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: hkn@ngi.no 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