Registration Dossier

Environmental fate & pathways

Adsorption / desorption

Currently viewing:

Administrative data

Link to relevant study record(s)

Reference
Endpoint:
adsorption / desorption
Remarks:
other: Sorption under field conditions in 18 sewage treatment plants
Type of information:
experimental study
Adequacy of study:
key study
Study period:
1 March - 31 December 2004
Reliability:
2 (reliable with restrictions)
Rationale for reliability incl. deficiencies:
other: Results from extensive field studies, based on a large database, not in compliance with GLP. Analytics with high level of quality control.
Principles of method if other than guideline:
In an extensive monitoring programme in 18 STPs in Italy (6), Spain (6), Greece (3) and Berlin, Germany (3) were sampled four times with an interval of at least 14 days. The ratio between the concentrations in effluent and on sludge was evaluated and results in an estimation of Kd under real world conditions.
GLP compliance:
no
Type of method:
other: evaluation of partitioning between effluent and sludge under real world conditions
Media:
other: activated sludge
Radiolabelling:
no
Test temperature:
The samples are taken from STPs during normal operation, under normal dry weather conditions in the period between March and September 2004. Thus the partitioning between effluent and sludge reflect the status under ambient temperature at the locations.
Analytical monitoring:
yes
Details on sampling:
Fifteen STPs in South Europe (S.E.), 6 in Italy, 6 in Spain and 3 in Greece were selected:
Italy: Ferrandina, Avellino, Roma Est, Martina Franca, Nova Siri, Matera
Spain: Rincon de Leon, Orihuela Costa, Monte Orgegia, Villajoyosa, Almoradi, Orhuela Ciudad
Greece: Lamia, Volos, Larissa
In additon three plants in Germany were sampled for comparison (North EU to South EU): in Berlin: Ruhleben, Wassne aus Dorf, Schonerlinde.

Sampling was conducted under normal dry weather conditions in the period between March and September 2004. In order to assess the temporal and spatial variability stack samples of activated sludge and effluent were taken at four times with intervals of at least 14 days for Italy, Spain and Greece and two times for Berlin. The final number of sludge samples was 60, while 59 effluent samples were studied (the first effluent sample of Larissa, Greece, was discarded because the system was not properly functioning).

The sampling and the sample preparation were conducted by different subcontractors in each country. For Germany the work was organized by Uwe Dünnbier, Berliner Wasserbetriebe Berlin, for Italy by Giovanni Tiravanti, Pilot Applied Technology Consulting, Palo del Colle, for Greece by Manos Dassenakis, University of Athens and for Spain by Ignacio Valor Labaqua Alicante, Spain.

Sampling procedure
All sampling, sample preparation and analyses were conducted according to the same protocols as described in the detailed sampling and pre-treatment instruction (Blok J , Kupper T, Bonnet C, Alencastro F de, Grandjean D, Tarradellas J. 2004. Sampling Instructions for Analysis of Polycyclic Musks in Activated Sludge and Sewage Treatment Plant Effluent. Ecole Polytechnique Federale de Lausanne).
Activated sludge was decanted after 0.5 hour settling on the site. The concentrated sludge was centrifuged and freeze dried at the local laboratory of the subcontractor.
Effluent was clarified by gravity settling for 0.5 hour on the site and the supernatant was extracted over a solid phase (SPE) by filtering through a Speed Disk (5 cm diameter). The speed disks were properly conditioned with methylene chloride/ diethylether followed by methanol and distilled water before filtration.
Sediment was freeze dried at the local laboratory of the subcontractor.

As OTNE is used in household cleaning, deodorants, detergents, hand soap, personal care products and perfumes, special care had to be taken not to contaminate the samples. The sampling and analytical technicians were instructed as much as possible to prevent cross contamination by refraining from the use of these products before sampling or analysis and to prevent any direct contact of the samples with skin or textile.

The substances have strong adsorbing properties, are moderately volatile and may be biodegradable and/or photodegradable by daylight. Therefore specific instructions were given to prevent the use of plastic sampling devices (scoop), containers or lids to which the substances may adsorb. Instructions were given to use devices of aluminium, stainless steel or glass, to cool or freeze samples and store in dark, to handle samples as quickly as possible and to seal final samples properly.
Specific sampling instructions were given to clean all equipment and containers properly before use by carefully washing with tap water, rinsing with de-ionized water, drying, rinsing with acetone and n-hexane and drying afterwards. Before taking and adding the sample to the transport container, the sampling devices (scoop, basket or mixer) and transport container were rinsed with the same activated sludge/effluent that was to be sampled. Lids of jars for transport containing plastics on the interior side were to be covered with aluminium foil to prevent any adsorption of the substances to plastic.

