performances of membrane bioreactors (MBR) and, in particular, of the TAMR systems that operate under thermophilic conditions (i.e., around 50°C) with high MLSS concentrations (up to 200 gTS L−1). The proper description of the rheological behaviour of sludge represents a useful and fundamental aspect that allows characterizing the hydrodynamics of sludge suspension devoted to the optimization of the related processes. Therefore, in this study, the effects on the sludge rheology produced by the biomass concentration, pH, temperature, and aeration were analysed.
The monitoring of biological wastewater treatment processes, whatever its final goal is (assessment of compliance with effluent legal emission values, evaluation of specific stages efficiency, calculation of the maximum treatment capacity and comparison with the actual treated load, mass balance for conventional and emerging pollutants, estimation of energy consumption, etc.), is based on physical, chemical, and biological analyses. Guidelines and technical documents define the main parameters together with the proper analytical methods in case of conventional and/or nonspecific pollutants, such as the COD (chemical oxygen demand), the BOD (biological oxygen demand), nitrogen and phosphorus compounds, surfactants, oils and greases, and suspended solids. The biological process is based on the activity of different populations of organisms set at growing levels of the detritus food web. Therefore, any tools enabling us to identify and measure the activity of these organisms provide valuable information for improving and strengthening the operation of wastewater treatment plants ([
The removal capacity of specific pollutants can be assessed by performing the measurement of metabolic activity, for example, by means of ammonia uptake rate (AUR) and nitrate uptake rate (NUR) tests [
Scientific literature reports hundreds of applications of conventional and innovative biological tools, as well as examples of investigations based on ecological criteria [
Rheological properties are crucial for activated sludge applications in wastewater treatment plants, since they severely impact the flow behaviour and many aspects that interfere with process performance and energy consumption, for example, sludge pumping, bioreactor hydrodynamics, mass transfer efficiency of aeration systems, sludge-water separation via settling, and filtration [
Basically, rheology of a sludge is defined by its viscous characteristics, which can be determined by the relationship between shear rate and shear stress, obtained through a rheological measurement (which imposes either shear rate or shear stress [
At lower concentration of the suspended matter, sludge's behaviour can be reasonably approximated to a Newtonian fluid characterized by a linear relation between shear stress and rate of deformation, with the proportionality coefficient being the fluid viscosity. The measured viscosity is rather independent from shear rate, at given values of temperature and pressure. The flow curve is therefore a straight line through the origin of the coordinate axes.
As the particulate concentration increases, the sludge deviates from Newtonian behaviour. For a non-Newtonian liquid, the measured viscosity becomes dependent on the rate of shear, and the so-called
The rheological characterization of non-Newtonian fluids leads to practical difficulties; different measurement protocols and devices (i.e., capillary rheometers as in [
The properties of activated sludge, such as MLSS concentration, particle size distribution and shape, interaction among particles, flocculation ability, and surface physicochemical characteristics, all have effects on rheology and, thus, on the kind of model most efficient in describing the sludge rheological profile and on the values of the preferable model parameters [
A number of studies demonstrate that also temperature significantly affects the rheological characteristics: sludge becomes progressively more fluid as the temperature increases due to thermal [
This paper reports the results of an integrated study performed on the sludge deriving from a real scale TAMR, supplied with pure oxygen (being the average dissolved oxygen concentration of the mixed liquor equal to 2.3 mg L−1, while minimum and maximum values equal 0.2 and 15.0 mg L−1, resp.) located at a facility for the treatment of solid and high-strength liquid wastes. The sludge has been characterized in terms of microbiological and rheological features. Actually, the authors found, previously, links between the amount of filamentous bacteria and zooglea clusters and viscosity values of mixed liquors from conventional activated plants treating municipal wastewater, thus suggesting the use of rheological tools to control dysfunctions caused by the proliferation of specific microorganism [
TAMRs operate at 45°C and are profitably applied for the treatment of high-strength wastewater deriving from food processing, pulp, and paper and pharmaceutical factories [
This research was aimed, therefore, at exploring the rheological and microbiological features of TAMR biomass, which remain almost obscure, in spite of the increasing knowledge gained about its excellent performances.
