;Cat. Res. Vol. 27, No. 5, pp. 829-838, 1993 0043-1354/93 $6.00+0.00 Printed in Great Britain. All fights reserved Copyright ~ 1993 Pergamon Press Ltd
BIOFLOCCULATION IN ACTIVATED SLUDGE:
AN ANALYTIC APPROACH
V. URBAn~ ~, J. C. BLOCK 2 0 and J. MANEM'
'C.I.R.S.E.E., Lyonnaise des Eaux-Dumez, 38 rue du President Wilson, 78230 Le Pecq, France and 2Laboratoire Sant~ et Environement, Facult6 de Pharmacie, 5 rue Albert Lebrun, 54000 Nancy, France
(First received February 1991; accepted in revised form November 1992)
Abatract--A study on the physico-chemical structure of activated sludge flocs was carried out to get a better insight in its relationship with sludge settleability. For this purpose, 16 sludge samples from different origins were analyzed in order to provide information with regard to their settleabifity, biomass and exocellular composition, surface characteristics and internal hydrophobicity. The presence of filamentous microorganisms was observed in all samples but was not always associated with poor settleability, supporting to some extent the idea of their role as a backbone in the floes. Relationships between the measured variables were studied through their linear correlations. A high amount of exocellular polymers (ECP) was associated with poor settling conditions. The DNA fraction and the C/N ratio of the ECP, had also a negative influence on the adsorption of a cationic molecule in the sludge samples. Finally, sludge settleabifity was described with a mathematical model which shows the opposition between ECP and the internal hydrophobic~ty of the floes. From the model, the positive role of hydrophobic interactions should provide a new approach in the understanding of flocculation mechanisms in activated sludge.
Key words---activated sludge, settleability, SVI, exocellular polymers, hydrophobicity, model, surface charge
INTRODUCTION
Poor settling of activated sludge is an important ecological as well as technical problem which often leads to a discharge of suspended solids in the environment and operating problems i n the waste- water treatment plant (WWTP) itself. A survey on bulking in biological WWTPs (Pujoi and Carder, 1989) showed that at least 25% o f the plants were concerned with settling problems (SVI ~ 150 or 200 ml.g-I).
The negative influence of filamentous micro- organisms on sludge settling is well-known (Wanner and Grau, 1989). Their control is still very difficult to achieve because of the diversity of species and thereby of environmental parameters which influence their (over-)growth. This biological problem hides the fact that there are few informations on the physico- chemical structure of activated sludge floc and its relationships to the sludge settling capacity.
Aggregation in activated sludge is an active process as cells are originally dispersed. This process is stochastic and heterotypic as fluid motion and differ- ent microbial species are involved (Calleja et al., 1984). The size o f the floes ranges from 20 to 200/tin according to Mueller et al. (1967) but Parker et al.
(1971) showed a bimodal size distribution from 0.5 to 5/~m and 25 to 3000 #m. Most of the methods assume one or two principal axes and sometimes a sphere-like shape, and because o f the three dimen- sional structure o f the flocs realistic sizing is still
difficult. The internal organization of the floc is also difficult to modelize and, for example, Li and Gan .caxczyk (1990) did not observe any uniform structure in stained sections of activated sludge flocs.
The overall floc structure is negatively charged and is the result of physico-chemical interactions between microorganisms (mainly bacteria), inorganic particles (sillicates, calcium phosphate and iron oxides), exocellular polymers and multivalent cations (Fig. 1).
Exocellular polymers (ECP) have two different origins, (1) from metabolism or lysis of microorgan- isms (proteins, DNA, polysaccharides and lipids) and (2) from the wastewater itself (e.g. cellulose, humic acids...). Their influence on sludge settling has been widely studied, but since there is no unified method for their extraction, it is always difficult to compare results from different studies. Thus, relationships between sludge settling and ECP (Magara et ai., 1976;
Chao and Keinath, 1979) and data about the ECP composition are sometimes contradictory (Sato and Ose, 1980; Horan and Eccles, 1986).
Since bacterial surfaces, ECP and eventually inor- ganic particles provide negative adsorption sites, the role of divalent cations in the floc stability must be emphasized. Divalent cations such as Ca 2 + and Mg 2 + are known to be involved in the chemical structure of bacterial aggregates (Belcourt et al., 1974; Eriksson and Aim, 1991) and biofflms (Turakhia eta/., 1983), because o f their ability to bind to negatively charged chemical groups. Steiner et al.
(1976) showed that different types of adsorption 829
830 V. Uz~L~q et al.
Bacten'a ; PO4n"
OH
N +
I h y d r ~ e m
Exocellular polymers
Inorganic particles
I B
Oivalem cations
CZ+
m
Fig. 1. Schematic representation of the activated sludge floc on an arbitrary scale of size.
q-
isotherms can be obtained with cations such as cobalt, calcium or copper and with a higher attinity for ECP alone than for the whole sludge. A difference of affinity between C a 2+ and Mg :+ ions for the ECP has also been demonstrated by Forster and Lewin (1972) and Forster and Dallas-Newton (1980). In spite of the importance of divalent cations in floccu- iation processes, there is a lack of information about either their accumulation in the exocellular struct~es of the sludge flocs or their affinity with specific constituents of the ECP.
Biological sludges are highly hydrated structures and little attention has been paid to the role of liydrophobicity in flocculation. As pointed out by Eriksson and Axberg (1981), the hydrophobic- hydrophilic balance should be considered an import- ant factor. Cell surfaces are known to exhibit hydro- phobic areas (Magnusson, 1980) and hydrophobic molecules such as lipids or proteins from the cells can be trapped into the flocs. Hydrophobicity has been poorly studied mainly because microorganisms in the sludge are aggregated and most of the methods for testing hydrophobicity of bacteria have been devel- oped for testing dispersed cultures. The contact angle technique has been used to relate the flocculation performance of activated sludge to its hydropbobic characteristic (Valin and Sutherland, 1982). Because hydrophobic surfaces have a natural tendency to avoid hydrated environments, the hydrophobicity o f activated sludge flocs must be determined from disaggregated samples.
This short literature review shows that the under- standing of the floc structure and its relationships with activated sludge settling is dependent on methods that are able to provide information on its chemical composition. Moreover there is still no single technique allowing a good description of the
complex structure of the flocs. Thus in this paper, 14 analytic methods are utilized to describe activated sludge from different origins. The results are dis- cussed such as to describe the physico-chemical struc- ture of the flocs and its relation with settleability.
