Registration Dossier

Environmental fate & pathways

Monitoring data

Currently viewing:

Administrative data

Endpoint:
monitoring data
Type of information:
experimental study
Adequacy of study:
weight of evidence
Reliability:
2 (reliable with restrictions)
Rationale for reliability incl. deficiencies:
other: Test procedures cannot be subsumed under a testing guideline, nevertheless are well documented and scientifically acceptable.

Data source

Reference
Reference Type:
other: Dissertation
Title:
Unnamed
Year:
1997

Materials and methods

Principles of method if other than guideline:
In order to evaluate the risk of laundry detergents FWAs, for which behaviour in the aquatic environment is known, were investigated.
An analytical method for the FWAs was developed as a prerequisite for the investigation of the aquatic environment. This method consists of the extraction of FWAs from both solid and aqueous samples with subsequent separation with reversed-phase HPLC. The method was then used to investigate the behaviour and fate of detergent-derived FWAs in natural waters.
The investigation was divided into three parts:
1. Occurrence and substance fluxes
2. Transformation processes
3. Enrichment in the benthos of a lake.
GLP compliance:
no
Type of measurement:
background concentration

Test material

Reference
Name:
Unnamed
Type:
Constituent

Study design

Details on sampling:
SAMPLES for ANALYTICAL METHOD
SEDIMENT
- Source: sediments from Greifensee, a small lake in Switzerland
- Sample collection: sediments were collected in February 1995 by means of a gravity coring device. A PVC tube (diameter, 6 cm) was pushed into the lake bottom, closed on the top and pulled out.
- Treatment: on shore, the core was sliced horizontally into 5 cm segments corresponding roughly to 10 years of sediment accumulation.
- Storage: the samples were freeze-dried, homogenized with mortar and pestle and stored in the dark at 4 °C.

LAKE
- Source: lake water was collected from the middle of Greifensee.
- Sample collection: samples collected in February 1995, 10 m below the surface, were used for recovery and precision measurements. The samples used as application examples were collected in different depths in July 1995 and January 1996.
- Storage: all samples were immediately placed in the dark after collection, transported to the laboratory and analysed within 2 h without filtration.

Results and discussion

Any other information on results incl. tables

1. ANALYTICAL METHODS

1 1. Separation

An HPLC column with a diameter of only 2 mm was applied for the separation of the FWA. This allowed the injection of smaller sample volumes and saved 60% on solvents. To achieve baseline-separated peaks, the ratio of acetonitrile and methanol in the solvent had to be adjusted for the new column.

Photoisomers of test item, which occur under environmental conditions could be separated.

To assign the measured peaks to the FWA isomers, a standard solution was irradiated for 10 min by the sun and analysed twice.

The first time, the postcolumn UV irradiation was switched off, so that only the fluorescent E-isomer gave a signal. The second time, the UV irradiation was turned on again to visualize the nonfluorescent Z-isomer. The peaks that appares only in the second measurements were assigned to the Z-isomers.

 

1.2. Detection and quantitation

Photoisomerization is a reversible process, and the photostationary state of a diluted solution depends only on the wavelength of the incident light and on the temperature. If a dilute solution of FWA is irradiated for a time sufficient to reach a constant isomer ratio (photostationary state), the resulting E:Z ratios are independent of the initial isomeric composition.

In the absence of light, isomerization does not occur, and isomer ratios remain constant.

The UV irradiation is applied on-line after HPLC separation and before fluorescence detection in order to minimize matrix effects. The isomer concentrations of all isomers are determined on the basis of the E- isomer only.

This is possible because:

(i) only the E-isomer exhibits fluorescence, and the Z-isomer gives virtually no signal, and

(ii) the E:Z ratios are constant after UV irradiation.

The high limits of quantitation (LOQ) prevented the application of previously reported methods for FWA determination in lake sediments and surface waters. As the LOQs are strongly affected by the variability of blanks, a major effort was made to identify the sources of high blank values. In the case of aqueous samples, the main source was found in the repeated use of enrichment equipment.

The highly sorbing FWA was carried over from one sample to another by filter assemblies and by the vacuum manifold. Replacing polypropylene funnels with glass funnels and eluting the C18 disks directly into glass vials instead of through the vacuum manifold prevented a major part of the carryover. Due to technical problems, the polypropylene filter assemblies could not be replaced. Their substitution would probably have lowered the LOQs even more. Nevertheless, LOQ could be lowered by factors of 10-100 (to 3 ng/l). The blank signals never exceeded 2% of the concentrations measured in surface waters for (Z)-test item. Only for (E)-test item were the blank values 10-20% of the concentrations measured in surface waters. The average of the measured blank values was subtracted from signals obtained for each sample.

