Registration Dossier

Environmental fate & pathways

Adsorption / desorption

Currently viewing:

Administrative data

Link to relevant study record(s)

Description of key information

 The Key study is the highly reliable study by Milinovic et al (2015) using five field-collected soils reported the Kd (Log Kd) for nonylphenol ranged from 24 to 1,059 ml g-1(1.4 to 3.0) and a Koc value of 11,060, with a strong relationship indicated between the soil organic carbon and adsorption. Desorption experiments showed nonylphenol was irreversibly sorbed onto the soils, demonstrating the significant potential for adsorption of nonylphenol to soils.
    
    
    

An array of different approaches have been considered to determine the predicted adsorption of nonylphenol to sludge within WWTP. This has included experimentally derived Koc and Kd values, modelled Koc and Kd values, and the application of those values into two different modelling approaches - SimpleTreat and based on equations detailed in the OECD 106 guidelines. All results indicate that nonylphenol will be strongly adsorbed to sewage and sewage treatment plant solids similarly to the absorption onto soil and sediment in the environment.  

Key value for chemical safety assessment

Koc at 20 °C:
11 060

Additional information

Adsorption in soil and sediments experimental data

Twelve reliable and reliable with restriction (Klimisch 1 and 2) studies are available for the assessment of adsorption/desorption of nonylphenol in soils and wastewater treatment works (WWTW).

A high-quality, Klimisch 1 rated study conducted in accordance with OECD 106 Adsorption Desorption guidelines, batch equilibrium method, conducted by Milinovic et al (2015) is considered the key study. Adsorption and desorption of both nonylphenol and nonylphenol monoethoxylate (NP1EO) was investigated in five soils with a range of physico-chemical characteristics, including organic carbon concentrations of between 0.2 and 9.4%. No losses, as a result of sorption or degradation were observed in the blank and control vessels The Kd (Log Kd) for nonylphenol ranged from 24 to 1,059 ml g-1(1.4 to 3.0) for the five different soils based on the linear model. The Kd (Log Kd) for NP1EO ranged from 51 to 740 ml g-1(1.8 to 3.0). Koc values of 11,060 and 6,880 were determined for nonylphenol and NP1EO respectively, with a strong relationship indicated between the soil organic carbon and adsorption. Desorption experiments were also conducted with linear desorption soil liquid distribution coefficients (KFdes) calculated at 130 to 1,467 ml g-1and 24 to 1,285ml g-1for nonylphenol and NP1EO, which indicated the chemicals were irreversibly sorbed onto the soils with the exception of a high desorption yield (45%) for NP1EO in the soil with the lowest organic carbon. These data demonstrate the significant potential for adsorption of nonylphenol and NP1EO to soils. A summary of the supporting study information is provided below.

Weston (1990) assessed the adsorption characteristics of branched 4-nonylphenol in a guideline study and to GLP. Adsorption was assessed in three soils of light brown/brown gravelly sandy silt or clay at a pH of 6.4 to 7.1, and organic carbon contents of 0.82 %, 10.2 %, and 8.6 %, respectively, in a batch equilibrium experiment. Kd values ranged between approximately 2,300 and 5,200 and calculated Log Koc values were between 4.35 and 5.69 (Koc 23,000 to 489,000). The wide range of Koc values derived indicates that factors other than organic carbon content of the soil, such as specific surface area, ionic exchange capacity and mineral constituents may have been important in determining the extent of the adsorption. Control experiments showed the recovery of nonylphenol from the test vessels (amount in soil plus amount measured in aqueous phase) were lower than expected. This may be due to adsorption of nonylphenol to the walls of the vessel and could lead to an overestimate of the adsorption being calculated. These values are comparable, although slightly higher than the range measured by Milinovic et al (2015).

Another supporting study by Roberts et al (2014) investigated the relationship between the fraction of organic carbon of soil and the equilibrium sorption partitioning coefficient of 4-nonylphenol. Five-point Freundlich isotherms were constructed from equilibrium sorption batch tests with four different soils and organic carbon fractions of 0.1, 2.1, 3.1 and 5.4%. For soils with organic carbon fraction between 2.1 % and 5.4% (excluding the soil classified as sand), the Kd values were calculated between 52 and 1783 l/kg for 4-nonylphenol. The Log Koc values derived from these results range from 2.7 to 4.5 (Koc 600 to 33,000). These values are comparable to the range measured by Milinovic et al (2015).