As environmental conditions are known to influence sludge settleability, characteristics of the WWTPs and wastewater composition are also related to the measured out variables.
MA'rKRL, U.S AND NEgTHODS Aczivazed sludge aamplMg
Sixteen activated sludge samples were taken from seven different WWTPs (A'G). Four of these plants (A, B, F and (3) were sampled once and the three others (C, D and E) four times on different days. The sludge samples from plants C, D and E are really different because large variatioM were observed for the different measured out variables. These large variations show that samples from the same plant but from di~erent times are really different.
Charactemties of each plant (mass loading rate Bx) and wutewater ¢ompmition (BOD/COD, COD/N and COD/P) were provided by the laboratory of Metz treatment plant (E) and from control quality measurements (A, B, C, D, F and G) by the LOREAT (technical a ~ ' _ ~ com- pany, Metz) (Table 1).
Sludges from plants C, D and E were analysed immedi- ately after umpling but those from plants A, B, F and G, sampled by the LOREAT Company the day before the analysis, were stored at +4°C overnight. The pmslble effect of low temperature on the variables was not investigated.
Micrmcopic examination
The relative density of filamentous becteria was estimated by means of a micrmeopic obeervation (x 320), aoeording to a qualitiative scale: + / - presence, + + high demity, + + + + proliferation.
Sezth~,~ zezz
The sludge volume after 30 miss of Settlinl (SV~e) was measured on sludge volumes of 100 ml (King and Forfcer, 1990). The sludge volume index (SVI, nd.g -t) represents
Bioflocculation in activated sludge
Table I. Characteristics of the wastewater treatment plants (WWTP) (Bx: mass loading rate) and composition of the influent wastewaters
WWTP Bx(kg-IBOD.kg-IVM.j -') Loading range BODICOD COD/N COD/P
A 0.61 High 0.49 9.3 25.4
B 0.03 Extended aeration 0.31 6.4 17.6
CI 0.03 Extended aeration 0.40 8.6 32.8
C2 0.03 Extended aeration 0.40 8.6 32.8
C3 0.03 Extended aeration 0.40 8.6 32.8
CA 0.03 Extended aeration 0.40 8.6 32.8
DI 0.14 Low 0.41 13.0 68.1
D2 0.40 Intermediate 0.41 13.0 68, |
D3 0.43 High 0.41 13.0 68.1
D4 0.26 Intermediate 0.41 13.0 68.1
El 0.42 High 0.38 5.2 16.5
E2 0.37 Intermediate 0.38 5.8 21.0
E3 0.42 High 0.38 6.5 25.0
FA 0.42 High 0.38 6.2 21.8
F 0.29 Intermediate 0.51 9.4 32.2
G 0.77 High 0.47 8.8 23.4
831
the ratio between SV30 (in ml. 1- t) and the dry matter content of the sludge (DM, g.l-~). As the standard procedure is based on a 1 1 volume of activated sludge, SVI values from 100ml (SVlea) or 1 1 (SVI 0 were compared. From five acti- vated sludge samples (Metz WWTP) (37% < SV30 < 72%
and 2.99g.i -t < D M < 4 . 0 9 g . I -m) deviations of - 1 6 to + 13% were found when comparing SVI0. ~ to SVI~, which is close to the uncertainty on the test itself.
Biomass analysis
The biomass composition was analysed according to four different analytic methods: the dry (DM) and volatile matter (VM) content of the sludge samples were determined after centrifugation (50 mi, 10 rain at 2500 g). The pellets placed in crucibles were successively dried (24 h at 105°C), calcined (3 h at 550°C) and weighed at each step. The volatile matter content (VM) was calculated by difference between the dry and mineral water values °and the results were expressed in grams per liter of sludge.
In order to provide a rapid estimate of the DM concen- tration of the sludge, a relation was established between optical density at 600 nm ( O D ~ ) of a sonicated sludge sample and its DM content (DM =5.40 × O D ~ - 0 . 1 2 ,
• = 0.97 and n -- 42). This relation has been calculated from triplicates on three different sludge samples, providing a range of DM concentrations between 0.72 and 4.628.1 -I.
The sludge samples (Metz WWTP) were concentrated then diluted with 0.22/~m-filtered secondary effluent. Each concentration (10 ml aliquotes) in a 20 ml plastic test tube, was sonicated (Bioblock sonicator, 50W 20 kHz) in an ice hath during two 15s periods interrupted by a 10s resting time. Then I ml was mixed with 9 ml of MilllQ water and the ODee0 (Kontron Uvikon 810) was measured. The OD~e-DM relationship gave results from - 9 . 4 to 28.6% of the DM content determined in this study with the drying and weighing procedure.
The C, H and N analysis have been performed on 105°C dried and ground sludge samples. Determinations were made in duplicate using a Carlo Erba C.H.N. analyser (Elemental analyser MOd.II06). The C, H and N values were expressed as rag. !-~ on the basis of the DM content of the sludge.
The total cell count in the sludge (Xt) was determined after souication, with an epifluorescence counting technique adapted from Malone and Caldwell (1983). In short, 10 ml of diluted sludge (1/5 in 0.2 #m-filtered MilliQ water) were sonicated as described above. The suspension was immedi- ately diluted from 1/500 to 1/2000, depending on the initial estimated dry matter concentration. In sterile test tubes, I ml of the diluted suspension was mixed with 1 mi of an acridine orange solution (0.01% final concentration) (Sigma A 6014). After a 30rain contact time, bacteria were recov- ered by vacuum filtration through 0,2 # m black polycarbon-
Wit 27/~,---G
ate membranes (Millipore GTBP 04700) which were covered with a small volume of filtered MilliQ water. The mem- branes were rinsed with 100 ml of MilliQ water and placed on a microscope slide with buffered glycerine for immuno- fluorescence (Diagnostic Pasteur 74.921) on top, and then covered with a glass slip. Fluorescent bacteria were counted at immersion ( × 1000) with a microscope under an u.v.-light source. Ten microscopic fields chosen at random were observed through a square eye-piece and the results were expressed in number of cells per liter of sludge,
Extraction and analysis of hydropholic ECP
The ECP were recovered in the aqueous phase of a sonicated sludge sample. The sludge was centrifuged (2500 g 10rain) to a final concentration about 8.5g.I -I (using the DM = f ( O D ~ ) relationship) in a 30 ml volume of MilliQ water. Then 20 ml of this concentrated sludge were soni- cared (as previously described for the cell counting pro- cedure) and mixed with 20 ml of MilliQ water. After a contact time of less than 60 rain, since preliminary assays have shown that cell lysis can occur, the hydrophilic ECP were recovered in the aqueous phase by centrifugation (20rain, 14,000g at +4°C) and frozen at - 2 6 ° C until chemical analysis is performed.
