**Meet the editor**

Associate Professor Jacques Oosthuizen received his Bachelor's degree, MMed Sci and PhD in Environmental Health. Dr Oosthuizen is the Director of the Occupational Health Research Group at Edith Cowan University, Perth, Australia. He coordinates the undergraduate Environmental Health and the Post Graduate Occupational Hygiene Program. Dr Oosthuizen is a certified

occupational hygienist, registered with the Australian Institute of Occupational Hygienists and he is a member of Environmental Health Australia and the Safety Institute of Australia. He is responsible for the development of a number of successful courses, consulting with the industry. He has also worked closely with the industry providing consultancy and performing contract research in various areas of health, safety and occupational hygiene (exposure assessment). Dr Oosthuizen has supervised numerous Master and PhD students and has a strong track record in applied occupational health research as evidenced by his conference proceedings, journal publications, book chapters and prior as well as current industry collaborative research grants.

Contents

**Preface IX** 

**Part 1 Water Quality 1** 

**Shipping Environmentally Competitive? 3** 

**of Organotins (OTs) and Heavy Metals (MHs) in the Freshwater and Marine Environments 21**

Peter P. Ndibewu, Rob I. McCrindle and Ntebogeng S. Mokgalaka

**Pollution Source Tracking in Surface and Groundwater 57** 

**Following Exposure to Particulate Matter Air Pollution 81** 

**Legislation** *Versus* **Health and Environmental Effects 103**  Klara Slezakova, Simone Morais and Maria do Carmo Pereira

Klara Slezakova, Simone Morais and Maria do Carmo Pereira

**Additives in Gasoline, a Laboratory Based Study 147** 

Harald M. Hjelle and Erik Fridell

Chapter 2 **Speciation Methods for the Determination** 

Chapter 3 **Use of** *Enterococcus,* **BST and Sterols for Poultry** 

Vesna Furtula, Charlene R. Jackson, Rozita Osman and Patricia A. Chambers

Chapter 4 **Understanding Human Illness and Death** 

Chapter 7 **The Potential Environmental Benefits of Utilising Oxy-Compounds as** 

Mihaela Neagu (Petre)

Erin M. Tranfield and David C. Walker

**Relevant Aspects and Health Impacts 125** 

Chapter 1 **When is Short Sea** 

**Part 2 Air Quality 79** 

Chapter 5 **Traffic-Related Air Pollution:** 

Chapter 6 **Indoor Air Pollutants:** 

### Contents

#### **Preface XI**

**Part 1 Water Quality 1** 


#### **Part 2 Air Quality 79**


#### **Part 3 Food Safety 177**

Chapter 8 **Studies on the Isolation of**  *Listeria monocytogenes* **from Food, Water, and Animal Droppings: Environmental Health Perspective 179**  Nkechi Chuks Nwachukwu and Frank Anayo Orji

#### **Part 4 New Technologies 197**

Chapter 9 **Linkages Between Clean Technology Development and Environmental Health Outcomes in Regional Australia 199**  Susan Kinnear and Lisa K. Bricknell

#### **Part 5 Health Impacts 225**

	- **Part 6 Environmental Justice 307**

### Preface

This book contains chapters that cover a diverse range of topics in contemporary environmental health practice. What makes this book unique is the emphasis on emerging research and translation of research into practice. Furthermore, the chapters cover issues from a range of different countries and contexts in both the developed and the developing world, thus providing an insight into the complexity of issues that face environmental health researchers and practitioners globally.

The chapters have been arranged according to topic into six broad sections that cover the state-of-the-art in current literature and introduce new and emerging methods. Some topics in the book include water, air, food, new technologies, health impacts and environmental justice.

> **Dr. Jacques Oosthuizen**  School of Exercise, Biomedical & Health Sciences, Edith Cowan University, Australia

**Part 1** 

**Water Quality**

**Part 1** 

**Water Quality**

**1** 

*1Norway 2Sweden* 

**When is Short Sea Shipping** 

Harald M. Hjelle1 and Erik Fridell2

*2IVL Swedish Environmental Research Institute* 

*Northern Maritime University,* 

*and Northern Maritime University,* 

**Environmentally Competitive?** 

*1Molde University College – Specialized University in Logistics and* 

Maritime transport is broadly accepted as an environmentally friendly mode of transport in terms of CO2 emissions, and is also receiving government support for promotion and development, often based on presumed performance along environmental dimensions.

There is really no debate about the superior comparative efficiency of ships with respect to fuel consumption when calculated per deadweight tonne along routes of similar length. However, the emission figures calculated per deadweight tonne is only relevant for bulk transports, and fuel consumption per cargo tonne is quite different for typical short sea shipping services based on container or RoRo technologies. Further, other emissions to air, like sulphur dioxide, nitrogen oxides and particles, are typically very high for shipping –

The case for short sea shipping as an environmentally-friendly mode of transport is no

The main competitors of such shipping services are rail and road transport. Considering realistic load factors – could the environmental friendly case for maritime transport still be made? This paper is based on the latest data for comparative environmental performance and presents a set of realistic European multimodal transport chains, and their environmental outputs, focusing on fuel consumption and CO2 emissions. Through this comparative analysis we differentiate the common comprehension of shipping being the indisputable green mode of cargo transport, and analyze necessary actions that need to be taken for short sea shipping to maintain its green label. Finally, perspectives on both

Short sea shipping (SSS) plays an important role in the market for regional freight transport in many areas of the world. It's relative importance compared to alternative land-based

longer self-evident under realistic assumptions, and needs deeper analysis.

**2. When is short sea shipping environmentally competitive?** 

**2.1 The competition between short sea shipping and land-based modes** 

especially when no abatement technologies are applied.

regulatory regime and technology are analysed.

**1. Introduction** 

### **When is Short Sea Shipping Environmentally Competitive?**

Harald M. Hjelle1 and Erik Fridell2

*1Molde University College – Specialized University in Logistics and Northern Maritime University, 2IVL Swedish Environmental Research Institute and Northern Maritime University, 1Norway 2Sweden* 

#### **1. Introduction**

Maritime transport is broadly accepted as an environmentally friendly mode of transport in terms of CO2 emissions, and is also receiving government support for promotion and development, often based on presumed performance along environmental dimensions.

There is really no debate about the superior comparative efficiency of ships with respect to fuel consumption when calculated per deadweight tonne along routes of similar length. However, the emission figures calculated per deadweight tonne is only relevant for bulk transports, and fuel consumption per cargo tonne is quite different for typical short sea shipping services based on container or RoRo technologies. Further, other emissions to air, like sulphur dioxide, nitrogen oxides and particles, are typically very high for shipping – especially when no abatement technologies are applied.

The case for short sea shipping as an environmentally-friendly mode of transport is no longer self-evident under realistic assumptions, and needs deeper analysis.

The main competitors of such shipping services are rail and road transport. Considering realistic load factors – could the environmental friendly case for maritime transport still be made? This paper is based on the latest data for comparative environmental performance and presents a set of realistic European multimodal transport chains, and their environmental outputs, focusing on fuel consumption and CO2 emissions. Through this comparative analysis we differentiate the common comprehension of shipping being the indisputable green mode of cargo transport, and analyze necessary actions that need to be taken for short sea shipping to maintain its green label. Finally, perspectives on both regulatory regime and technology are analysed.

#### **2. When is short sea shipping environmentally competitive?**

#### **2.1 The competition between short sea shipping and land-based modes**

Short sea shipping (SSS) plays an important role in the market for regional freight transport in many areas of the world. It's relative importance compared to alternative land-based

When is Short Sea Shipping Environmentally Competitive? 5

Road Rail Oil Pipeline Inland ww Sea Air

Fig. 2. Freight transport activity in EU27, billion tonne-kilometres

1995 1997 1999 2001 2003 2005

Fig. 3. Average annual growth rates of transport modes in EU27

Source: DG Energy and Transport

0

200

400

600

800

1000

**Billion tkm**

1200

1400

1600

1800

2000

Figures from Eurostat 2009

modes is, however, quite different in different regions. Whereas SSS along with inland waterways represents 40% of the intra EU27 transports and more than 60% of the total tonnekilometres in China, the equivalent market-share in the US and Russia is much smaller (Figure 1). To some extent such differences in market shares could be explained by geographical characteristics like the length of the coast-line compared to land area and population, or by the characteristics of natural inland waterways and coastal waters. Such factors may be a natural explanation for the low market share of SSS in Russia – and the equivalently high market-shares in Japan and China. However, it is harder to see how such factors could explain the very different market-shares of EU27 versus the USA. Both have a long coastline and some natural inland waterways. Differences in policy-regimes and the quality of alternative land-based infrastructure are factors that might explain the higher market-share of SSS in Europe compared to the USA.

Fig. 1. Short sea shipping market-shares in 2006 Compilation: Eurostat 2009

From the mid 1990s to 2003 short sea shipping in Europe largely kept up with the growth rates of road transport (Figure 3), but in the years from 2003 to 2006 there has been a significantly lower growth in SSS relative to road transport (Figure 2). The average annual growth rates for road transport in EU27 from 1995 to 2006 was 3.5%, whereas the equivalent figure for SSS was 2.7%.

modes is, however, quite different in different regions. Whereas SSS along with inland waterways represents 40% of the intra EU27 transports and more than 60% of the total tonnekilometres in China, the equivalent market-share in the US and Russia is much smaller (Figure 1). To some extent such differences in market shares could be explained by geographical characteristics like the length of the coast-line compared to land area and population, or by the characteristics of natural inland waterways and coastal waters. Such factors may be a natural explanation for the low market share of SSS in Russia – and the equivalently high market-shares in Japan and China. However, it is harder to see how such factors could explain the very different market-shares of EU27 versus the USA. Both have a long coastline and some natural inland waterways. Differences in policy-regimes and the quality of alternative land-based infrastructure are factors that might explain the higher

From the mid 1990s to 2003 short sea shipping in Europe largely kept up with the growth rates of road transport (Figure 3), but in the years from 2003 to 2006 there has been a significantly lower growth in SSS relative to road transport (Figure 2). The average annual growth rates for road transport in EU27 from 1995 to 2006 was 3.5%, whereas the equivalent

market-share of SSS in Europe compared to the USA.

Fig. 1. Short sea shipping market-shares in 2006

Compilation: Eurostat 2009

figure for SSS was 2.7%.

Fig. 3. Average annual growth rates of transport modes in EU27 Figures from Eurostat 2009

When is Short Sea Shipping Environmentally Competitive? 7

industry in Europe. The legal emissions of NOX and particles are also much higher for shipping than for trucks. This could be attributed to the very different policy regimes for

The global nature of the shipping industry makes it harder to regulate than the trucking business. Regulation must be imposed on a supranational scale to be efficient. This is also true to some extent for road transport, but the degree of national control is much higher on the road networks than for international waters. In Europe this means that the environmental performance of trucks has been improved significantly over the past decades through a series of emission standards gradually reducing emissions of CO, NO, HC and particles (Figure 4). From 2013 the Euro 6 limits will apply with further cuts in NOX, HC and PM emissions. Sulphur emission levels have also been significantly reduced through stricter regulations of the sulphur content of diesel oil. The reductions in fuel use and CO2 emissions

Trains can use either diesel or electricity. In the former case the situation is similar to that of trucks, although the specific emissions of NOX and PM (per work of the engine) are somewhat higher for a modern train engine compared to a truck. From 2012 the emission limits in the EU will be similar to that of a Euro 5 truck. There are no direct emissions from an electric engine. However, for a fair comparison with other modes of transport one should consider the emissions that arise from electricity production. For CO2 this means that the emissions vary significantly with the actual source of the electricity - from negligible for

International shipping has not been subjected to similar regulations over the same period of time, but emissions to air was introduced to the global regulatory regime through the Annex VI of the IMO Marpol convention in 2007. Emissions of CO2 from international shipping were exempted from the Kyoto protocol due to the complexity of allocating emission to the individual partner states. Lately, the Marpol Annex VI regulations have become stricter, especially in the so-called Environmental Control Areas (ECAs). These areas can be for either SO2 (SECAs), NOX (NOX-ECAs) or both. Currently The Baltic Sea, The North Sea and The English Channel are SECAs and the North American coasts will be both SECAs and NOX-ECAs in 2012. The sulphur content in the fuel is currently (2011) limited to 3.5% worldwide and to 1.0% in SECAs. The sulphur restrictions will be further tightened to 0.5% worldwide from 2020 and in SECAs to 0.1% from 2015. The regulation for NOX is also gradually tightened, although through another regulatory instrument, - the NOX-code, applying to marine engines. Engines delivered at present must comply with Tier 1 regulations. From 2012 Tier 2 regulations, giving a cut of about 20%, will apply. In NOX-ECAs Tier 3 regulations apply from 2016, representing a cut in NOX emissions of about 80% compared with Tier 1. The allowed emission for a slow-speed engine will then be 3.4 g/kWh. No specific regulations for particle emissions are implemented for marine engines. Vessels have become more fuel efficient over the past decades, but the most significant advances were made in the late 1970s and the 1980s, triggered by significant increases in bunker prices. Some national regulations have been imposed, e.g. an environmentally differentiated fairway due system in Sweden and a NOX tax in Norway. The European

**2.3 The regulatory regime of shipping vs. land-based transport modes** 

these alternative modes of transport.

have not been as substantial.

hydropower to relatively large for coal-power.

#### **2.2 Short sea shipping as an instrument for greening freight transport**

Since the 1970s European national and EU transport policy papers have had a relatively high focus on moving cargo from road to sea, inland waterways and rail. Partly the rationale for such a policy has been based on the environmental performance of SSS compared to road transport. In general shipping has been regarded "the green mode" of freight transport – often substantiated by empirical data on average energy use per tonne-kilometre and corresponding emission figures. Sometimes such figures have been based on energy use per deadweight tonne, calculated for big wet or dry bulk vessels. Such figures would typically show that shipping is 10-20 times more energy efficient than relevant road transport alternatives (IMO 2009) calculated per tonne-kilometre. This is why land-based modes like road and rail transport normally would not be competitive to maritime transport when it comes to the transport of commodities like iron ore or oil and chemicals, unless the sea leg is significantly longer than the land leg.

The relevant competition for the SSS industry is therefore not so much in the bulk markets, but in the markets for loose and unitized cargo (containers, trailers, pallets). The relevant vessels for such transports are general cargo vessels, container or RoRo vessels, partly in the business of feeding cargo to and from the deep sea, intercontinental, routes and partly transporting cargo within the continent. The environmental performance of these vessels is very different from the bulk vessels, mainly for three reasons. *Firstly*, the payload capacity relative to the size of the vessel is significantly lower than that of bulk vessels. *Secondly*, these vessels are typically designed for, and operated at, significantly higher operating speeds compared to the bulk vessels, and *thirdly* these vessels are operating in liner operations where average shipment sizes are much smaller than in the bulk market, necessitating a demanding consolidation activity in order to fill the available cargo capacity of the vessels. The latter factor normally means that the average load factor of such vessels may be lower than that of the bulk vessels. However, the scope for attracting back-haul cargoes – thus avoiding return trips in ballast – is definitely better for the general cargo, container and RoRo vessels than that of the bulk ships. This may mean that the average roundtrip cargo utilization does not have to be lower compared to bulk operations – which very often are operated with empty back-hauls.

For the RoRo and container industry there is an additional fourth factor – which may be called "the double load factor problem" of these modes (Hjelle 2010). The fact that containers and trailers transported are not always carrying cargo – and may be only partly filled – effectively means that the relevant load factor of such vessels is a multiple of two load factors. The number of containers / trailers compared to the container / trailer capacity – *and* the typical cargo load factor of containers and trailers. Statistics showing a 70% load factor of RoRo vessels often mean that on average 7 out of 10 available lanemetres are occupied by trucks and trailers. If these trailers have a load factor of 60%, then the relevant load factor of the RoRo vessel is not 70%, but 42%.

All of these factors (with the potential exclusion of the third one) contribute to a significantly lower fuel efficiency for relevant SSS vessels than for bulk vessels.

The level of CO2 emissions will follow the fuel efficiency, but emissions of particles, SO2 and NOX are very different for trucks and ships. Under current regulations the shipping industry is allowed to use fuels with much higher sulphur content than the trucking

Since the 1970s European national and EU transport policy papers have had a relatively high focus on moving cargo from road to sea, inland waterways and rail. Partly the rationale for such a policy has been based on the environmental performance of SSS compared to road transport. In general shipping has been regarded "the green mode" of freight transport – often substantiated by empirical data on average energy use per tonne-kilometre and corresponding emission figures. Sometimes such figures have been based on energy use per deadweight tonne, calculated for big wet or dry bulk vessels. Such figures would typically show that shipping is 10-20 times more energy efficient than relevant road transport alternatives (IMO 2009) calculated per tonne-kilometre. This is why land-based modes like road and rail transport normally would not be competitive to maritime transport when it comes to the transport of commodities like iron ore or oil and chemicals, unless the sea leg is

The relevant competition for the SSS industry is therefore not so much in the bulk markets, but in the markets for loose and unitized cargo (containers, trailers, pallets). The relevant vessels for such transports are general cargo vessels, container or RoRo vessels, partly in the business of feeding cargo to and from the deep sea, intercontinental, routes and partly transporting cargo within the continent. The environmental performance of these vessels is very different from the bulk vessels, mainly for three reasons. *Firstly*, the payload capacity relative to the size of the vessel is significantly lower than that of bulk vessels. *Secondly*, these vessels are typically designed for, and operated at, significantly higher operating speeds compared to the bulk vessels, and *thirdly* these vessels are operating in liner operations where average shipment sizes are much smaller than in the bulk market, necessitating a demanding consolidation activity in order to fill the available cargo capacity of the vessels. The latter factor normally means that the average load factor of such vessels may be lower than that of the bulk vessels. However, the scope for attracting back-haul cargoes – thus avoiding return trips in ballast – is definitely better for the general cargo, container and RoRo vessels than that of the bulk ships. This may mean that the average roundtrip cargo utilization does not have to be lower compared to bulk operations – which

For the RoRo and container industry there is an additional fourth factor – which may be called "the double load factor problem" of these modes (Hjelle 2010). The fact that containers and trailers transported are not always carrying cargo – and may be only partly filled – effectively means that the relevant load factor of such vessels is a multiple of two load factors. The number of containers / trailers compared to the container / trailer capacity – *and* the typical cargo load factor of containers and trailers. Statistics showing a 70% load factor of RoRo vessels often mean that on average 7 out of 10 available lanemetres are occupied by trucks and trailers. If these trailers have a load factor of 60%, then the relevant

All of these factors (with the potential exclusion of the third one) contribute to a significantly

The level of CO2 emissions will follow the fuel efficiency, but emissions of particles, SO2 and NOX are very different for trucks and ships. Under current regulations the shipping industry is allowed to use fuels with much higher sulphur content than the trucking

**2.2 Short sea shipping as an instrument for greening freight transport** 

significantly longer than the land leg.

very often are operated with empty back-hauls.

load factor of the RoRo vessel is not 70%, but 42%.

lower fuel efficiency for relevant SSS vessels than for bulk vessels.

industry in Europe. The legal emissions of NOX and particles are also much higher for shipping than for trucks. This could be attributed to the very different policy regimes for these alternative modes of transport.

#### **2.3 The regulatory regime of shipping vs. land-based transport modes**

The global nature of the shipping industry makes it harder to regulate than the trucking business. Regulation must be imposed on a supranational scale to be efficient. This is also true to some extent for road transport, but the degree of national control is much higher on the road networks than for international waters. In Europe this means that the environmental performance of trucks has been improved significantly over the past decades through a series of emission standards gradually reducing emissions of CO, NO, HC and particles (Figure 4). From 2013 the Euro 6 limits will apply with further cuts in NOX, HC and PM emissions. Sulphur emission levels have also been significantly reduced through stricter regulations of the sulphur content of diesel oil. The reductions in fuel use and CO2 emissions have not been as substantial.