In the case of sediment samples, blanks were already very low.

The LOQ was lowered by a factor of 2-5 due to the use of an HPLC column with a smaller diameter (2 mm) and a smaller pore size (3 µm). The measured LOQ is 10 µg/kg.

Limits of Quantitation's for (E)-test item in sediment and water samples

Sediments (n. 8)* Water (n.5)**
Added amount 24.0 4.13
Average meas. value 22.6 5.13
Standard Deviation 1.05 0.3
LOQ 10.5 3.0

*In µg of test item/kg of dry matter.

** In ng of test item/l.

Recoveries were between 93 and 100% and between 87 and 95% for solid and aqueous samples, respectively. The recoveries reported may only represent an upper limit. Especially in the case of sediments, one can assume that FWA will show a different binding behaviour to particles in the benthos of a lake compared to artificially spiked matrices. FWAs at the bottom of a lake have interacted with the particles for decades, whereas in the laboratory experiment interaction lasted only 15 minutes.

Recoveries of (E)-test item in sediment and water samples

Sediments (n. 3) Water (n.8)
Background conc. 0* 264* 8.5**
Spiked 129* 304* 49.6**
Recovery (%) 99 93 88
Standard Deviation (%) 3 2 2
*In µg of test item/kg of dry matter.

** In ng of test item/l.

The precisions of the individual FWA determinations were within 3-7% (relative standard deviation) for sediment extracts and 1-12% for aqueous extracts. Confidence intervals (95%) were generally below 8% of the mean concentration.

Precision of individual isomer measurements (n = 8)

Sediments*

Mean value conc. Standard deviation 95% confidence interval
Conc. % Conc. %
(E)-test item 521.8 17.5 3 ± 41.4 ± 8
(Z)-test item 94.0 3.3 4 ± 7.7 ± 8

Water**

(E)-test item 8.5 1.0 12 ± 2.4 ± 28
(Z)-test item 77.5 1.1 1 ± 2.6 ± 3

*In µg of test item/kg of dry matter.

** In ng of test item/l.

The low limits of quantitation, good recoveries, and excellent precision of individual measurements allow for the determination of individual FWA isomers, even at the low ambient concentrations found in lake sediments and surface waters.

 

1.3. Applications

In the lake water, two different situations can be observed:

(i) in summer, when the lake is stratified, and

(ii) in winter, when it is mixed

In summer, FWA concentrations are lower, due to stronger sunlight and presumably faster rates of photodegradation. As a consequence of stratification, concentrations in the sunlit top layer are lower than concentrations in the layers below 5 m, where no light is available. A second effect is the change in isomer ratios. Isomer ratios are a function of temperature and light conditions. First, the isomerization rate from E- to Z-isomers depends on temperature, while the reverse isomerization from Z- to E-isomers is temperature-independent. Due to lower temperatures in winter (3 vs 25 °C at the surface), the equilibrium of isomers is, therefore, shifted toward (E)-test item.

Second, in both summer and winter, the isomer ratios in deeper waters are shifted towards (Z)-test item, because longer wavelengths are better transmitted by lake water and favour the Z-isomers.

In the sediments of Greifensee, a general decrease of FWA concentrations is observed with increasing distance from the mouth of the river that delivers FWA into the lake. E-Isomers have higher sorption coefficients and, therefore, are transported efficiently to the lake bottom than Z-isomers. Consequently, the ratios of E- to Z-isomers are about a factor of 10 higher in sediments than in water.

For the same reason the isomer ratios in the sediments decrease with increasing distance from the river mouth.

Test item concentrations in water and sediments of Greifensee, Switzerland

Water (ng/l)
(E)-test item (Z)-test item total (E) (Z)
Summer*
Surface 6 47 53 0.13
10 m depth 5 54 59 0.10
20 m depth 8 65 73 0.12
Winter**
Surface 17 81 98 0.21
10 m depth 11 81 92 0.14
20 m depth 12 83 95 0.15
Sediment (µg/kg)***
0.2 km**** 1220 200 1420 6.1
1.4 km**** 590 200 790 3.0
3.1 km**** 470 180 650 2.6

*July 24, 1995.

**January 10, 1996.

***Topo layer (0 -5 cm), February 27, 1995.

****Distance from mouth of River Aa in Uster.

1.4. Conclusion

The lake water and sediment concentrations of individual isomers can be determined accurately by reversed-phase HPLC followed by postcolumn UV irradiation and fluorescence detection.