This aqueous extract was analysed for its polysaccharide, protein, DNA, Ca 2+ , Mg 2+ composition and its C/N ratio.
Polysaccharides in the ECP were measured on diluted thawed samples (1/10 in MilliQ water) according to the phenol-sulfuric method of Dubois et al. (1956) with glucose as standard. Proteins were measured immediately after extraction according to the method of Bradford (1976) with BSA as standard. DNA in the extract was determined according to the method of Deriaz et al. (1949) with calf thymus D N A as standard. All these determinations have been done in triplicate and the results expressed in rag (equivalents of standard) per liter of sludge.
The elemental C and N composition (in %) of the ECP was determined after lyophiliTation of the aqueous extract and only the C/N ratio was calculated.
The C,a 2+ and Mg 2+ contents were measured in the aqueous extract containing the ECP by atomic absorption spectrophotometry (Perkin-Ehner 2380) with the standard addition method to overcome the interference of the organic matrix in the samples.
Sludge surface characteristics
The negative sludge surface charges can be estimated by the adsorption of cationic molecules such as ruthenium red (RR+). The pnx:edure was adapted from Figueroa and Silverstein (1987). Adsorption of RR + ( M W = 576.5g.mol -m) was measured at equilibrium with a con- stant sludge concentration and increasing dye concen- trations. The adsorptio n data can be approximated by the
Frenndfich model which relates the adsorbed concentration of the molecule (Z) divided by the amount of dry matter (m), to the free concentration (Ce) in the form of X/m •K.Ce tl". The adsorption constant l/n shows the non-linearity of the model and K is related to a maximum of adsorption when l/n is close to zero. Preliminary exper- iments showed that adsorption of R R + can be studied with dry matter concentrations near 0.25 g.l -~ and R.R + final concentrations between 0.025 and 0.15 g. l- i.
The dry matter contents of the sludge sample were estimated (DM = f ( O D ~ ) relationship) and the sludge diluted to 0.25g.1 -I with 0.22/~m-filtered secondary waste- water (250 ml final volume). In 1 50 mi glass flasks, 18 mi of this dilution were mixed with 2 ml of R R + solutions from 0.25 to I g.l -~. In other flasks the diluted sludge was replaced by filtered wastewater in order to check for a possible adsorption onto dissolved material in the liquid.
All the flasks were placed on a rotary table at 250 rpm at room temperature (~25°C). After three hours of contact (Figueroa and Silverstein, 1987) the flasks containing the sludge were centrifuged for 10 rain at 2500g and the OD at 530 nm ()~m RR+) was measured out in the supernatants.
The adsorbed (mmol.kg -t DM) and free (mmol.l - l ) concentrations of RR + were calculated from a calibration curve. The linear regression analysis on the logarithm of the adsorbed concentrations vs free concentrations of RR + gives the value of l/n (slope) and In K (y axis intercept).
Determination of an internal hydrophobicity in the sludge The salt aggregation test developed by Lindahi et al.
(1981) for testing hydrophobicity of pure bacterial strains was adapted to activated sludge flocs.
Flocculation in a sonicated sludge sample mixed with increamg concentrations of ammonium sulfate in phos- phate buffer was measured. The ammonium sulfate influ- ence was determined from OD~eo measurements after a 3 h contact time with reference to a control without ammonium sulfate. Preliminary experiments showed that the difference between the ODeee of the control and the test is a linear function of the ammonium sulfate concentrations: the in- ternal hydrophobieity of the flocs (IHB) was expressed by the slope of the linear regression analysis.
The sludge dry matter concentration was set up to 2.5 g.l -j and then sonicated as previously described for the ECP extraction procedure. In short, I mi samples of soni- cated sludge were poured in 20 mi plastic test tubes which contained 9 ml of ammonium sulfate in phosphate buffer (0.07 M, pH 7 + 0.4) at 0.3, 0.75, 1.5, 2.25 and 2.5 M final concentrations. After 3 h at room temperature, the tubes were stirred by hand and OD~ e were measured.
A high value of IHB is equivalent to a large aggregation in response to a small increase in the ammonium sulfate concentration. It can be explained either by a low density of
negative charges 0tydrophilic groups) or by a high density of hydrophobic sites inside the flocs.
Reproducibility of the methods
For practical reasons, analysis on activated sludge samples were not replicated. Preliminary experiments showed that the reproducibility of the analytic methods described above is quite acceptable. As an example, stan- dard deviations of thirteen parameters measured in repficate on sludge samples from Metz WWTP are shown in Table 2.
All coefficients of variation with the exception of IHB and Xt (14%, duplicate tests) are lower than 10%. This 14%
value may arise from the coarse state of dispersed sludge samples from which turbidity (ODeeo) or total cell numbers are measured.
Statistical analysis of data
A statistic software (Statview II run on Macintosh SI) was used to analyze the data (16 observations, 16 variables) (Table 3). The linear correlation matrix was computed and significant linear relationships between variables were con- sidered when coe~clents of correlation were higher t h a n 0.50 (~ = 5%, n ffi 16) (Table 4). T h e linear relationships were checked graphically in order to prevent from situations where dispersion around the regression line is high.
Data were also computed by means of multiple regression analysis (stepwise method with 0c and ~ equal to 5%) (Box et aL, 1978). In order to avoid autocorrelations, care has been taken to use absolute units, i.e. rna~ units per liter of sludge, except for data from the settling, adsorption and flocculation tests.
RESULTS AND DISCUSSION
The chemical structure o f activated sludge floes has been studied o n sixteen sludge samples f r o m seven different wastewater treatm¢nt plants. In this study, 22 variables were analyzed. M a s s loading rates (Bx) and wastewater c o m p o s i t i o n ( B O D / C O D , N / C O D , P / C O D ) (see T a b l e 1) were obtained f r o m control quafity measurement. T h e SVIo.~ values, floc c o m p o - sition (biomass analysis, exocellular hydrophilic constituents) a n d surface characteristics (cationic a d s o r p t i o n a n d internal hydrophobicity) (see Table 3) were determined with the m e t h o d s described herein.