A synthetic description of the studied waste treatment facility, together with sampling and analytical procedures, is presented.
The facility is located in Northern Italy and treats 60,000 t/y of solid and liquid high-strength COD wastes (except mutagens and carcinogens); since fifteen years ago, this technology has been investigated in detail by [
Figure
Process scheme of the studied facility.
Before entering the biological tank, metals are removed by means of a chemical-physical treatment which leads to their precipitation as ammonium salts, hydroxides, and phosphates (pH equal to 11 units). The TAMR surface is 267 m2 wide; net volume is about 1000 m3.
Pure oxygen is supplied through static mixers, yielding oversaturation of the sewage, which is recirculated within them. Ultrafiltration unit includes two parallel lines consisting in a feeding pump, a recirculation pump, and three channels of ceramic membranes (each containing 99 tubular membranes having 25 channels). Pores size allows retaining molecules larger than 0.3 microns and with a molar mass higher than 300 kDa. Minimum and maximum operation pressure values are 3 and 5 bar, respectively.
The average concentrations of pollutants inlet to the TAMR are 25,000 mg COD L−1, 1200 mg TN L−1, and 700 mg TP L−1. The TAMR performances, in terms of COD, TN, and TP removal yields, are 78%, 80%, and 90%, respectively. The high removal yields for TP, as reported by [
Samples taken from the TAMR were submitted to investigations aimed to define their microbial profiles and rheological behaviour.
Mixed liquor samples were taken from TAMR and freighted immediately to the laboratory at 4°C. After adding absolute ethanol (Sigma Aldrich), (1 : 1) mixed liquor was pipetted onto a glass slide and allowed to dry at 46°C for 15 minutes; afterwards, samples were submitted to
Besides, observation of floc features and identification and abundance assessment of filamentous bacteria were performed according to [
Mixed liquor samples were also submitted to the analysis of the microfauna for the calculation of the Sludge Biotic Index (SBI) according to [
The rheological experiments were carried out using rheometer RC20 (RheoTec) with a configuration CC25DIN of coaxial cylinders (Figure
Experimental apparatus: coaxial cylinder CC25DIN (a) and rheometer RC20 (b).
The working principle of the instrument is based on this aspect: the sliding of the sludge in the cavity between the coaxial cylinders, due to the spindle rotation with a fixed share rate, while the external cylinder is held, requires a torque that is measured by the instrument. Several rotational tests with different controlled shear rate (CSR) were performed to obtain the rheological equation:
The temperature was controlled by the use of water in a beaker placed on a heating magnetic stirrer (Figure
In Table
Rheological tests performed: operative conditions.
Test # | Operative conditions | Shear rate |
|||
---|---|---|---|---|---|
Average pH | Average temperature [°C] | Aeration | |||
Sample A |
1 | 6.7 | 44.5 | N | 700, 800, 900, 1000 |
2 | 6.7 | 45.2 | Y | 700, 800, 900, 1000 | |
3 | 7.7 | 45.1 | N | 700, 800, 900, 1000 | |
4 | 8.3 | 45.3 | Y | 750, 800, 900, 1000 | |
5 | 9.0 | 45.1 | N | 750, 800, 850, 900, 1000 | |
6 | 9.1 | 44.9 | Y | 800, 850, 900, 1000 | |
|
|||||
Sample B |
7 | 7.0 | 45.4 | N | 500, 600, 750, 900, 1000 |
8 | 6.7 | 45.3 |
|
500, 600, 750, 900, 1000 | |
9 | 6.6 | 45.0 | Y | 500, 600, 750, 900, 1000 | |
10 | 7.8 | 44.9 | N | 500, 600, 750, 900, 1000 | |
11 | 8.3 | 45.2 |
|
500, 600, 750, 900, 1000 | |
12 | 8.4 | 45.3 | Y | 500, 600, 750, 900, 1000 | |
13 | 8.9 | 45.0 | N | 500, 600, 750, 900, 1000 | |
14 | 9.0 | 45.3 |
|
500, 600, 750, 900, 1000 | |
15 | 9.1 | 45.0 | Y | 500, 600, 750, 900, 1000 |
N: no aeration;
The biomass samples withdrawn from the TAMR were delivered to the laboratory within 120 min after sampling. They were stored (for 180 min) at 45°C under different conditions: a first subsample was not submitted to aeration; a second subsample was kept in aeration conditions by an air compressor at lab scale (1
Each rheological test was performed with fixed shear rates that were maintained for 300 seconds; shear rate was increased step by step as reported in Table
Both the sludge samples presented in Table
It is worth noting that in the present study a limited number of shear rates has been investigated (i.e., four values for tests 1–4 and 6, and five in case of tests 7–15). This aspect, in principle, may represent an inherent limitation of the work that, however, has the merit to provide a first insight into the rheological behaviour of a highly concentrated thermophilic sludge. Furthermore, the preliminary findings (see Section
This chapter gathers the information obtained from the microbiological and the rheological analyses and sets them in the scientific literature context.