2. RIVER MONITORING

2.1. FWA Concentrations and Loads in Swiss Rivers

Based on the measured test item concentrations, the investigated rivers can be divided into two groups: the two rivers of group III (small rivers with highly populated catchment areas) and the two sites, that are below the FWA manufacturing plants show relatively high FWA concentrations. All the other sites have clearly smaller concentrations. Dilution seemingly has a strong impact on the observed concentrations. The latter fact is derived from the observation that concentrations are directly correlated with water discharge at all stations, i.d. high FWA concentrations at low discharge rates and vice versa. On the other hand, a comparison with the temperatures in the investigated rivers showed no direct correlation to the FWA concentrations.

Test item loads and content in the DOC (stations below manufacturing plants are in bold)

River Location Yearly load (kg/y)* Content in the DOC (ppm)**
Group I
Rhine Diepoldsau 120 10
Saane Gümmenen 160 40
Rhone Porte du Scex 320 50
Group II
Aare Bern 170 20
Aare Hagneck 490 50
Rhine Rekingen 1000 30
Rhine Weil 20000 240
Rhone Chancy 980 50
Group III
Thur Andelfingen 260 40
Glatt Rheinsfelden 150 130

*Average of 13 measurements; **Average load/average DOC load (BOWAL, 1995)

Test item loads ranged from 120 kg at location Diepoldsau (0.3 x 106 inhabitants in the catchment area) to 20000 kg at the location Weil (6.9 x 106 inhabitants). At all sampling sites (except the ones below the FWA manufacturing plant), loads of test item were between 0.3 and 4.2 mg/day/inhabitant throughout the year. These relatively small differences in FWA loads during the year or from one site to another are attributed to unknown factors. The average of all values (except Weil) was used to calculate the average contribution of Swiss households. Averaged over the investigated year, and based on a population of 7.06 x 106 in Switzerland, per capita test item load was 1.8 (± 0.5) mg/day/inhabitant.

This corresponds to a total yearly load for Switzerland of 4.6 (± 1.3) t (95 % confidence interval calculated from the standard deviation). In other words, 13 (±4)% of the FWAs that were consumed during the year can be found in the rivers. The amount that is discharged by sewage treatment plants to surface waters, is probably greater, since FWA loads are reduced in surface waters between the discharge points and the sampling sites due to sorption/sedimentation and photochemical degradation.

ln contract to the value we obtained, the value of 22% discharge to surface waters reported by Poiger et al. (1997) is based on a ten day in investigation in one sewage treatment plant (STP) only. As the discharge from STPs may vary significantly depending on weather conditions any type of STP, the two values are in good agreement.

In contrast to the uniform behaviour of most of the investigated rivers much higher per capita FWA loads were measured at the stations below FWA manufacturing plant. At sampling site Weil, below the plant manufacturing the test item, an average of 8.0 (± 1.3) mg test item/day/inhabitant was found throughout the year. Assuming that the average per capita contribution of households is the same throughout Switzerland, an amount of approximately 6.2 (± 1.4) mg/day/inhabitant (15.7 (± 3.5) t per year) must have originated from the manufacturing process of the test item.

Test item discharge to surface waters during the manufacturing process

Weil site
Test item per capita loads (mg/day/inhabitant)
Below manufacturing plant 8.0± 1.3
All other sites 1.8± 0.5
Contribution from manufacturing plants
In units of mg/day/inhabitant 6.2± 1.4
In units of year 15.7± 3.5

2.2 FWA Mass Balance for Switzerland

An FWA mass balance was developed for Switzerland in order to achieve an overall assessment of the environmental fate of FWAs.

Source, pathways and sinks are described below.

a) Source and Pathways of FWAs

Estimates of the detergent producers (Eugster, 1997; Gehri, 1995) showed a consumption of 59 t of the test item and of another FWA in 1995. The confidence interval was estimated at ± 5 t. Since test item is only used in laundry detergents, there are two possible pathways to surface waters:

1. Directly from FWA manufacturing plants: during the manufacturing process, FWA precipitated from an aqueous solution by adding sodium chloride to the solution followed by filtration. A small fraction of FWA remains in the liquor and is discharged into the industrial wastewater and eventually to surface waters. By attributing the difference between the average FWA load in Swiss rivers and the FWA load below FWA manufacturing plant to the manufacturing process, the FWA discharge to surface waters by manufacturing plant is estimated as 15.7 t of test item in 1995.