General characteristics o f the plants and the sludge samples
A c c o r d i n g to mass loading rates (Table 2), plants B and C I - 4 are operated at extended aeration Table 2. Mean values, standard deviations (SD) and ~ t s of variation (c.v,) for some of
the variables from the analysis of dilfemnt activated sludges (Metz WWTP)
Variables Mean SD c.v. % Replicatm
DM (g-I -i 1.89 0.08 4 3
VM (g'l-') 1.56 0.03 1.9 3
Xt.10 I: (I-') 4.11 0.58 14 2
Carbon (g.l-') 0.86 6.24.10 -3 < I 3
Hydrogen (g-I-') 0.II 8.13'i0 -3 <1 3
Nitrogen (g.l") 0 . 1 8 4.25.10 -3 < I 3
Exocellular polysaccharides (ms,l -i) 35.30 1.79 5.1 3
Exocellular proteim (rag-1 -I) 212.13 24.94 11.7 3
Exoeeilular DNA (mg.I -I) 16.70 0.31 i.9 3
SVI~ I (ml.g -I) 372 17 4.6 3
IHB. 102 2.23 0.28 14 2
l/n 0.24 0.01 4.2 3
K 837 1 0.12 3
DM and VM: dry and volatile matter content of the slmllge; Xt: total number of cells; SVt,:
sludge volume index; IHB: interred hydrophobicity of biological floc~ l/n and £: admrption contants from the Freundich model.
Bioflocculation in activated sludge 833
. . . ~ . . . ~ g ~
i " ' " ~ ~ ~ 4 ~
~ N ~ - ~ ~ ~
IC ?1
~ ~ ~ " ~ • . ~ : . . .
~ ' ~ " ~ •
~ " ~ ~ ~
" 0 ~ 0 ~ 0 ~ 0 0 ~ 0 ~
l=l
~ o ~ o ~ o - - . ~ . . . - - . =
~ ~ . . . . ~ . . .
conditions, plants D I - 4 and E l - 4 between intermedi- ate and high loading, plant F at intermediate and plant G at high loading conditions. The overall composition of the incoming wastewatcr shows a large but typical range of BOD/COD ratio from 0.31 to 0.51. The major types of nutrients (C, N and P) are assumed to be balanced as the COD/N and COD/P ratios are about 8 and 27, respectively.
According to our qualitative scale, microscopic examination of the sludges always revealed the pres- ence of filamentous structures even with low SVI0.~
values (sludges F, G, DI, D2, D3 and D4) (Table 4) and ten sludges out of sixteen showed SVIo. ~ values higher than 150 ml-g-m which is considered to be a limit before bulking (Forster, 1985a). The presence of filaments in each sample supports to some extent the role of filamentous structures as a backbone for the structure of the floes (Sezgin et al., 1978).
Biomass analysis
On the basis of the lowest and highest values obtained from biomass analysis, the 16 sludge samples contained 1.8-6.6gDM.I -~ with organic fractions from 55 to 85% and bacterial cell numbers from 4 to 12.1012.1 -~.
The cell dry weight calculated on the basis of the volatile matter concentration and the bacterial cell number was found to be fairly constant with an average value of 0.40 (_+0.06).10-ng (n = 16).
This value is quite comparable to those found by Herbert (1971) with pure cultures of Escherichia coli (0.12-0.41.10.-12g) between resting phase and log phase. Herbert's values were calculated on a plate count determination basis which gives lower results than the epifiuorescvnce counting method and thus overestimate the cell dry weight. As our results and those from Herbert are comparable, microorganisms in activated sludge could be considered an association between cells and organic ¢xocellular compounds ("cells + ECP").
The bacterial cell number is strongly correlated to the amount of volatile matter (VM) (r = 0.92) (Fig. 2) and C, H and N amount (r > 0.87) of the sludge.
These five variables can be gathered in a homo- geneous group which describes the organic fraction of the biomass. These strong linear correlations suggest that it is not necessary to determine all these vari- ables. However, correlations between the C/N ratio or the amount of elemental nitrogen in the sludge and constituents of the ECP (proteins, Ca 2 + ), settlcability (SVI0.~) or sludge surface characteristics (adsorption constant, K) (see Table 4) will provide information that cannot be obtained from a single method such as the dry matter determination.
ECP and divalent cation~
Extraction procedure. The amount of ECT rep- resents a small fraction of the activated sludge mass since even with a drastic extraction procedure such as heat treatment (Beccari et al,, 1980; Clarke and
834 V. URBAIN et ol.
Table 4. Linear coeff~ents of correlation statistcally significant at a 0.95 probability level (~ = 5%), i.e. r > 0.5 (abbreviations can be found in "Materials and Methods")
DM VM C H N C/N Xt EPS Proteins DNA C/N Ca 2+ Mg 2+ SVI K IHB COD/N
DM 1.00 . . . . . . . . . . .
VM 0.95 1.00 . . . . . . . . . . .
C 0.95 0 . 9 9 1 . 0 0 . . . . . . . . . .
H 0,97 0 . 9 7 0.97 i.00 . . . . . . . . . . "
N 0.86 0 . 9 0 0.91 0.94 1.00 . . . . . . . . . .
C / N - - - - - - I .oo . . . . . . . . .
Xt 0.87 0 . 9 2 0.90 0.91 0.88 - - 1.00 . . . . . . . .
EPS - - 0.70 - - - - 1.00 . . . . .
Proteins - - - - 0 . 6 3 - - 0.85 1 . 0 0 . . . .
DNA - - 0.56 0.57 0.62 0.81 - - 0.67 0.94 0.82 1.00 . . . . .
C / N . . . . . . 0.63 - - 1.00 . . . .
Ca 2+ - - - - --0.54 - - 0.88 0.93 0.85 -0.62 1.00 - - - - - -
Mg 2+ - - 0.55 0.57 0.58 0.78 - - 0.60 0.86 0.73 0.92 -0.59 0 . 8 3 1.00 - -
SVI - - 0.56 - - - - 0.85 0.64 0.74 - - 0.61 0.63 1.00 - -
K - - -0.50 0.60 . . . . . 0.55 0.52 - - -0.69 - - 1.00 - - - -
IHB . . . . . . . . . . . . 1.00 --
COD/N . . . . . 0.59 . . . . . . 0.67 1 . 0 0
COD/P . . . . . . . . . . . 0.95
Forster, 1982) ECP still accounts for less than 14%
of the sludge dry weight. The exocellular organic matrix in biological floes is really heterogeneous and requires a disaggregation step for recovery. ECP comes from cell structures, either because of ceil lysis, metabolic excretion, or leakage of exocellular con- stituents and also from the adsorption of organic matter (cellulose, humic acids ...). ECP and poly- mers attached to the outer membrane (e.g. lipo- polysaccharides in gram negative species) or to the peptidoglycan (e.g. teichoic acids in gram positive species) are too close to ensure that only non- covalently bound exocellular structures (ECP) are removed by a defined extraction procedure. There- fore, the basic requirement of an extraction pro- cedure must be the lowest extent of cell lysis, i.e. of intracellular polymers contamination.