Trains can use either diesel or electricity. In the former case the situation is similar to that of trucks, although the specific emissions of NOX and PM (per work of the engine) are somewhat higher for a modern train engine compared to a truck. From 2012 the emission limits in the EU will be similar to that of a Euro 5 truck. There are no direct emissions from an electric engine. However, for a fair comparison with other modes of transport one should consider the emissions that arise from electricity production. For CO2 this means that the emissions vary significantly with the actual source of the electricity - from negligible for hydropower to relatively large for coal-power.

International shipping has not been subjected to similar regulations over the same period of time, but emissions to air was introduced to the global regulatory regime through the Annex VI of the IMO Marpol convention in 2007. Emissions of CO2 from international shipping were exempted from the Kyoto protocol due to the complexity of allocating emission to the individual partner states. Lately, the Marpol Annex VI regulations have become stricter, especially in the so-called Environmental Control Areas (ECAs). These areas can be for either SO2 (SECAs), NOX (NOX-ECAs) or both. Currently The Baltic Sea, The North Sea and The English Channel are SECAs and the North American coasts will be both SECAs and NOX-ECAs in 2012. The sulphur content in the fuel is currently (2011) limited to 3.5% worldwide and to 1.0% in SECAs. The sulphur restrictions will be further tightened to 0.5% worldwide from 2020 and in SECAs to 0.1% from 2015. The regulation for NOX is also gradually tightened, although through another regulatory instrument, - the NOX-code, applying to marine engines. Engines delivered at present must comply with Tier 1 regulations. From 2012 Tier 2 regulations, giving a cut of about 20%, will apply. In NOX-ECAs Tier 3 regulations apply from 2016, representing a cut in NOX emissions of about 80% compared with Tier 1. The allowed emission for a slow-speed engine will then be 3.4 g/kWh. No specific regulations for particle emissions are implemented for marine engines.

Vessels have become more fuel efficient over the past decades, but the most significant advances were made in the late 1970s and the 1980s, triggered by significant increases in bunker prices. Some national regulations have been imposed, e.g. an environmentally differentiated fairway due system in Sweden and a NOX tax in Norway. The European

When is Short Sea Shipping Environmentally Competitive? 9

the fuel consumption is known. Nitrogen oxides are formed in the engine and the emissions will depend on the type of engine and on the presence of NOX after-treatment systems. PM comprises a number of different types of particles and the emissions will depend on engine

For shipping the tabulated emission factors are usually in the form of mass of emission per energy for propulsion from a specific engine. These are normally divided into slow speed, medium speed and high speed engines. Further, the emission factors depend on the type of fuel used; residual oil or gasoil and sulphur content (Cooper and Gustafsson 2004). The emission from a specific vessel thus depends on the engine power and fuel type. In reality the emissions per transported amount of goods and distance will vary significantly depending on the type of ship (tanker, container, general cargo, RoRo etc) and the ships' size. The emissions factor for CO2, for a ship that carries cargo up to its payload, can vary from 1.2 g/tonne-km for a large tanker, to 250 g/tonne-km for a small RoRo ship. Once the emission factor for CO2 is established, emission factors for other substances can be obtained through the relationship with fuel consumption. However, the emissions of NOX, PM and HC may vary significantly

Within the Greenhouse gas working groupof the IMO, a design index for CO2 emissions are being developed for different types of ships (IMO 2009). These are expressed as functions of the deadweight tonnage (dwt) for emissions in g/tonne-nm, and are based on data from a large number of ships. In order to get emission factors for the transported cargo, the relationship between dwt and payload needs to be known as well as typical load factors. The former relationship has been presented in the Clean Ship index (The Clean Shipping Project). In the calculations presented below we have used the specific emission factors presented by Cooper and Gustafsson (2004) as implemented in the model documented in NTM Working Group Goods and Logistics (2008) and NTM (2009). These are obtained from a large number of measurements and correspond well with other reports (see, e.g Whall et al.(2002)). The emission factors for SO2 and PM are adjusted for the sulphur content in the fuel both inside and outside the SECA regions. However, to get the emissions from a specific ship the power used needs to be known. Here we have used the CO2 indexes from IMO and then calculated the corresponding emissions for NOX, PM and SO2. The relationship between dwt and payload used (The Clean Shipping Project) are 0.95 for tanker, 0.8 for container ship and 0.5

**3.2 Empirical evidence on fuel consumption and emissions for road transport** 

The emissions from trucks for the transport of a specific cargo will depend on the size of the truck, the emission classification, the fuel used, driving conditions and the load factor. The emissions of NOX and PM decrease significantly the newer the truck is (see Figure 4). The emission of SO2 will depend on the sulphur content in the diesel which is now at a maximum of 10 wt-ppm in Europe. The CO2 emissions are lower the larger the truck is, when considering emissions per mass of transported goods. The fuel consumption and thus the CO2 emissions will also depend on the type of driving. Within the European Artemis project (Andre 2005) emission factors are available for a large number of trucks and driving

**3.1 Empirical evidence on fuel consumption and emissions for short sea vessels** 

from engine to engine depending on model and maintenance level.

type, fuel quality and after-treatment system.

for RoRo ships.

Commission currently considers implementing emissions from the shipping industry into its cap and trade system of CO2 emissions.

Fig. 4. Truck engine emission standards in Europe Source: EC DG Energy & Transport

The international regulatory regime of maritime transport is moving quite slowly due to the demanding process of reaching the necessary consensus among nations. Adding to this sluggishness of new regulations is the fact that the penetration lead time of technological advances is much longer for ships than for trucks. The average age of a typical short sea vessel trading in European waters is probably around 15 years (Hjelle 2010), whereas a typical long distance truck in Western Europe has an average age of 4 years. This means that the Euro 5 standard, and in a few years Euro 6, will shortly be representative of the fleet of long distance trucks.

#### **3. The environmental performance of vessels, trucks and trains**

In order to compare different alternatives for transporting goods one needs to obtain emission factors expressed as mass of emitted substance per transported amount of goods and distance (functional unit), i.e. an emission factor with units like g/tonne-km. This requires knowledge of emissions per km for the specific vehicle/vessel and the mass of the cargo transported. The latter is often expressed as the maximum possible load multiplied with a load factor.

In this paper we compare the emissions of carbon dioxide (CO2), sulphur dioxide (SO2), nitrogen oxide (NOX) and particulate matter (PM) for different transport alternatives. CO2 emissions are directly obtained from the fuel consumption. The emitted SO2 is formed from sulphur present in the fuel and can easily be obtained if the sulphur content in the fuel and

Commission currently considers implementing emissions from the shipping industry into

CO (g/kWh) HC (g/kWh) NOx (g/kWh) PM (g/kWh)

The international regulatory regime of maritime transport is moving quite slowly due to the demanding process of reaching the necessary consensus among nations. Adding to this sluggishness of new regulations is the fact that the penetration lead time of technological advances is much longer for ships than for trucks. The average age of a typical short sea vessel trading in European waters is probably around 15 years (Hjelle 2010), whereas a typical long distance truck in Western Europe has an average age of 4 years. This means that the Euro 5 standard, and in a few years Euro 6, will shortly be representative of the fleet of

1988 1992 1995 1999 2005 2008

Euro 0 Euro 1 Euro 2 Euro 3 Euro 4 Euro 5

In order to compare different alternatives for transporting goods one needs to obtain emission factors expressed as mass of emitted substance per transported amount of goods and distance (functional unit), i.e. an emission factor with units like g/tonne-km. This requires knowledge of emissions per km for the specific vehicle/vessel and the mass of the cargo transported. The latter is often expressed as the maximum possible load multiplied

In this paper we compare the emissions of carbon dioxide (CO2), sulphur dioxide (SO2), nitrogen oxide (NOX) and particulate matter (PM) for different transport alternatives. CO2 emissions are directly obtained from the fuel consumption. The emitted SO2 is formed from sulphur present in the fuel and can easily be obtained if the sulphur content in the fuel and

**3. The environmental performance of vessels, trucks and trains** 

its cap and trade system of CO2 emissions.

Fig. 4. Truck engine emission standards in Europe

Source: EC DG Energy & Transport

long distance trucks.

0

2

4

6

8

10

**g per kWh**

12

14

16

18

with a load factor.

the fuel consumption is known. Nitrogen oxides are formed in the engine and the emissions will depend on the type of engine and on the presence of NOX after-treatment systems. PM comprises a number of different types of particles and the emissions will depend on engine type, fuel quality and after-treatment system.

#### **3.1 Empirical evidence on fuel consumption and emissions for short sea vessels**

For shipping the tabulated emission factors are usually in the form of mass of emission per energy for propulsion from a specific engine. These are normally divided into slow speed, medium speed and high speed engines. Further, the emission factors depend on the type of fuel used; residual oil or gasoil and sulphur content (Cooper and Gustafsson 2004). The emission from a specific vessel thus depends on the engine power and fuel type. In reality the emissions per transported amount of goods and distance will vary significantly depending on the type of ship (tanker, container, general cargo, RoRo etc) and the ships' size. The emissions factor for CO2, for a ship that carries cargo up to its payload, can vary from 1.2 g/tonne-km for a large tanker, to 250 g/tonne-km for a small RoRo ship. Once the emission factor for CO2 is established, emission factors for other substances can be obtained through the relationship with fuel consumption. However, the emissions of NOX, PM and HC may vary significantly from engine to engine depending on model and maintenance level.

Within the Greenhouse gas working groupof the IMO, a design index for CO2 emissions are being developed for different types of ships (IMO 2009). These are expressed as functions of the deadweight tonnage (dwt) for emissions in g/tonne-nm, and are based on data from a large number of ships. In order to get emission factors for the transported cargo, the relationship between dwt and payload needs to be known as well as typical load factors. The former relationship has been presented in the Clean Ship index (The Clean Shipping Project).

In the calculations presented below we have used the specific emission factors presented by Cooper and Gustafsson (2004) as implemented in the model documented in NTM Working Group Goods and Logistics (2008) and NTM (2009). These are obtained from a large number of measurements and correspond well with other reports (see, e.g Whall et al.(2002)). The emission factors for SO2 and PM are adjusted for the sulphur content in the fuel both inside and outside the SECA regions. However, to get the emissions from a specific ship the power used needs to be known. Here we have used the CO2 indexes from IMO and then calculated the corresponding emissions for NOX, PM and SO2. The relationship between dwt and payload used (The Clean Shipping Project) are 0.95 for tanker, 0.8 for container ship and 0.5 for RoRo ships.

#### **3.2 Empirical evidence on fuel consumption and emissions for road transport**

The emissions from trucks for the transport of a specific cargo will depend on the size of the truck, the emission classification, the fuel used, driving conditions and the load factor. The emissions of NOX and PM decrease significantly the newer the truck is (see Figure 4). The emission of SO2 will depend on the sulphur content in the diesel which is now at a maximum of 10 wt-ppm in Europe. The CO2 emissions are lower the larger the truck is, when considering emissions per mass of transported goods. The fuel consumption and thus the CO2 emissions will also depend on the type of driving. Within the European Artemis project (Andre 2005) emission factors are available for a large number of trucks and driving

When is Short Sea Shipping Environmentally Competitive? 11

CO2 emissions are directly correlated with use of fossil fuels. The most fuel efficient among the cases in Fig. 5 is the big tanker vessel, with a CO2 emission of 4 grams per tonne-km. At the other end of the scale is the truck/trailer combination with a CO2 emission of 63 grams per tonne-km. The RoRo vessel is marginally better with an equivalent figure of 53 grams. The CO2 emissions from the electric train with the EU25 energy mix is 24 grams per tonnekm. The container feeder vessel performs much better than the RoRo-vessel at 37 grams per

The comparatively very high SO2 emissions from the vessels range from 0.024 grams for the large tanker to 0.32 grams for the RoRo-vessel while it is only 80 μg/tonne-km for the truck. This is despite the fact that we have assumed that the fuel quality is according to the SECAregulations of 1.0% sulphur content. Future stricter limits for sulphur content will to some extent make short sea shipping SO2 emissions come closer to those of the alternative modes,

European trucks (Euro 4 and Euro 5 standard) have relatively low particle emissions2. No other mode has lower PM emissions. NOX emissions are also low for truck transport, only beaten by the large tanker and the electric train. Further, a Euro 6 truck would have an

Comparative figures like these are often presented in policy papers as a rationale for promoting short sea shipping as an alternative to land based modes of transport. Sometimes the figures presented are quite different from one setting to another. One late example is the figures presented in Chapter 9 in the IMO MEPC (IMO 2009) report. Here the CO2 emissions of a wide range of vessels are presented along with figures for road and rail. As a benchmark for the figures presented in Figure 5, we present a subset of figures representing

additional cut in NOX emissions by around 90% compared with the Euro 4 truck.

**Vessel / Vehicle Total CO2-efficiency (g/tonne-km)** 

Table 1. CO2 emissions per tonne-km for alternative freight transport modes according to

1 The container feeder vessel performs better than the RoRo vessel, but it should be noted that the weight of the container itself is included when the calculations have been made. 2 This applies to exhaust PM. Trucks will also generate resuspended particles from road dust and wear.

The data for the oil tanker used here is 3.7 g/tonne-km, as compared to 4.4 g/tonne-km in the IMO MEPC-report. The latter is an average for tankers between 120 000 and 200 000 dwt, whereas the tanker considered in this paper is a 125 000 tonner. This discrepancy may be partly explained by the fact that the model used yields a load factor of 55% for crude

CO2 emissions per tonne-km from this paper in Table 1.

Crude oil tanker 120'-200' dwt 4.4 Container 1000-1999 TEU 32.1 Container 0-999 TEU 36.3 RoRo 2000+ lm 49.5

Road freight 150 (80-180) Rail 10-119

IMO MEPC (2009). Compiled from the text and various tables.

tankers, whereas the IMO MEPC-report applies 48%.

tonne-km1.

but not beat them.

conditions. For example, the CO2 emission per km for a typical Euro 4 truck of around 19 m length and capable of loading 26 tonnes of goods vary from 700 g/km (urban driving) to 580 g/km (rural) for an empty truck and from 1380 g/km to 1080 g/km for a fully loaded truck. In the calculations made here a load factor of 60% is used and the calculations are made for rural driving.

#### **3.3 Empirical evidence on fuel consumption and emissions for rail freight**

For diesel rail engines the data on fuel consumption is very limited and in this review the procedure of EcoTransIT was used to calculate emissions. In the case of electrical engines the CO2 emissions depend on the source of electricity. For calculations an electricity mix for EU 25 obtained from EcoTransIT was utilized (Knörr 2008).

#### **3.4 Realistic load factors and realistic speed are crucial elements in the comparative analysis**

In Figure 5 the emissions per tonne-kilometre are presented for the alternative modes of transport included in this paper. These are estimated based on realistic load factors for the various modes as presented above. For the RoRo vessel a load factor of 44% is used, for the container feeder 48%, for the tanker 55%, for trains 50% and for the truck/trailer 60%. The load factor for the RoRo and container vessel represents the relation with the transported goods and the payload and takes into account both the weight of the trucks themselves and that containers are assumed to have a fill factor of 60%.

Fig. 5. Emissions per tonnekilometre for the alternative freight transport modes. CO2 emissions in kg/tkm. NOX, PM, and SO2 emissions in g/tonne-km.

conditions. For example, the CO2 emission per km for a typical Euro 4 truck of around 19 m length and capable of loading 26 tonnes of goods vary from 700 g/km (urban driving) to 580 g/km (rural) for an empty truck and from 1380 g/km to 1080 g/km for a fully loaded truck. In the calculations made here a load factor of 60% is used and the calculations are made for

For diesel rail engines the data on fuel consumption is very limited and in this review the procedure of EcoTransIT was used to calculate emissions. In the case of electrical engines the CO2 emissions depend on the source of electricity. For calculations an electricity mix for

**3.4 Realistic load factors and realistic speed are crucial elements in the comparative** 

In Figure 5 the emissions per tonne-kilometre are presented for the alternative modes of transport included in this paper. These are estimated based on realistic load factors for the various modes as presented above. For the RoRo vessel a load factor of 44% is used, for the container feeder 48%, for the tanker 55%, for trains 50% and for the truck/trailer 60%. The load factor for the RoRo and container vessel represents the relation with the transported goods and the payload and takes into account both the weight of the trucks themselves and

**3.3 Empirical evidence on fuel consumption and emissions for rail freight** 

Fig. 5. Emissions per tonnekilometre for the alternative freight transport modes.

CO2 emissions in kg/tkm. NOX, PM, and SO2 emissions in g/tonne-km.

EU 25 obtained from EcoTransIT was utilized (Knörr 2008).

that containers are assumed to have a fill factor of 60%.

rural driving.

**analysis** 

CO2 emissions are directly correlated with use of fossil fuels. The most fuel efficient among the cases in Fig. 5 is the big tanker vessel, with a CO2 emission of 4 grams per tonne-km. At the other end of the scale is the truck/trailer combination with a CO2 emission of 63 grams per tonne-km. The RoRo vessel is marginally better with an equivalent figure of 53 grams. The CO2 emissions from the electric train with the EU25 energy mix is 24 grams per tonnekm. The container feeder vessel performs much better than the RoRo-vessel at 37 grams per tonne-km1.

The comparatively very high SO2 emissions from the vessels range from 0.024 grams for the large tanker to 0.32 grams for the RoRo-vessel while it is only 80 μg/tonne-km for the truck. This is despite the fact that we have assumed that the fuel quality is according to the SECAregulations of 1.0% sulphur content. Future stricter limits for sulphur content will to some extent make short sea shipping SO2 emissions come closer to those of the alternative modes, but not beat them.

European trucks (Euro 4 and Euro 5 standard) have relatively low particle emissions2. No other mode has lower PM emissions. NOX emissions are also low for truck transport, only beaten by the large tanker and the electric train. Further, a Euro 6 truck would have an additional cut in NOX emissions by around 90% compared with the Euro 4 truck.

Comparative figures like these are often presented in policy papers as a rationale for promoting short sea shipping as an alternative to land based modes of transport. Sometimes the figures presented are quite different from one setting to another. One late example is the figures presented in Chapter 9 in the IMO MEPC (IMO 2009) report. Here the CO2 emissions of a wide range of vessels are presented along with figures for road and rail. As a benchmark for the figures presented in Figure 5, we present a subset of figures representing CO2 emissions per tonne-km from this paper in Table 1.


Table 1. CO2 emissions per tonne-km for alternative freight transport modes according to IMO MEPC (2009). Compiled from the text and various tables.

The data for the oil tanker used here is 3.7 g/tonne-km, as compared to 4.4 g/tonne-km in the IMO MEPC-report. The latter is an average for tankers between 120 000 and 200 000 dwt, whereas the tanker considered in this paper is a 125 000 tonner. This discrepancy may be partly explained by the fact that the model used yields a load factor of 55% for crude tankers, whereas the IMO MEPC-report applies 48%.

<sup>1</sup> The container feeder vessel performs better than the RoRo vessel, but it should be noted that the

weight of the container itself is included when the calculations have been made. 2 This applies to exhaust PM. Trucks will also generate resuspended particles from road dust and wear.

