Passive sampling and analysis of cyclic siloxanes. Chris Sparham 27 May 2009 NORMAN expert group meeting

Transcription

1 Passive samplin and analysis of cyclic siloxanes Chris Sparham 27 May 2009 NORMAN expert roup meetin

2 Safety & Environmental Assurance Centre Addresses Consumer, Occupational and Environmental safety and supports Sustainability

3 Aims of presentation Properties of siloxanes n Why is Unilever interested in these materials? Blank issues quality control Water monitorin Sediment monitorin Why identify passive samplin approach Passive samplin n Choice limitations n Desin n Initial work

4 Cyclic Volatile Methylsiloxanes (cvms) Cyclic Siloxanes n D4, D5, D6 Clear, volatile, low molecular weiht liquids with very low water solubility Used in the manufacture of larer siloxane polymers and as base fluid or frarance carrier for a rane of consumer applications includin - n Personal care (skin, hair, deodorancy) n Household care products Beneficial physical and chemical properties n very low mammalian toxicity (D5) n adequate evaporation rate n low surface tension n no odour

5 Decamethylcyclopentasiloxane (D5) in products

6 Properties of D5 relevant to analysis Low solubility in water n 17 µ/l at 23 C O Si O Si Hih Henry s law constant Si O n Lo K AW ~ 3 Hih Lo P n Lo K ow = 8.0 O Si O Si Chemical Formula: C 10 H 30 O 5 Si 5 Exact Mass:

7 Monitorin data Environmental monitorin data are required to compare with modellin/theoretical predictions about fate of D5 In particular it s partitionin behaviour between environmental compartments n Therefore, analysis in surface water n Analysis in sediment n Lack of data on the local scale i.e. UK rivers Post-doctoral projects n Michael McLachlan and colleaues at ITM Stockholm (time trends, spatial distribution and bioaccumulation) n Passive samplin (Portsmouth University)

8 Blank issues Source of D5 contamination Septum GC column Lab Air HS Vial Caps Personal Care Products (Analysts) Solution Low bleed Septa used and injector temperature reduced Also on-column preferred, septumless head or Merlin PEG based (non-siloxane) columns used Use of clean air enclosure Lab air ( 0.35 µ/m 3 D5) Enclosure air ( 0.04 µ/m 3 D5) Critical for low level water and sediment analysis Butyl not silicone Do not wear!

9 Last year

10 This year

11 GC/MS equipment Combi/Pal auto sampler/6890 GC/5975 MS

12 Surface water analysis Headspace vials placed on Gerstel multi purpose sampler (MPS2) Aitated and heated at 80 C for 10 min 1 ml aliquots sampled usin a as tiht syrine Separation: Ailent 6890 GC split/splitless injector (heated to 150 C) 10:1 Column flow: 1 ml/min with helium carrier as Column: 30 m x 0.25 mm J & W DB-Wax column with film thickness 0.5 μm Oven Conditions: 40 C for 4 min, heat to 150 C at 8 C/min, held for 5 min Detection: Ailent 5973 MSD operated in electron impact (EI) mode & sinle ion monitorin (SIM) with ions m/z 73, 267 and 355 (D5)and m/z 75, 270 and 360 ( 13 C 5 - D5 internal standard) Calibration: Curve constructed (rane n/l) usin internal standards Recovery of mass-labelled spikes: Typically 90 % for all the samples analysed. Recoveries decreased when levels of DOC and SS were hih LOD: Currently 3 to 6 n/l based on backround levels of D5

13 River Great Ouse samplin sites

14 Samplin and Analysis of Water Samples taken in the field usin a sub-surface sampler usin a 1 Litre lass bottle.

15 Analysis in field Water samples prepared for analysis in headspace vials out in the field (includin addition of internal standard). Microlitre syrines used for spikin with the needle inserted below liquid surface.

