session i contaminated sediments risk …...grain shapes, elastic and heterogenous. low dry density...
TRANSCRIPT
Session I
Contaminated Sediments
– Risk Assessment and
Remediation
Part I
Chair: Peter Harms-Ringdahl
M.Sc., Environmental Consultant, EnviFix, Sweden
TREASURE: Novel and innovative methods used to characterize
two of Sweden’s contaminated fiberbank sediment sites
Ian Snowball
Professor, Uppsala University, Sweden
Remediation of Oskarshamn hamn:
challenges and lessons learned from contractor point of view
Bart van Renterghem
Regional Manager, Envisan, Belgium
Quantification of microplastics in sediments from benthic and coastal environments
Heidi Knutsen
M.Sc., NGI, Norway
Nontarget analysis of sediment samples from Copenhagen, Denmark
Josephine Lübeck
Ph.D.-student, University of Copenhagen, Denmark
TREASURE
Novel and innovative methods used to characterize two of Sweden’s contaminated fiberbank sediment sites
Ian Snowball, Prof., Natural Resources and Sustainable Development Uppsala University, Sweden
Anna Apler, PhD student, Geological Survey of Sweden
Anna-Karin Dahlberg, Dr, Swedish University of Agricultural Sciences
Paul Frogner-Kockum, Dr, Swedish Geotechnical Institute
Gunnel Göransson, Dr, Swedish Geotechnical Institute
Per Hall, Prof., University of Gothenburg, Sweden
Sarah Josefsson, Dr, Geological Survey of Sweden
Mikhail Kononets, University of Gothenburg, Sweden
Achim Kopf, Prof., University of Bremen
Hjördis Löfröth, Dr, Swedish Geotechnical Institute
Catherine Paul, Dr, Lund University, Sweden
Matt O’Regan, Dr, Stockholm University, Sweden
Karin Wiberg, Prof., Swedish University of Agricultural Sciences
Lovisa Zillén, Dr, Geological Survey of Sweden
An August morning, near Kramfors in Ångermanälven (Photo taken by I. Snowball, 2015)
NORDROCS 2018 – Helsingör 3-6 September
Sweden’s paper and pulp industry in the distant past
A plume of suspended, contaminant-laden woody materials.
100 years use of metals in processing.
Persistent organic pollutants (e.g. PCBs, DDT) used 1930-1970’s.
Ångermanälven estuary
2
SGU surveyed ~212 km2 of seafloor along the coast of Norrland (Apler et al. 2014, Norrlin et al. 2016, Larsson et al. 2017)
A total area 2.6 km2 is covered with about 45 banks of pure wood ”chips” and cellulose fibers, and some timber.
Designated as FIBERBANKS
Larger areas of the seafloor (~27 km2) also contaminated by this waste is designated as FIBER-RICH sediment.
Fiberbanks and Fiber-rich sediment = fibrous sediments See poster by Norrlin et al. Abstract pp220-221.
High BOD Anoxic Sulphide oxidation by Beggiatoa sp. Metals and POPs
3
TOP
BOTTOM
One cannot “see” how thick these deposits are with normal geophysical methods (e.g. seismics) due to gas.
An example of what a core of fiberbank can look like
4
Natural sediment
Fiberbank
TREASURE (2015-2018) Targeting Emerging Contaminated Sediments Along The Uplifting
Northern Baltic Coast of Sweden For Remediation
Financed by the Swedish Research Council FORMAS, as part of the SGI TUFFO* initiative to
expedite the remediation of contaminated areas.
*Teknikutveckling & Forskning inom Förorenade Områden
6 components
1. Coordination 2. Fieldwork 3. Geotechnical 4. Chemical/biological 5. Risk assessment 6. Recommendations
for management.
Sandviken 1928
Väja 2015
“3D” Images by Jim Hedfors (SGI), using SGU data.
TREASURE study sites
6
TREASURE methods
For geotechnical results see poster by Löfroth. Abstract pp167-168.
FF-CPTu (free-fall dynamic cone penetrator)
Unique to MARUM. Measures geotechnical parameters, rapidly. Identification of layers in-situ.
Combined with hydro-acoustic surveys and ground truthing obtained from cores of sediment to estimate volumes of ”fiberbanks” and some strength parameters.