When is Short Sea Shipping Environmentally Competitive? 13

environments. In a setting where truck transport is compared to short sea shipping such settings are not very relevant as the maritime transport alternatives would compete against long haul truck/trailer combinations rather than distribution vehicles in urban settings. This might explain the fact that our figure lies below the lower bound of the IMO

Our rail alternatives yield CO2 emission figures between 24.3 (electric) and 42.6 (diesel) g/tonne-km. The IMO MEPC (2009) study refers to six different studies, yielding a range between 10 to 119 g/tonne-km. The lower figure stems from the long and slow moving bulk trains in the USA, and the upper limit stems from a top-down calculation based on data for the EU region provided by Eurostat. Our data based on the EcoTransIT (Knörr 2008) model lie within these limits, but are significantly lower than the top-down calculations based on Eurostat data. Among the sources cited by the IMO MEPC-report, our figures are quite close to the ones based on US container trains (35-50 g/tonne-km). Further, it can be pointed out that an electric train using exclusively hydro electricity would in our calculations have a

We have chosen four typical intra-European trade links which are quite different with respect to the comparative distances for alternative modes of freight transport (Figure 6). The first case is Gothenburg (Sweden) to Rotterdam (The Netherlands), which is a relatively short distance by sea, and somewhat longer by road and rail. The second case, Helsinki (Finland) to Genoa (Italy) is the longest one, and a case where the sea-link is significantly longer than the road and rail alternatives. Rotterdam to LeHavre (France) is a link where the sea-leg is almost parallel to the road and rail alternatives, which means that this third case will mainly be affected by differences in emissions per tonne-km for the alternative modes. Finally, the last case is Gothenburg to Aberdeen (Scotland). This case represents an alternative where short sea shipping has a very significant comparative advantage distancewise. Road and rail alternatives for this case are three times as long as the maritime

These four geographical cases are then combined with alternative modes, also with some different varieties within the broad modal categories of sea, road and rail transport. The sea transport alternatives included in this analysis are a 10 000 dwt RoRo-vessel, a 6000 dwt container feeder vessel and a 125 000 dwt tanker. The latter one would typically be used for shuttle transports from offshore oil production sites to refineries, and thus road and rail transport is no realistic alternative to the tanker. We have included this vessel here more as a reference to illustrate how typical calculations of emissions per dwt for large bulk vessels will be very different to such figures for typical short sea cargo vessels. For rail transport, we have included one diesel train alternative, and one electric train with a typical mix of electricity production for the EU. Finally, we have included one typical long distance truck/trailer combination (19 meter) with a Euro 4 engine. As the average age of such trucks in Western Europe will in the area of 4-5 years (Sandvik 2005), this will be a representative

CO2 emission of 0.004 g/tonne-km with a load factor of 0.5.

**4.1 Four cases and seven modal alternatives** 

**4. Comparing alternative modes on typical short sea legs** 

MEPC figures.

transport alternative.

engine type.

The two container vessels from the IMO MEPC-report yields a CO2 emission level of 32.1- 36.3 g/tonne-km. The 13 000 dwt container vessel included in our analysis would typically carry 1000 TEUs, and emits 37.3 g/tonne-km – which is somewhat higher than the IMO MEPC figures. According to the text in the IMO MEPC-report the cargo capacity of the container vessels is based on an assumed 7 tonnes per container. The 70% load factor applied in the IMO MEPC-report is probably calculated as a percentage of this figure, meaning that the assumed net cargo on a 1000 TEU vessel would be 4 900 tonnes. This is similar to our assumption which is based on a cargo capacity of 10 400 tonnes for the 13 000 dwt container feeder vessel, and a load factor of 48%, yielding 4 992 tonnes of cargo.

In our case study we have included a RoRo vessel of 10 000 dwt, emitting 52.7 g/tonne-km. This is slightly higher than the 2000+ lm RoRo-vessel in the IMO MEPC-figures above – which yields 49.5 g/tonne-km. We have applied a load-factor of 44%. This is a combination of the truck load factor and the "lanemeter loadfactor" – see Hjelle (2010) on the double load-factor problem of RoRo shipping. We have also corrected the net cargo carrying capacity of the vessel for the difference between the gross and payload weight of the truck/trailer (40 tonnes vs 26 tonnes).

The IMO MEPC figures are based on an assumption of a cargo capacity of 2 tonnes per lanemeter for the RoRo vessels. The IMO report does not state weather the term "cargo" means net cargo, or a gross term in the form of the combination of truck/trailer and cargo. A plausible interpretation would be that one has assumed only unaccompanied trailers with a payload of 26 tonnes and a lanemeter footprint of 13 meters, which yields 2 tonnes per lanemeter as the maximum net cargo capacity. In most operations one would have a mix of accompanied and unaccompanied trailers. One will also have to allow some extra space for stowage, which means that a more plausible figure probably would be in the area of 1.6 tonnes per lanemeter as a maximum capacity limit. The 2 tonnes applied in the IMO MEPC figures implies that the average lanementer capacity of the 2000+ lm category is 2577 lanemeters. According to the calculations above this corresponds to a cargo carrying capacity of 4123 tonnes. If 70% of the lanemeters are utilized on average, and the truck has an average load factor of 60%, the combined loadfactor of 42% means that this vessel category on average carries 1732 tonnes of cargo.

In our calculations we have applied the IMO GHG group's CO2 index for a 10 000 dwt RoRo ship which is 15.1 g/tonne-km when it is full. Such a vessel is assumed by us to have a payload of 5000 tonnes (including the own weight of the trucks and trailers, 3250 tonnes without). As indicated above, we have applied a combined load factor (representing both lanemeter utilization and truck payload utilization) of 44%. Based on this we end up with a CO2 emission factor that is close to the one reported in the IMO MEPC-report.

For road freight the IMO MEPC (IMO 2009) report refers to seven different sources/studies, and concludes with an average figure of 150 g/tkm and a range from 80 to 180 g/tkm. Based on the Artemis model we end up with 63.1 g/tkm for our 19m truck/trailer combination with a load factor of 0.6. Since the IMO publication only briefly refers to external sources, it is not quite clear which settings all of these figures stem from, neither the implied load factors. It is clear though, that some of the referred sources include figures representative for smaller trucks and trucks operating in more urban

The two container vessels from the IMO MEPC-report yields a CO2 emission level of 32.1- 36.3 g/tonne-km. The 13 000 dwt container vessel included in our analysis would typically carry 1000 TEUs, and emits 37.3 g/tonne-km – which is somewhat higher than the IMO MEPC figures. According to the text in the IMO MEPC-report the cargo capacity of the container vessels is based on an assumed 7 tonnes per container. The 70% load factor applied in the IMO MEPC-report is probably calculated as a percentage of this figure, meaning that the assumed net cargo on a 1000 TEU vessel would be 4 900 tonnes. This is similar to our assumption which is based on a cargo capacity of 10 400 tonnes for the 13 000

dwt container feeder vessel, and a load factor of 48%, yielding 4 992 tonnes of cargo.

truck/trailer (40 tonnes vs 26 tonnes).

category on average carries 1732 tonnes of cargo.

In our case study we have included a RoRo vessel of 10 000 dwt, emitting 52.7 g/tonne-km. This is slightly higher than the 2000+ lm RoRo-vessel in the IMO MEPC-figures above – which yields 49.5 g/tonne-km. We have applied a load-factor of 44%. This is a combination of the truck load factor and the "lanemeter loadfactor" – see Hjelle (2010) on the double load-factor problem of RoRo shipping. We have also corrected the net cargo carrying capacity of the vessel for the difference between the gross and payload weight of the

The IMO MEPC figures are based on an assumption of a cargo capacity of 2 tonnes per lanemeter for the RoRo vessels. The IMO report does not state weather the term "cargo" means net cargo, or a gross term in the form of the combination of truck/trailer and cargo. A plausible interpretation would be that one has assumed only unaccompanied trailers with a payload of 26 tonnes and a lanemeter footprint of 13 meters, which yields 2 tonnes per lanemeter as the maximum net cargo capacity. In most operations one would have a mix of accompanied and unaccompanied trailers. One will also have to allow some extra space for stowage, which means that a more plausible figure probably would be in the area of 1.6 tonnes per lanemeter as a maximum capacity limit. The 2 tonnes applied in the IMO MEPC figures implies that the average lanementer capacity of the 2000+ lm category is 2577 lanemeters. According to the calculations above this corresponds to a cargo carrying capacity of 4123 tonnes. If 70% of the lanemeters are utilized on average, and the truck has an average load factor of 60%, the combined loadfactor of 42% means that this vessel

In our calculations we have applied the IMO GHG group's CO2 index for a 10 000 dwt RoRo ship which is 15.1 g/tonne-km when it is full. Such a vessel is assumed by us to have a payload of 5000 tonnes (including the own weight of the trucks and trailers, 3250 tonnes without). As indicated above, we have applied a combined load factor (representing both lanemeter utilization and truck payload utilization) of 44%. Based on this we end up with a

For road freight the IMO MEPC (IMO 2009) report refers to seven different sources/studies, and concludes with an average figure of 150 g/tkm and a range from 80 to 180 g/tkm. Based on the Artemis model we end up with 63.1 g/tkm for our 19m truck/trailer combination with a load factor of 0.6. Since the IMO publication only briefly refers to external sources, it is not quite clear which settings all of these figures stem from, neither the implied load factors. It is clear though, that some of the referred sources include figures representative for smaller trucks and trucks operating in more urban

CO2 emission factor that is close to the one reported in the IMO MEPC-report.

environments. In a setting where truck transport is compared to short sea shipping such settings are not very relevant as the maritime transport alternatives would compete against long haul truck/trailer combinations rather than distribution vehicles in urban settings. This might explain the fact that our figure lies below the lower bound of the IMO MEPC figures.

Our rail alternatives yield CO2 emission figures between 24.3 (electric) and 42.6 (diesel) g/tonne-km. The IMO MEPC (2009) study refers to six different studies, yielding a range between 10 to 119 g/tonne-km. The lower figure stems from the long and slow moving bulk trains in the USA, and the upper limit stems from a top-down calculation based on data for the EU region provided by Eurostat. Our data based on the EcoTransIT (Knörr 2008) model lie within these limits, but are significantly lower than the top-down calculations based on Eurostat data. Among the sources cited by the IMO MEPC-report, our figures are quite close to the ones based on US container trains (35-50 g/tonne-km). Further, it can be pointed out that an electric train using exclusively hydro electricity would in our calculations have a CO2 emission of 0.004 g/tonne-km with a load factor of 0.5.

#### **4. Comparing alternative modes on typical short sea legs**

#### **4.1 Four cases and seven modal alternatives**

We have chosen four typical intra-European trade links which are quite different with respect to the comparative distances for alternative modes of freight transport (Figure 6). The first case is Gothenburg (Sweden) to Rotterdam (The Netherlands), which is a relatively short distance by sea, and somewhat longer by road and rail. The second case, Helsinki (Finland) to Genoa (Italy) is the longest one, and a case where the sea-link is significantly longer than the road and rail alternatives. Rotterdam to LeHavre (France) is a link where the sea-leg is almost parallel to the road and rail alternatives, which means that this third case will mainly be affected by differences in emissions per tonne-km for the alternative modes. Finally, the last case is Gothenburg to Aberdeen (Scotland). This case represents an alternative where short sea shipping has a very significant comparative advantage distancewise. Road and rail alternatives for this case are three times as long as the maritime transport alternative.

These four geographical cases are then combined with alternative modes, also with some different varieties within the broad modal categories of sea, road and rail transport. The sea transport alternatives included in this analysis are a 10 000 dwt RoRo-vessel, a 6000 dwt container feeder vessel and a 125 000 dwt tanker. The latter one would typically be used for shuttle transports from offshore oil production sites to refineries, and thus road and rail transport is no realistic alternative to the tanker. We have included this vessel here more as a reference to illustrate how typical calculations of emissions per dwt for large bulk vessels will be very different to such figures for typical short sea cargo vessels. For rail transport, we have included one diesel train alternative, and one electric train with a typical mix of electricity production for the EU. Finally, we have included one typical long distance truck/trailer combination (19 meter) with a Euro 4 engine. As the average age of such trucks in Western Europe will in the area of 4-5 years (Sandvik 2005), this will be a representative engine type.

When is Short Sea Shipping Environmentally Competitive? 15

Fig. 7. Emissions of alternative freight transport modes. One shipment of 1000 tonnes from Gothenburg to Rotterdam

Fig. 8. Emissions of alternative freight transport modes. One shipment of 1000 tonnes from Helsinki to Genoa

Fig. 6. Distances for alternative OD-pairs and modes (kilometres)

#### **4.2 Emissions to air for the alternative cases**

Putting these alternative modes into realistic settings, differences in relative distances also comes into play. In Figure 7 the environmental performance of the alternative modes are presented for the Gothenburg-Rotterdam link. This is a link where the sea-leg is somewhat shorter than the road and rail alternatives. This makes the RoRo and container liner alternatives the winners along with the electric train, regarding CO2 emissions. The truck/trailer combination yields CO2 emissions that are more than twice as high as those of the RoRo-vessel, and 4-5 times that of the container vessel.

Even with the distance advantage for the shipping alternative – the emissions of SO2 are significantly higher from the SSS alternatives than for road and rail. The picture is more mixed for NOx and PM emissions. The container feeder performs much better than the diesel-train regarding NOx, whereas the RoRo alternative is comparable to the diesel train. Regarding PM-emissions both train alternatives are of the same order of magnitude as the container vessel, but yield a lower emission level compared to RoRo transport. As we have pointed out earlier, European trucks have very low particle emissions compared to alternative modes.

Fig. 6. Distances for alternative OD-pairs and modes (kilometres)

the RoRo-vessel, and 4-5 times that of the container vessel.

Putting these alternative modes into realistic settings, differences in relative distances also comes into play. In Figure 7 the environmental performance of the alternative modes are presented for the Gothenburg-Rotterdam link. This is a link where the sea-leg is somewhat shorter than the road and rail alternatives. This makes the RoRo and container liner alternatives the winners along with the electric train, regarding CO2 emissions. The truck/trailer combination yields CO2 emissions that are more than twice as high as those of

Even with the distance advantage for the shipping alternative – the emissions of SO2 are significantly higher from the SSS alternatives than for road and rail. The picture is more mixed for NOx and PM emissions. The container feeder performs much better than the diesel-train regarding NOx, whereas the RoRo alternative is comparable to the diesel train. Regarding PM-emissions both train alternatives are of the same order of magnitude as the container vessel, but yield a lower emission level compared to RoRo transport. As we have pointed out earlier, European trucks have very low particle emissions compared to

**4.2 Emissions to air for the alternative cases** 

alternative modes.

Fig. 7. Emissions of alternative freight transport modes. One shipment of 1000 tonnes from Gothenburg to Rotterdam

Fig. 8. Emissions of alternative freight transport modes. One shipment of 1000 tonnes from Helsinki to Genoa

When is Short Sea Shipping Environmentally Competitive? 17

Fig. 10. Emissions of alternative freight transport modes. One shipment of 1000 tonnes from Gothenburg to Aberdeen

emission regulations are put in place.

a powerful green-house gas.

**5. Technological and political perspectives on green shipping** 

**5.1 The scope for technology-based reductions of emissions from short sea shipping**  The picture presented here will be altered in the future as new engine and exhaust conversion technologies are introduced. If we look at regulations already in place we can note that a Euro 6 truck would reduce the emissions of NOX by 90% and by PM with 50% compared with the truck used in our calculations. For the ships most of the routes we studied are already within SECAs. Here the emissions of SO2 will be reduced by 67% by 2015 compared with today. Also the PM emissions are then probably reduced by about 80% by 2015 in SECAs. If the ship were to have Tier 3 engines the NOX emissions would be reduced by about 80%. The use of natural gas as fuel would give even further reductions in all three substances. The train with diesel engines will show better performance as the new

When it comes to CO2 emissions all three transport modes have the potential for reductions through increasing the load factors. For the ships, significant improvements can be obtained through reducing the ships' speed, since the fuel consumption is strongly dependent on speed. All three modes also have the possibility to use alternative fuels. Natural gas should give a 25% reduction in CO2 emissions but may increase the emissions of methane which is

Fig. 9. Emissions of alternative freight transport modes. One shipment of 1000 tonnes from Bremen to Le Havre

The Helsinki-Genoa case illustrates the effects of cases where the sea leg is significantly longer than the land-based alternatives. With such a big distance-disadvantage the SSS modes will lose along all environmental dimensions. Still, it may be interesting to note that our "reference" tanker vessel is more energy efficient than the land based modes even with such a huge difference in distances. The train alternatives are preferable to the truck alternative with respect to CO2 emissions, but the picture is more mixed for other emission types.

The Bremen-Le Havre case (Figure 9) would be a typical project for the Motorways of the Seas programme of the EU, since this would be a service that might relieve traffic congestion on parallel road (and rail) networks. Would it also be good case along pure emissions-to-air dimensions? As always the maritime transport alternatives performs poorly with respect to SO2 emissions – and also with respect to NOX and PM when compared to truck transport. The container feeder emits much less CO2 than the two rail alternatives, whereas the RoRo vessel emits more CO2 than the electric rail alternative and somewhat less than the diesel train. Both SSS services perform better with respect to CO2 than the truck/trailer combination.

Finally, our Gothenburg-Aberdeen case represents the other extreme, compared to the third case. Here the SSS-alternatives have a very large distance advantage compared to road and rail. Even with this advantage SO2 emissions are high for the container and RoRoalternatives. This is also true for the NOX emissions for RoRo relative to the road transport alternative. The energy use is of course much lower for the vessels than for the road and rail alternatives.

The Helsinki-Genoa case illustrates the effects of cases where the sea leg is significantly longer than the land-based alternatives. With such a big distance-disadvantage the SSS modes will lose along all environmental dimensions. Still, it may be interesting to note that our "reference" tanker vessel is more energy efficient than the land based modes even with such a huge difference in distances. The train alternatives are preferable to the truck alternative with

The Bremen-Le Havre case (Figure 9) would be a typical project for the Motorways of the Seas programme of the EU, since this would be a service that might relieve traffic congestion on parallel road (and rail) networks. Would it also be good case along pure emissions-to-air dimensions? As always the maritime transport alternatives performs poorly with respect to SO2 emissions – and also with respect to NOX and PM when compared to truck transport. The container feeder emits much less CO2 than the two rail alternatives, whereas the RoRo vessel emits more CO2 than the electric rail alternative and somewhat less than the diesel train. Both SSS services perform better with respect to CO2 than the truck/trailer

Finally, our Gothenburg-Aberdeen case represents the other extreme, compared to the third case. Here the SSS-alternatives have a very large distance advantage compared to road and rail. Even with this advantage SO2 emissions are high for the container and RoRoalternatives. This is also true for the NOX emissions for RoRo relative to the road transport alternative. The energy use is of course much lower for the vessels than for the road and rail

respect to CO2 emissions, but the picture is more mixed for other emission types.

Fig. 9. Emissions of alternative freight transport modes. One shipment of 1000 tonnes from Bremen to Le Havre

combination.

alternatives.

Fig. 10. Emissions of alternative freight transport modes. One shipment of 1000 tonnes from Gothenburg to Aberdeen

#### **5. Technological and political perspectives on green shipping**

#### **5.1 The scope for technology-based reductions of emissions from short sea shipping**

The picture presented here will be altered in the future as new engine and exhaust conversion technologies are introduced. If we look at regulations already in place we can note that a Euro 6 truck would reduce the emissions of NOX by 90% and by PM with 50% compared with the truck used in our calculations. For the ships most of the routes we studied are already within SECAs. Here the emissions of SO2 will be reduced by 67% by 2015 compared with today. Also the PM emissions are then probably reduced by about 80% by 2015 in SECAs. If the ship were to have Tier 3 engines the NOX emissions would be reduced by about 80%. The use of natural gas as fuel would give even further reductions in all three substances. The train with diesel engines will show better performance as the new emission regulations are put in place.