2. Through households, institutions and industrial laundries: measurements in rivers of the study reported here indicate a total discharge of more than 8 (± 2) t of FWAs to surface waters in 1995.

b) Sinks of FWA

Different processes reduce the amount of FWA in sewage effluents and eventually in surface waters. They are called "sinks" here, although only one of them (sorption onto particles in surface waters followed by sedimentation) is a final sink. The four main categories are:

1. Fabrics: the purpose of detergent-FWAs is to replace FWA on fabrics, which were photochemically degraded during wearing. One can therefore assume that fabrics are an important "sink" for FWAs. Nevertheless it is difficult to assess the amount of FWA that is really adsorbed onto fabrics during the washing process, because this fraction depends on the material of the fabrics and on the way the fabrics are washed (temperature, ratio of fabrics and water, detergent, water quality). A crude estimation of mis "sink" was made by Poiger et al, (1997) by attributing the FWA load lost before introduction to the sewage treatment facility (56%) to fabrics. For 1995 this would result in an amount of 33 t. Since the basis of this data is uncertain, the confidence interval was arbitrarily estimated at 30 ± 10 t.

2. Sewage sludge: according to Poiger et al. (1997), 32% of the consumed FWAs are eliminated from the wastewater during sewage treatment due to sorption onto sewage sludge. For 1995 this corresponds to 19 t, 6 tof which are incinerated. The rest is applied on agricultural land (13 t), where the further fate has not yet been investigated. Nevertheless, photochernical degradation is likely to occur. Since these data are based on one sewage treatment plant only, the confidence interval is arbitrarily estimated at ± 50%.

3. Lakes: two processes reduce the FWA concentrations in lakes. FWAs can be (i) degraded photochemically (Kramer et al, 1996) and (ii) sorbed onto particles and sedimented to the lake bottom (Stoll et al., 1997). At this stage it is not possible to assess the relative importance of the two processes.

4. Export out of Switzerland: part of the catchment areas of the sampling site Weil lay outside of Switzerland. The per capita load calculated for this site was multiplied only by the number of inhabitants of the Swiss catchment area (5.2 x 106; Jakob et al., 1994), reflecting only the fraction of FWA originating from Switzerland. On the other hand, this result was enlarged by 4%, in order to include the Swiss inhabitants not living in the catchment area. In this way the amount of FWA exported out of Switzerland was assessed as 15 t foe the year 1995.

2.3. FWAs as Molecular Marker for Production Wastewater

Except for the station directly downstream from FWA manufacturing plant, the average FWA concentrations correlated reasonably well (correlation coefficients of 0.95) with the chloride concentrations (BUWAL, 1995). As has been demonstrated in other studies (Muller et al., 1996; Jakob et al., 1994), chloride can be used in Switzerland as an anthropogenic indicator because the present chloride levels in rivers are significantly higher than the natural chloride levels anticipated from weathering. Test item can be applied as molecular markers for production wastewater.

Concentrations of dissolved organic carbon (DOC; BUWAL, 1995) also correlated well with the measured FWA concentrations (correlation coefficient: 0.76). Consequently the results show that the station at which the FWA represented the greatest proportion of the organic carbon is that below the manufacturing plant (240 ppm). However, in contrast to the correlation with chloride, one additional site has also higher FWA contents in the DOC than the others River 3 (Glatt at Rheinsfelden) is by far the smallest of the investigated rivers (with an average tlow of 12 m3/s) and has a highly populated catchment area. This special situation seems to be reflected in the FWA/DOC ratio.

2.4. Elimination of FWA in Lakes

In lakes, FWA concentrations in the water column are reduced by three processes, photodegradation (Kramer et al., 1996), sorption/sedimentation

(Stoll et al., 1997), and flushing. Nevertheless, per capita loads measured at the locations of group Il rivers are not significantly smaller than the ones of the other locations. The investigated river stations seem to be too far away from the lakes, so that the influence of the lakes is small compared to other factors.

However, in a direct comparison of stations located before and after a lake, an influence of the lake can be documented by the FWA loads. Two such examples have been investigated within this study, Lake Constance (1A before and 1B after the lake) and Lake Geneva (6A before and 6B after the lake). The FWA loads measured at stations before and after these lakes are listed in the table below and show that FWA per capita loads are

effectively decreased in lakes.

Test item loads before (A) and after (B) lakes. Calculated as average of 13 measurements; 95% confidential interval derived from standard deviation.

Station load (t/y) per capita load (mg/day/inhabitant)
Costance 1A 0.1 1.0± 0.1
Costance 1B 1.0 1.1± 0.2
Geneva 6A 0.3 3.2± 0.4
Geneva 6B 1.0 1.9± 0.4

The average per capita load decreases in Lake Geneva and remains constant in Lake Constance. FWA has more time to react photochemically or to sediment to the bottom. Test item sorbs to particulate matter (Stoll, 1997).

The observed changes in FWA loads are not only due to degradation and sedimentation since, in both lakes investigated, other processes also influence the differences of the FWA loads. Additional rivers and sewage effluents flow into the two lakes and rivers between the sampling locations, so that only about half of the water analysed at the second locations (1B and 6B) also passed by the first locations (1A and 6A). Furthermore, the average residence time of the water in the Lakes Geneva and Constance is several years (11 y and 4 y, respectively). This means that the loads measured before and after the lake during one period do not correspond to each other. Thus, only a rough comparison can be made and the conclusions of this paragraph have a speculative character.