The sludge appeared extremely thick, due to the great suspended solids concentration maintained in the TAMR (150–200 g L−1). Figure
Micrograph of the raw sample (magnification: 100x).
Based on the characterization rules of [
These findings tally with the outcome reported in the scientific literature: a poor (or even lacking) flocculation is usually seen to occur under thermophilic conditions [
About the amount of filamentous bacteria, however, [
The paper [
Table
Results of the application of the VIT gene probes.
Analysed target microorganism(s) | Physiological features | Share of each group in relation to the overall bacteria population (%) | |
---|---|---|---|
Species | Group | ||
|
Group with high amount of heterotrophic organisms (e.g., genera |
1 | |
|
|||
|
Group consisting of filamentous bacteria similar to |
n.d. | |
|
|||
|
Group consisting of the |
77 | |
|
|||
|
Group consisting of floc forming bacteria and |
76 |
|
|
|||
|
Group consisting of heterotrophic, mixotrophic, and autotrophic microorganisms (e.g., genera |
15 | |
|
|||
|
Filamentous and nonfilamentous bacteria. Common in municipal and industrial WWTPs with a relatively high sludge age. Favoured by scarce dissolved oxygen concentration | n.d. | |
|
|||
|
Filamentous sulphur bacterium, favoured by nutrient deficiency. Adapted to low dissolved oxygen concentration | n.d. | |
|
|||
|
Filamentous sulphur bacterium common in municipal and industrial WWTPs. favoured by nutrient deficiency. Adapted to low dissolved oxygen concentration | n.d. | |
|
|||
|
Group including most sulphate reducing bacteria (e.g., families Desulfobacteraceae, Desulfobulbaceae, and Desulfovibrionaceae) | <1 | |
|
|||
|
Group of filamentous organisms with increasing impact on municipal and industrial WWTPs, due to their contribution to foaming and bulking events, very stable against mechanical stress (genus |
<1 | |
|
|||
|
Filamentous bacterium; contribution to sludge bulking. Favoured by nutrient deficiency | n.d. | |
|
|||
|
Filamentous bacterium. Aerobic and anaerobic conditions | n.d. | |
|
|||
Cytophaga-Flexibacter subphylum |
Group containing filamentous and floc forming bacteria. Mainly in WWTPs with nutrients removal (e.g., genera |
4 | |
|
|||
|
Filamentous bacterium; contribution to sludge bulking. Favoured by high sludge age, low dissolved oxygen concentration, and high ammonia concentration | n.d. | |
|
|||
|
Group, mainly in municipal WWTPs; members of this group are involved in the ANAMMOX process. Adapted to low dissolved oxygen concentration (e.g., genera |
n.d. | |
|
|||
Candidatus |
Filamentous bacterium; contribution to sludge bulking | n.d. | |
|
|||
|
Group of GRAM positive bacteria with a high DNA GC content. Several problematic filamentous bacteria like |
n.d. | |
|
|||
|
Filamentous bacterium; in WWTPs with nutrients removal. Contribution to sludge bulking, floating, and foaming. Strong hydrophobic characteristics | n.d. | |
|
|||
Nocardioforms | Group containing filamentous and floc forming bacteria. Contribution to sludge bulking, floating, and foaming. Strong hydrophobic characteristics | n.d. | |
|
|||
Candidatus |
Typical chain structures. Contribution to sludge bulking. In municipal WWTPs | n.d. | |
|
|||
|
Group of GRAM positive bacteria with a low DNA GC content. Several fermentative bacteria like the genera |
n.d. | |
|
|||
Candidatus |
Filamentous bacterium. Contribution to sludge bulking | n.d. | |
|
|||
|
Group recently detected, including filamentous bacteria like Eikelboom type 0041 | n.d. | |
|
|||
|
For example, genera |
n.d. | |
|
|||
|
For example, genera |
n.d. | |
|
|||
|
Responsible for the biological phosphorus removal (e.g., Candidatus |
n.d. | |
|
|||
|
Candidatus |
n.d. |
Summary of rheological tests and results of linear regression analysis.