When it comes to CO2 emissions all three transport modes have the potential for reductions through increasing the load factors. For the ships, significant improvements can be obtained through reducing the ships' speed, since the fuel consumption is strongly dependent on speed. All three modes also have the possibility to use alternative fuels. Natural gas should give a 25% reduction in CO2 emissions but may increase the emissions of methane which is a powerful green-house gas.

When is Short Sea Shipping Environmentally Competitive? 19

Our case studies illustrate that short sea shipping operations, represented by RoRo and container services may very well deserve their "green label" when compared to alternative modes with respect to CO2 emissions. This conclusion is valid under what we consider realistic operating environments with respect to vessel operation speeds and achieved load factors, and when the shipping leg is not much longer than the distances of the land based modes. This conclusion holds at least for the container vessels, but the advantage of RoRo operations versus truck transport may be marginal – and is highly dependent on the

The short sea shipping alternative does generally *not* deserve a "green label" when SO2, NOX and PM emissions are considered. Although some improvements are in the pipeline through the stricter Marpol Annex VI regulations, the maritime transport alternatives will still not be able to compete with road transport along these dimensions unless new fuels

We have applied quite large feeder vessels, a 10 000 dwt RoRo vessel and a 13 000 dwt container vessel, in our case studies. Smaller vessels will generally yield higher emissions

We have illustrated that the use of realistic load factors is crucial in a comparative analysis like this. Applying load factors related to the cargo *capacity* of the vessel measured in tonnes will not yield a realistic setting, especially for RoRo vessels. All emissions should be attributed to the net cargo transported – as is the intention of IMOs proposed Energy

The recent work of the IMO MEPC points out that there is a very significant potential for reductions of CO2 emissions from ships – but that many of the possible technological and

This paper is partly financed through the Northern Maritime University (NMU) project. The

Andre, J.-M. (2005). "Vehicle emission measurement collection of the ARTEMIS datatbase."

Cooper, D. and T. Gustafsson (2004). Methodology for calculatin emissions from ships: 1. Update of emission factors. IVL Reports. Gothenburg, IVL Svenska Miljöinstitutet. Hjelle, H. M. (2010). The double load-factor problem of Ro-Ro shipping. 12th World

Hjelle, H. M. (2010). "Short Sea Shipping's green label at risk." Transport Reviews 30(5): 617-

organizational measures are dependent on efficient policy regimes to come into play.

An earlier version of this paper was presented at IAME 2010 in Lisbon, Portugal.

http://www.inrets.fr/ur/lte/publications/publications-pdf/Joumard/

prevailing market situation and the resulting load factors achieved.

(LNG) are introduced or abatement technologies are installed.

NMU project is partly funded by the EC Interreg IVB programme.

per tonne-km and may therefore be less competitive.

Efficiency Operational Index (EEOL).

Retrieved Feb 23, 2010, from

A3312reportJMALTE0504.pdf

Conference on Transport Research. Lisbon.

**7. Acknowledgements** 

**8. References** 

640.

**6. Conclusions** 

The Second IMO Greenhouse Gas Study (IMO 2009) points to a number of technology-based options for improving the energy efficiency of vessels. Partly these are related to improved design (concepts, hull and superstructure, power and propulsion systems) and improved operations (fleet management, logistics, incentives, voyage optimization, energy management). The combined potential for reductions of CO2 emission from these technologies is estimated to be between 25 and 75%. To reach the upper bound of this range, reductions in operating speed would be necessary.

#### **5.2 The potential impact of future regulatory actions**

Some of the technological options mentioned above will be financially attractive to the shipowners, and theoretically there should be no need for regulatory actions to put them to work. Other technologies need regulatory support in the form of regulations or incentives. Such policies could be categorized into market-based instruments, command-and-control instruments and voluntary measures (Table 2). Within these categories one could think of different concrete instruments and ways of benchmarking environmental performance. Currently benchmarks like the Energy Efficiency Operational index (EEOI) and the Eneregy Efficiency Design Index (EEDI) are candidates for benchmarking the CO2 emissions of vessels within the IMO discussions.


\* METS - Maritime emissions trading scheme.

† ICF - International Compensation Funf

Table 2. Overview of policies to limit or reduce emissions of greenhouse gases from ships Source: IMO (IMO 2009)

As noted above, such global regulatory regimes are to a large extent dependent on achieving consensus among many nations which makes the international regulatory regime related to shipping more sluggish than the equivalent regimes applied to land based modes. This is one of the reasons why the EU "threatens" to take unilateral action by including CO2 emissions from international shipping into the EU trading regime for CO2-quotas.

#### **6. Conclusions**

18 Environmental Health – Emerging Issues and Practice

The Second IMO Greenhouse Gas Study (IMO 2009) points to a number of technology-based options for improving the energy efficiency of vessels. Partly these are related to improved design (concepts, hull and superstructure, power and propulsion systems) and improved operations (fleet management, logistics, incentives, voyage optimization, energy management). The combined potential for reductions of CO2 emission from these technologies is estimated to be between 25 and 75%. To reach the upper bound of this range,

Some of the technological options mentioned above will be financially attractive to the shipowners, and theoretically there should be no need for regulatory actions to put them to work. Other technologies need regulatory support in the form of regulations or incentives. Such policies could be categorized into market-based instruments, command-and-control instruments and voluntary measures (Table 2). Within these categories one could think of different concrete instruments and ways of benchmarking environmental performance. Currently benchmarks like the Energy Efficiency Operational index (EEOI) and the Eneregy Efficiency Design Index (EEDI) are candidates for benchmarking the CO2 emissions of

Table 2. Overview of policies to limit or reduce emissions of greenhouse gases from ships

emissions from international shipping into the EU trading regime for CO2-quotas.

As noted above, such global regulatory regimes are to a large extent dependent on achieving consensus among many nations which makes the international regulatory regime related to shipping more sluggish than the equivalent regimes applied to land based modes. This is one of the reasons why the EU "threatens" to take unilateral action by including CO2

reductions in operating speed would be necessary.

vessels within the IMO discussions.

\* METS - Maritime emissions trading scheme. † ICF - International Compensation Funf

Source: IMO (IMO 2009)

**5.2 The potential impact of future regulatory actions** 

Our case studies illustrate that short sea shipping operations, represented by RoRo and container services may very well deserve their "green label" when compared to alternative modes with respect to CO2 emissions. This conclusion is valid under what we consider realistic operating environments with respect to vessel operation speeds and achieved load factors, and when the shipping leg is not much longer than the distances of the land based modes. This conclusion holds at least for the container vessels, but the advantage of RoRo operations versus truck transport may be marginal – and is highly dependent on the prevailing market situation and the resulting load factors achieved.

The short sea shipping alternative does generally *not* deserve a "green label" when SO2, NOX and PM emissions are considered. Although some improvements are in the pipeline through the stricter Marpol Annex VI regulations, the maritime transport alternatives will still not be able to compete with road transport along these dimensions unless new fuels (LNG) are introduced or abatement technologies are installed.

We have applied quite large feeder vessels, a 10 000 dwt RoRo vessel and a 13 000 dwt container vessel, in our case studies. Smaller vessels will generally yield higher emissions per tonne-km and may therefore be less competitive.

We have illustrated that the use of realistic load factors is crucial in a comparative analysis like this. Applying load factors related to the cargo *capacity* of the vessel measured in tonnes will not yield a realistic setting, especially for RoRo vessels. All emissions should be attributed to the net cargo transported – as is the intention of IMOs proposed Energy Efficiency Operational Index (EEOL).

The recent work of the IMO MEPC points out that there is a very significant potential for reductions of CO2 emissions from ships – but that many of the possible technological and organizational measures are dependent on efficient policy regimes to come into play.

#### **7. Acknowledgements**

This paper is partly financed through the Northern Maritime University (NMU) project. The NMU project is partly funded by the EC Interreg IVB programme.

An earlier version of this paper was presented at IAME 2010 in Lisbon, Portugal.

#### **8. References**

Andre, J.-M. (2005). "Vehicle emission measurement collection of the ARTEMIS datatbase." Retrieved Feb 23, 2010, from

 http://www.inrets.fr/ur/lte/publications/publications-pdf/Joumard/ A3312reportJMALTE0504.pdf


**2** 

**Speciation Methods for the Determination** 

**the Freshwater and Marine Environments** 

Peter P. Ndibewu, Rob I. McCrindle and Ntebogeng S. Mokgalaka

*Tshwane University of Technology,* 

*South Africa*

**of Organotins (OTs) and Heavy Metals (MHs) in** 

Our primary goal for the development of analytical methods is their application in environmental monitoring to achieve good assessment of the contamination situation in freshwater and marine environments. As clearly stated in the endocrine disrupting contaminant (EDCs) program strategic plan for health related water issues (HRWI) of the Republic of South Africa (Version 1.2B, 7/02/2001), one of the objectives in the water research field is to protect aquatic ecosystems and human health based on sound science and defensible data through developing and validation of appropriate methods and by investigating the sources, persistence and effects of potential EDCs in water to support the risk assessment process and contribute towards a trustworthy environmental policy for

Research goals around the globe in this area have focused on the development of speciation methods for the determination of organotins and heavy metal pollutants in both freshwater and marine environments. Research and development over the years has provided reliable and sensitive analytical techniques that can be used for water research analysis, monitoring and health risks assessment including sampling, testing and validation, although some challenges still exist in regard to the availability of efficient and cost-effective sampling

The procedures for method modification and development vary depending on the properties of the chemical, possible interferences, the desired sampling medium, the desired analytical technique, sensitivity required, and similar factors (Ombaba and Barry, 1992). The following are questions, which have to be considered and answered by any method

Is the detection limit sufficiently low to provide meaningful data, especially when

 Will expected interferences produce false positive, false negative or biased results? If possible, can the results be verified by comparison with an accepted procedure?

 Can the analyte be collected by and removed from the sampling media? What are the collection and recovery factors and are they acceptable?

adjusted for collection and recovery factors?

**1. Introduction** 

techniques.

endocrine disrupting contaminants.

modification or development:

IMO (2009). Second IMO GHG Study 2009. London, IMO.


### **Speciation Methods for the Determination of Organotins (OTs) and Heavy Metals (MHs) in the Freshwater and Marine Environments**

Peter P. Ndibewu, Rob I. McCrindle and Ntebogeng S. Mokgalaka *Tshwane University of Technology, South Africa*

#### **1. Introduction**

20 Environmental Health – Emerging Issues and Practice

Knörr, W. (2008). EcoTransIT: Ecological Transport Information Tool. Environmental

NTM Working Group Goods and Logistics (2008). Environmental data for international

Sandvik, E. T. (2005). Environmental impacts of intermodal freight transport. MFM Report.

Whall, C., D. Cooper, et al. (2002). Quantification of emissions from ships associated with

NTM (2009). "The NTM methodology in brief." Retrieved February 23, 2010, from

Methodology and Date. Update 2008. Heidelberg, Germany, Institut für Energie

cargo sea transport. Calculation methods, emission factors, mode-specific issues.

ship movements between ports in the European Community. Northwich, UK,

IMO (2009). Second IMO GHG Study 2009. London, IMO.

und Umweltforschung Heidelberg GmbH.

http://www.ntm.a.se/english/eng-index.asp.

Molde, Møreforsking Molde: 40 s.

Stockholm.

Entec UK Limited.

Our primary goal for the development of analytical methods is their application in environmental monitoring to achieve good assessment of the contamination situation in freshwater and marine environments. As clearly stated in the endocrine disrupting contaminant (EDCs) program strategic plan for health related water issues (HRWI) of the Republic of South Africa (Version 1.2B, 7/02/2001), one of the objectives in the water research field is to protect aquatic ecosystems and human health based on sound science and defensible data through developing and validation of appropriate methods and by investigating the sources, persistence and effects of potential EDCs in water to support the risk assessment process and contribute towards a trustworthy environmental policy for endocrine disrupting contaminants.

Research goals around the globe in this area have focused on the development of speciation methods for the determination of organotins and heavy metal pollutants in both freshwater and marine environments. Research and development over the years has provided reliable and sensitive analytical techniques that can be used for water research analysis, monitoring and health risks assessment including sampling, testing and validation, although some challenges still exist in regard to the availability of efficient and cost-effective sampling techniques.

The procedures for method modification and development vary depending on the properties of the chemical, possible interferences, the desired sampling medium, the desired analytical technique, sensitivity required, and similar factors (Ombaba and Barry, 1992). The following are questions, which have to be considered and answered by any method modification or development:


Speciation Methods for the Determination of Organotins (OTs)

Shrader *et al.*, 1983; Leal *et al.*, 1995).

and marine structures (Brian, 1991).

at concentrations of a few ng L-1 (Fent, 1996; Fatoki *et al.*, 2000).

(Leal *et al.*, 1995).

and Heavy Metals (HMs) in the Freshwater and Marine Environments 23

Although the total concentration of an element is still useful to know, and sometimes essential, the determination of species is necessary to fully understand the biogeochemical and toxicological behavior of the metals. Pollution can influence aquatic life, either directly or indirectly in several ways. By pH changes (increase in acidity); decreasing dissolved oxygen (most common index for pollution); toxicity; mechanical injury to gills (for example from silt); thermal change of medium; killing food organisms through pH change or thermal changes; destruction of spawning grounds (FAO, 1978); shell malformation in oysters (Bayona and Cai, 1994), imposex in gastropods (Kuballa *et al*., 1995); mortality of the larvae of mussels (Jiang *et al.*, 2000) and fish poisoning (Cai and Bayona, 1995; Stab *et al*., 1992;

More specifically, maritime and coastal areas, as well as freshwater are definitely amongst today`s prominent endangered ecosystems. Industrialization and other human activities have caused major changes in these reservoirs` water quality, both inland and marine (Leal *et al.*, 1995). Dumping at sea and maritime-based transport activities are mostly responsible for this problem. Polluting loads emptied into the aquatic environments are of various nature and types depending on the point or non-point source though some, like the heavy metals, occur naturally (Lucinda *et al*., 1983; Fatoki *et al*., 2000; Ndibewu *et al.*, 2002). The point sources are essentially discharges of sewages and industrial effluents, and are easily identifiable and controllable (Maenpa *et al.*, 2002; Lucinda *et al.*, 1983). The non-point sources arise in part from natural phenomena, for example, soil erosion; irrigation return flows; outflow from fish farms, and are often diffuse, and so difficult to identify and to control

**3. Occurrence and ecotoxicity of heavy metals, TBT and other organotins** 

Unlike methyltin, which may be formed naturally in the environment, TBT is exclusively of anthropogenic origin (MCkie, 1987; Fent, 1996). This is why its occurrence in the aquatic environment has been directly attributed to its application as an antifouling agent. TBT residues in the sediments of harbors, marinas and shipping channels has been found to be considerably higher typically in the range of about 200 – 1000 g kg-1 (Balls, 1987). Progressive introduction of organic groups at the tin atom produces increasing biological activity (Bayona and Cai, 1994). Organotin compounds with three alkyl groups attached to the tin atom, such as tributyltin (TBT), triphenyltin and tricyclohexyltin, have found wide applications as antifouling agents in marine paints formulations, bactericides in cooling water (MCkie, 1987; Fatoki, 2000), agricultural fungicides and acaricides (Leal *et al.*, 1995). The most import of these is TBT, which is used in marine paints as an effective means of the growth of fouling organisms such as tubeworms, barnacles and mussels on seafaring vessels

Meech *et al.* (1998) has shown that TBT is acutely toxic to a variety of fresh water species at concentrations down to 0.1 L-1. TBT is particularly toxic (Fent, 1996; Reisch, 1996; Meech *et al.*, 1998) to mollusks (oysters) and gastropods. The decline of dog whelk populations on various coasts of France and UK has been attributed to the occurrence of TBT in these waters (Fent, 1996). Chronic toxic effects on oysters in the form of shell deformation (Fent, 1996) and marine gastropods in the form of sterilization of females have been reported occurring

This work is partitioned into two sub-sections covering the organotins (OTs) and the heavy metals which are toxic, (TMs) and in most cases are carcinogens. In the heavy metals group, only a few of them, known to cause serious health hazards are fully discussed. These are mercury (Hg), cadmium (Cd), arsenic (As), lead (Pb) and zinc (Zn), all known as endocrine disrupting contaminants (EDCs) (Fatoki and Ngassoum, 2000; HRWI, 2001; Ndibewu *et al.*, 2002). Other toxic metals including chromium (Cr) and vanadium (V) will be briefly mentioned in our discussion.

The first part of this chapter discusses speciation analysis of organotins by liquid-liquid (Espadaler et *al*., 1997; Jiang et *al*., 2000; Mueller, 1984) and microsolid phase extraction methods (Mueller, 1987) followed by sodium tetrahydroborate (Jiang et *al*., 2000), sodium tetraethylborate (Cai and Bayona, 1995; Thomaidis et *al*., 2001; Ceulemans and Adams, 1995; Pereira et *al*., 1999) and the Grignard`s reagents (Chau et *al*., 1996; Ceulemans and Adams, 1995; Krull et *al*., 1985; Lucinda, 1983) derivatization. Separation and detection is usually accomplished using the GC-FPD/GC-AAS techniques (Fatoki et *al.*, 2000). In the liquidliquid extraction phase, solvents such as tropolone, hexane-soxhlet and/or diethyl ether have been used for water (Fatoki and Ngassoum, 2000; Mueller, 1987; Leal et *al*., 1995; Abalos et *al*., 1997), sediment samples (Fatoki et *al*., 2000; Abalos et *al*., 1997; Krull et *al*., 1985) and the biota (Kan-atireklap et *al*., 1998). In the derivatization step, two techniques have been used. Firstly, hybridization reactions using sodium tetrahydroborate as the reagent (Abalos et *al*., 1997) was used. Alternatively, derivatization technique based on alkylation reactions employ two reagents namely: the Grignard`s reagents (methylation or ethylation) and sodium tetraethylborate (Fatoki and Ngassoum, 2000; Cai and Bayona, 1995). While the GC-FPD (Fatoki *et al*., 2000; Richardson and Gangolli, 1994) and GC-AAS techniques (Fatoki *et al*., 2000) can be used for the speciation of various organotins compounds, elemental Tin (Sn) is analyzed using flame AAS (Quevauviller *et al*., 1989) in water and sediment samples and the biota.

For the determination of cadmium (Cd), mercury (Hg), arsenic (As), and zinc (Zn), while an ion chromatography-hydride generation-atomic absorption (HG-AAS) procedure (Wade *et al*., 1988) has been used for speciation of As, Cd and Zn are usually determined using flame AAS spectrometry (Lucinda *et al*., 1983; Maenpa *et al*., 2002), and Hg analyzed using the cold vapor technique (CVAAS) (Shrader *et al*., 1983; Willis, 1965). More recently, Fatoki *et al*. (2000) has used GC-FPD for the determination of tributyltin concentrations in the coastal water and freshwater sediments from both the Port Elizabeth and East London harbors in South Africa, which contributed to resources for building regulatory data in that part of the world.

#### **2. Background**

Aquatic pollution is a major cause in the decline of resources from water. It is, thus, important to monitor the condition of water. A major concern is the need to develop accurate, reliable and efficient speciation methods for the determination of the polluting compounds within ultra-low detectable ranges. Those known so far to be particularly toxic to the aquatic ecosystems are the organotins (Fent, 1996; Mueller, 1987) and the heavy metals (Cai and Bayona, 1995; Lucinda *et al*., 1983). The term "speciation" in analytical chemistry refers to the separation and quantification of the different oxidation states or chemical forms of a particular element (http://www.frontiergeosciences.com/ebru/).