2.5 Conclusions

The present study shows that per capita loads of test item do not vary significantly in the investigated rivers, neither due to changing seasons of the year nor to different catchment areas. Especially the occurrence of lakes in the catchment area did not have an influence on the measured per capita FWA loads. Obviously, only limited conclusions can be drawn from this monitoring study concerning processes affecting the fate of FWA in surface waters. Averaged over the year, about 13% of the FWAs consumed in Switzerland are found in the rivers and were attributed to private households. Another 30% are discharged by FWA manufacturing plants directly into the rivers. Therefore, test item can serve as molecular marker for production wastewaters of the manufacturing plants. A good parameter for this purpose is the FWA concentration relative to the chloride concentration, as FWA load and number of inhabitants in the catchment area correlate significantly in rivers which are not influenced by effluents from FWA manufacturing plants.

3. LAKE MODEL

This study is based on three research concepts that are reflected in the text below:

First, every relevant process was evaluated independently. In particular, reliable parameters were determined for the FWA input from

STP effluents, for the sorption of FWA to particles and the sedimentation and for the photochemical degradation of FWA.

Second, FWA concentration profiles were measured in the lake during one year.

Third, the actual model was set up based on the two previous steps.

3.1. Evaluation of FWA Inputs from STP Effluents

ln order to assess the loading of FWA into Greifensee for every day of the year investigated, total daily loads were measured on 13 days. These total daily loads were correlated with the total daily sewage (sum of the water that was treated in all STPs of the catchment area). The correlation determined by linear regression showed good linearity with correlation coefficients of 0.94.

This indicates that FWA is eliminated from sewage in STPs more effectively when less water is treated (based on the assumption that the daily FWA amount discharged to STPs is constant). This observation can be explained, as FWA is eliminated from sewage treated in STPs exclusively by sorption onto sewage sludge[1]. Sludge is diluted during rainy periods which - in combination with the constant sorption coefficient leads to higher loads in the effluent. The correlation equations were assumed to be representative for the entire investigated year, since samples were collected on all days of the week, in all seasons and under dry as well as rainy conditions. Calculated loading was 114 ± 27 g/d for test item, averaged over the whole investigation period. The value obtained by this procedure was directly used in the model.

 

3.2. Evaluation of Sarption onto Particles and Sedimentation

The FWA concentrations in the dried particles collected in sediment traps range from 0.2 to 2.2 mg/kg for test item. Sorption isotherm was determined by correlating the concentration (Cs) with the FWA concentrations in the water of the corresponding depth at the end of the collection period of particles (CW) and then used to assess apparent solid-water distribution ratios Kd.

Previously sorbed molecules lead to a modification of the surface which favours further sorption. This seems to be the case for FWA in the concentration range investigated ( 0.1 µg/l). Negative exponents (and smaller Kd-values) have been reported for higher concentrations ( 10 mg/l), where a negative surface charge is built up with increasing adsorption, making the surface less attractive to other negatively charged molecules[2]. For the computer model, average distribution ratios were determined and average FWA concentration (0.088 µg/l). Average log Kd value (l/kg) is 3.9, with a 95% confidence interval of 0.6. This relatively big confidence interval is due to a large variety in composition and size of in situ particles and their respective settling velocity. In addition, particles are formed in some layers and dissolved in others.

Input parameters for the computer simulation, confidence intervals are 95%

Loading from inflow (g/d)*
minimum (October) 67± 26
maximum (June) 239± 26
average of 13 periods 114± 27
Sorption/sedimentation
log kd(l/kg) 3.9± 0.6
Photochemical degradation in the perfectly mixed top meter, k (d-1)*
minimum (October) 0.002
maximum (June) 0.036
average of 13 periods 0.019

*Individual values for every period were used for the computer simulation (minimum and maximum given as examples), The average is given only for infomation and was not used for calculations.

3.3. Evaluation of Photochemical Degradation

Maximum photodegradation rate coefficients under natural light conditions have been determined; these were converted into daily rates proportionally to the flux of photons for every particular day. The new rates were subsequently reduced by a factor corresponding to the fraction of light reflected at the air-water interface (6%, constant throughout the year[3]) and by a factor corresponding to attenuation by particles in the water (5- 29%, estimated by the reduction of visible light in the top meter of Greifensee).