In spite of several comparative studies (Can" and Gancarczyk, 1974; Kiffand Thompson, 1979; Brown and Lester, 1980; Rudd et al., 1982) there is still no unified method for the ECP extraction. Shearing effect of ultracentrifugation is not an effective method for activated sludge (Novak and Haugan, 1981), whereas chemical or heating procedures are too hard
on cell structures. Therefore a physical method such as sonication was recommended by Kiff and Thompson 0979).
In our experimental conditions, preliminary assays with ultrasonic treatment have shown that a dead level of the total number of cells in sonicated samples was reached. According to our method ECP consists of readily extractible macromolecules which are mostly hydrophilic as they are recovered in an
a q u e o u s e x t r a c t .
T h e r e s u l t s o f t h e a n a l y s i s o f e x o c e l l u l a r p o l y - s a c c h a r i d e s , D N A a n d p r o t e i n s in t h e 16 a c t i v a t e d s l u d g e s a m p l e s a r e r e p o r t e d in T a b l e 6 ( l o w e s t a n d h i g h e s t v a l u e s ) a n d compared w i t h r e s u l t s f r o m o t h e r studies. T h e d a t a f r o m K i f f a n d T h o m p s o n (1979) a r e in t h e r a n g e o f o u r v a l u e s f o r D N A , b u t t h e r e s u l t s o f B r o w n a n d L e s t e r (1980) c a n b e c o m p a r e d t o o u r r e s u l t s o n l y f o r p o l y s a c c h a r i d e s a n d p r o t e i n s w i t h t h e s t e a m i n g p r o c e d u r e . T h e r a t i o b e t w e e n s t e a m i n g a n d s o n i c a t i o n p r o c e d u r e s in B r o w n a n d L e s t e r ' s s t u d y r a n g e s b e t w e e n 18 a n d 25 w h i c h s u g g e s t s t h a t t h e i r u l t r a s o n i c p r o c e d u r e ( u l t r a s o n i c b a t h : 2 0 W , 2 rain) is t o o w e a k f o r a g o o d e x t r a c t i o n . B r o w n a n d L e s t e r (1980) u s e d t h e p r o t e i n / c a r b o h y d r a t e r a t i o f o r t h e i r c h o i c e o f s t e a m i n g a s t h e b e s t m e t h o d f o r t h e e x t r a c - t i o n o f E C P f r o m a c t i v a t e d s l u d g e b u t f o r R u d d e t al.
6 0
5
~
4 o O A Be2 c3 c 4 D!
D2
1 3 6 9 12 D3 134
XI.1012.1-1 El
E2 Fig. 2. Linear relation between the m o u n t o f volatile matter E3 (VM) and the total number o f cells (Xt) in the 16 sludge FA
samples. F
Table 5. Qualitative scale for the presem~ of filamentous micro- organisms in the sludge samples ( + / - - , p r e u e n c e , + + ,-high
density and + + + + = proliferation) Sludge SVI0. I (ml.g -~) Density of filaments
315 + / -
182 + + - F +
201 + + + +
223 + + + +
201 + + + +
204 + + + +
139 + / -
76 + / -
69 + +
107 + / -
188 + +
279 + +
264 + +
288 + +
84 + / -
Bioflocculation in activated sludge
Table 6. Comparison of different results on the analysis of cxocellular polymers (ECP)
835
ECP at mg. g- t (on dry matter basis)
Reference Extraction procedure P o l ~ DNA Proteins
Our study Sonication 6.50-24.84 9.49-29.66 6.16--163.44
Brown and Letter (1980) Sonication 0.80 + 0.14 0.20 + 0.07 3.20 + 0.68
Steaming 20.30+3.60 3.70+0.31 77.10+12.90
Kiff and Thompson (1979) Sonication ND 17-28 ND
Brown and Lester: 9 replicates; Kiff and Thompson: 5 replicates; our study: minimum and maximum values; ND:
not determined.
(1982) this ratio remains fairly constant (near 3/1) irrespective of the severity of the extraction pro- cedure. Since in our study this ratio varies from 0.77 to 6.25 it seems dependent rather on the composition of the exoceilular matrix than on the degree of lysis caused by the method.
Composition of the ECP aqueous extract. Only polysaccharides, D N A and proteins were analyzed in the ECP extract although other macromolecules such as phospholipids, giycolipids.., may be found in the exo~llular organic matrix of the floes. It is difficult to classify these polymers for their relative abundance in the ECP because their molar concentration is not known. In addition, they are measured as stan- dard equivalents (glucose, BSA . . . ) which arc surely not representative of their true composition. This is probably the reason for contradictions in the literature. For example, Sato and Ose (1980) found more nucleic acids (DNR + RNA) than proteins and polysaccharides, Horan and Eccles (1986) more poly- saccharides than nucleic acids and proteins and in our study we found more proteins than D N A and polysaccharides.
The three chemical types of polymers analyzed in the 16 sludges are strongly correlated with each other (r > 0.82) which means that they form an organic matrix in a more or less constant ratio: the strongest positive linear correlation is obtained between D N A and polysaceharides (r ffi 0.94) because the standard deviation on their mean ratio is low ( D N A / E P S = I . 1 9 + 0 . 1 8 , c . v . = 1 5 % ) . The ratio between proteins and D N A or polysaccharides shows a greater variability, 2.31 + 1.45 (c.v. = 63%) and 2.61 + 1.60 (c.v. = 62%) with linear coefficients of correlation equal to 0.82 and 0.85 respectively (Fig. 3).