This work is partitioned into two sub-sections covering the organotins (OTs) and the heavy metals which are toxic, (TMs) and in most cases are carcinogens. In the heavy metals group, only a few of them, known to cause serious health hazards are fully discussed. These are mercury (Hg), cadmium (Cd), arsenic (As), lead (Pb) and zinc (Zn), all known as endocrine disrupting contaminants (EDCs) (Fatoki and Ngassoum, 2000; HRWI, 2001; Ndibewu *et al.*, 2002). Other toxic metals including chromium (Cr) and vanadium (V) will be briefly

The first part of this chapter discusses speciation analysis of organotins by liquid-liquid (Espadaler et *al*., 1997; Jiang et *al*., 2000; Mueller, 1984) and microsolid phase extraction methods (Mueller, 1987) followed by sodium tetrahydroborate (Jiang et *al*., 2000), sodium tetraethylborate (Cai and Bayona, 1995; Thomaidis et *al*., 2001; Ceulemans and Adams, 1995; Pereira et *al*., 1999) and the Grignard`s reagents (Chau et *al*., 1996; Ceulemans and Adams, 1995; Krull et *al*., 1985; Lucinda, 1983) derivatization. Separation and detection is usually accomplished using the GC-FPD/GC-AAS techniques (Fatoki et *al.*, 2000). In the liquidliquid extraction phase, solvents such as tropolone, hexane-soxhlet and/or diethyl ether have been used for water (Fatoki and Ngassoum, 2000; Mueller, 1987; Leal et *al*., 1995; Abalos et *al*., 1997), sediment samples (Fatoki et *al*., 2000; Abalos et *al*., 1997; Krull et *al*., 1985) and the biota (Kan-atireklap et *al*., 1998). In the derivatization step, two techniques have been used. Firstly, hybridization reactions using sodium tetrahydroborate as the reagent (Abalos et *al*., 1997) was used. Alternatively, derivatization technique based on alkylation reactions employ two reagents namely: the Grignard`s reagents (methylation or ethylation) and sodium tetraethylborate (Fatoki and Ngassoum, 2000; Cai and Bayona, 1995). While the GC-FPD (Fatoki *et al*., 2000; Richardson and Gangolli, 1994) and GC-AAS techniques (Fatoki *et al*., 2000) can be used for the speciation of various organotins compounds, elemental Tin (Sn) is analyzed using flame AAS (Quevauviller *et al*., 1989) in

For the determination of cadmium (Cd), mercury (Hg), arsenic (As), and zinc (Zn), while an ion chromatography-hydride generation-atomic absorption (HG-AAS) procedure (Wade *et al*., 1988) has been used for speciation of As, Cd and Zn are usually determined using flame AAS spectrometry (Lucinda *et al*., 1983; Maenpa *et al*., 2002), and Hg analyzed using the cold vapor technique (CVAAS) (Shrader *et al*., 1983; Willis, 1965). More recently, Fatoki *et al*. (2000) has used GC-FPD for the determination of tributyltin concentrations in the coastal water and freshwater sediments from both the Port Elizabeth and East London harbors in South Africa, which contributed to resources for building regulatory data in that part of the

Aquatic pollution is a major cause in the decline of resources from water. It is, thus, important to monitor the condition of water. A major concern is the need to develop accurate, reliable and efficient speciation methods for the determination of the polluting compounds within ultra-low detectable ranges. Those known so far to be particularly toxic to the aquatic ecosystems are the organotins (Fent, 1996; Mueller, 1987) and the heavy metals (Cai and Bayona, 1995; Lucinda *et al*., 1983). The term "speciation" in analytical chemistry refers to the separation and quantification of the different oxidation states or chemical forms of a particular element (http://www.frontiergeosciences.com/ebru/).

mentioned in our discussion.

water and sediment samples and the biota.

world.

**2. Background** 

Although the total concentration of an element is still useful to know, and sometimes essential, the determination of species is necessary to fully understand the biogeochemical and toxicological behavior of the metals. Pollution can influence aquatic life, either directly or indirectly in several ways. By pH changes (increase in acidity); decreasing dissolved oxygen (most common index for pollution); toxicity; mechanical injury to gills (for example from silt); thermal change of medium; killing food organisms through pH change or thermal changes; destruction of spawning grounds (FAO, 1978); shell malformation in oysters (Bayona and Cai, 1994), imposex in gastropods (Kuballa *et al*., 1995); mortality of the larvae of mussels (Jiang *et al.*, 2000) and fish poisoning (Cai and Bayona, 1995; Stab *et al*., 1992; Shrader *et al.*, 1983; Leal *et al.*, 1995).

More specifically, maritime and coastal areas, as well as freshwater are definitely amongst today`s prominent endangered ecosystems. Industrialization and other human activities have caused major changes in these reservoirs` water quality, both inland and marine (Leal *et al.*, 1995). Dumping at sea and maritime-based transport activities are mostly responsible for this problem. Polluting loads emptied into the aquatic environments are of various nature and types depending on the point or non-point source though some, like the heavy metals, occur naturally (Lucinda *et al*., 1983; Fatoki *et al*., 2000; Ndibewu *et al.*, 2002). The point sources are essentially discharges of sewages and industrial effluents, and are easily identifiable and controllable (Maenpa *et al.*, 2002; Lucinda *et al.*, 1983). The non-point sources arise in part from natural phenomena, for example, soil erosion; irrigation return flows; outflow from fish farms, and are often diffuse, and so difficult to identify and to control (Leal *et al.*, 1995).

#### **3. Occurrence and ecotoxicity of heavy metals, TBT and other organotins**

Unlike methyltin, which may be formed naturally in the environment, TBT is exclusively of anthropogenic origin (MCkie, 1987; Fent, 1996). This is why its occurrence in the aquatic environment has been directly attributed to its application as an antifouling agent. TBT residues in the sediments of harbors, marinas and shipping channels has been found to be considerably higher typically in the range of about 200 – 1000 g kg-1 (Balls, 1987). Progressive introduction of organic groups at the tin atom produces increasing biological activity (Bayona and Cai, 1994). Organotin compounds with three alkyl groups attached to the tin atom, such as tributyltin (TBT), triphenyltin and tricyclohexyltin, have found wide applications as antifouling agents in marine paints formulations, bactericides in cooling water (MCkie, 1987; Fatoki, 2000), agricultural fungicides and acaricides (Leal *et al.*, 1995). The most import of these is TBT, which is used in marine paints as an effective means of the growth of fouling organisms such as tubeworms, barnacles and mussels on seafaring vessels and marine structures (Brian, 1991).

Meech *et al.* (1998) has shown that TBT is acutely toxic to a variety of fresh water species at concentrations down to 0.1 L-1. TBT is particularly toxic (Fent, 1996; Reisch, 1996; Meech *et al.*, 1998) to mollusks (oysters) and gastropods. The decline of dog whelk populations on various coasts of France and UK has been attributed to the occurrence of TBT in these waters (Fent, 1996). Chronic toxic effects on oysters in the form of shell deformation (Fent, 1996) and marine gastropods in the form of sterilization of females have been reported occurring at concentrations of a few ng L-1 (Fent, 1996; Fatoki *et al.*, 2000).

Speciation Methods for the Determination of Organotins (OTs)

environment of the system (Connell *et al*., 1984).

and Heavy Metals (HMs) in the Freshwater and Marine Environments 25

alleviate teething pain in infants (Nriagu, 1996). Once emitted, metals can reside in the environment for hundreds of years or more (Nriagu, 1996), while causing immediate or long term damage depending on the concentration released. Evidence of human exploitation of heavy metals has been found in the ice cores in Greenland and seawater in the Antarctic (Nriagu, 1996). The lead contents of ice layers deposited annually in Greenland show a steady rise that parallels the mining renaissance in Europe, reaching values 100 times the natural background level in the mid-1990s (http://h2osparc.wq.ncsu.edu/info/hmetals.html). Mining itself, not only of heavy metals but also of coal and other minerals, is another major route of exposure. Despite some noted improvements in worker safety and cleaner production, mining remains one of the most hazardous and environmentally damaging industries (Nriagu, 1996; Maenpa *et al.*, 2002). In Bolivia, toxic sludge from a zinc mine in the Andes had killed aquatic life along a 300-kilometer stretch of river systems as of 1996 (http://h2osparc.wq.ncsu.edu/info/hmetals.html). It also threatened the livelihood and health of 50,000 of the region's subsistence farmers (Nriagu, 1996; Fent, 1996). Uncontrolled smelters have produced some of the world's only environmental "dead zones" where little or no vegetation survives. For instance, toxic emissions from the Sudbury, Ontario, and nickel smelter have devastated 10,400 hectares of forests downwind of the smelter (Nriagu, 1996). All heavy metals exist in surface waters in colloidal, particulate, and dissolved phases, although dissolved concentrations are generally low (Kennish, 1992). The colloidal and particulate metal may be found in (1) hydroxides, oxides, silicates, or sulfides; or (2) adsorbed to clay, silica, or organic matter. The soluble forms are generally ions or unionized organometallic chelates or complexes. The solubility of trace metals in surface waters is predominately controlled by the water pH, the type and concentration of ligands on which the metal could adsorb, and the oxidation state of the mineral components and the redox

Heavy metals in surface water systems can be from natural or anthropogenic sources. Currently, anthropogenic inputs of metals exceed natural inputs. Excess metal levels in surface water may pose a health risk to humans and to the environment (Nriagu, 1996). Considering that heavy metals are natural constituents of the Earth's crust, they are present in varying concentrations in all ecosystems and human activities have drastically changedthe biogeochemical cycles and balance of some of these heavy metals. The main anthropogenic sources of heavy metals are various industrial sources (Shrader *et al.*, 1983; Fatoki, 2000) including present and former mining activities (http://www.msceast.org/hms/), foundries and smelters (http://www.osha.gov/SLTC/cadmium/index.html), and diffuse sources such as piping (http://www.msceast.org/hms/), constituents of products, combustion byproducts, traffic (Shrader *et al.*, 1983), etc. Relatively, volatile heavy metals and those that become attached to airborne particles can be widely dispersed on very large scales. Heavy metals conveyed in aqueous and sedimentary transport enter the normal coastal biogeochemical cycle and are largely retained within near-shore and shelf regions (http://www.msceast.org/hms/). The toxicity of these metals has also been documented throughout history: Greek and Roman physicians diagnosed symptoms of acute lead poisoning long before toxicology became a science (Nriagu, (1996). Today, much more is known about the health effects of heavy metals. Exposure to heavy metals has been linked with developmental retardation, various cancers, kidney damage, and even death in some instances of exposure to very high concentrations. Exposure to high levels of mercury, gold,

Unlike TBT and organotins, metals are unique environmental and industrial pollutants in that they are found naturally distributed in all phases of the environment. The term "heavy metals" is generally interpreted to include those metals from periodic table groups IIA through VIA. The semi-metallic elements boron, arsenic, selenium, and tellurium are often included in this classification. At trace levels, many of these elements are necessary to support life. Heavy metals are elements having atomic weights between 63.546 and 200.590g (Kennish, 1992), and a specific gravity greater than 4.0 (Connell *et al*., 1984). Living organisms require trace amounts of some heavy metals, including cobalt, copper, iron, manganese, molybdenum, vanadium, strontium, and zinc (Nriagu, 1996). Excessive levels of essential metals, however, become toxic and may build up in biological systems, and become a significant health hazard (Brickman, 1978). Non-essential heavy metals of particular concern to surface water systems are cadmium (Cd), chromium (Cr), mercury (Hg), lead (Pb), arsenic (As) and antimony (Sb) (Kennish, 1992).

During the last two decades, considerable attention has been given to problems concerning negative effects of heavy metals (HMs) on various ecosystems in different environmental media (Lucinda, 1983; Nriagu, 1996). The heavy metals rated among most of the environmental risk pollutants (Cai and Bayona, 1995) requires that, fast, accurate and reliable analytical techniques suitable for their assessment and for their determination in environmental samples at trace levels be developed. In the class of the heavy metal ecotoxicants, Hg, Cd, As and Zn are considered fairly hazardous because of their high toxicity (Schrader *et al.*, 1983; Nriagu, 1996; Fent, 1996). These metal species actually occur in the environment at sub ultra-low trace concentrations level (Cai and Bayona, 1995; Lucinda *et al.,* 1983). Therefore, accurate and sensitive determination techniques are of fundamental interest for the assessment of the effectiveness of regulatory control measures.

Heavy metals are stable and persistent environmental contaminants since they cannot be degraded or destroyed. Therefore, they tend to accumulate in soils, seawater, freshwater, and sediments (Schrader *et al*., 1983; http://www.osha.gov/SLTC/cadmium/index.html). Excessive levels of metals in the marine environment can affect marine biota and pose risk to human consumers of seafood (http://www.msceast.org/hms/). Heavy metals are also known to have adverse effects on the environment and human health (Schrader *et al*., 1983). Numerous field observations also indicate a significant increase of HM concentrations in agricultural and forest soils as well as in marine and inland water sediments. This increase is frequently observed in remote areas thousands of kilometers away from major anthropogenic sources and can be explained by transboundary atmospheric long-range transport only (http://www.msceast.org/hms/). An assessment of the potential ecological and health risks associated with atmospheric fluxes of heavy metals requires an understanding of the relationships between sources of emission to the atmosphere and the levels of concentrations measured in ambient air and precipitation (Ikeda *et al*., 1996).

Since the industrial revolution, the production of heavy metals such as Pb, Cu, and Zn has increased exponentially (Lucinda, 1983; Maguire *et al.*, 1982). Between 1850 and 1990, production of these three metals increased nearly 10-fold, with emissions rising in tandem (Maguire *et al*., 1982). The heavy metals have been used in a variety of ways for at least 2 millennia (Lu *et al*., 1996; Lucinda, 1983; Meech *et al.*, 1998). For example, lead has been used in plumbing, and lead arsenate has been used to control insects in apple orchards. The Romans added lead to wine to improve its taste, and mercury was used as a salve to

Unlike TBT and organotins, metals are unique environmental and industrial pollutants in that they are found naturally distributed in all phases of the environment. The term "heavy metals" is generally interpreted to include those metals from periodic table groups IIA through VIA. The semi-metallic elements boron, arsenic, selenium, and tellurium are often included in this classification. At trace levels, many of these elements are necessary to support life. Heavy metals are elements having atomic weights between 63.546 and 200.590g (Kennish, 1992), and a specific gravity greater than 4.0 (Connell *et al*., 1984). Living organisms require trace amounts of some heavy metals, including cobalt, copper, iron, manganese, molybdenum, vanadium, strontium, and zinc (Nriagu, 1996). Excessive levels of essential metals, however, become toxic and may build up in biological systems, and become a significant health hazard (Brickman, 1978). Non-essential heavy metals of particular concern to surface water systems are cadmium (Cd), chromium (Cr), mercury

During the last two decades, considerable attention has been given to problems concerning negative effects of heavy metals (HMs) on various ecosystems in different environmental media (Lucinda, 1983; Nriagu, 1996). The heavy metals rated among most of the environmental risk pollutants (Cai and Bayona, 1995) requires that, fast, accurate and reliable analytical techniques suitable for their assessment and for their determination in environmental samples at trace levels be developed. In the class of the heavy metal ecotoxicants, Hg, Cd, As and Zn are considered fairly hazardous because of their high toxicity (Schrader *et al.*, 1983; Nriagu, 1996; Fent, 1996). These metal species actually occur in the environment at sub ultra-low trace concentrations level (Cai and Bayona, 1995; Lucinda *et al.,* 1983). Therefore, accurate and sensitive determination techniques are of fundamental

Heavy metals are stable and persistent environmental contaminants since they cannot be degraded or destroyed. Therefore, they tend to accumulate in soils, seawater, freshwater, and sediments (Schrader *et al*., 1983; http://www.osha.gov/SLTC/cadmium/index.html). Excessive levels of metals in the marine environment can affect marine biota and pose risk to human consumers of seafood (http://www.msceast.org/hms/). Heavy metals are also known to have adverse effects on the environment and human health (Schrader *et al*., 1983). Numerous field observations also indicate a significant increase of HM concentrations in agricultural and forest soils as well as in marine and inland water sediments. This increase is frequently observed in remote areas thousands of kilometers away from major anthropogenic sources and can be explained by transboundary atmospheric long-range transport only (http://www.msceast.org/hms/). An assessment of the potential ecological and health risks associated with atmospheric fluxes of heavy metals requires an understanding of the relationships between sources of emission to the atmosphere and the levels of concentrations measured in ambient air and precipitation (Ikeda *et al*., 1996).

Since the industrial revolution, the production of heavy metals such as Pb, Cu, and Zn has increased exponentially (Lucinda, 1983; Maguire *et al.*, 1982). Between 1850 and 1990, production of these three metals increased nearly 10-fold, with emissions rising in tandem (Maguire *et al*., 1982). The heavy metals have been used in a variety of ways for at least 2 millennia (Lu *et al*., 1996; Lucinda, 1983; Meech *et al.*, 1998). For example, lead has been used in plumbing, and lead arsenate has been used to control insects in apple orchards. The Romans added lead to wine to improve its taste, and mercury was used as a salve to

interest for the assessment of the effectiveness of regulatory control measures.

(Hg), lead (Pb), arsenic (As) and antimony (Sb) (Kennish, 1992).

alleviate teething pain in infants (Nriagu, 1996). Once emitted, metals can reside in the environment for hundreds of years or more (Nriagu, 1996), while causing immediate or long term damage depending on the concentration released. Evidence of human exploitation of heavy metals has been found in the ice cores in Greenland and seawater in the Antarctic (Nriagu, 1996). The lead contents of ice layers deposited annually in Greenland show a steady rise that parallels the mining renaissance in Europe, reaching values 100 times the natural background level in the mid-1990s (http://h2osparc.wq.ncsu.edu/info/hmetals.html). Mining itself, not only of heavy metals but also of coal and other minerals, is another major route of exposure. Despite some noted improvements in worker safety and cleaner production, mining remains one of the most hazardous and environmentally damaging industries (Nriagu, 1996; Maenpa *et al.*, 2002). In Bolivia, toxic sludge from a zinc mine in the Andes had killed aquatic life along a 300-kilometer stretch of river systems as of 1996 (http://h2osparc.wq.ncsu.edu/info/hmetals.html). It also threatened the livelihood and health of 50,000 of the region's subsistence farmers (Nriagu, 1996; Fent, 1996). Uncontrolled smelters have produced some of the world's only environmental "dead zones" where little or no vegetation survives. For instance, toxic emissions from the Sudbury, Ontario, and nickel smelter have devastated 10,400 hectares of forests downwind of the smelter (Nriagu, 1996). All heavy metals exist in surface waters in colloidal, particulate, and dissolved phases, although dissolved concentrations are generally low (Kennish, 1992). The colloidal and particulate metal may be found in (1) hydroxides, oxides, silicates, or sulfides; or (2) adsorbed to clay, silica, or organic matter. The soluble forms are generally ions or unionized organometallic chelates or complexes. The solubility of trace metals in surface waters is predominately controlled by the water pH, the type and concentration of ligands on which the metal could adsorb, and the oxidation state of the mineral components and the redox environment of the system (Connell *et al*., 1984).