The finding is further supported by Liet al(2013) in which the adsorption characteristics of nonylphenol was investigated in a soil–water system. Batch equilibrium studies were performed to identify the sorption of nonylphenol. The sorption isotherm was fitted to the Freundlich model. For a soil with an organic matter content of 19.1 g/kg (1.91%) a Kd was calculated to be 1,785 l/kg. The Koc derived from this value is 91,675 (Log Koc of 5.0). These values are comparable to the range measured by Milinovic et al (2015).

A supporting study by Feiet al(2011) included batch adsorption tests with nonylphenol on sediments. The sediments used were artificial sediments amended with rich sediment organic matter (SOM) content which was incubated for more than 120 days to allow diagensis of the sediment to occur. The sediment was sampled every week and new batch sediment adsorption tests were conducted. The results show that while the amount of organic matter loaded in the sediment decreased by nearly 80% during incubation, the adsorption tests with nonylphenol displayed an increase in Koc from 5,000 to 35,000 l/kg (Log Koc of 3.7 to 4.5). Organic matter diagenesis increased the adsorption behaviour of sediment, as the SOM residue has an increasing affinity and partition capacity for organic contaminants. These values are comparable to the range measured by Milinovicet al, 2015.

These results are further supported by the review by Melcer et al (2007) on the adsorption properties of nonylphenol. A range of Koc values were reported with the majority within the range of approximately 10,000 to 50,000 l/kg. Melcer et al (2007) assume that the broad range of Koc values for nonylphenol is likely due to the inherent analytical challenges or differences in experimental procedures.

The key study results are also supported by Dinget al(2014) who considered adsorption to sediment, including the role of sediment biofilms for adhering nonylphenol using batch adsorption experiments. The results indicated a Log Koc of 3.09 (Koc of 1,230) at pH7 at room temperature and a Log Kd of 2.2 (Kd of 172.4).

Adsorption in soil and sediments modelling data

The EU Risk Assessment Report on 4-nonylphenol (branched) and nonylphenol (EURAR 2002) recommends the use of a Koc of 5,360 l/kg based on the European Union System for the Evaluation of Substances (EUSES) model. The members states ran the model was run using the Log Kow of 4.48. There is evidence that the experimental values in the EURAR (2002) are overestimated due to adsorption of nonylphenol to the test vessel. As a result, the EURAR (2002) recommends using an estimated value (EUSES) of 5,360 l/kg, although it is possible that the actual adsorption onto soil and sediment may be higher than the estimated value, possibly due to factors other than organic carbon content being important in the process.

The sorption model for phenols, benzonitriles is set out in the Technical Guidance Document (TGD) of Risk Assessment prepared by the European Chemicals Bureau (ECB), Part III, Chapter 4, p. 26. The domain of the model is phenols, benzonitriles with a log Kow range of 0.5 to 5.5 and was derived based on 24 data points.  The Registrant ran the model using a Log Kow of 5.4 (at 23 °C and pH 5.7; see IUCLID section 4.6 (Spilker, 2009)) the calculated Koc value based on the equation for phenols and benzonitriles set out in the TGD is 14,388:

Log Koc = 0.57*Log Kow + 1.08 =

              = 0.57*5.4+1.08 = 4.158

The corresponding Koc is 14,388.