SPE filters and filtration devices, standard solutions and silicagel samples were purchased, prepared and distributed by the analytical laboratory (OMEGAM) to each of the contracted sampling laboratories.

The effluent samples on the SPE filters and the freeze dried sludge and sediment samples were sent to the analytical laboratory OMEGAM in Amsterdam for further analyses.
Details on matrix:
The matrices in the study were the activated sludge and the effluent in the STP. Technical details on the design and process operation of the STPs were collected according to a questionnaire. The results are summarised below under "Any other information on materials and methods".
Details on test conditions:
Selected STPs in S.E. are operated with the normal activated sludge process, treating mainly municipal sewage for more than 10.000 inhabitants. Process conditions and design characteristics were collected from the operators. Technical data on the design and process operation of the sewage treatment plants were collected by the subcontractors for sampling from the STP operators, according to a questionnaire. The details are given in "Any other information on materials and methods".
Computational methods:
RATIO BETWEEN THE CONCENTRATION IN SLUDGE AND IN EFFLUENT
Based on the theory of equilibrium partitioning for the adsorption of OTNE from water to the organic fraction of sludge, a constant ratio (Kd) between Csludge/Cwater may be expected after normalization of the concentrations in sludge for the total organic carbon fraction (TOC). This theory is tested here for the adsorption of OTNE from water to the organic fraction of sludge.

The log Kd expressed as log (103 * Csludge/Ceffluent) can be found by a linear plot of:

Log (Ceffluent) = Log (Csludge) + 3 - log Kd

where Csludge is expressed in mg/kg and Ceffluent in µg/l
Key result
Type:
log Koc
Value:
4.12
Adsorption and desorption constants:
Based on the evaluation of the pair-wise ratios Clsudge/Ceffluent data and the observation that there was no significant correlation between Kd and TOC (p = 0.31), the log Kd = 3.53 (for the 18 STPs in this study, including the 3 plants from Berlin). If the log Kd-value is constant and independent of the OTNE concentration, the relation between log Csludge and log Ceffluent is a straight line with an angle of 45º.
In fact a better curve fitting was achieved by the line
log Ceffl = -0.294 + 0.733 * log Csludge
or
(Csludge) ^ 0.733 = 1.95 * Ceffluent with a significant deviation from the angle of 45º line (p=0.0003). This relation is the non-linear Freundlich equation with n=0.733.

The deviation from a constant partition coefficient implies that at low OTNE concentrations the effluent concentration were relatively high (small Kd) and at high overall OTNE concentration the effluent concentration were relatively small (high Kd). Apparently the sludges appeared to have a higher adsorptive potential at higher concentrations. This phenomenon is shown in the figure attached to this study record.

Recovery of test material:
Analytical and local recovery data for OTNE including coefficients (CV) in % (based on deuterated AHTN)
Sludge Effluent
Analytical recovery n=21 n=21
Range 89.5 – 127.0 61.9 – 134.0
Mean 102.0 96.3
CV 10.9 14.9
Local recovery n=14 n=80
Range 66 - 101 44 - 127
Mean 83 82
CV 13 20

Blanks sludge
In general blank values were below 0.1 mg/kg, but incidentally a contamination up to 1.2 mg/kg was observed.

Blanks effluent
Blanks were generally <0.1 µg/l, but at two occasions 0.14 µg/l OTNE was measured.

Sludge Effluent
CV of analytical recoveries
n= 21 10.9 14.9
CV of local recoveries
n = 14 sludge
n = 80 effluent* 13 20
CV of duplicates
n= 8 17 24
-----------------------------------------------
An analysis of variance of the effluent and sludge concentrations showed that the observed variance was mainly due to local differences, to lower extent to variation in time. The variation between sludge samples due to analytical error is only 4%, whereas for effluent samples the analytical error contributes for 9 % in the variability.
Concentration of test substance at end of adsorption equilibration period:
not applicable. The concentrations in the effluent ranged between 0.5 - 7 mg/l. Concentrations on sludge were between 2 and 25 mg/l.

The ratio between the concentration in effluent and the concentration in sludge, expressed as the log Kd-value is fairly constant for all the data of this monitoring study (66 data pairs). 80% of the log Kd-values is between 3.2 and 3.75, which is on a linear scale only a factor of 3.5. When all log Kd-values are considered together the extreme values for OTNE varied between 2.98 and 4.18 which on a linear scale represents a factor of 16.

 

Various hypotheses were assessed to explain this variability of Kd-values in relation to:

1.     OC content of sludge

High or low Kd-values did not correlate with high and low fractions of organic carbon in the sludge.

2.     Lipid content in organic fraction of sludge

Theoretically a higher organic sludge load (the amount of BOD per day that is loaded to an amount of activated sludge) could give a higher lipid content in the sludge and hence a correlation could be expected between Kd and the sludge load. However no correlation was found between Kd and the sludge load.