Test # | Average |
Average pH | Aeration | MLSS |
|
|
|
|
|
---|---|---|---|---|---|---|---|---|---|
|
|
|
|
|
|
|
|
||
1 | 44.5 | 6.7 | N | 150 | 0.0 | 2.760 |
|
0.997 | 0.0296 |
2 | 45.2 | 6.7 | Y | 0.0 | 2.909 |
|
0.971 | 0.0944 | |
3 | 45.1 | 7.7 | N | 0.0 | 2.976 |
|
0.965 | 0.1066 | |
4 | 45.3 | 8.3 | Y | 0.0 | 3.120 |
|
0.967 | 0.0905 | |
5 | 45.1 | 9.0 | N | 0.0 | 2.138 |
|
0.923 | 0.0711 | |
6 | 44.9 | 9.1 | Y | 0.0 | 2.925 |
|
0.953 | 0.0706 | |
|
|||||||||
7 | 45.4 | 7.0 | N | 190 | 2.35 | 5.307 |
|
0.986 | 0.2107 |
8 | 45.3 | 6.7 |
|
1.95 | 3.901 |
|
0.992 | 0.1146 | |
9 | 45.0 | 6.6 | Y | 2.05 | 3.508 |
|
0.989 | 0.1199 | |
10 | 44.9 | 7.8 | N | 1.90 | 4.433 |
|
0.998 | 0.0646 | |
11 | 45.2 | 8.3 |
|
1.65 | 3.452 |
|
0.994 | 0.0875 | |
12 | 45.3 | 8.4 | Y | 1.70 | 3.590 |
|
0.992 | 0.1091 | |
13 | 45.0 | 8.9 | N | 1.40 | 4.585 |
|
0.998 | 0.0651 | |
14 | 45.3 | 9.0 |
|
1.60 | 5.468 |
|
0.998 | 0.0791 | |
15 | 45.0 | 9.1 | Y | 1.25 | 5.221 |
|
0.998 | 0.0852 |
Pie chart of the detected taxa shares.
Betaproteobacteria were the dominant group in the sludge sample. They were detected with a remarkable high share of 77% of the total viable bacteria. No filaments or ammonium oxidizing bacteria belonging to this class were observed, as found by [
The share of members of the Cytophaga-Flexibacter subphylum was 4% of the total viable flora, which corresponds more or less to typical values in activated sludge, which can reach a percentage of 10%. The thin rod-shaped cells showed low fluorescence signals. Filaments like
Alphaproteobacteria were detected with a share of only 1% and appeared completely as coccus-shaped uniform cells. Shares of up to 20% of this group are quite normal for industrial activated sludge.
Members of the
Deltaproteobacteria, including most sulphate-reducing bacteria, were represented as uniform single cells, with a share of <1%. Values of up to 8% for this group can be found in the mixed liquor of plants treating industrial wastewater. All other main bacteria groups including filamentous bacteria were not detected.
High values for total cell counts (dead and alive) were determined, with
Figures
Micrographs (magnification: 1000x) of the flocs under phase contrast and epifluorescence. Comparison of the same floc area. Hybridization target: viable bacteria and Betaproteobacteria. (a) Phase contrast. Identical microscopic fields under fluorescence: (b) detection of all viable bacteria and (c) analysis with a specific probe for Betaproteobacteria.