A Koc of 14,388 (log Koc of 4.16) may be considered representative of typical soils based on measured Kow using the QSAR set out in the TGD since it is within the boundaries of this model. Based on the available literature on adsorption of nonylphenol to soil and sediment (Milinovicet al, 2015; Fei, 2011; Roberts, 2014; Melcer, 2007; and Weston, 1990) the geometric and arithmetic means for Koc (Log Koc) are estimated at 4.19 (15,443) and 4.25 (17,792), respectively.  The overall conclusion based on both experimental data and the calculated value of Koc is that nonylphenol will be strongly adsorbed to soils and sediments. Evidence from experimental studies indicated varied relationships between adsorption and organic carbon content. The key study Milinovicet al,2015 demonstrated a strong relationship between organic carbon content and adsorption, in contrast to the Weston (1990) study which indicated adsorption to soil may be influenced by factors other than organic carbon content, such as the concentration of the test substance

Adsorption in wastewater treatment plants

The adsorption of nonylphenol onto solids in sewers and wastewater treatment plants (WWTP) was also evaluated based on peer reviewed literature. In one report (Claraet al, 2007), several surfactants were monitored in treated and untreated sewage in nine municipal WWTPs in western Austria. The nine sampled WWTPs were representative in size and applied treatment technology to a variety of such plants. The study investigated a range of substances including nonylphenol. Nonylphenol was analysed separately in the liquid phase and in the solid phase and was present in all analysed influents in concentrations between 1 and 35 µg/l. Effluent concentrations were notably lower than the measured influent concentrations. Kd values for nonylphenol varied between 500 and 6,600 l/kg and were within the same range (and higher) as the Kd values of Milinovicet al(2015) at 297 to 1,059 l/kg. As the organic content of solids in the influent was not measured, no determination of Koc was possible.

In another report (Sekelaet al, 1999), 4-nonylphenol was monitored in treated effluent from a municipal WWTP in Vancouver, Canada. Samples were collected with a continuous flow centrifuge to concentrate the solids. The clarified liquid and concentrated solids were analysed for 4-nonylphenol. The concentration of 4-nonylphenol in the solids was reported to be 73µg/g and the concentration in the clarified water to be 2.9µg/L. A Log Koc value of 4.7 was reported, with a calculated Koc of 25,172 l/kg. The organic content of the solids was not reported but has been estimated from the data at approximately to be 50%. The Koc reported here is far higher than that reported by Milinovicet al(2015), but is likely due to the very high organic carbon content in the sludge.

Another study (Luet al, 2015) evaluated the isomer selectivity of 19 nonylphenol isomers in a laboratory-scale continuous flow conventional activated sludge bioreactor under various operational conditions. The concentration of the nonylphenol isomers was analysed in the dissolved phase in the solids and effluent from the bioreactors. From these data, the Kd for the isomers were calculated. The Kd values for the different isomers ranged from 9,500 to 16,000 l/kg. The corresponding Log Koc values were reported from 4.02 to 4.23. These values are again higher than the adsorption coefficients measured by Milinovic et al (2015), but are also likely due to the high organic matter expected in the activated sludge bioreactors compared to soils.

 

A supporting study by Bina et al (2017) considered the adsorption rates of nonylphenol; within the different treatment processes from WWTPs in Iran, including a range of different treatment processes and a range of different sewage inputs. The received sewage types included urban, commercial and hospital sources. Treatment was based on Activated Sludge (AS) and Moving Bed Biofilm Reactor (MBBR) processes. Log Koc and Log Kd values were calculated from the field samples at between 4.15 to 4.63 and 3.55 to 4.03, respectively. Binaet al(2017) also considered the relative importance of nonylphenol biodegradation to compared to adsorption, as a removal process, suggesting biodegradation accounted for greater than 50% and adsorption between 22 to 32%. These values are again higher than the adsorption coefficients measured by Milinovicet al(2015), but are also likely due to the high organic matter expected in the AS and MBBR treatment processes.

Based on the available literature on adsorption of nonylphenol to sewage and WWTP matrices only (Clara, 2007; Lu, 2015; Sekala et al, 1999; Bina et al, 2017), the minimum, maximum, geometric mean and arithmetic mean Log Kd (Kd) are 2.7 (503), 4.4 (25,172), 3.9 (7,140) and 3.6 (3,890), respectively. The minimum, maximum, geometric mean and arithmetic mean Log Koc (Koc) are 4.02 (10,471), 4.7 (50,119), 4.29 (19,712) and 4.3 (19,953), respectively.  These are higher than the Kd and Koc based assessment of adsorption to soil and sediment only.