During summer, periodic stratification of the top meter is likely to occur. Thus, the rate of vertical mixing is slow compared to photodegradation. ln order to compensate for smaller DSBP concentrations in the top meter, and in order to achieve a balanced model, DSBP photodegradation rates from March until September were decreased by 10 -30% as the simulation model requires a completely mixed epilimnion. The effect for test item is five times smaller and was, therefore, neglected.

Average degradation rate estimated in such a way for the top meter of Greifensee (degradation below 1 meter is virtually zero, as light with wavelengths below 400 nm is completely absorbed within this layer) is 0.019. The value obtained by this procedure was directly used in the model.

3.4. Seasonal Variation of Measured FWA in Greifesee

As indicated by the temperature profiles (dotted lines), the lake was stratified between May and November, followed by complete overturn in winter. In winter, with less sunlight and less particles in the lake, total FWA concentrations generally increased. ln summer, test item concentrations exhibit a maximum in the thermocline, whereas concentrations in the epilimnion and in the hypolimnion are distinctly lower. These findings are consistent with the assumption that test item is eliminated from the water column primarily by two processes, photodegradation in the epilimnion and sorption/sedimentation. Two hypotheses have been formulated about the origin of these peaks:

(I) they are a consequence of subsurface loading of the FWA. This would mean that a part of the water discharged to the lake surface

through tributaries and STPs is transported immediately to the depth with the corresponding temperature (and density). During summer this theoretical intrusion depth is 4-8 m from the lake surface. During winter the temperature in the lake is always smaller than in the tributaries, thus giving no evidence for a subsurface loading;

(II) a second possibility is that the peaks in the thermocline are caused by particle dynamics, because a part of the sinking particles is dissolved in the thermocline, thus transporting FWA from the epilimnion to the thermocline.

 

3.5. Computer Simulation, Model Validation

Two simulation models were established and calibrated with test item, model A with subsurface loading and model B without subsurface loading.

In both models, sorption coefficients of test item were kept constant and particulate organic carbon values (POC) were used as measured in the lake. Subsurface loadings for model A were applied from June to October with increasing depths from 4-6 m to 6-8 m, corresponding to the depths with the same temperature as the tributaries. Best results with the simulation model were obtained by assuming between 10% (June and October) and 40% (July/August) of the total loadings to enter the lake in the mentioned depths.

A control was achieved by comparing the measured and modelled FWA sedimentation fluxes at depths of 10 and 30 m. Considering the large confidence interval determined for the sorption coefficients (factor 4), measured and modelled data are in good correspondence. However, there seems to be a systematically error yielding to big values in the simulation model. A possible explanation is that the distance of the sedimentation traps from the lake surface was depending on the water level of the lake, thus varying from 9 to 10 m and from 29 to 30 m, respectively.

In summer, when the sedimentation velocity of model A has a strong gradient around 10 m depth, a reduction of the trap depth to 9 m causes a reduction of settling FWA of 25%. In fact, the depth of the upper sedimentation trap was 9.1 - 9.3 m during the periods with the biggest error (Apr. 3 - May 30 and Jun. 26 - Aug. 21).

First order rate constants averaged for whole lake (photodegradation and flushing) and calculated by model A (sedimentation), and their relative contribution to yhe total FWA reduction

Kmin (10-3/d)* Kmax (10-3/d)** Kaverage (10-3/d)*** Contribution to total FWA reduction ***
Photodegradation 0.9 8.8 4.4 49%
Sorption/sedimentation 0.9 4.9 2.5 27%
flushing 1.1 4.8 2.1 24%

*Perods 8/9 (October 16 -December 13); **Period 3 (May 30 -June 26); ***Averaged over the whole year (April 1995 -April 1996).

3.6. Computer Simulation, Transport by Sedimenting Particles

Calculated sedimentation velocities of particles (dots) had to be modified for the computer simulation (solid line) in order to achieve the measured depth profiles. There is a good qualitative agreement with both calculated and adapted sedimentation velocities being highest between 10 and 25 m lake depth. While the quantitative differences are small in the epilimnion, they can be interpreted as being due to incorrect assumptions about the suspended particulate matter concentration in the hypolimnion.

By using the measured POC-values for calculations, the value of particle matter concentration is over-estimated, as a part of the particles in the water column are not sinking and thus not transporting FWA to the ground of the lake. In order to compensate for over-estimated particle matter concentration values, the particle settling velocities had to be increased.

3.7. Computer Simulation, Photodegradation in Greifensee

The computer simulation that could successfully be applied, is a confirmation for

(i) the photodegradation rates measured in a laboratory[4] and

(ii) their transformation to “effective" degradation rates in a lake by taking into account the water mixing and the absorption of photons by the

water body.