Hydro-acoustic, FF-CPTu and coring methods
Practical limitations due to low density and thick fiberbanks studied
SGU Corer (6 m) MARUM FF-CPTu
Sub-seafloor hydro-acoustic signal
Bad Good
Core & CPTu data
Bad
Good
TREASURE fiberbanks are a geotechnical challenge See posters by Löfroth et al. Abstract pp167-168,
Lehoux et al. Abstract pp223-224.
Physical properties contrast with mineral-rich sediments that most geotechnical methods were designed for. Abnormal grain shapes, elastic and heterogenous.
Low dry density (500 kg m3) High porosity (80%)
Low shear strength (0.7-4 kPa) Shear strength increases only marginally
with reduced water content.
9
Väja
SGI “tipster”
Q (Sandviken): 250 000 m3 fiberbank (max 8 m thick) area 48 000 m2
R (Väja): 200 000 m3 fiberbank (max 12 m thick) area 94 000 m2
Thicknesses and volumes have not been systematically estimated. Still largely ”guesstimates”
But, our 2 study sites have large volumes and are thick
10
TREASURE has taught us that better methods for estimating areas and volumes are needed. There are 45 known sites.
Fibrous sediments, biota and bioaccumulation of POPs Presented by Josefsson et al. Abstract pp90-91.
Benthic biota not found at fiberbanks, only bacteria.
The invasive polychaete Marenzelleria spp. is common in fiber-rich sediments.
The predatory crustacean Saduria entomon is also present.
Bioaccumulation of POPs, and evidence of biomagnification
11
Benthic lander (Gothenburg University & SGI)
Benthic lander (BFC) See poster by Paul Frogner-Kockum. Abstract pp227-228
µmol/m2/d
Measured flux of metals higher than diffusions models predict from
concentrations
But for some metals that exist in high concentrations (e.g. Pb), no fluxes were detected.
These observations imply that risk assessment should not be based solely on total sediment concentrations.
Ideally complemented by flux measurements.
Idea: bacteria are present that are performing natural remediation.
Mercury resistance merA PAH degradation Gene for α-subunit of PAH degrading
oxygenase
Naphthalene degradation nahAc PCB degradation bphA
We know the genes that some bacteria use to transform chemicals found in the fiberbanks. We search specifically for those genes to see if natural remediation is occuring. We look to see if these bacteria are in the fiberbanks.
Genes (Catherine Paul, Lund University)
Bla
nk
11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
FRS FB FB FRS S FRS FRS FB FRS FRS FB S FR
S
FB S
Detection of the gene nahAc by polymerase chain reaction (PCR)
Gene as a dark band in the DNA
Väja Kramfors Sv Krf Sandviken
This gene is found in bacteria that can break napthalene, a small polycyclic hydrocarbon (PCH). We only detect it in FB and FRS samples, associating it with natural
breakdown of hydrocarbons in these three fiberbanks.
What have we learnt so far?
1. Low density, high organic and gas contents. Fiberbank sediments” are atypical and challenging to physically characterize.
2. Sites studied are biochemically very active: the microbial induced production of gas (~methane, carbon dioxide, hydrogen sulphide) is ”enormous”.
3. Fauna living in nearby fiber-rich sediments are bioaccumulating the persistant organic pollutants (POPs), including the ”legacy” banned substances.
4. While metals are bound to the organics in the anoxic environment there is potential for release if oxidation occurs.
TREASURE (2015-2018) Targeting Emerging Contaminated Sediments Along The Uplifting
Northern Baltic Coast of Sweden For Remediation
TREASURE (2015-2018) Risk of sediment and contaminant transport
See poster by Göransson et al. Abstract pp209-210
A probabilistic approach to assess the likelihood for each of
the dispersion processes to occur within a given time
frame.
For risk assessment and/or risk
management method.
TREASURE (2015-2018) Targeting Emerging Contaminated Sediments Along The Uplifting
Northern Baltic Coast of Sweden For Remediation
I. Snowball (UU), H. Löfröth (SGI), L. Zillén (SGU), S. Josefsson (SGU), G. Göransson (SGI), P. Frogner-Kockum (SGI), A. Svensson (SGU),