16 River Great Ouse (water) Series1 D5 concentration (n L -1 ) LOQ Tyrinham Bride (n = 3) Felmersham (n = 2) Sharnbrook Mill (n = 1) Priory Park (n = 3) Great Barford Bride (n = 1)

17 Headspace GC/MS

18 Sediment samplin old rab Samples collected usin stainless steel rab samplin device

19 Sediment samplin- heavier rab

20 Samplin and analysis of sediment Surface sediment passed throuh 2 mm stainless steel sieve into acetone-rinsed 1 L jars (PTFE-lined caps)

21 Samplin and analysis of sediment Transfer to centrifue tubes in the field

22 Accelerated solvent extraction (ASE)

23 Sediment analysis Centrifued and triplicate 2.5 ± 0.02 aliquots dried with hydromatrix (~ 6 ) Spike ASE cell (33 ml) with internal standard ( 13 C 5 -D5) 100 C, 1500 psi, heat 5 min, static 5 min, flush vol 50 % (33 ml cell), pure time 150 s Extract once usin ethyl acetate (~ 50 ml) On-column (OC) injection (5 µl) onto fused silica pre-column (10 m, 0.53 mm) Zebron ZB 5HT INFERNO 30 m x 0.25 mm (0.25 µm film) Calibration 0 to 20 n/ml (LOQ 0.1 n/ml) n D5 m/z 73, 267,355; IS m/z 360 (SIM)

24 Sediment data D5 concentration (µ k-1) D5 (µ -1 0C) Great Barford:1 (n=3) Priory Park:3 (n=5) Priory Park:2 (n=5) Priory Park:1 (n=3) Sharnbrook Mill:3 (n=9) Sharnbrook Mill:2 (n=9) Sharnbrook Mill:1 (n=9) Felmersham:3 (n=9) Felmersham:2 (n=9) Felmersham:1 (n=9) Tyrinham Bride:3 (n=3) Tyrinham Bride:2 (n=3) Tyrinham Bride:1 (n=3) D5 concentration (µ k -1 dw) D5 ( µ -1 OC)

25 Quality Control D5 Field Blank (Sanford Lake) IS recovery ( 13 C 5 -D5) a 8 30 μ/k, dw (typically < 10 μ/k, dw) % (n = 64) Full procedural (from sievin on river bank To analysis by GC/MS Spiked ASE cell compared to calibration standards (averaed as appropriate) Typical native D5 recovery (spike 548 μ/k, dw) 89 ± 7 (n = 3) Spiked in sediment and left to contact a Data obtained in the rane 37 to 114 µ/k, dw prior to clean air cabinet

26 Conclusions River Great Ouse All sediment results reported without blank correction n Sample concentrations > 3 x field blank (LOQ) Rane of D5 concentrations in sediment = 75 to 1448 µ/k Rane of D5 concentrations in water = 7 to 57 n/l Relationship between D5 present and the amount of OC (~ 20 μ -1 OC typical) Consider Total accessible fraction of D5 in sediment

27 Practical approaches for estimatin bioavailability Earliest approach by JJ Pinatello examined the reversible sorption of aliphatic halocarbons in soil n Desorption by repetitive batch extraction (water extraction) n Desorption with continuous removal of contaminant from the aqueous phase (e.. Tenax extraction) More recent investiators have compared Tenax extraction with pore-water concentrations estimated usin SPME

28 Batch desorption (water) Sediment containin D5 886 µ/k, dw (solids 30%, OC 3.2%) 20 wet weiht in centrifue tube (foil covered cap) ~70 ml of 0.01 M CaCl 2 (no headspace) Place on roller at 20 C for 24 h Centrifue at 3000 rpm (20 min) n Or until supernatant is clear Analyse D5 concentration in supernatant usin HS GC/MS n Remove all supernatant and replace 0.01 M CaCl 2 n Analyse for DOC as required Repeat each day