“Effective" photodegradation coefficient in Greifensee, averaged over the whole lake volume, vary from 0.9 to 8.8 x 10-3/d. This means that less than 70% of test item is degraded within 28 days by photochemical processes, even in summer with the strongest sunshine. Test item is therefore not ready photodegradable in Greifensee following the definition about degradability given by an EC directive[5]. Note that this statement is somewhat arbitrary, as the EC directive does not specify the "aquatic environment".

lf only the photic zone of Greifensee is considered as "aquatic zone", as was done by Kaschig et al.[6], test item (in spring and summer) is ready photodegradable. However, it is more appropriate to define the whole lake as “aquatic environment".

4. LAKE SEDIMENT

4.1. FWA as Molecular Marker for Domestic Waste Water

Gravity cores were collected at 16 different locations of Greifensee. Every core was analysed from the sediment water interface down to a layer where no more FWA could be found (normally below 20 cm), The concentrations were integrated over the whole core in order to obtain total FWA inventory (in mg FWA/m2) for every core. The sample sites were divided into 3 groups, being under the influence of tributary Aabach (group A) and being under the influence of tributary Aa (group B towards south, group C towards north). FWA inventory was generally found to decrease with increasing distance from wastewater inputs.

Since (E)-test item has higher sorption coefficient, it sediment faster to the lake bottom than the corresponding (Z)-test item. Hence (E)-test item is enriched compared to (Z)-test item areas close to the tributaries.

 

FWA Inventory* in sediment of Greifensee, Switzerland

Distance from tributary (E)-isomer (Z)-isomer Total (E) (Z)
Group A (Aabach)
0.3 km 243.4 21.9 265.2 11.1
0.6 km 35.6 6.6 42.2 5.4
0.7 km 14.9 3.3 18.2 4.5
1.7 km 19.7 5.6 25.2 3.5
Group B (Aa towordsv south)
0.2 km 157.9** 15.0** 172.9** 10.5
0.5 km 40.3 8.1 48.4 5
1 km 28.5 6.7 35.2 4.2
1 km 27.5 5.7 33.2 4.8
Group C (Aa towards north)
0.2 km 157.9** 15** 172.9** 10.5
1.1 km 49.8 7.6 57.4 6.6
1.4 km 27.7 5.6 33.3 4.9
1.8 km 24.3 6.4 30.7 3.8
2.2 km 21.4 5.5 26.9 3.9
2.2 km 26.5 6.6 33.1 4
2.3 km 20.9 5.9 26.8 3.6
2.9 km 18.9 4.8 23.7 4
3.1 km 21.1 6.1 27.2 3.5

*: Inventory of cores from the sediment/water interface down to a depth where no more FWA could be detected, unit: mg/m2

**: Minimum value, FWA could still be detected in the lowest layer analysed (40 cm).

4.2. Sedimentary Archive and Emission History of FWAs

One freeze core was taken in the middle of the lake. Layers corresponding to periods of two years were analyzed. The obtained vertical concentration and flux profiles of FWA in the sediments of Greifensee document the emission history of the FWA: test item is first found in sediment layers that were deposited in 1965. This finding match the very limited information which was available from the FWA manufacturing industry. Concerning the general development of FWA concentrations in the sediment, the two expected trend changes in 1967/71 and in 1981/84 can be observed. Laundry detergent use increased continuously during this century, but FWA discharge to the environment, as recorded in the sediment, only increased until 1971 and then more or less stagnated until 1983, after which a decrease in FWA concentrations can be observed.

The first change in 1971 corresponds to the beginning of sewage treatment in 1967/71. The second one in 1983 reflects the addition of direct filtration to sewage treatment in 1981/84.

Sediment Core Collected by Means of 2 Freeze Core in Greifensee nn October 24, 1995.

year of deposition lower limit (cm) thickness (cm) water content (% mass) sedimentation (g/cm3·2y)
1993/94 1.46 1.46 89.1 0.246
1991/92 2.38 0.92 80.1 0.174
1989/90 3.36 0.97 80.2 0.233
1987/88 4.29 0.93 78.2 0.265
1985/86 5.18 0.90 76.4 0.205
1983/84 5.82 0.63 73.2 0.214
1981/82 6.71 0.89 73.5 0.299
1979/80 7.49 0.78 69.3 0.287
1977/78 8.20 0.71 69.1 0.250
1975/76 8.90 0.70 68.9 0.273
1973/74 9.76 0.86 69.2 0.275
1971/72 10.45 0.69 71.6 0.231
1969/70 11.25 0.80 72.8 0.255
1967/68 11.94 0.69 76.5 0.185
1965/66 12.84 0.90 72.6 0.264
1963/64 13.49 0.66 71.3 0.210
1961/62 14.03 0.54 69.1 0.220
1959/60 14.70 0.67 66.1 0.211