M. Holmen (SGI), M. O’Regan (SU), K. Wiberg, SLU), A. Apler (SGU/UU). Missing: C. Paul (LU), A. Kopf MARUM).
Licentiat thesis (2018) by Anna Apler, SGU
Overskrift her Navn på oplægsholder Navn på KU-enhed
For at ændre ”Enhedens navn” og ”Sted og dato”: Klik i menulinjen, vælg ”Indsæt” > ”Sidehoved / Sidefod”. Indføj ”Sted og dato” i feltet for dato og ”Enhedens navn” i Sidefod
Josephine Lübeck, Guilherme L. Alexandrino and Jan H. Christensen
Analytical Chemistry
Department of Plant and Environmental Sciences
University of Copenhagen
Denmark
Non-target GC×GC-HR-MS analysis of sediment samples from Copenhagen, Denmark
UNIVERSITY OF COPENHAGEN DEPARTMENT OF PLANT AND ENVIRONMENTAL SCIENCES
DEPARTMENT OF COMPUTER SCIENCE
GANDALF project
Non-target Fingerprinting Analysis and GIS Visualisation of Contaminants
A New Paradigm for Chemical Impact Assessment in Urban Development
Comprehensive characterisation of contamination
Source identification and apportionment within urban environments
2
Organic contaminants Elemental species
GC(×GC)-MS ICP-MS
(GPC×)-SFC-MS LC-ICP-MS
LC-IMS-MS SFC-ICP-MS
ChemFing
11/09/2018 3
Many samples Few variables
Few samples Many variables
Level 3: Non-target analysis with advanced analytical tools
Level 1: Target analysis & preliminary classification
Iterative sampling
Iterative sampling
Level 2: Source allocation based on hydrocarbon fingerprint
GANDALF project
4
Sampling site
Utterslev Mose (UttM)
Fortress channel (FSK)
5
Sampling
Composite samples (5 increments)
Single samples
In-depth samples
Oxic vs. anoxic samples
0-30 cm
GANDALF project
Sample preparation
6
1. Polar 2. Nonpolar
Temperature [°C] 50 100
Pressure [psi] 1,500 1,500
Solvent MeOH:H2O (1:1) DCM:acetone (3:1)
Cycles 2 2
Pre-heat time [min]
2 2
Static time [min] 10 10
Flush volume [%] 70 70
Purge time [sec] 60 60
Selective pressurised liquid extraction (sPLE)
Chemical analysis
7
Parameters Values
Column setup 1D: ZB-5 (60 m) 2D: ZB-50 (1.5 m)
Split mode 1:5
Initial T [°C] 60
Ramp [°C/min] 7.5
Max T [°C] 310 (hold: 15 min)
Flow [mL/min] 1.5
Injection volume [µL] 1
Modulation time [s] 6
Initial T (Hot jet) [°C] 120
Max T (Hot jet) [°C] 350
Initial T (2nd oven) [°C] 70
Max T (2nd oven) [°C] 320
RIS MSD [°C] 315
ZX2 Cooled Loop Modulation GANDALF project
GC×GC-TOFMS chromatogram
8
PCBs
>2,000 compounds detected
122 NIST library hits >90% using ”Unknown analysis” from Agilent
FSK-8S0-2
Alkenes/
Cycloalkenes
Alkanes p-Cresol
2-Chloronaphthalene
Fluoranthene
Pyrene
Dibenzofuran
C3-DBT
C2-DBT
C1-DBT
Naphthalene
Anthracene
C1-Nap
C2-Nap
11/09/2018 9
Two exemplary GC×GC plots (same scale): FSK 2-b (upper) and UttM 9B (lower).
DEHP
FSK-2b
UttM-9B
Large concentration
differences
GC×GC-TOFMS chromatograms
The pixel-based approach
10
*.CDF files
Raw files
2D (rt1,rt2)
Phase-correction
Decrease column bleeding Remove sulphur ions
- m/z 207.05, 207.10,
281.05, 281.20, 133.00, ... - m/z 32.065, 64.130,
96.195, 128.260, …
before rt alignment
after rt alignment
w = 1/RSD(QCs)
X(rt1,rt2) .*w
Downscale pixels highly affected by noise
DATA MODELING WITH PCA
PC model: Map of samples
11/09/2018 11
UttM
Large concentration
differences
Chemical interpretation PC2 loading plot. Closure effect and contamination level
11/09/2018 12
PCBs
Column bleeding
One ring aromatics
Naphthalenes
Alkanes
4-tert-Butylphenol
Sesquiterpenes
BHT
2,4-di-tert-pentylphenol
PC model: Map of samples - Interpretation
11/09/2018 13
Increasing gasoline
input
UttM
PC model: Map of samples - Interpretation
11/09/2018 14
UttM
Increasing natural input
Close to channel/higher anthropogenic input
15
The pixel-based approach is:
• The human brain’s way to perform non-target analysis.