Inside the organic exoceUular matrix of the floes, divalent cations (Ca 2 + and M g 2 +) m a y act as bridg- ing agents as they are correlated with the analyzed polymers (r >I 0.73). This bridging role is supported by the work of Eriksson and A i m 0 9 9 0 on the effect of E D T A on sludge characteristics (change of the dewatering property of activated sludge and release of ECP). Ca 2 + ions are measured in higher amounts 0.62-72 mg.l- l of sludge) than M g 2+ ions (1.26-5.63 mg.l-t of sludge) and this difference m a y be related to their affinity for ECP. Forster and Lewin (1972) have shown a higher binding of C a 2+ rather than M S # + ions on E C P recovered from activated sludge and Forster (1985b) hypothesized that the
ionic size of cations may influence their binding ability to charged (carboxyl) and uncharged groups (hydroxyl) in the ECP: thus as calcium ions are "larger" than magnesium ions (Handbook of Chemistry and Physics, 1990), their binding may be favoured. The linear correlations between exocellular Mg 2+ and D N A (r--0.92), bacterial cell number (Xt, r -- 0.60) or volatile matter concentration (VM, r --0.55) or between exocellular Ca 2+ and proteins (r = 0.94) or the C/N ratio of the sludge (r ffi - 0 . 5 4 ) may indicate a higher "affinity" o f Mg 2+ ions for D N A and Ca 2 + ions for proteins than for the other polymers. Additional experiments are needed to de- termine the mechanisms by which these species are involved in flocculation.
There is no correlation between exocellular polysaccharides or proteins and organic biomass variables such as VM and Xt. This means that the floc structure does not stem from the association of equal amounts of elementary subunits (e.g. [bacteria, ECP, inorganic particles ...] x n) as it is suggested from our schematic representation of the floc structure (see Fig. 1).
Origin of the ECP constituents. The origin of the exocellular D N A from lysis of dead cells is supported by the correlations with organic biomass variables such as Xt (r = 0.67), VM (r = 0.56) and the C, H, N composition of the sludges (r >I 0.57).
600
"-[
O polysacx~harides
A DNA
O &
O A
° , ° ,
0 30 60 90 120
poly~ccharldes o¢ DNA (mg.r 1) Fig. 3. Relations between the amount of exo~llular pro-
teins, polysaccharides or DNA in the 16 sludge rumples.
836 V. Ut~m~ et al.
The linear correlation between proteins and the C/N ratio of the ECP (r = - 0 . 6 3 ) is not surprising as proteins represent one of the nitrogenous constituent of the exocelhilar organic matrix in activated sludge.
The origin of proteins and polysaccharides from cell metabolism is supported by the negative relationship between exopolysaccharides (EPS) and the C O D / N ratio. Inspite of the r value of - 0 . 5 9 , there is a high dispersion around the regression line and the linear relationship must be considered only as a trend. This relation can not be explained by an inducing effect of a nitrogen limitation on EPS synthesis (Sutherland, 1977) as the COD/N ratio is not growth limiting.
Influence of ECP and divalent cations on sludge surface and settling
The linear coefficients of correlation between K and the N composition of the sludges (r = 0.60) or the C/N ratio of the ECP (r = 0.52) show that nitrogenous constituents which are accessible at the surface of the flocs do not provide adsorption sites for a cationic dye (ruthenium red). Exocellular DNA too (r = -0.55), seems to hinder adsorption of ruthenium red as it is negatively correlated with K.
Divalent cations may be able to reduce the number of negative charges at the surface of the flocs as there is a negative linear correlation between the amount of Mg 2+ ions measured out in the ECP extract and K (r = -0.70), but this hypothesis may proceed from the relation of this divalent cation with DNA.
All the analyzed constituents o f the ECP are positively correlated with SVI0a (r > 0.61), high con- centrations of polysaccharides, proteins . . . resulting in a worsening of sludge settleability. This negative influence has already been described in the literature in the case of EPS (Table 7). Results from different studies are not easily comparable, but irrespective of the experimental conditions, there is in most cases a positive linear relationship between SVI and ECP.
The exception comes from Goodwin and Forster's study who found a negative linear relationship be- tween SVI0a and ECP. The authors assumed that when the sludge settled poorly, EPS became less amenable to extraction. A more likely explanation refers to the units for ECP in terms of their total sugars content on the basis of total organic carbon from ECP. The lack of linear correlation between SVI and ECP in the study of Chao and Keinath (1979) is quite surprising in reference to the five other relation- ships, but the extraction procedure and the range of SVI values are really different from the other studies.
The exocelhilar organic matrix in the floes plays an important role in their surface characteristics. Both the nitrogenous content of the sludge, exocellular D N A and Mg 2 + ions are able to reduce the negative charge of the sludge surface, as they are negatively correlated with the adsorption capacity o f the sludge for a cationic dye (expressed as K). Owing to the fact that low values of these variables are related to low SVI0a values, a high surface charge (high values of K) may not be inconsistent with good settling con- ditions. This statement agrees with the results of Pavoni et al. (1972) who showed that the surface charge reduction is not the prime mechanism in bioflocculation because polymers are able to bridge the cells either electrostatically or physically. Like- wise, Ries and Meyers (1968) have shown with a synthetic system, that charge neutralization and bridging could not act simultaneously. Neither did Barber and Veenstra (1986) find any relationship between electrophoretic mobility of sludge particles and SVI of sludge samples from full-scale plants.
However, the electrophoretic mobility of activated sludge panicles has been shown to be directly (Forster, 1968), as well as inversely (Magara et al.,
1976) related to SVI.
Internal hydrophobicity and E C P related to settling capacity
Multiple regression analysis has been performed on data with SVIoa (ml. g - ~) as a dependent variable and five independent variables, namely the sludge C/N ratio (g. g-~) and volatile matter content (VM, g. 1- ~), K (dimensionless), IHB ( I 0 2 O D ~ unit. mol- ' of ammonium sulfate) and exocellular polysaccha- rides (EPS, m g . l - t ) .
In the range of SVIoa values between 69 and 315m1.g -~ and with two variables out of five, a model in the form of SVIoa = 3.3 x [EPS] - 35.1 x IHB + 178 with a coefficient of correlation equal to 0.89 can be computed. The sign associated to these coefficients shows that at low concentration of EPS and when the internal hydrophobicity is high, settle- ability is improved. If EPS are kept in the model, they are in fact only the "leader" of a larger group of other variables which characterize the hydrophilic ECP.
Except for proteins, when DNA, Mg 2 + or Ca 2 + ions are introduced instead o f EPS, similar models are obtained. However, it must be pointed out that this model is relevant to a small number of values and should be verified by the analysis of a larger group of sludges.