Heavy metals in surface water systems can be from natural or anthropogenic sources. Currently, anthropogenic inputs of metals exceed natural inputs. Excess metal levels in surface water may pose a health risk to humans and to the environment (Nriagu, 1996). Considering that heavy metals are natural constituents of the Earth's crust, they are present in varying concentrations in all ecosystems and human activities have drastically changedthe biogeochemical cycles and balance of some of these heavy metals. The main anthropogenic sources of heavy metals are various industrial sources (Shrader *et al.*, 1983; Fatoki, 2000) including present and former mining activities (http://www.msceast.org/hms/), foundries and smelters (http://www.osha.gov/SLTC/cadmium/index.html), and diffuse sources such as piping (http://www.msceast.org/hms/), constituents of products, combustion byproducts, traffic (Shrader *et al.*, 1983), etc. Relatively, volatile heavy metals and those that become attached to airborne particles can be widely dispersed on very large scales. Heavy metals conveyed in aqueous and sedimentary transport enter the normal coastal biogeochemical cycle and are largely retained within near-shore and shelf regions (http://www.msceast.org/hms/). The toxicity of these metals has also been documented throughout history: Greek and Roman physicians diagnosed symptoms of acute lead poisoning long before toxicology became a science (Nriagu, (1996). Today, much more is known about the health effects of heavy metals. Exposure to heavy metals has been linked with developmental retardation, various cancers, kidney damage, and even death in some instances of exposure to very high concentrations. Exposure to high levels of mercury, gold,

Speciation Methods for the Determination of Organotins (OTs)

by the following mechanisms:

**3.1 Environmental effects** 

metal ions into the overlying water.

inputs).

and Heavy Metals (HMs) in the Freshwater and Marine Environments 27

cadmium toxicity may result in brain damage. Metal fume fever may result from acute exposure with flu-like symptoms of weakness, fever, headache, chills, sweating and muscular pain. Acute pulmonary edema usually develops within 24 hours and reaches a maximum by three days. If death from asphyxia does not occur, symptoms may resolve within a week. Chronic cadmium poisoning can cause eventual death. The most serious consequence of chronic cadmium poisoning is cancer (lung and prostate). The first observed chronic effect is generally kidney damage, manifested by excretion of excessive (low molecular weight) protein in the urine. Cadmium also is believed to cause pulmonary emphysema and bone disease (osteomalcia and osteoporosis). The latter has been observed in Japan ("itai-itai" disease) where residents were exposed to cadmium in rice crops irrigated with cadmium-contaminated water. Cadmium may also cause anemia, teeth discoloration (Cd forms CdS) and loss of smell (anosmia) (http://www.osha.gov/SLTC/cadmium/index.html). Arsenic ingestion can cause severe toxicity through ingestion of contaminated food and water. Ingestion causes

vomiting, diarrhea, and cardiac abnormalities (Viessman and Hammer, 1985).

The behavior of metals in natural waters is a function of the substrate sediment composition, the suspended sediment composition, and the water chemistry (Nriagu, 1996). Sediment composed of fine sand and silt will generally have higher levels of adsorbed metal than will quartz, feldspar and carbonate-rich sediment. Metals also have a high affinity for humic acids, organo-clays, and oxides coated with organic matter (Connell *et al*., 1984). The water chemistry of the system controls the rate of adsorption and desorption of metals to and from sediment. Adsorption removes the metal from the water column and stores the metal in the substrate. Desorption returns the metal to the water column, where recirculation and bioassimilation may take place. Metals may be desorbed from the sediment if the water experiences increase in salinity, decreases in redox potential, or decreases in pH controlled

 Salinity increase: Elevated salt concentrations create increased competition between cations and metals for binding sites. Often, metals will be driven off into the overlying water. (Estuaries are prone to this phenomenon because of fluctuating river flow

 Redox potential decrease: A decreased redox potential, as is often seen under oxygen deficient conditions, will change the composition of metal complexes and release the

 pH decrease: A lower pH increases the competition between metal and hydrogen ions for binding sites. A decrease in pH may also dissolve metal-carbonate complexes,

Aquatic organisms may be adversely affected by heavy metals in the environment. The toxicity is largely a function of the water chemistry and sediment composition in the surface water system, as clearly detailed under the section "Environmental fate/Mode of transport". Slightly elevated metal levels in natural waters may cause the following sublethal effects in aquatic organisms: (1) histological or morphological change in tissues; (2) changes in physiology, such as suppression of growth and development, poor swimming performance,

releasing free metal ions into the water column (Connell *et al.*, 1984).

and lead has also been associated with the development of autoimmunity, in which the immune system starts to attack its own cells, mistaking them for foreign invaders (Nriagu, 1996; http://www.mercurypolicy.org/). Autoimmunity can lead to the development of diseases of the joints and kidneys, such as rheumatoid arthritis, or diseases of the circulatory or central nervous systems. Despite abundant evidence of these deleterious health effects, exposure to heavy metals continues and may increase in the absence of concerted policy actions. Mercury is still extensively used in gold mining in many parts of Latin America. Arsenic, along with copper and chromium compounds, is a common ingredient in wood preservatives. Lead is still widely used as an additive in gasoline. Increased use of coal in the future will increase metal exposures because coal ash contains many toxic metals and can be breathed deeply into the lungs. For countries such as China, India and South Africa, which continue to rely on high-ash coal as a primary energy source, the health implications are ominous.

Mercury is a toxic metal that is liquid at room temperature (http://h2osparc.wq.ncsu.edu/info/hmetals.htm). Exposure to mercury is known to cause permanent damage to the brain, nervous system, and kidneys (http://www.mercurypolicy.org). Pregnant women are particularly vulnerable as mercury may damage the developing fetus. While mercury is released naturally from rocks, soil, and volcanoes, human activities have boosted atmospheric levels to some three times above preindustrial levels, the experts say. Estimates vary, but the UNEP group of experts says some 5,000 to 10,000 metric tons of mercury are thought to enter the atmosphere every year and 50 to 75 percent of it from human activities (http://h2osparc.wq.ncsu.edu/info/hmetals.htm). The main human source of mercury emissions is coal combustion from electrical power plants and industrial, commercial and residential burners. Other sources include municipal solid waste incineration, mining of non-ferrous metals, and artisanal gold mining (http://www.mercurypolicy.org).

Interest in the biogeochemical cycle (Shrader *et al.*, 1983) of mercury in the environment has dramatically increased in recent years due to the observation that mercury accumulates in aquatic organisms. Moreover, methylmercury becomes magnified in the upper tropic levels as a result of bioaccumulation, from dietary intake of organisms containing methylmercury (http://www.frontiergeosciences.com/ebru/). It has been demonstrated that mercury can be methylated in the environment and bioconcentrated in the biota (Cai and Bayona, 1995). Ingestion of fish muscle is an important exposure pathway of mercury to humans (Cai and Bayona, 1995). Studying mercury in environmental systems requires a very sensitive method as typical mercury levels in aquatic environments range from 0.5 to 5.0 ng L-1 (http://www.frontiergeosciences.com/ebru/). Total mercury permissible in the environment is 0.005 mg L-1 (FAO, 1978). The high toxicity of methylmercury has been well recognized in fish (Cai and Bayona, 1995; Lucinda *et al.*, 1983; Wade *et al.*, 1988) and ingestion of fish muscle is an important exposure pathway of mercury for humans (Cai and Bayona, 1995).

Cadmium may interfere with the metallothionein's ability to regulate zinc and copper concentrations in the body. Metallothionein is a protein that binds to excess essential metals to render them unavailable when cadmium induces metallothionein's activity binding it to copper and zinc, disrupting the homeostasis levels (Kennish, 1992). Cadmium is used in industrial manufacturing processes and is a byproduct of the metallurgy of zinc. Acute

and lead has also been associated with the development of autoimmunity, in which the immune system starts to attack its own cells, mistaking them for foreign invaders (Nriagu, 1996; http://www.mercurypolicy.org/). Autoimmunity can lead to the development of diseases of the joints and kidneys, such as rheumatoid arthritis, or diseases of the circulatory or central nervous systems. Despite abundant evidence of these deleterious health effects, exposure to heavy metals continues and may increase in the absence of concerted policy actions. Mercury is still extensively used in gold mining in many parts of Latin America. Arsenic, along with copper and chromium compounds, is a common ingredient in wood preservatives. Lead is still widely used as an additive in gasoline. Increased use of coal in the future will increase metal exposures because coal ash contains many toxic metals and can be breathed deeply into the lungs. For countries such as China, India and South Africa, which continue to rely on high-ash coal as a primary energy source, the health implications

Mercury is a toxic metal that is liquid at room temperature (http://h2osparc.wq.ncsu.edu/info/hmetals.htm). Exposure to mercury is known to cause permanent damage to the brain, nervous system, and kidneys (http://www.mercurypolicy.org). Pregnant women are particularly vulnerable as mercury may damage the developing fetus. While mercury is released naturally from rocks, soil, and volcanoes, human activities have boosted atmospheric levels to some three times above preindustrial levels, the experts say. Estimates vary, but the UNEP group of experts says some 5,000 to 10,000 metric tons of mercury are thought to enter the atmosphere every year and 50 to 75 percent of it from human activities (http://h2osparc.wq.ncsu.edu/info/hmetals.htm). The main human source of mercury emissions is coal combustion from electrical power plants and industrial, commercial and residential burners. Other sources include municipal solid waste incineration, mining of non-ferrous metals, and artisanal gold mining

Interest in the biogeochemical cycle (Shrader *et al.*, 1983) of mercury in the environment has dramatically increased in recent years due to the observation that mercury accumulates in aquatic organisms. Moreover, methylmercury becomes magnified in the upper tropic levels as a result of bioaccumulation, from dietary intake of organisms containing methylmercury (http://www.frontiergeosciences.com/ebru/). It has been demonstrated that mercury can be methylated in the environment and bioconcentrated in the biota (Cai and Bayona, 1995). Ingestion of fish muscle is an important exposure pathway of mercury to humans (Cai and Bayona, 1995). Studying mercury in environmental systems requires a very sensitive method as typical mercury levels in aquatic environments range from 0.5 to 5.0 ng L-1 (http://www.frontiergeosciences.com/ebru/). Total mercury permissible in the environment is 0.005 mg L-1 (FAO, 1978). The high toxicity of methylmercury has been well recognized in fish (Cai and Bayona, 1995; Lucinda *et al.*, 1983; Wade *et al.*, 1988) and ingestion of fish muscle is an important exposure pathway of mercury for humans (Cai and

Cadmium may interfere with the metallothionein's ability to regulate zinc and copper concentrations in the body. Metallothionein is a protein that binds to excess essential metals to render them unavailable when cadmium induces metallothionein's activity binding it to copper and zinc, disrupting the homeostasis levels (Kennish, 1992). Cadmium is used in industrial manufacturing processes and is a byproduct of the metallurgy of zinc. Acute

are ominous.

Bayona, 1995).

(http://www.mercurypolicy.org).

cadmium toxicity may result in brain damage. Metal fume fever may result from acute exposure with flu-like symptoms of weakness, fever, headache, chills, sweating and muscular pain. Acute pulmonary edema usually develops within 24 hours and reaches a maximum by three days. If death from asphyxia does not occur, symptoms may resolve within a week. Chronic cadmium poisoning can cause eventual death. The most serious consequence of chronic cadmium poisoning is cancer (lung and prostate). The first observed chronic effect is generally kidney damage, manifested by excretion of excessive (low molecular weight) protein in the urine. Cadmium also is believed to cause pulmonary emphysema and bone disease (osteomalcia and osteoporosis). The latter has been observed in Japan ("itai-itai" disease) where residents were exposed to cadmium in rice crops irrigated with cadmium-contaminated water. Cadmium may also cause anemia, teeth discoloration (Cd forms CdS) and loss of smell (anosmia) (http://www.osha.gov/SLTC/cadmium/index.html). Arsenic ingestion can cause severe toxicity through ingestion of contaminated food and water. Ingestion causes vomiting, diarrhea, and cardiac abnormalities (Viessman and Hammer, 1985).

The behavior of metals in natural waters is a function of the substrate sediment composition, the suspended sediment composition, and the water chemistry (Nriagu, 1996). Sediment composed of fine sand and silt will generally have higher levels of adsorbed metal than will quartz, feldspar and carbonate-rich sediment. Metals also have a high affinity for humic acids, organo-clays, and oxides coated with organic matter (Connell *et al*., 1984). The water chemistry of the system controls the rate of adsorption and desorption of metals to and from sediment. Adsorption removes the metal from the water column and stores the metal in the substrate. Desorption returns the metal to the water column, where recirculation and bioassimilation may take place. Metals may be desorbed from the sediment if the water experiences increase in salinity, decreases in redox potential, or decreases in pH controlled by the following mechanisms:


#### **3.1 Environmental effects**

Aquatic organisms may be adversely affected by heavy metals in the environment. The toxicity is largely a function of the water chemistry and sediment composition in the surface water system, as clearly detailed under the section "Environmental fate/Mode of transport". Slightly elevated metal levels in natural waters may cause the following sublethal effects in aquatic organisms: (1) histological or morphological change in tissues; (2) changes in physiology, such as suppression of growth and development, poor swimming performance,

Speciation Methods for the Determination of Organotins (OTs)

**3.3 Health effects** 

teratogenesis (USEPA, 1987).

food chain.

and Heavy Metals (HMs) in the Freshwater and Marine Environments 29

Ingestion of metals such as lead (Pb), cadmium (Cd), mercury (Hg), arsenic (As), barium (Ba), and chromium (Cr), may pose great risks to human health. Trace metals such as lead and cadmium will interfere with essential nutrients of similar appearance, such as calcium (Ca2+) and zinc (Zn2+). Amongst the heavy metals pollution, mercury pollution has become a global problem (Schrader *et al.*, 1983) because of its occurrence from natural anthropogenic sources, and its biogeochemical processes (Cai and Bayona, 1995; Coello *et al*., 1996). As public awareness regarding the toxicity and the environmental impact of mercury contamination increases, speciation analytical methods developed, are required to distinguish between organic and inorganic forms of mercury. The determination and monitoring of mercury and arsenic is a special concern in the field of mine works and food engineering respectfully (Leal *et al.*, 1995; Nriagu, 1996). It has been reported (Leal *et al.*, 1995; Fatoki, 2000) that mercury can be methylated in the environment and bioconcentrated in the biota. Mercury poses a great risk to humans, especially in the form of methylmercury. When mercury enters water, it is often transformed by microorganisms into the toxic methyl mercury form. Symptoms of acute poisoning are pharyngitis, gasteroenteritis, vomiting, nephritis, hepatitis, and circulatory collapse. Chronic poisoning is usually a result of industrial exposure or a diet consisting of contaminated fish (mercury is the only metal that will bioaccumulate). Chronic poisoning may cause liver damage, neural damage, and

The Global mercury assessment working group of the United Nations Environment Programme (UNEP) had in the past concluded a week long meeting in Geneva (2002) with the recommendation that governments negotiate a treaty to limit the amount of mercury traded worldwide. In the meantime, countries should reduce mercury risks by cutting or eliminating the production and consumption of the chemical by substituting other products and processes. Mercury has been widely used in consumer products because it is an excellent conductor of electricity and is highly malleable. Products containing mercury include thermometers, dental fillings, fluorescent lamps and other electrical equipment, and some batteries. Mercury is used in several types of instruments common to electric utilities, municipalities and households, such as switches, barometers, meters, temperature gauges, pressure gauges and sprinkler system contacts. It has been used as an ingredient in some pesticides and biocides, certain pharmaceuticals, and cosmetics such as skin lightening creams. In some countries, mercury has ritual religious uses. People are most likely to be exposed to mercury by eating fish or shellfish contaminated with methylmercury, and many jurisdictions have issued fish consumption warnings based on the presence of mercury in fish (Cai and Bayona, 1995; Abalos *et al.*, 1997; Fatoki *et al.*, 2000; Ndibewu *et al*., 2002). People can be exposed when breathing vapours in air from spills, incinerators, and industries that burn fuels containing mercury (Nriagu, 1996). Mercury can be released from dental work or medical treatments and dental or health service workers can be exposed from breathing contaminated workplace air or skin contact during use in the workplace. When placed in landfills, mercury can slowly seep into groundwater or evaporate into the air. It can travel over long distances and persist in the environment for lengthy periods of time. Two studies released in March (2002) (http://h2osparc.wq.ncsu.edu/info/hmetals.html) show that mercury generated by fossil fuel burning power plants is falling from the sky in Antarctica and in the Arctic, and is entering the

changes in circulation; (3) change in biochemistry, such as enzyme activity and blood chemistry; (4) change in behaviour; (5) and changes in reproduction (Connell *et al*., 1984). Many organisms are able to regulate the metal concentrations in their tissues. Fish and the crustacea can excrete essential metals, such as copper, zinc, and iron that are present in excess. Some can also excrete non-essential metals, such as mercury and cadmium, although this is usually met with less success (Connell *et al.,* 1984).

Research has shown that aquatic plants and bivalves are not able to successfully regulate metal uptake (Connell *et al*., 1984). Thus, bivalves tend to suffer from metal accumulation in polluted environments. In estuarine systems, bivalves often serve as biomonitor organisms in areas of suspected pollution (Kennish, 1992). Shell fishing waters are closed if metal levels make shellfish unfit for human consumption. In comparison to freshwater fish and invertebrates, aquatic plants are equally or less sensitive to cadmium, copper, lead, mercury, nickel, and zinc. Thus, the water resource should be managed for the protection of fish and invertebrates, in order to ensure aquatic plant survival (USEPA, 1987). Metal uptake rates will vary according to the organism and the metal in question. Phytoplankton and zooplankton often assimilate available metals quickly because of their high surface area to volume ratio. The ability of fish and invertebrates to adsorb metals is largely dependent on the physical and chemical characteristics of the metal (Kennish, 1992). With the exception of mercury, little metal bioaccumulation has been observed in aquatic organisms (Kennish, 1992). Metals may enter the systems of aquatic organisms via three main pathways: (1) Free metal ions that are absorbed through respiratory surface (e.g. gills) are readily diffused into the blood stream, (2) Free metal ions that are adsorbed onto body surfaces are passively diffused into the blood stream, and (3) Metals that are sorbed onto food and particulates may be ingested, as well as free ions ingested with water (Connell *et al*., 1984).

#### **3.2 Irrigation effects**

Irrigation water may transport dissolved heavy metals to agricultural fields. Although most heavy metals do not pose a threat to humans through crop consumption, cadmium may be incorporated into plant tissue. Accumulation usually occurs in plant roots, but may also occur throughout the plant (De Voogt *et al*., 1980). Most irrigation systems are designed to allow for up to 30 percent of the water applied to not be absorbed and to leave the field as return flow. Return flow either joins the groundwater or runs off the field surface (tail water). Sometimes tail water must be rerouted into streams because of downstream water rights or a necessity to maintain stream flow. However, usually the tail water is collected and stored until it can be reused or delivered to another field (USEPA, 1993a). Tail water is often stored in small lakes or reservoirs, where heavy metals can accumulate as return flow is pumped in and out. These metals can adversely impact aquatic communities. An extreme example of this is the Kesterson reservoir in the San Joaquin Valley, California, which received subsurface agricultural drain water containing high levels of selenium and salts that, had been leached from the soil during irrigation. Studies in the Kesterson reservoir revealed elevated levels of selenium in water, sediments, terrestrial and aquatic vegetation, and aquatic insects. The elevated levels of selenium were cited as relating to the low reproductive success, high mortality, and developmental abnormalities in embryos and chicks of nesting aquatic birds (Schuler *et al*., 1990).