• Excellent to obtain an overview of sample similarities and differences and for profiling analysis (used since 2005).
• A clever way of data handling of very large datasets.
Non-target contaminant fingerprinting in sediments:
• Very fresh results with room for improvement
• Preconcentrate the Utterslev Mose samples
• Remove elemental sulphur during data analysis
• Improve pre-processing
• Identification work (e.g. with NIST; homologous series)
• To be combined with LC-, SFC- and LC-ICP-MS data
Take home messages
11/09/2018 16
GANDALF project
11/09/2018 17
GANDALF project
Geostatistics of
public data on urban
soil & sediment
contamination in the
City of Copenhagen
Nemanja Milosevic et al.
Poster at NORDROCS
Quantification of microplastic in
sediments from benthic and
coastal environments Hans Peter H. Arp1,2, Heidi Knutsen1, Emma Jane Wade1, Linn Merethe Brekke Olsen1, Sabnam Mahat1,3 and Øyvind Lilleeng1,3 Contact: [email protected] 1NGI, Oslo; 2NTNU, Trondheim; 3NMBU, Ås, Norway
Helsingør, Denmark September 5th
Where do you find the most plastic in the ocean?
Cozar et al. PNAS 2014; Eriksen et al. Plos One 2014; Eunomia, Plastics in the Marine Environment, 2016
150 million tons
0
50
100
150
TotalEmitted
Floating Beaches Seafloor
Mill
ion
s o
f to
ns
Estimated amounts of plastic in the ocean
0.27 million tons
1.4 million tons
25 - 65 million tons Microplastic concentrations
Hidalgo-Ruz et al. ES&T 2012
What causes plastic to sink?
3 Linn Merethe Brekke Olsen
1. Sinking due to high density
Photo: seegraswiese commons.wikemedia
Floaters: LDPE, HDPE, PP, EVA, etc.
Sinkers: PET, PVC, Nylon, Polyacrylate, Polycarbonate, Polyacetate, etc.
https://phys.org/news/2016-07-solution-plastic-continents.html www.dailymail.co.uk/news/article-4690526/The-bottom-ocean-filled-plastic-bottles.html
2. Sinking by organism uptake / faecal pellet
Cole et al. Environ. Sci. Technol. 2016, 50, 3239−3246
3. Sinking assisted by photo-oxidation and mechanical
stress
δ=0
UV
O2
δ-
δ- δ-
Changes: surface area, negative charges on the surface number of fragments
δ-
UV exposed PS particle
4. Sinking by biofilm growth
4mm LDPE granule with little biofilm
4mm LDPE granule with biofilm
Rummel et al., Environ. Sci. Technol. Lett., 2017, 4(7), pp 258-267
5. Sinking by aggregation – marine snow
4 mm LDPE and other granules aggregated with biofilm
Methods for quantification of microplastic in
sediment Common steps
1. Density separation for removal of minerals
2. Filtering
3. Chemical digestion to remove organic material
4. Quantification: weight, number of particles, polymer type etc.
All steps might contaminate the sample and affect the size-distribution of the particles
NGIs method – Step 1: Density separation
Bauta Microplastic-Sediment Separator
(BMSS)
Inspired by the Münich Microplastic Sediment Separator (Imhof et al.)
ZnCl2:CaCl2-solution: ρ ≥ 1,52 g/mL:
4.4: 3.6: 2 kg (ZnCl2:CaCl2:H2O) (Hudgins, C. M., 1964)
NGIs method – Step 1: Density separation
ZnCl2:CaCl2 solution (ρ ≥ 1,52 kg/L)
Plastic, charcoal and organic matter with ρ < 1,52 float, other materials with ρ > 1,52 sink
NGIs method – Step 2: Filtering
Wait 24 h
Attach separation chamber
Filter
Steel mesh filter (45 µm)
NGIs method – Step 2: Filtering
Directly from Bauta to a steel filter «tea bag»
Steel filter (45 µm)
NGIs method – Step 3: Chemical oxidation
2-step chemical oxidation – Enzyme free procedure. Method soon published (Olsen et al., in prep.)
Exothermic reaction Removes 98 ± 4 % cellulose, up to 100 % of chitin, 0% plastic pellets and 0-4 % plastic fibre
NGIs method – Step 3: Chemical oxidation
.
Before After
Note: Plastic is not the
only material with a low
density that can survive
this procedure.
Notably charcoal,
porous glass/ceramic &
sea shells can survive.