3.     Dispersed solids

Apparent differences between Kd-values may also result from the sampling procedure. If a small amount of finely dispersed solids or dissolved humic acids is trapped in the SPE filters and extracted together with the amount dissolved in water, this will reduce the apparent Kd. With a log Kd of 4.0 an amount of 100 mg solids per litre effluent could cause a doubling of the total (dissolved and sorbed) amount in the effluent. However, such a large amount of solids in the observed effluent is very unlikely. Samples were not taken unless the effluent was of good quality and suspended solids were removed by settling. It is estimated that the largest amount of solids that could be trapped is between 5 and 10 mg/l, so this could not create the apparent difference between Kd-values.

4.     Dissolved Organic Carbon

In addition to suspended solids, dissolved organic macromolecules also result in variation of Kd values. The amount of dissolved humic acids or other organic residues in the effluent can be estimated on the basis of reported COD. For the plants invalues of effluent COD are between 40 and 60 mg/l. A difference of 50 mg COD/l is still possible. If we assume that this COD consists mainly of humic acids and that humic acids have the same adsorption properties as the suspended solids, a variation of the COD in the effluent of 50 mg/l could be responsible for a variation of the Kd by a factor 1.5. Unfortunately the data set does not allow one to determine a correlation between effluent COD and Kd-values, because in many cases the COD is not measured as a routine and only BOD values were presented.

5.     Non-linear adsorption

Another possible reason for the variability of Kd-values follows from the observation that Kd is not linear. As figure 1 shows, Kd-values for OTNE are higher at higher concentration levels on the sludge by a factor of 16 on a linear scale. This phenomenon is opposite to a normal adsorption isotherm, where saturation occurs at higher concentrations. The phenomenon is well known for catalyst particles where the reaction on the surface interacts with the diffusion to the centre of the particles. In this case it may be explained by the combined interaction of diffusion and biodegradation.

Assuming a considerable primary biodegradation in the flocs, the diffusion process from the solution into the core of the flocs will be partly counteracted by this biodegradation. As a result the centre of the floc may not be available for sorption. If the biodegradation rate is low and constant, whereas the concentration levels of OTNE are different, this will lead to a variable fraction of degradation and this in turn may influence the fraction of the sludge that participates in the sorption process. At a high concentration level the migration of the substance through the peripheral layer of sludge flocs towards the centre of the flocs may be more complete than at a low concentration.

Thus at a high concentration a larger fraction of the solids participates in the equilibrium partitioning, giving a relative high Kd, whereas at a low concentration only a peripheric layer of the sludge floc participates in the partitioning, giving a relative low Kd.

Both hypothetical effects, adsorption to water soluble organic macromolecules (4) and incomplete participation of the floc centre to partitioning (5), could coincide and the combined effect could explain a factor 1.5 * 2.3= 3.5 difference of Kd on a linear scale.

Validity criteria fulfilled:
not applicable
Conclusions:
The mean Log Kd-value based on concentrations measured in sludge and effluent was 3.52. Using the mean fraction of organic carbon in the activated sludge, Koc was estimated. Log Koc = 4.12.
Executive summary:

Kd was derived from measurements in 60 paired samples of sludge and effluent in 18 sewage treatment plants in southern and northern Europe. The sampling procedure and analyses were carried out with a high level of quality control and the variation due to analytical error was low. The ratio between the concentration in effluent and the concentration in sludge, expressed as the log Kd-value is fairly constant for all the data of this monitoring study. 80% of the values is between 3.2 and 3.75, which is on a linear scale only a factor of 3.5. When all log Kd-values are considered together the extreme values for OTNE varied between 2.98 and 4.18 which on a linear scale represents a factor of 16.

The mean Kd-value was used to estimate log Koc, using the mean fraction of organic carbon in the sludge, resulting in log Koc = 4.12.

Description of key information

Kd was derived from measurements in 60 paired samples of sludge and effluent in 18 sewage treatment plants in southern and northern Europe. The sampling procedure and analyses were carried out with a high level of quality control and the variation due to analytical error was low. The ratio between the concentration in effluent and the concentration in sludge, expressed as the log Kd-value is fairly constant for all the data of this monitoring study. 80% of the values is between 3.2 and 3.75, which is on a linear scale only a factor of 3.5. When all log Kd-values are considered together the extreme values for OTNE varied between 2.98 and 4.18 which on a linear scale represents a factor of 16. The mean Kd-value was used to estimate log Koc, using the mean fraction of organic carbon in the sludge, resulting in log Koc = 4.12.

Key value for chemical safety assessment

Koc at 20 °C:
12 589

Additional information

[LogKoc: 4.12]