Considering all data collectively relating to adsorption to soil, sediment and sewage matrices, the minimum, maximum, geometric mean and arithmetic mean Log Kd (Kd) are 1.7 (52), 4.4 (25,172), 3.1 (1,213) and 3.3 (2,139), respectively. The minimum, maximum, geometric mean and arithmetic mean Log Koc (Koc) are 3.09 (1,230), 5.69 (489,000), 4.24 (17,333) and 4.27 (18,790), respectively. 

WWTP nonylphenol removal pathway

Adsorption of chemicals to sludge in WWTP can provide a removal pathway for chemicals from industrial discharges entering the WWTP; this is particularly so for chemicals with highly adsorptive properties such as nonylphenol.  Since this is considered an important removal route for nonylphenol, a sensitivity analysis has been carried out using the data in the literature to determine the expected adsorption rates to sludge matrices within WWTPs. The sensitivity analysis has been carried out based on two different modelling approaches, applying the minimum, maximum, arithmetic and geometric adsorption coefficients from the literature to each approach (Kd and Koc). 

Modelling approach 1: bespoke approach based on the OECD 106 guideline analysis

An evaluation of the data described above from the peer reviewed literature on nonylphenol adsorption to sludges and other WWTP matrices was carried out by applying methods detailed in the OECD 106 guideline for adsorption/desorption and its Appendix OECD 106 equations, specifically equations (4) and (5).  It has not been possible to copy the equations into this dossier, but it is hoped that ECHA can refer to the equations in the Appendix to OECD 106 whilst reading the discussion below.

The calculations do not take into account other processes such as biotic and abiotic degradation, volatilisation etc.

As per the sensitivity analysis, the following parameters have been changed in the equations to understand the potential and relative influence on adsorption of nonylphenol in a WWTP: 

  • The Vo/msoil is used as a surrogate for the Suspended Solids (SS) concentration in the wastewater treatment plant, sometimes referred to as the mixed liquor suspended solids concentration. Two SS concentrations have been considered, 1,000 mg/L and 2,500 mg/L here. A concentration of 1,000 mg/L is considered at the lower end of concentrations expected to occur in WWTP (Hammer and Hammer, 1996) and therefore enables a conservative calculation of the removal mechanism via adsorption to sludge. A concentration of 2,500 mg/L is the concentration required in activated sludge vessel in the OECD 303A Activated Sludge Simulation guidelines (OECD, 2001).
  • a range of Kd values based on those detailed in the peer reviewed literature relating to soils, sewage and sediment [1]. Inclusion of those relating to soil is considered a precautionary approach as the Kd and Koc values determined for the sewage and WWTP process matrices were generally higher than those for soils and sediments.

Applying the range of Kd values (1.7 to 4.4) from the literature, in a WWTP with a mixed liquor SS concentration of 1,000 mg/L, the adsorption to the sludge without accounting for degradation processes was calculated to range from 5 to 96%. The arithmetic and geometric means were calculated at 55 and 68%, respectively. Maintaining the same parameters other than a mixed liquor concentration of 2,500 mg/L, the adsorption to the sludge was calculated to range from 12 to 98%. The arithmetic and geometric means were calculated at 75 and 84%, respectively. Based on these results, the percentage of nonylphenol discharged to the environment in effluent would range from 95 to 4%, with arithmetic and geometric means of 45 and 32% for a treatment plant with a 1,000 mg/L MLSS, or 88 to 2% with arithmetic and geometric means of 25 and 16% discharged for a treatment plant with a 2,500 mg/L MLSS. This range is very wide from almost complete retention to almost no retention, although using only the Kd values determined for sewage matrices the percentage of nonylphenol adsorbed to the sludge is greater. This approach is very simplistic and does not take account of other processes such as biodegradation and volatilisation occurring in the WWTP.

Modelling approach two: SimpleTreat

To provide a more holistic understanding of the behaviour (including adsorption) of nonylphenol in a WWTP, predictions have been made using the model SimpleTreat version 4.0 (http://www.rivm.nl/en/Topics/S/Soil_and_water/SimpleTreat). Changing the parameters in this model further develops the sensitivity analysis for understanding the impact of different variables on nonylphenol adsorption and the scenario which is considered most likely. 