NGIs method – Step 4: Visual microscopy
Visual microscopy to describe shape, colour and size group
A B
D C
A: overview of processed sample B: representative picture of sample showing many granules (< 500 µm) C: observed, red granule (>1000 µm) among smaller granules (<500 µm) D: white fibre < 1000 µm
http://www.miljodirektoratet.no/Documents/publikasjoner/M976/M976.pdf
NGIs method – Step 5: FT-IR microscopy
FT-IR microscopy to determine polymer type and size
NGIs method - Quality assurance
Method blanks: follow all steps
as the regular analysis, but with
no sediment.
Important to correct for plastics
appearing in the samples from:
─ The lab and lab clothes
─ Equipment
─ Chemicals and salt solution
A
D C
B
45 µm
NGIs method - Quality assurance
Spiked blanks: adding a known amount of microplastic – fibres, granules and/or powders - to a «clean» sediment (sediment remaining in Bauta after density separation), followed by quantification.
Give information about the precision.
Recovery correction factor
Results: Microplastic in sediments on the Norwegian Continental Shelf
Maximum weight of MP particles – collaboration with DNV GL
20
Northern North Sea (30 ± 40 mg mMP/kg)
Barents Sea (30 ± 20 mg mMP/kg)
Central North Sea (90 ± 100 mg mMP/kg).
mMP/kg = max.
microplast / kg dw,
other particles with
a similar density
and resistence to
digestion also
included
Central North Sea = 6 800 ± 7 600 mMP items/kg
Northern North Sea 2 500 ± 2 900 mMP items/kg
Barents Sea 2 400 ± 1 300 mMP items/kg
We are now working with FT-IR to obtain more accurate results
Preliminary results: Microplastic in sediments on the Norwegian
Continental Shelf. FT-IR results – collaboration with DNV GL
Preliminary results: Microplastic in sediments on the Norwegian
Continental Shelf. FT-IR results – collaboration with DNV GL
Preliminary results: Microplastic in sediments on the Norwegian
Continental Shelf. FT-IR results – collaboration with DNV GL
Example from one sample (work in progress):
Results: Microplastic in sediments on the Norwegian Continental Shelf
Comparison with other studies – collaboration with DNV GL
mg/kg rare in the literature; average for Norwegian coastal shelf similar to “Ship-breaking yard” i India (due to method differences/non microplastics?) Number of items/kg in general more than other studies, but matches well with a recent study from the arctic (42-6595).
24
Location Location
specification
Particle size
range
Separation
technique Measured concentration Reference
India Ship-breaking yard 1.6µm – 5 mm
Sieving and
density
separation: NaCl
(30%)
81.4 mg/kg Reddy et al. (2006)
NW Pacific Deep sea trench 300 µm – 5 mm Sieving 60 – 2 020 items/m2 Fischer et al. (2015)
Belgium Continental Shelf 38 µm – 1 mm Unknown 97.2 items/kg Claessens et al.
(2011)
Worldwide Deep sea 5 µm – 1 mm
Sieving and
density
separation:
NaI (1.6 g cm-3)
50 items/m2 Van Cauwenberghe
et al. (2013)
Arctic Deep sea 10 µm – 5mm
Density
separations: ZnCl
(1.7 −1.8 g cm−3)
42 - 6 595 items/kg dry
3200 – 247 400 items /m2
Bergmann et al.
(2017)
Norway Oslo sediment 45 µm – 5 mm
plus fibers
Density
separation:
ZnCl2:CaCl2
(1.57 g/mL)
20 – 90 mg/kg Mahat (2017)
Norway
Reference areas in
the Norwegian
coastal shelf
Unknown Unknown 6.3 – 300.5 items/kg Booth et al. (2017)
(MAREANO)
Norway Norwegian
Continental Shelf
45 µm – 5 mm
plus fibers
Density
separation:
ZnCl2:CaCl2
(1.53 g/mL)
< LOD – ≤ 410 (60) mmg/kg
< LOD – ≤ 3 200 (480) mmg/m2
≤ 180 – ≤ 31000 (4 900) max items/kg
≤ 700 – ≤ 250000 (37 000) max items/m2
Møskeland et al.
(2018)
Note:
Different methods
quantify different MP
fractions and have
different biases.
NB! Maximum concentrations
Thank you for your
attention!
Acknowledgement to JPI Oceans WEATHER-MIC,
Research Council of Norway Grant 257433/E40
NORGES GEOTEKNISKE INSTITUTT NGI.NO
#påsikkergrunn