#### **3.3 Health effects**

28 Environmental Health – Emerging Issues and Practice

changes in circulation; (3) change in biochemistry, such as enzyme activity and blood chemistry; (4) change in behaviour; (5) and changes in reproduction (Connell *et al*., 1984). Many organisms are able to regulate the metal concentrations in their tissues. Fish and the crustacea can excrete essential metals, such as copper, zinc, and iron that are present in excess. Some can also excrete non-essential metals, such as mercury and cadmium, although

Research has shown that aquatic plants and bivalves are not able to successfully regulate metal uptake (Connell *et al*., 1984). Thus, bivalves tend to suffer from metal accumulation in polluted environments. In estuarine systems, bivalves often serve as biomonitor organisms in areas of suspected pollution (Kennish, 1992). Shell fishing waters are closed if metal levels make shellfish unfit for human consumption. In comparison to freshwater fish and invertebrates, aquatic plants are equally or less sensitive to cadmium, copper, lead, mercury, nickel, and zinc. Thus, the water resource should be managed for the protection of fish and invertebrates, in order to ensure aquatic plant survival (USEPA, 1987). Metal uptake rates will vary according to the organism and the metal in question. Phytoplankton and zooplankton often assimilate available metals quickly because of their high surface area to volume ratio. The ability of fish and invertebrates to adsorb metals is largely dependent on the physical and chemical characteristics of the metal (Kennish, 1992). With the exception of mercury, little metal bioaccumulation has been observed in aquatic organisms (Kennish, 1992). Metals may enter the systems of aquatic organisms via three main pathways: (1) Free metal ions that are absorbed through respiratory surface (e.g. gills) are readily diffused into the blood stream, (2) Free metal ions that are adsorbed onto body surfaces are passively diffused into the blood stream, and (3) Metals that are sorbed onto food and particulates

may be ingested, as well as free ions ingested with water (Connell *et al*., 1984).

Irrigation water may transport dissolved heavy metals to agricultural fields. Although most heavy metals do not pose a threat to humans through crop consumption, cadmium may be incorporated into plant tissue. Accumulation usually occurs in plant roots, but may also occur throughout the plant (De Voogt *et al*., 1980). Most irrigation systems are designed to allow for up to 30 percent of the water applied to not be absorbed and to leave the field as return flow. Return flow either joins the groundwater or runs off the field surface (tail water). Sometimes tail water must be rerouted into streams because of downstream water rights or a necessity to maintain stream flow. However, usually the tail water is collected and stored until it can be reused or delivered to another field (USEPA, 1993a). Tail water is often stored in small lakes or reservoirs, where heavy metals can accumulate as return flow is pumped in and out. These metals can adversely impact aquatic communities. An extreme example of this is the Kesterson reservoir in the San Joaquin Valley, California, which received subsurface agricultural drain water containing high levels of selenium and salts that, had been leached from the soil during irrigation. Studies in the Kesterson reservoir revealed elevated levels of selenium in water, sediments, terrestrial and aquatic vegetation, and aquatic insects. The elevated levels of selenium were cited as relating to the low reproductive success, high mortality, and developmental abnormalities in embryos and

**3.2 Irrigation effects** 

chicks of nesting aquatic birds (Schuler *et al*., 1990).

this is usually met with less success (Connell *et al.,* 1984).

Ingestion of metals such as lead (Pb), cadmium (Cd), mercury (Hg), arsenic (As), barium (Ba), and chromium (Cr), may pose great risks to human health. Trace metals such as lead and cadmium will interfere with essential nutrients of similar appearance, such as calcium (Ca2+) and zinc (Zn2+). Amongst the heavy metals pollution, mercury pollution has become a global problem (Schrader *et al.*, 1983) because of its occurrence from natural anthropogenic sources, and its biogeochemical processes (Cai and Bayona, 1995; Coello *et al*., 1996). As public awareness regarding the toxicity and the environmental impact of mercury contamination increases, speciation analytical methods developed, are required to distinguish between organic and inorganic forms of mercury. The determination and monitoring of mercury and arsenic is a special concern in the field of mine works and food engineering respectfully (Leal *et al.*, 1995; Nriagu, 1996). It has been reported (Leal *et al.*, 1995; Fatoki, 2000) that mercury can be methylated in the environment and bioconcentrated in the biota. Mercury poses a great risk to humans, especially in the form of methylmercury. When mercury enters water, it is often transformed by microorganisms into the toxic methyl mercury form. Symptoms of acute poisoning are pharyngitis, gasteroenteritis, vomiting, nephritis, hepatitis, and circulatory collapse. Chronic poisoning is usually a result of industrial exposure or a diet consisting of contaminated fish (mercury is the only metal that will bioaccumulate). Chronic poisoning may cause liver damage, neural damage, and teratogenesis (USEPA, 1987).

The Global mercury assessment working group of the United Nations Environment Programme (UNEP) had in the past concluded a week long meeting in Geneva (2002) with the recommendation that governments negotiate a treaty to limit the amount of mercury traded worldwide. In the meantime, countries should reduce mercury risks by cutting or eliminating the production and consumption of the chemical by substituting other products and processes. Mercury has been widely used in consumer products because it is an excellent conductor of electricity and is highly malleable. Products containing mercury include thermometers, dental fillings, fluorescent lamps and other electrical equipment, and some batteries. Mercury is used in several types of instruments common to electric utilities, municipalities and households, such as switches, barometers, meters, temperature gauges, pressure gauges and sprinkler system contacts. It has been used as an ingredient in some pesticides and biocides, certain pharmaceuticals, and cosmetics such as skin lightening creams. In some countries, mercury has ritual religious uses. People are most likely to be exposed to mercury by eating fish or shellfish contaminated with methylmercury, and many jurisdictions have issued fish consumption warnings based on the presence of mercury in fish (Cai and Bayona, 1995; Abalos *et al.*, 1997; Fatoki *et al.*, 2000; Ndibewu *et al*., 2002). People can be exposed when breathing vapours in air from spills, incinerators, and industries that burn fuels containing mercury (Nriagu, 1996). Mercury can be released from dental work or medical treatments and dental or health service workers can be exposed from breathing contaminated workplace air or skin contact during use in the workplace. When placed in landfills, mercury can slowly seep into groundwater or evaporate into the air. It can travel over long distances and persist in the environment for lengthy periods of time. Two studies released in March (2002) (http://h2osparc.wq.ncsu.edu/info/hmetals.html) show that mercury generated by fossil fuel burning power plants is falling from the sky in Antarctica and in the Arctic, and is entering the food chain.

Speciation Methods for the Determination of Organotins (OTs)

and Heavy Metals (HMs) in the Freshwater and Marine Environments 31

also transports dissolved metals. Although dissolved metals are primarily transported in overland flow, some underground transport is possible (Nriagu, 1996). Metals that are introduced to the unsaturated zone and the saturated zone will most likely not be transported a long distance. Dissolved metals that are carried below the land surface will readily sorb to soil particles or lithic material in the unsaturated zone and the saturated zone (Nriagu, 1996). Metals introduced into the atmosphere may be carried to the land surface by precipitation and dry fallout. Additionally, because metals readily sorb to many sediment

Zehra Aydin (2002) reported that by the early 1970s, there was clearly a need to promote better use and management of the seas and their resources which imposed a call on the international community to begin negotiating a comprehensive treaty on the law of the sea. What is remarkable is that, these laws had diversified in time to fit specific country`s standards and regulatory needs. For continual assessment, there had then been a growing need to develop suitable analytical tools to assess organotins and heavy metals. In response to this trend, countries with advanced economy began research in this area long ago. Today, a substantial body of knowledge on OTs and heavy metals in waters of the developed countries of Europe, America, Asia and Oceania has evolved. However, data are very scanty for developing nations' water environments (Bryan, 1991; ATRP Corp-U.S-EPA, 2000, 2001

Particularly, the ecotoxicological effects of TBT and other tri-organotin (Leal *et al.*, 1995) compounds in the aquatic environment have caused much concern in recent years leading to the control or banning of their use in a few developed countries (Jiang *et al*., 2000). At present, it is doubtful if specific legislation exists controlling the use of TBT in many, if not all developing countries (Samson and Shenker, 2000). This is primarily due to the lack of supporting data on the occurrence and impact of TBT in these countries. Tributyltin has been described as "the most toxic substance ever deliberately introduced into the natural waters" (Jiang *et al.*, 2000; Leal *et al.*, 1995; Thomaidis *et al.*, 2001). Owing to its extremely toxic effects to aquatic life at low concentrations, TBT and other forms of organotin such as triphenyltin are legislatively banned to be used as antifouling paints from since the late 1980s in most European and North American countries (Jiang and Yang, 2000). The first regulatory and legislative control on the use of TBT was only adopted in France in 1982 followed by UK in 1985 (Meech *et al.*, 1998). Most of the control measures introduced since then involved banning the use of TBT in marine boats of less than 25 m length (Ceulemans and Adam, 1995). For marine water, the UK adopted an environmental quality target of 20 ng L-1 TBT in 1985 and environmental quality standard of 2 ng L-1 TBT was proposed in 1989 (Cai *et al*., 1994). The US Environmental Protection Agency's proposed limits for TBT in fresh and marine waters were 26 ng L-1 (4-day average) and 10 ng L-1 (4-day average), respectively, (Dirkx *et al.*, 1994). The Canadian Council of the Ministers of Environment derived an Interim Water Quality Guideline of 8 ng L-1 TBT in estuarine or seawaters for the

Levels of TBT of the order of a few hundred ng L-1 have been reported in coastal waters with heavy marine traffic, such as ports, marinas and dockyards, as compared to open water

types, wind-borne sediment is a potential route for metal transport (Nriagu, 1996).

**7. Regulatory measures applied to TBT and organotins** 

protection of aquatic life (Cai and Bayona, 1995).

and 2002).

Cadmium is an extremely toxic metal commonly found in industrial workplaces, particularly where any ore is being processed or smelted. Due to its low permissible exposure limit (PEL), over exposures may occur even in situations where cadmium is only in trace quantities in the parent ore or smelter dust. Cadmium is used extensively in electroplating, although the nature of the operation does not generally lead to overexposures. Several deaths from acute exposure have occurred among welders who have unsuspectingly welded on cadmium-containing alloys and among silver solders. Cadmium is also found in industrial paints and may represent a hazard when spray applied. Operations involving removal of cadmium paints by scraping or blasting may similarly pose a significant hazard. Cadmium emits a characteristic brown fume (CdO) upon heating, which is relatively non-irritating, and thus, does not alarm the exposed individual (Maenpa *et al.*, 2002; Meech *et al.*, 1998).

#### **4. Pollution source – Points of TBT and organotins**

Organotin compounds have found many important industrial and agricultural applications for more than three decades (Prudente *et al.*, 1999; Leal *et al.*, 1995). These include the use of mono-methyl tins, mono-butyltins and di-butyltins as stabilizers in polyvinyl chloride (PVC) and as catalysts in industrial processes. Organotin compounds with three alkyl groups attached to the tin atom, such as tributyltin (TBT), tri-phenyltin and tricyclohexyltin, have found wide applications as antifouling agents in marine paints formulations, bactericides in cooling water, agricultural fungicides and acaricides (Leal *et al.*, 1995), as previously mentioned. Most import of TBT is used in marine paints as an effective means of the growth of fouling organisms such as tubeworms, barnacles and mussels on seafaring vessels and marine structures (Abalos *et al.*, 1997; Fatoki, 2000).

#### **5. Pollution source – Points of heavy metals (HMs)**

Nonpoint sources of heavy metals pollution are mostly natural. Chemical and physical weathering of igneous and metamorphic rocks and soils often releases heavy metals into the sediment and into the air. Other contributions include the decomposition of plant and animal detritus, precipitation or atmospheric deposition of airborne particles from volcanic activity, wind erosion, forest fire smoke, plant exudates, and oceanic spray (Kennish, 1992). Anthropogenic sources are contributed by surface runoffs from mining operations usually has a low pH and contains high levels of metals such as iron, manganese, zinc, copper, nickel and cobalt. The combustion of fossil fuels pollutes the atmosphere with metal particulates that eventually settle to the land surface. Urban stormwater runoffs often contain metals from roadways and atmospheric fallout (Connell *et al*., 1984). Currently, anthropogenic inputs of metals exceed natural inputs. Point sources include domestic wastewater effluents which contains metals from metabolic wastes, corrosion of water pipes, and consumer products. Industrial effluents and waste sludges may substantially contribute to metal loading (Connell *et al*., 1984).

#### **6. Mode of transport and environmental fate of HMs**

Transport occurs mostly in water and air. Water can transport metals that are bound to sediment particles. The primary route for sediment-metal transport is overland flow. Water

Cadmium is an extremely toxic metal commonly found in industrial workplaces, particularly where any ore is being processed or smelted. Due to its low permissible exposure limit (PEL), over exposures may occur even in situations where cadmium is only in trace quantities in the parent ore or smelter dust. Cadmium is used extensively in electroplating, although the nature of the operation does not generally lead to overexposures. Several deaths from acute exposure have occurred among welders who have unsuspectingly welded on cadmium-containing alloys and among silver solders. Cadmium is also found in industrial paints and may represent a hazard when spray applied. Operations involving removal of cadmium paints by scraping or blasting may similarly pose a significant hazard. Cadmium emits a characteristic brown fume (CdO) upon heating, which is relatively non-irritating, and thus, does not alarm the exposed individual (Maenpa

Organotin compounds have found many important industrial and agricultural applications for more than three decades (Prudente *et al.*, 1999; Leal *et al.*, 1995). These include the use of mono-methyl tins, mono-butyltins and di-butyltins as stabilizers in polyvinyl chloride (PVC) and as catalysts in industrial processes. Organotin compounds with three alkyl groups attached to the tin atom, such as tributyltin (TBT), tri-phenyltin and tricyclohexyltin, have found wide applications as antifouling agents in marine paints formulations, bactericides in cooling water, agricultural fungicides and acaricides (Leal *et al.*, 1995), as previously mentioned. Most import of TBT is used in marine paints as an effective means of the growth of fouling organisms such as tubeworms, barnacles and mussels on

Nonpoint sources of heavy metals pollution are mostly natural. Chemical and physical weathering of igneous and metamorphic rocks and soils often releases heavy metals into the sediment and into the air. Other contributions include the decomposition of plant and animal detritus, precipitation or atmospheric deposition of airborne particles from volcanic activity, wind erosion, forest fire smoke, plant exudates, and oceanic spray (Kennish, 1992). Anthropogenic sources are contributed by surface runoffs from mining operations usually has a low pH and contains high levels of metals such as iron, manganese, zinc, copper, nickel and cobalt. The combustion of fossil fuels pollutes the atmosphere with metal particulates that eventually settle to the land surface. Urban stormwater runoffs often contain metals from roadways and atmospheric fallout (Connell *et al*., 1984). Currently, anthropogenic inputs of metals exceed natural inputs. Point sources include domestic wastewater effluents which contains metals from metabolic wastes, corrosion of water pipes, and consumer products. Industrial effluents and waste sludges may substantially

Transport occurs mostly in water and air. Water can transport metals that are bound to sediment particles. The primary route for sediment-metal transport is overland flow. Water

*et al.*, 2002; Meech *et al.*, 1998).

**4. Pollution source – Points of TBT and organotins** 

seafaring vessels and marine structures (Abalos *et al.*, 1997; Fatoki, 2000).

**5. Pollution source – Points of heavy metals (HMs)** 

contribute to metal loading (Connell *et al*., 1984).

**6. Mode of transport and environmental fate of HMs** 

also transports dissolved metals. Although dissolved metals are primarily transported in overland flow, some underground transport is possible (Nriagu, 1996). Metals that are introduced to the unsaturated zone and the saturated zone will most likely not be transported a long distance. Dissolved metals that are carried below the land surface will readily sorb to soil particles or lithic material in the unsaturated zone and the saturated zone (Nriagu, 1996). Metals introduced into the atmosphere may be carried to the land surface by precipitation and dry fallout. Additionally, because metals readily sorb to many sediment types, wind-borne sediment is a potential route for metal transport (Nriagu, 1996).

#### **7. Regulatory measures applied to TBT and organotins**

Zehra Aydin (2002) reported that by the early 1970s, there was clearly a need to promote better use and management of the seas and their resources which imposed a call on the international community to begin negotiating a comprehensive treaty on the law of the sea. What is remarkable is that, these laws had diversified in time to fit specific country`s standards and regulatory needs. For continual assessment, there had then been a growing need to develop suitable analytical tools to assess organotins and heavy metals. In response to this trend, countries with advanced economy began research in this area long ago. Today, a substantial body of knowledge on OTs and heavy metals in waters of the developed countries of Europe, America, Asia and Oceania has evolved. However, data are very scanty for developing nations' water environments (Bryan, 1991; ATRP Corp-U.S-EPA, 2000, 2001 and 2002).

Particularly, the ecotoxicological effects of TBT and other tri-organotin (Leal *et al.*, 1995) compounds in the aquatic environment have caused much concern in recent years leading to the control or banning of their use in a few developed countries (Jiang *et al*., 2000). At present, it is doubtful if specific legislation exists controlling the use of TBT in many, if not all developing countries (Samson and Shenker, 2000). This is primarily due to the lack of supporting data on the occurrence and impact of TBT in these countries. Tributyltin has been described as "the most toxic substance ever deliberately introduced into the natural waters" (Jiang *et al.*, 2000; Leal *et al.*, 1995; Thomaidis *et al.*, 2001). Owing to its extremely toxic effects to aquatic life at low concentrations, TBT and other forms of organotin such as triphenyltin are legislatively banned to be used as antifouling paints from since the late 1980s in most European and North American countries (Jiang and Yang, 2000). The first regulatory and legislative control on the use of TBT was only adopted in France in 1982 followed by UK in 1985 (Meech *et al.*, 1998). Most of the control measures introduced since then involved banning the use of TBT in marine boats of less than 25 m length (Ceulemans and Adam, 1995). For marine water, the UK adopted an environmental quality target of 20 ng L-1 TBT in 1985 and environmental quality standard of 2 ng L-1 TBT was proposed in 1989 (Cai *et al*., 1994). The US Environmental Protection Agency's proposed limits for TBT in fresh and marine waters were 26 ng L-1 (4-day average) and 10 ng L-1 (4-day average), respectively, (Dirkx *et al.*, 1994). The Canadian Council of the Ministers of Environment derived an Interim Water Quality Guideline of 8 ng L-1 TBT in estuarine or seawaters for the protection of aquatic life (Cai and Bayona, 1995).