SimpleTreat is a component of EUSES and a standard exposure assessment tool in REACH (ECHA, 2016). This model simulates the fate of trace organic chemicals in a treatment plant to provide estimates of the percentage discharged via the effluent and the sludge. The model requires input of a range of physico-chemical parameters relating to the chemical and an appropriate first order biodegradation rates. Many additional parameters are set as either default or can be amended, where required. The Registrant ran six scenarios to consider the influence of variable adsorption rates on the retention of nonylphenol on the sewage sludge and the discharge of nonylphenol via the effluent. The molecular weight, octanol water coefficient (Kow), vapour pressure, solubility, and the biodegradation rate remained the same across each scenario based on the CSR [2]. Examination of the impact of using the default Henry’s Law Constant (8.7) versus the Henry’s Law Constant as per the CSR (11.02) was made, however this had very little influence on the predicted WWTP retention of nonylphenol in the modelled outputs. The model was run with six different Koc values, including the default SimpleTreat value, the minimum, maximum, arithmetic and geometric means obtained from the peer reviewed literature, in addition to using the Koc value determined solely on the key study by Milinovic et al (2015). Again, the Koc values used included all data from the literature relating to sediments, soil and sewage matrices [3]. The model provides a total percentage elimination of nonylphenol from the wastewater, which includes elimination at three treatment stages (the primary settler, the aerator and the solids liquid separator) as a result of volatilisation, adsorption, stripping and biodegradation. The model also provides a percentage emission to the effluent.

The scenario modelling the lowest Koc (minimum Koc) determined in the literature related to adsorption to a sediment and a calculated Koc of 1,230 l/kg (Ding et al, 2015). The total percentage nonylphenol eliminated from the wastewater was calculated at 64.13 %, with 35.87% released in the effluent. Of the total eliminated from the wastewater approximately 11.7% is predicted adsorbed to the sludge, 43% biodegraded and 9% volatilised or stripped.

The results for each of the other scenarios (geometric mean Koc, arithmetric mean Koc, maximum Koc, default SimpleTreat Koc and based on the key study Koc by Milinovicet al, 2015) provided predictions of greater than 78.9% total elimination from the wastewater with some as great as 97% elimination. The maximum of predicted nonylphenol discharged in the effluent was 21%. Between 60 and 95% of the eliminated nonylphenol is predicted to be removed via the sludge (adsorbed) with the remainder accounted for via biodegradation and volatilisation. Considering the result from the key study alone, based on a Koc of 11,060, the total elimination from the wastewater was predicted to be approximately 79%, of which approximately 50% adsorbed to the sludge, 24% was biodegraded and about 5% lost to volatilisation and stripping. Approximately 20% was predicted to be released via the effluent. This scenario is considered conservative as this is based on a Koc value determined using soils which generally have a lower organic matter component than activated sewage sludge. Organic matter has been indicated in various studies to greatly influence adsorption (Milinovic et al, 2015).

The results of the sensitivity analysis (based on both the simple modelling approach as per the OECD 106 guidelines and the SimpleTreat model) indicate the adsorption coefficient is the factor with the greatest influence on adsorption of nonylphenol in WWTPs. The scenario indicating the greatest release of nonylphenol via effluent (hence low adsorption) relate to a low Kd or Koc value (1.7 and 1230). Modelling of this scenario using the OECD 106 approach, based on a Log Kd of 1.7, resulted in prediction of between 88 to 95% discharge to the environment. Modelling of a similar scenario using SimpleTreat, based on a Koc of 1230, predicted a worst case of approximately 36.7% discharge to the environment. The SimpleTreat model includes adsorption, biodegradation and volatilisation in the treatment processes. For low adsorption coefficient scenario, elimination of nonylphenol via biodegradation was far more important (approximately 43%) than in scenarios with higher adsorption coefficients. Use of OECD 106 model does not incorporate biodegradation, which if relied on for the low adsorption scenarios would lead to an overestimation of the likely discharge to a receiving environment. This is less of an issue for the scenarios modelling higher adsorption coefficients, where biodegradation is less of a driver. These low adsorption coefficients are not considered realistic of WWTPs where the concentration of organic matter is high and will facilitate adsorption.