Table 7. Relationships between the amount of exocellular polymers (ECP) and sludge settlcability from the literature and from our results Reference SVI range (ml.g -I) ECP units Linear rgeression analysis n •
Our study 69-315 mg ge" i - ' SVI - 3.46ECP + 14.90 16 0.89
Forster (1971) 46-190 gge.100g DM-' SV1 - 3.04ECP + 48.99 II 0.84
Mapra et at. (1976) 37-152 g TOC.100 g DM-' SVI ,- 4.2$ECP- 6.28 3 0.99
Kiff (1978) 65-400 ms ECP- 100 S DM - ' SVI m 57.60ECP - 27.34 4 0.90
Chao and K©inath (1979) 100-725 mgge. 100 g D M - ' No ~,lation - - --
Goodwin and Fortser (1985) 17~-345 m g g e . 1 0 0 S T O C - ' SVI - - 19.03ECP + 6.47 8 0.87
ge: glucose equivalents; DM: dry matter content of the sludge; TOC: total organic carbon; n: number of data.
Bioflocculation in activated sludge 837 In a highly hydrated structure such as a biological
sludge, little attention has been paid to the role of hydrophobicity in flocculation. Valin and Sutherland (1982) correlated flocculation in activated sludge with hydrophobicity on the basis of contact angle measurements, but this method describes rather an external surface characteristic, as the sludge is not dispersed before analysis, than the hydrophobic interaction inside the floe. Biological sludges are composed of many different species of micro- organisms and Singh and Vincent (1987) have iso- lated one Pseudomonas sp, strain which has proved hydrophobic only when grown in a diluted medium, this hydrophobic character being associated with the capacity to form aggregates.
This opposition between hydrophilic and hydro- phobic interactions has been shown by Wrangstadh et al. (1986) with a marine Pseudomonas sp.: the production of an EPS by the cells under starvation conditions was associated with a decrease in their hydrophobicity measured by their adsorption onto a hydrophobic surface, the subsequent release of this EPS leading to an increase in cell surface hydro- phobicity.
The hydrophobicity of Corynebacterium glutamicum is higher when grown in phosphate-saturated con- ditions compared with cells grown in phosphate- depleted conditions (Bftchs et aL, 1988). Phosphate limitations induce the synthesis of teichuronic acids in place of teichoic acids in Gram positive bacteria (Duckworth, 1977). According to Bfichs et ai. results, additional changes may occur in the cell wail structure as these two kinds of polymers are mainly hydrophilic.
Hydrophobic interactions between cell surfaces are probably mediated mainly by proteic constituents of bacteria such as fimbriae. Many bacteria, either Gram positive or negative, are known to possess these kinds of structures which are responsible for their adhesive properties (Jones, 1977). The role of a surface protein in the hydrophobicity of Aeromonas salmonieida (Parker and Munn, 1984) or in the adhesion of Vibrio proteolytica to hydrophobic surfaces (Paul and Jeffrey, 1985) are demonstrative of the influence of proteic exo~llular structures in the hydrophobicity of bacteria.
CONCLUSION
Settleability of activated sludge can be described, in the investigated range of SVI0.1 values, by a stochastic model with only variables describing the hydrophobic and hydrophilic interactions inside the structure of biological floe.
The schematic view of flocs which result from the aggregation o f multiple elementary sub-units is open to criticism as there is not a constant ratio between each of their constituents, namely cells and exocellular polymers.
Filamentous bacteria have always been observed in our study and it may be suggested that they provide a backbone for the overall floc structure. However,
their overgrowth is always associated with settling problems.
Exocellular polymers from metabolism, cell lysis or wastewater are involved in the formation of a three- dimensional matrix or gel where divalent cations, at least Ca 2 + and Mg 2 +, act as bridging agents with probably specific affinities for each kind of exoceilular polymer.
In such a highly hydrated system as biological sludges, internal hydrophobic bondings are involved in flocculation mechanisms and their balance with hydrophilic bondings determines the sludge settling properties. From the model for SVIo,, one may assume that if the famous exopolysaccharides are really needed in the floc structure, hydrophohic areas in between the cells act as essential adhesives.
REFERENCES
Barber J. B. and Veenstra J. N. (1986) Evaluation of biological sludge properties influencing volume reduction.
J. Wat. Poilut. Control Red. Sg, 149-156.
Beccari M., Mappelli P. and Tandoi V. (1980) Relationship between bulking and physicochemical-biological proper- ties of activated sludge. Biotectmol. Bioeng. 22, 969-979.
Belcourt A., Kroely F. and Frank R. M. (1974) Effet de divers cations ct des variations de pH sur ragrtgation bacttrienue exptrimentale. J. Biol. Bucc. 2, 189-198.
Box G. E. P,, Hunter W. G. and Hunter J, S. (1978) Statistics for Experimenters, an Introduction to Design, Data Analysis and Model Building. Wiley, New York.
Bradford M. M. (1976) A rapid and sensitive method for the quantification of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72, 248-254.
Brown M. J. and Lester J. N. (1980) Comparison of bacterial extracellular polymer extraction methods. Appl.
MicrobioL BiotechnoL 40, 179-185.
Bfichs J. N., Mozes N., Wandrey C, and Rouxhet P. G.
(1988) Cell adsorption control by culture conditions.
Appl. Microbiol. Biotechaoi. 29, i 19-128.
Calleja G. B. et al. (group report) (1984) In Microbial Adhesion and Aggregation (Edited by Marshall K. C.), pp. 303-321. Springer, Berlin.
Cart D. F. and Ganczarczyk J. J. (1974) Activated sludge exoceilular material extraction methods and problems.
Proc. 9th Can. Syrup, War. Pollut. Res. Can., pp. 250--261.
Chao A. C. and Kcinath T. M. (1979) Influence of process loading intensity on sludge clarification and thickening characteristics. War. Re$. 13, 1213-1223.
Clarke A. R. and Forster C. F. (1982) Biopolymer yields from activated sludge and their relation to the operation of treatment plant. Biotechnoi. I, ell. 4, 655-660.
Deriaz R. E., Stacey M., Teece E. G. and Wigg/ris L. F.
(1949) The Dische reaction for 2-deoxypentoses. 3. Chem.
Soc. 14, 1222-1232.
Dubois M., Giiles K. A., Hamilton J. K., Rebel's P. A. and Smith F. (i 956) Colorimetric determination of sugars and related substances. Anal. Chem. 28, 350-352.
Duckworth M. (1977) Teichoic acids. In Surface Carbo- hydrates of the Prokaryotic Cell (Edited by Sutherland I. W.), pp. 177-208. Academic Press, New York.
Eriksson L. and Aim B. (1991) Study of flocculation mech- anisms by observing effects of a complexing agent on activated sludge properties. War. ScL Tech. 24, 21-28.
Eriksson L. and Axberg C. (1981) Direct influence of wastewater polluants on flocculation and sedimentation behavior in biological wastewater treatment. I Model system Facherichia coil B. War. Res. 15, 421-431.