FWA Concentrations in a Sediment Core Collected by Means of a Freeze Core in Greifensee on October 24, 1995.

year of deposition (E)-isomer (Z)-isomer
1993/94 473.7 200.9
1991/92 429.0 235.5
1989/90 456.6 214.7
1987/88 536.4 160.2
1985/86 651.5 127.1
1983/84 484.3 112.8
1981/82 895.4 134.4
1979/80 887.6 128.0
1977/78 911.6 169.4
1975/76 748.5 151.6
1973/74 933.5 179.3
1971/72 1197.9 165.9
1969/70 707.2 119.4
1967/68 231.8 64.8
1965/66 16.4 27.1
1963/64 0.0 < 10.0
1961/62 0.0 0.0
1959/60 0.0 0.0

4.3 Conclusions

Detergent-derived FWA could be determined in sediment cores of Greifensee. The known properties of FWA indicate a conservative behaviour in the sediments. Moreover, historic events like the introduction of a new FWA or the improvement of sewage treatment that can be documented by the measured FWA concentration profiles in the sediment make processes that would affect the FWA concentrations in the sediment core unlikely. Because FWA is strongly sorbed to particles, sedimentation is a rapid process compared to the horizontal mixing in Greifensee, and, hence, concentrations in the sediments can provide information about the transport of particles in the lake. Another property that could provide information, is the ratio of E- and Z-isomers of the FWA, since exposure to sunlight causes isomerization and different isomers show different sorption behaviour.

Reference

[1] Poiger, T.; Field, J. A.; Field, T. M.; Sicgrist, H.; Gigcr, W. Water Res. (in press)

[2] Kramer, J. B. Ph.D. Thesis, ETH Zurich, N0. 11934, 1996.

[3] Sauberer, F. Mitt. int. Ver. Limnol, 1962, 11, 1-77.

[4] Kramer, J. B.; Canonica, S.; Hoigné, J.; Kaschig, J, Environ. Sci. Technol. 1996, 30, 2227-2234.

[5] European Community Commision Directive 93/2l/EEC (67/548/EEC, 18th ATP) OJ No Ll l0/20, 1993, Apr. 27.

[6] Kaschig, J.; l-lochbcrg, R.; Zeller, M. Presented at the 4th World Surfaclanr Congress, Barcelona 1996.

Applicant's summary and conclusion

Conclusions:
Per capita loads of test item do not vary significantly in the investigated rivers, neither due to changing seasons of the year nor to different catchment areas. Especially the occurrence of lakes in the catchment area did not have an influence on the measured per capita FWA loads.
Averaged over the year, about 13% of the FWAs consumed in Switzerland are found in the rivers and were attributed to private households. Another 30% are discharged by FWA manufacturing plants directly into the rivers.
The enrichment of FWA in the benthos of Greifensee was investigated; depth profiles from several locations of the lake benthos indicate that FWA are resistant to alteration after deposition.
Executive summary:

In order to evaluate the risk of laundry detergents, it is important to know the fate and behaviour of every individual component in the aquatic environment, therefore three important detergent-FWAs (for which behaviour in the aquatic environment is known) were investigated: DAS 1 (a diaminostilbene), DSBP (a distyrylbiphenyl) and BLS (a bleach stable compound).

An analytical method for the FWAs was developed as a prerequisite for the investigation of the aquatic environment. This method consists of the extraction of FWAs from both solid and aqueous samples with subsequent separation with reversed-phase HPLC.

The method was then used to investigate the behaviour and fate of detergent-derived FWAs in natural waters.

The investigation was divided into three parts:

1. Occurrence and substance fluxes

2. Transformation processes

3. Enrichment in the benthos of a lake.

For investigating the occurrence and the substance fluxes of the FWAs, a monitoring program was conducted in Swiss rivers during one year. Concentrations for the test item measured in the rivers were mostly between 10 and 120 ng/l. On the basis of these measurements, a mass balance was developed, indicating that 13% of the FWAs being used are discharged to surface waters.

The transformation processes of the FWAs in Greifensee, a small lake in Switzerland, were investigated in order to estimate the photodegradation, sorption/sedimentation and flushing rates. On this basis, measured seasonal variations of FWA concentration depth profiles were modelled with a computer simulation. The successful modelling is a confirmation of the estimated processes and shows that photochemical degradation is the most important process for FWAs in Greifensee. However, this process is so slow that FWAs have to be considered as not readily degradable in Greifensee.

The enrichment of FWAs in the benthos of Greifensee was investigated; depth profiles from several locations of the lake benthos indicate that FWAs are resistant to alteration after deposition. The concentrations measured in every particular layer can, therefore, be attributed to the inputs occurring at the time of sediment deposition.