The results of the SimpleTreat modelling indicate that at a worst case approximately 36% will be discharged via the effluent into a receiving body, although a more realistic scenario indicates around 20%. Some of the nonylphenol will be degraded but most will be adsorbed to the sludge (60 to 95%) and not enter the downstream aquatic environment.

Overall conclusion on the adsorption potential of nonylphenol in WWTW and soils

The highly reliable study by Milinovic et al (2015) using five field-collected soils reported the Kd (Log Kd) for nonylphenol ranged from 24 to 1,059 ml g-1(1.4 to 3.0) and a Koc value of 11,060, with a strong relationship indicated between the soil organic carbon and adsorption. Desorption experiments showed nonylphenol was irreversibly sorbed onto the soils, demonstrating the significant potential for adsorption of nonylphenol to soils.

An array of different approaches have been considered to determine the predicted adsorption of nonylphenol to sludge within WWTP. This has included experimentally derived Koc and Kd values, modelled Koc and Kd values, and the application of those values into two different modelling approaches - SimpleTreat and based on equations detailed in the OECD 106 guidelines. All results indicate that nonylphenol will be strongly adsorbed to sewage and sewage treatment plant solids similarly to the absorption onto soil and sediment in the environment. 

Assessment approach

Predicted release to water (%)

Predicted release to sludge (%)

Predicted degradation (%)

Released to air (%)

Total elimination from wastewater (%)

OECD 106 approachab

 

 

Minimum Log Kd

88-95

5-12

Not determined

5-12

Maximum Log Kd

2-4

96-98

96-98

Geometric mean Kd

16-32

68-84

68-84

Arithmetric mean Kd

25-45

55-75

55-75

Key study, Milinovic et al (2015) (based on high and low Log Kd)

29-66

44-71

44-71

SimpleTreat approach

 

 

Minimum Log Koc

35.87

11.71

43.4

9.02

64.13

Maximum Log Koc

2.95

95.34

1.43

0.27

97.05

Geometric mean Koc

16.59

60.56

19.0

3.84

83.41

Arithmetric mean Koc

16.45

62.49

18.51

2.56

83.55

Key study, Milinovic et al (2015)

20.36

50.98

23.81

4.85

79.64

 

CSR modelling

 

 

CSR values

25.8

55.53

15.73

2.87

74.2

Notes:

a – the OECD approach does not account for degradation or volatilisation processes.

b – the ranges are based on a SS concentration of 1 and 2.5 g/L

Footnotes

[1]: Kd values for soils with organic carbon contents of <0.2% are not considered representative of sewage sludge characteristics and have been excluded from the calculations (Roberts, 2014 and Milinovicet al, 2015)

[2]:Molecular weight – 220.351 g/M

Octanol water coefficient (Kow) – 251189, based on a Log Kow of 5.4 (Spilker, 2009)

Vapour pressure – 0.3 Pa (Staley, 2009)

Water solubility – 5.7 mg/L at 25˚C, pH 6-7 (Spilker, 2009)

Biodegradation rate – 0.1 hr-1 rate constant (based on classification as inherently biodegradable)

 [3]: Kd values for soils with organic carbon contents of <0.2% are not considered representative of sewage sludge characteristics and have been excluded from the calculations (Roberts, 2014 and Milinovic et al, 2015)

  

References:

ECHA (2016). Guidance on information requirements and Chemical Safety Assessment Chapter R.16: Environmental exposure assessment. Version 3.0 February 2016.

Hammer, MJ and Hammer MJ Jr. 1996. Water and Wastewater Technology, 3rdEd. Prentice Hall, Ohio.

OECD, 2001. OECD Guidelines for the testing of chemicals, Section 3. Test No, 303: Simulation Test – Aerobic Sewage Treatments -A: Activated Sludge Units; B: Biofilms.

Spiker 2009. Property Measured/Assessed: Partition coefficient n-octanol/water. Study report number 0648 08 00090.

Staley, C. 2009. Property Measured/Assessed: 7.5 Vapour Pressure. Study report number 1030-2009110108-003A.

[LogKoc: 4.04]