29 Batch desorption (water) - practical

30 Batch desorption (water) - plots Cumulative D 5 desorbed from sediment (n) D5 desorption from sediment Time days FE1 FE2 FE3 Initially 5351 n D5, < 4% desorbed (from modelled value, 205 d)

31 K oc apparent After 24 days repeated extraction < 3.5 % of D5 in the sediment was desorbed into the aqueous phase. After equilibration (> 9 days) C water (n/l) was determined (Dissolved Oranic Carbon < 1 m/l) At this point D5 measurements were freely dissolved In the last 4 days of the test the lo Koc values raned from Total accessible fraction of D5 very low

32 K oc apparent calculations

33 Why identify passive samplin? Time-weihted averae (TWA), freely dissolved concentrations in surface water n Lower detection limit possible 4 Lake with no direct effluent input, only atmospheric deposition 4 Urbanised catchment and estuary (sea water) Potential use in sediment to establish freely dissolved pore water concentration

34 Collaboration with Portsmouth University Research team led by Graham Mills, Richard Greenwood n Post doc Enamul Hoque

35 Passive Samplin Concentrations of dissolved D5 in environmental waters is often low, near or below the detection limit of bottle samplin combined with GC-MS techniques (headspace) One way forward is to increase the mass available for analysis by usin passive samplers A rane of passive samplin devices is available, but not all are suitable for this application because of the unusual properties (hih lok OW, and hih volatility) of this compound

36 Passive Samplin: types SPMDs are not suitable because of the need to extract the analyte from the triolein receivin phase and clean up and concentrate the extract n This would lead to losses Chemcatcher, membrane-enclosed sorptive coatin sampler (MESCO) are not suitable because of low uptake rates for compounds as non-polar as D5 Sheet samplers without diffusion limitin membrane n Silicone rubber sheets have hih samplin rates but contain D5 monomers, and polymers that break down to produce D5 durin handlin (particularly at hiher temperatures used in cleanin before deployment) and so it is difficult to obtain low blank levels n This can limit level of detection

37 Passive samplin: types Polyoxymethylene (POM) is a more crystalline plastic, and whilst havin low contamination with D5 has slower uptake rates than are available with the silicone rubber sheets Low density polyethylene (LDPE) provides a compromise, and this was selected as a startin point for the development of a passive sampler at Portsmouth

38 LDPE Passive Sampler LDPE sheet samplers have been used for compounds (e.. lare PCBs and PAHs) of similar polarity, but lower Henry s constant This material is available in a rane of thicknesses and in lare sheets and so can be cut to provide the necessary area for samplin, and shape for convenient handlin Can be readily extracted in a rane of solvents

39 LDPE Passive Sampler LDPE sheets are enerally contaminated with D5 from either contact durin manufacture or uptake from the air durin storae Cleanin is necessary prior to use n Thermal desorption n Solvent extraction steps

40 LDPE: PRCs Potential for loadin performance reference compounds (PRCs) onto LDPE prior to deployment to provide an in situ calibration, thus overcomin problems associated with the impact of variation in turbulence and temperature durin deployment Loadin usin solutions of D5 and PRCs in Methanol : water (50 : 50, v/v) Extraction in solvents n MeOH compatible with headspace analysis n Hexane direct injection (lare volume)

41 Overall conclusions GC/MS methods described suitable for measurin total D5 concentrations in R Great Ouse (UK) surface water and sediment n Good control of blanks (clean air cabinet) Total accessible fraction in sediment very low n Repetitive batch extraction from water (< 3.5% of D5 desorbed in 24 days) n LoK OC determined in the rane 5.5 to 5.9 Passive samplin usin LDPE identified n LOQ 0.1 to 1 n/l (taret) n Enable monitorin below current LOQ

42 Acknowledements Project leader Roer van Emond (Environment) Analytical Team members n Colin Hastie n Sean O Connor n Nicola Bettles Environment team n Dave Gore Severn Trent Labs (Rakesh Kanda) Dow Cornin