Micrographs (magnification: 1000x) of the flocs under phase contrast and epifluorescence. Comparison of the same floc area. Hybridization target: viable bacteria and
Micrographs (magnification: 1000x) of the flocs under phase contrast and epifluorescence. Comparison of the same floc area. Hybridization target: viable bacteria and Gammaproteobacteria. (a) Phase contrast. Identical microscopic fields under fluorescence: (b) detection of all viable bacteria and (c) analysis with a specific probe for Gammaproteobacteria.
Several analytical models have been proposed to mimic the rheological behaviour of biological sludge, each with different degree of complexity depending on the number of the parameters contained [
In the paper [
Proper fitting of the experimental data analysed in this work (MLSS concentrations in the range 150–190 g L−1) was obtained with
Several statistical descriptors can be adopted to measure the accuracy of prediction of rheological models and the reliability of their estimated parameters [
The method of ordinary least squares (OLS) was adopted for linear regression for fitting the dependence of shear stress versus shear rate.
As well known, only two of the three above-mentioned model parameters can be reliably estimated through linear regression. Therefore, in the following statistical analysis, it was assumed as fitting parameters the flow behaviour index
In some works, the yield stress
In each test, when carrying out the linear regression of the experimental data with respect to the unknown parameters
The obtained results of the statistical analysis are summarized in Table
The results obtained for both MLSS concentrations show that the flow behaviour index
For all the tests carried out, the apparent viscosity increased, at any shear rate, with the solid content; moreover, the differences in apparent viscosity measured at different shear rates increased with solid content, in accordance with [
Anyway, such a dilatant behaviour may be related to the fact that in this work the sludge samples were tested at very high temperature of about 45°C, which is unusual if one considers that generally the testing temperature does not exceed 35°C [
In order to point out possible temperature effects on the obtained results, the work by [
These authors have shown that (for a given solid concentration) yield stress depends exponentially on the inverse of the temperature. Therefore, an increase of the testing temperature seems consistent with a relevant reduction of the yield stress.
It is worth noting that in all tests with suspended solid concentration of MLSS = 150 g L−1 when imposing a shear rate below 100 s−1 the measured shear stress was under the minimum value that can be resolved by the rheometer (see Section
At this stage of investigation, the flow properties of all the sludge samples were tested at a shear rate greater than or equal to 500 s−1. This is due to technical reasons related to the modelling of particular operating conditions of concern in the treatment plant. From a theoretical point of view, the experimental points below the shear rate value of 500 s−1 may influence the extrapolation of the yield stress for the analysed samples; for this reason, further studies will be carried out to investigate this aspect.
In the work by [
Suitable values were obtained for the regression coefficient in all tests. The best fitting of the adopted mathematical model is obtained for the higher MLSS concentration of 190 g L−1, as confirmed by the values of
Concerning the lower MLSS concentration of 150 g L−1, the minimum value of
In the paper by [
Shear stress versus rate of deformation (MLSS 150 g L−1): comparison between experimental and calculated values.
Shear stress versus rate of deformation (MLSS 190 g L−1). Lower right-hand panel compares calculated values for pH condition around 6.6–7.0
Shear stress versus rate of deformation (MLSS 190 g L−1). Lower right-hand panel compares calculated values for pH condition around 7.8–8.4
Shear stress versus rate of deformation (MLSS 190 g L−1). Lower right-hand panel compares calculated values for pH condition around 8.9–9.1
Figures
Figure
Obtained results seem however reasonable, since aeration of the sludge may lead to bubble entrapment into the bioaggregate that increases the mean distance between suspended solid particles and consequently weaken the interactions between them; this lowers both apparent viscosity and shear stress at any shear rate. Such an effect is clearly evident at the highest investigated pH (lower panel).
Figures
In those tests where pH ranged between 6.6 and 7.0 (Figure
Increasing the pH around 7.8–8.4 (Figure
Considering the maximum value of the pH = 9.0 (Figure
On the contrary, the flow curve of the half aerated sample (test 14, air
Based on the observation of the interpolated flow curves, the influence of aeration on the rheological behaviour of the samples appears to be not univocal: the reduction of the apparent viscosity does not take place at any shear rate for every investigated pH value. Furthermore, in those cases, where the obtained shear stress differences are extremely slight, such a discrepancy may be also ascribed to possible uncertainties caused by the reduced number of experimental samples available for interpolating the flow curves. Hence, further investigations are required so as to assess the influence of aeration with a higher degree of reliability.