838
Figueroa L. A. and Silverstein J. A. (1987) Ruthenium red adsorption method for measurement of extracellular polysaccharides in sludge flocs. Biotechnol. Bioeng. 33, 941-947.
Forster C. F. (1968) The surface of activated sludge particles in relation to their settling characteristics. Feat. Res. 2, 767-776.
Forster C. F. (1971) Activated sludge surfaces in relation to the sludge volume index. ;Vat. Res. 5, 861-870.
Forster C. F. (1985a) Factors involved in the settlement of activated sludge. I Nutrient and surface polymers.
War. Res. 19, 1259-1264.
Forster C. F. (1985b) Factors involved in the settlement of activated sludge. II The binding of polyvalent metals.
Wat. Res. 19, 1265-1271.
Forster C. F. and Dallas-Newton J. (1980) Activated sludge settlement, some suppositions and suggestions.
Wat. Pollut. Control 338-351.
Forster C. F. and Lewin D. C. (1972) Polymer interactions at activated sludge surfaces. Effi. Wat. Treat. J. 12, 520-525.
Goodwin J. A. S. and Forster C. F. (1985) A further examination into the composition of activated sludge surfaces in relation to their settlement characteristics.
Wat. Res. 19, 527-533.
Handbook of Chemistry and Physics (1990) 71st edn, 12-1 (Edited by Lide D. R.). CRC Press, Boston.
Herbert D. (1971) The chemical composition of micrc~
organisms as a function of their environment. In Methods in Microbiology (Edited by Norris J. R. and Ribbons D. W.), pp. 391-416. Academic Press, New York.
Horan N. J. and Eccles C. R. (1986) Purification and characterization of extracellular polysaccharide from actio rated sludges. Wat. Res. 20, 1427-1432.
Jones G. W. (1977) The attachment of bacteria to the surfaces of animal cells. In Microbial Interactions (Edited by Reissig J. L.), pp. 139-176. Halsted Press, New York.
Kiff R. J. (1978) A study of factors affecting bioflocculation in activated sludge process. War. Pollut. Control 77, 464-470.
Kiff R. J. and Thompson R. T. (1979) Physical methods for the extraction of bacterial exopolymers from activated sludge biomass. BiotechnoL Lett. 1, 183-186.
King R. O. and Forster C. F. 0990) Effects of sonication on activated sludge. Enz. Microb. Technol. 12, 109-115.
Li D. H. and Ganczarczyk J. J. (1990) Structure of activated sludge floes. BiotechnoL Bioeng. 35, 57-65.
Lindafil M. A., Faris A., Wadstrom T. and Hjerten S.
(1981) A new test based on "salting out" to measure relative surface hydrophobicity of bacterial cells. Biochim.
Biophys. Acta 677, 471-476.
Magara Y., Nambu S. and Utosawa K. (1976) Biochemical and physical properties of an activated sludge on settling characteristics. War. Res. 10, 71-77.
Magnusson K. E. (1980) The hydrophobic effect and how it can be measured with relevance to cell-cell interactions.
Scand. J. Infect. Dis. 24, 131-134.
Malone J. A. and Caldwell D. E. (1983) Evaluation
/
of surface colonization kinetics in continuous culture.
MicrobioL EcoL 9, 299-305.
Mueller J. A., Morand J. and Boyle W. C. (1967) Floc sizing techniques. App/. Microbiol. 15, 125-134.
Novak J. T. and Haugan B. E. (1981) Polymer extraction from activated sludge. J. Wat. Pollut. Control Fed. $3, 1420-1424.
Parker N. D. and Munn C. B. (1984) Increased cell surface hydrophobicity associated with possession of an ad- ditional surface protein by Aeromonas salmonicida.
F E M S Microbiol. Lett. 21, 233-237.
Parker D. S., Kaufman W. J. and Jenkins D. (1971) Physical conditioning of the activated sludge floe. J. Wat. Pollut.
Control Fed. 43, 1817-1833.
Paul J. H. and Jeffrey W. H. (1985) Evidence for separate adhesion mechanisms for hydrophilic and hydrophobic surfaces in Vibrio proteolytica. Appl. Erwiron. Microbial.
50, 431-437.
Pavoni J. L., Tenney M. W. and Echelberger W. F. Jr (1972) Bacterial exocellular polymers and biological flocculatiun.
J. Wat. Poilut. Control Fed. 44, 414--431.
Pujol R. and Canler J. P. (1989) Le foisonnement des boues activ&~s: situation du probitY'me en France. T.S.M. L'eau 1, 19-24.
Ries H. E. Jr and Meyers B. L. (1968) Flocculation mechan- ism: charge neutralization and bridging. Science 160, 1449-1450.
Rudd T., Sterrit R. and Lester J. N. (1982) Extraction of extraceUular polymers from activated sludge. Eur. J.
Microbioi. BiotechnoL 16, 23-27.
Sato T. and Ose Y. 0980) Floc-forming substances extracted from activated sludge by sodium hydroxyde solutions. War. Res. 14, 333-338.
Sezgin M., Jenkins D. and Parker D. S. 0978) A unified theory of filamentous bulking. J. Wat. Pollut. Control Fed. 50, 362-381.
Singh K. K. and Vincent W. S. (1987) Clumping character- istics and hydrophobic behavior of an isolated bacterial strain from sewage sludge. Appi. Microbiol. Biotechnol.
25, 396-398.
Steiner A. E., McLaren P. A. and Forster C. F. (1976) The nature of activated sludge flocs. Wat. Res. 10, 25-30.
Sutherland I. (T~d.) (1977) In Surface Carbohydrates of the Prokaryotic Cell, pp. 84--I 14. Academic Press, New York.
Turakhia M. H., Cooksey K. E. and Characklis W. G.
0983) Influence of a calcium-specific chelant on biofilm removal. Appl. Environ. Microbiol. 46, 1236-1238.
Valin S. D. and Suthcrland D. J. (1982) Predicting bio- flocculation: new developments in the application of flocculation theory. Environ. Technol. Lett. 3, 363-374.
Wanner J. and Grau P. (1989) Identification of filamentous microorganisms from activated sludge: a compromise between wishes, needs and possibilities. ;Vat. ICes. 23, 883-891.
Wrangstadh M., Conway P. L. and Kjeileberg S. (1986) The production and release of an extracellular polysaccharide during starvation of a marine Pseudomonas sp. and the effect thereof on adhesion. Appl. Microbiol. 145, 220-227.