Figure
Shear stress versus rate of deformation: influence of pH for nonaerated samples.
In the case of MLSS = 150 g L−1, it can be seen that increasing the pH value from 6.7 (test 1) to about 7.7 (test 3) caused a negligible variation of the shear stress and apparent viscosity took place at any shear rate. Further increase of the pH value to 9.0 (test 5) induced, at shear rates lower than 730 s−1, a negligible increment with respect to previous curves of both shear stress and apparent viscosity, while above 730 s−1 the shear stress decreased progressively together with the shear rate growth.
Considering the case MLSS = 190 g L−1, the observed behaviour was monotonic because, increasing pH values from 7.0 (test 7) to 8.9 (test 13), the shear stress and apparent viscosity diminished at any shear rate below 1000 s−1.
Figure
Shear stress versus rate of deformation: influence of pH for aerated samples.
For the case MLSS = 150 g L−1 (left side panel), it can be seen that passing from pH = 6.7 (test 2) to pH = 8.3 (test 4) a negligible variation of both shear stress and apparent viscosity took place. At pH = 9.1 (test 6), the corresponding flow curve was always below the other two curves, resulting in a significant reduction of shear stress at higher shear rate values.
The analysis of the aerated samples with MLSS 190 g L−1 (right side panel) showed that, increasing the pH value with respect to test 9, the reduction of the shear stress was significant at shear rates below 800 s−1, while above this value the flow curves corresponding to pH = 8.4 (test 12) and pH = 9.1 (test 15) tended to grow faster and overcome the curve corresponding to pH = 6.6 (test 9).
Figure
Shear stress versus rate of deformation: influence of pH for 1/2 aerated samples.
The pH influence on rheological behaviour might be ascribed to the bacteria aggregation extent: [
In a subsequent study [
The samples studied in [
The population profile differed clearly from a typical profile of municipal as well as industrial wastewater treatment plants. Notable is the missing of several of the major bacteria groups and the low diversity within the present major bacteria groups. Beside the considerably dominating Betaproteobacteria, only a few other main bacteria groups were present showing only very limited diversity within the groups. The remaining organisms seemed to be adapted to the extreme conditions applied and the obtained population profile was very different from typical profiles from other WWTPs with “normal” temperature conditions. All in all, the extreme temperature conditions seem to trigger the surviving of the (fittest) “most adapted” organisms to the conditions applied. Also, the almost complete absence of filamentous bacteria is a typical feature of thermophilic aerobic sludge treatment and can lead to sludge losses due to the inability to form stable sludge flocs.
This early phase of the study was focused on a limited number of shear rate values: nevertheless, the first results provide a valuable insight into the rheological behaviour of a highly concentrated thermophilic sludge. Furthermore, the trend of the flow curves shows a certain degree of reliability, as corroborated by the regression coefficients.
The influence of some relevant operative parameters (i.e., pH, temperature, biomass concentration, and aeration within the biological reactor) on the rheological behaviour of the activated sludge was confirmed, thus providing valuable guidelines for an efficient management of the wastewater treatment process. The effects of aeration were sometimes not univocal and reveal an opposing behaviour depending on the values of the other parameters. This aspect may be related to possible uncertainties connected to the low number of samples available for the interpolation of the flow curves, thus resulting in small discrepancy between them. Therefore, it will be better investigated within a future study.
The rheological behaviour affects significantly the treatment processes of TAMR systems because, with respect to conventional processes involving activated sludge, the major part of the energy demand is connected to the use of membranes and is strongly affected by the suspended solid concentration and the biomass viscosity.
The analyses carried out show that the rheological behaviour of TAMR biomass is mainly affected by MLSS concentration.
The sludge exhibited shear-thickening (dilatant) behaviour; at MLSS concentration of 150 g L−1, the estimated yield
The authors declare that there are no conflicts of interest regarding the publication of this paper.