Levels of TBT of the order of a few hundred ng L-1 have been reported in coastal waters with heavy marine traffic, such as ports, marinas and dockyards, as compared to open water

Speciation Methods for the Determination of Organotins (OTs)

draft treaty to ban the use of TBTs on all hulls worldwide.

compounds to human and marine lives including the aquatic ecosystems.

volatile derivatives, (iii) pre-concentration (iv) clean-up and, (v) determination.

such as South Africa.

and Heavy Metals (HMs) in the Freshwater and Marine Environments 33

triphenyltins (TPTs), diphenyltins (DPTs) and monophenyltins (MPTs). Amongst them, TBT is acutely toxic to a variety of freshwater species at concentrations down to 0.1 µg L-1 (Prudente *et al.*, 1999; Chau *et al.*, 1996). Indeed, this toxicity limits level has been checked for only some marinas in South Africa (Fatoki *et al*., 2000). Although, the lack of research in this area cannot be over stressed as an underlining factor in setting regulatory limits in developing countries, BT contamination should be regarded as a global pollution problem (Maguire *et al.*, 1982) particularly in countries where no regulation has been implemented

We have mentioned that TBT and organotins once were the preferred universally available biocides for marine coatings (Mueller, 1987) and large amounts were used (Leal *et al.*, 1995; Kuballa *et al.*, 1995) for pleasure boats, large ship or vessels, docks and fish-nets, lumber preservatives and slimicides in cooling systems, as an effective antifouling agent in paints. Also, that dibutyltins (DBTs) and monobutyltins (MBTs) were mostly used as stabilizers in polyvinyl chlorides and as catalysts in the production of polyurethane foams, silicones, and in other industrial processes (Cai and Bayona, 1995; Fent, 1996). Finally, it suffices to recall that another well-known source (Leal *et al.*, 199) is their use in the manufacture of fungicides, acaricides and insecticides for use in agriculture. And after years of use, their effects on marine environments have brought about actions to limit the use of TBTs through laws and regulations (Kan-atireklap *et al.*, 1997; Jiang *et al.*, 2000; Fatoki *et al.*, 2000), especially in developed countries. With the growing concern regarding the impact on the environment with threats to aquatic life (Kuballa *et al.*, 1995; Kumar *et al.*, 1993) and human health (Fatoki, 2000), an international stakeholders' process (Murmansk-2000) considered a

Actually, considerable work has been done in the area of techniques development for organotins speciation analysis elsewhere (Mueller, 1987; Prudente *et al.*, 1999; Cai and Bayona, 1995; Ikeda *et al.*, 1996; Jiang *et al.*, 2000; Meech *et al.*, 1998; Abalos *et al.*, 1997) while research efforts in most developing countries, including South Africa are still at their first endeavors at the moment (Fatoki *et al.,* 2000; Ndibewu *et al.*, 2002). Thus, lack of continual consistent research work in this field and lack of monitoring really present potential danger both to the aquatic biota and man. As reported for the case of man-generated heavy metals discharge into the environment (Lucinda *et al.*, 1983; http://www.mercurypolicy.org/), OTs, disperse on a global scale by long-range atmospheric transport and deposit into colder regions could cause an environmental disaster. If the lack of research interest in this area on a global scale stays as such, this will lead to no regular monitoring of water environments in order to avoid potential danger that can be caused by these endocrine-disrupting

Generally, any procedure for speciation analysis consists of five successive steps (Fatoki, 2000): (i) extraction of the analytes from the sample matrix, (ii) derivatization to form the

In the first step, extraction is critical, meaning that the choice of a particular extraction technique is also critical. Two extraction methods are popularly used (Abalos *et al.*, 1997), namely: the liquid-liquid (LLE) and the solid phase extraction (SPE) extraction techniques. Liquid-liquid extraction methods often require a large amount of hazardous solvents and tend to be replaced by the solid phase extraction (SPE) procedures (Fent, 1996). The

where TBT is found to be near or below 10 ng L-1 or less for Europe and North America and Hong Kong (Evans *et al.*, 2000; Forstner, 1983; Maguire *et al.*, 1982).

Fatoki, (2000) have reported a preliminary study of TBT in waters and sediments from some major ports in South Africa (Port Elizabeth and East London). This study indicates significant contamination of the East London and Port Elizabeth Harbors' aquatic environment with TBT (Fatoki *et al.*, 2000). This study also indicates contamination levels of 5.5 ng L-1 - 22.7 ng L l-1 (water samples) and 1.8 ng g-1 - 26.2 ng g-1 (sediments) for Port Elizabeth. The figures for East London are 3.3 – 49.9 ng L-1 and 3.5 – 1103.1 ng g-1 for water and sediments, respectively. As such, this should be viewed as danger to the biota in the aquatic system.

#### **8. Regulatory measures applied to heavy metals in marine environments**

In 1998, in Aarhus (Denmark), 36 Parties to the Convention on Long-Range Transboundary Air Pollution signed the Protocol on HMs. The Protocol was aimed at the elimination, restriction on use, and reduction of HM emissions to the environment. An integrated program for the inter-comparison study of mercury models was developed (http://www.msceast.org/hms/).

The US Food and Drug Administration (FDA) has set an action level of 1g g-1 (wet mass) for fresh fish (Cai and Bayona, 1995; Lucinda *et al.*, 1983), Canada and several US states have set consumption limits for fish at 0.5g g-1. The European Union (UE) has set environmental quality objectives of 0.3g g-1 (wet mass), 1g L-1 for continental water, 0.5g L-1 for estuarine water, and 0.3g L-1 for coastal water as total mercury (Cai and Bayona, 1995). In addition to a legally binding mercury treaty, the Global Mercury Assessment Working Group (http://www.mercurypolicy.org) urges governments to establish a non-binding global program of action, and strengthen cooperation among countries on information sharing, risk communication, assessment and related activities. The Working Group recommended immediate action to enhance outreach to vulnerable groups, such as pregnant women and provide technical and financial support to developing countries and to countries with economies in transition. Increased research, monitoring, data collection on the health, environmental aspects of mercury and development of environmentally friendly alternative chemicals to this one are among the group's recommendations. Some countries have seriously taken action to deal with mercury pollution while others still take it slightly, especially the developing countries. The European Union had faced a bill of up to 330 million Euros (US\$324 million) to dispose safely of excess mercury stocks from an obsolete method of chlorine production (http://www.mercurypolicy.org/). The U.S. Senate passed legislation in early 2002 (http://h2osparc.wq.ncsu.edu/info/hmetals.html) banning the sale of mercury fever thermometers anywhere in the United States. Similarly, in the same year, the U.S. Environmental Protection Agency proposed changing waste regulations for computers, televisions and mercury containing equipment to discourage the flow of these materials to municipal landfills and incinerators (http://www.mercurypolicy.org).

#### **9. An overview of speciation methods and determination techniques of heavy metals, TBT and other organotins in seawaters**

Organotins or butyl compounds (BCs) which are generally of interest for speciation include the tributyltins (TBTs), dibutyltins (DBTs) and monobutyltins (MBTs), as well as the

where TBT is found to be near or below 10 ng L-1 or less for Europe and North America and

Fatoki, (2000) have reported a preliminary study of TBT in waters and sediments from some major ports in South Africa (Port Elizabeth and East London). This study indicates significant contamination of the East London and Port Elizabeth Harbors' aquatic environment with TBT (Fatoki *et al.*, 2000). This study also indicates contamination levels of 5.5 ng L-1 - 22.7 ng L l-1 (water samples) and 1.8 ng g-1 - 26.2 ng g-1 (sediments) for Port Elizabeth. The figures for East London are 3.3 – 49.9 ng L-1 and 3.5 – 1103.1 ng g-1 for water and sediments, respectively. As such, this should be viewed as danger to the biota in the

**8. Regulatory measures applied to heavy metals in marine environments** 

In 1998, in Aarhus (Denmark), 36 Parties to the Convention on Long-Range Transboundary Air Pollution signed the Protocol on HMs. The Protocol was aimed at the elimination, restriction on use, and reduction of HM emissions to the environment. An integrated program for the inter-comparison study of mercury models was developed

The US Food and Drug Administration (FDA) has set an action level of 1g g-1 (wet mass) for fresh fish (Cai and Bayona, 1995; Lucinda *et al.*, 1983), Canada and several US states have set consumption limits for fish at 0.5g g-1. The European Union (UE) has set environmental quality objectives of 0.3g g-1 (wet mass), 1g L-1 for continental water, 0.5g L-1 for estuarine water, and 0.3g L-1 for coastal water as total mercury (Cai and Bayona, 1995). In addition to a legally binding mercury treaty, the Global Mercury Assessment Working Group (http://www.mercurypolicy.org) urges governments to establish a non-binding global program of action, and strengthen cooperation among countries on information sharing, risk communication, assessment and related activities. The Working Group recommended immediate action to enhance outreach to vulnerable groups, such as pregnant women and provide technical and financial support to developing countries and to countries with economies in transition. Increased research, monitoring, data collection on the health, environmental aspects of mercury and development of environmentally friendly alternative chemicals to this one are among the group's recommendations. Some countries have seriously taken action to deal with mercury pollution while others still take it slightly, especially the developing countries. The European Union had faced a bill of up to 330 million Euros (US\$324 million) to dispose safely of excess mercury stocks from an obsolete method of chlorine production (http://www.mercurypolicy.org/). The U.S. Senate passed legislation in early 2002 (http://h2osparc.wq.ncsu.edu/info/hmetals.html) banning the sale of mercury fever thermometers anywhere in the United States. Similarly, in the same year, the U.S. Environmental Protection Agency proposed changing waste regulations for computers, televisions and mercury containing equipment to discourage the flow of these

materials to municipal landfills and incinerators (http://www.mercurypolicy.org).

**metals, TBT and other organotins in seawaters** 

**9. An overview of speciation methods and determination techniques of heavy** 

Organotins or butyl compounds (BCs) which are generally of interest for speciation include the tributyltins (TBTs), dibutyltins (DBTs) and monobutyltins (MBTs), as well as the

Hong Kong (Evans *et al.*, 2000; Forstner, 1983; Maguire *et al.*, 1982).

aquatic system.

(http://www.msceast.org/hms/).

triphenyltins (TPTs), diphenyltins (DPTs) and monophenyltins (MPTs). Amongst them, TBT is acutely toxic to a variety of freshwater species at concentrations down to 0.1 µg L-1 (Prudente *et al.*, 1999; Chau *et al.*, 1996). Indeed, this toxicity limits level has been checked for only some marinas in South Africa (Fatoki *et al*., 2000). Although, the lack of research in this area cannot be over stressed as an underlining factor in setting regulatory limits in developing countries, BT contamination should be regarded as a global pollution problem (Maguire *et al.*, 1982) particularly in countries where no regulation has been implemented such as South Africa.

We have mentioned that TBT and organotins once were the preferred universally available biocides for marine coatings (Mueller, 1987) and large amounts were used (Leal *et al.*, 1995; Kuballa *et al.*, 1995) for pleasure boats, large ship or vessels, docks and fish-nets, lumber preservatives and slimicides in cooling systems, as an effective antifouling agent in paints. Also, that dibutyltins (DBTs) and monobutyltins (MBTs) were mostly used as stabilizers in polyvinyl chlorides and as catalysts in the production of polyurethane foams, silicones, and in other industrial processes (Cai and Bayona, 1995; Fent, 1996). Finally, it suffices to recall that another well-known source (Leal *et al.*, 199) is their use in the manufacture of fungicides, acaricides and insecticides for use in agriculture. And after years of use, their effects on marine environments have brought about actions to limit the use of TBTs through laws and regulations (Kan-atireklap *et al.*, 1997; Jiang *et al.*, 2000; Fatoki *et al.*, 2000), especially in developed countries. With the growing concern regarding the impact on the environment with threats to aquatic life (Kuballa *et al.*, 1995; Kumar *et al.*, 1993) and human health (Fatoki, 2000), an international stakeholders' process (Murmansk-2000) considered a draft treaty to ban the use of TBTs on all hulls worldwide.

Actually, considerable work has been done in the area of techniques development for organotins speciation analysis elsewhere (Mueller, 1987; Prudente *et al.*, 1999; Cai and Bayona, 1995; Ikeda *et al.*, 1996; Jiang *et al.*, 2000; Meech *et al.*, 1998; Abalos *et al.*, 1997) while research efforts in most developing countries, including South Africa are still at their first endeavors at the moment (Fatoki *et al.,* 2000; Ndibewu *et al.*, 2002). Thus, lack of continual consistent research work in this field and lack of monitoring really present potential danger both to the aquatic biota and man. As reported for the case of man-generated heavy metals discharge into the environment (Lucinda *et al.*, 1983; http://www.mercurypolicy.org/), OTs, disperse on a global scale by long-range atmospheric transport and deposit into colder regions could cause an environmental disaster. If the lack of research interest in this area on a global scale stays as such, this will lead to no regular monitoring of water environments in order to avoid potential danger that can be caused by these endocrine-disrupting compounds to human and marine lives including the aquatic ecosystems.

Generally, any procedure for speciation analysis consists of five successive steps (Fatoki, 2000): (i) extraction of the analytes from the sample matrix, (ii) derivatization to form the volatile derivatives, (iii) pre-concentration (iv) clean-up and, (v) determination.

In the first step, extraction is critical, meaning that the choice of a particular extraction technique is also critical. Two extraction methods are popularly used (Abalos *et al.*, 1997), namely: the liquid-liquid (LLE) and the solid phase extraction (SPE) extraction techniques. Liquid-liquid extraction methods often require a large amount of hazardous solvents and tend to be replaced by the solid phase extraction (SPE) procedures (Fent, 1996). The

Speciation Methods for the Determination of Organotins (OTs)

toxic solvents and acids used.

extracting agent (Tanabe *et al.*, 1998).

diethyl dithiocarbamate (DDTC) (Fatoki et al, 2000) are also convenient.

and Heavy Metals (HMs) in the Freshwater and Marine Environments 35

of low polar organotin complexes with tropolone or diethyldithiocarbanate (DDTC). Tropolone is preferred to DDTC (Dirkx *et al.*, 1994), as under acidic condition, this undergoes decomposition, giving rise to extractable interference (http://h2osparc.wq.ncsu.edu/info/hmetals.html). Sample preparation procedures before analyses, such as liquid-liquid extraction of organotin chelates with fresh tropolone or

Non-polar solvent plus acids are used in complex matrices' extraction (Tanabe *et al.*, 1998). For example, sediment sample is treated with hydrochloric acid with shaking or sonification, followed by sequential solvent extraction (Wade *et al.*, 1988). Hydrobromic acid or acetic acid or a mixture is also used (Tolosa *et al.*, 1992; Martin *et al.*, 1994). Sonification has become the most widely used stirring method for sediment matrix, whereas, energymixing methods are used for biotic materials. Mixtures of solvents have been used to increase the polarity of the medium, hexane-ethyl acetate, hexane -diethyl ether, and chloroform-ethyl acetate (Fatoki, 2000). The salting out effect or ion-pairing effect is used to increase the efficiency of extraction of organotins (OTs) from aqueous phase to the organic phase, when HCl is used (Dirk *et al.,* 1992; Tao *et al.*, 1999). Polar solvents have also been used to achieve extraction. The polar solvents which have been used are aqueous (i) HCl (Ceulemans and Adams, 1995); (ii) HCl or HOAc in polar organic solvents (MeOH, acetone) (Kuballa *et al.*, 1995; Cai *et al.*, 1993); (iii) acetic acid (Quevauviller, 1996); (iv) net polar organic solvents (MeOH, DCM-MeOH, butanol, MeOH-EtOAc (Han and Weber, 1988; Apte and Gardner, 1998); (v) polar organic solvents in basic conditions (Pawliszyn, 1997). In this case, sonification is used in most procedures. Very recently, a focused microwave field has been introduced to reduce extraction time from hours to several munites (Donard *et al.*, 1995). In some cases, after the acid or polar solvent extraction, a liquid-liquid extraction (LLE) with a non-miscible solvent (benzene, CH3Cl-DCM, ETOAc-MeOH, DCM, hexane, cyclohexane, toluene, hexane-ETOAc) is used to recover OTs from the extract. Several authors (Abalos *et al.*, 1997) have used tropolone and salting out effect to increase the solubility of OTs in the organic solvent. Quite recently, as already mentioned, a more environmentally-friendly extraction technique developed is the supercritical fluid extraction (SFE). Advantages of these methods are shortened extraction time and limited amount of

For biological samples analysis, tetramethyl ammonium (TMAH) hydrolysis is currently applied (Fatoki *et al.*, 2000), above room temperature (60oC) for several hours, for example 1- 2 h. (Leal *et al.*, 1995). The TMAH hydrolysis can be reduced from hours to minutes when the digestion is carried out under focused microwave irradiation. OTs are isolated from the hydrolyzed tissue by hexane (LLE) in the presence of tropolone. Alternatively, after a pH adjustment, simultaneous extraction derivatization with NaBEt4 is used to reduce the numbers of LLE, compared to Grignard reagents. Also, ethanolic- KOH at 60oC for 90 min or NaOH at 40oC for 20 min followed by pH adjustment and LLE has also been applied to the determination of OTs from biotic matrices (Nagase *et al.*, 1998). The digestion time in basic extraction conditions is critical due to the lack of stability of mono- and di-organotin compounds. Basic and enzymatic hydrolysis methods, which are restricted to biotic samples, lead to tissue solubilization. This makes the embedded organotin more available to

In the case of heavy metals speciation, many extraction techniques have been reported with differences in methods approach depending on the aqueous, solid or gaseous nature of the

advantage of SPE include a higher pre-concentration factor and ease of application in the field and in on- line systems, while a drawback has been observed in that only filtered samples can be analyzed. More recently, the solid phase microextraction (SPME) technique has been applied to organotin speciation (Attar, 1996; Abalos *et al.*, 1997). SPME offers an attractive alternative method, which minimizes some problems associated with other methods (Fatoki *et al*., 2000).

Using Liquid-liquid extraction methods, many non-polar and polar solvents have been used for extraction in water, sediment and biological samples. During early techniques applications for organotins determination in water samples, speciation analysis were based on acidification (hydrochloric acid-HCL, hydrobromic acid-HBr or acetic acid-HOAc), to release alkyl tin compounds from the sample matrix, then converting them to the halides or acetate forms (Forstner, 1983). Relatively high polarity solvents are now being used for extraction. These methods (Abalos *et al*., 1997) succeeded for TBT, TPT and tricyclohexyl (TCyT) and failed for other species due to their high polarity.

For sediment analysis with regard to organotins speciation, acid leaching (Kan-atireklap *et al.*, 1997; Chau *et al.*, 1995; Tanabe *et al.*, 1998) to release organotin compounds from sediment was the basic approach in early use. Hydrochloric acids (HCl), Hdrobromic acid (HBr) and acetic acids (HOAc) are used. This is done in an aqueous or methanolic medium by sonification, stirring, shaking or Soxhlet extraction with an organic solvent (Dirkx *et al.*, 1994). As mentioned above, organotin compounds are not involved in biogeochemical process. They rather bind onto the surface of the sediment, hence, the complete dissolution of the later prior to the analysis is, therefore, not considered necessary. Extraction yield is increased by the addition of complexing agents such as tropolone or DDTC (Fatoki *et al.,*  2000). While the tri- and di-substituted compounds can be extracted quantitatively, only about 60 % or less of the mono-substituted compounds are recovered. Two approaches have been evaluated to improve the extraction efficiency of mono- and di-organotin species: (i) the addition of complexing agents (e.g. diethylammonium-diethylthiocarbamate, DEA-DDC or (DDC), and (ii) alkylation in a reaction cell with Grignard reagent prior to the extraction (Fatoki *et al.,* 2000; Abalos *et al.*, 1997). On one hand, recoveries obtained by the first approach are satisfactory for di- and tri-organotin species but a clean-up step is usually needed. On the other hand, the second method yields satisfactory recoveries only for TBT and TPT. Further developments are needed to bring these methods to routine analysis. Apparently, no reliable and efficient method for extracting all organotins from sediment has yet been developed (Abalos *et al.*, 1997).

Hexane, benzene, toluene or dichloromethane (DCM) are non-polar solvents used for the extraction of organotins with complexing agent (Cai and Bayona, 1995; Jiang *et al.*, 2000; Abalos *et al.*, 1997). The efficiency with which butyltins are extracted from spiked sediment with non-polar solvents in the presence of complexing agents is satisfactory. In contrast, very poor recovery is obtained with monobutyl tin and dibutyltin with DCM without a complexing agent (Abalos *et al.*, 1997). With volatile solvents such as hexane, hexaneacetone, dichloromethane (Abalos *et al.*, 1997; Tanabe *et al.*, 1998; Dirkx *et al.,* 1992), Soxhlet extraction is applied without complexing agent (Willis, 1965), since the more polar solvents are incompatible with the Grignard reagents used later for derivatization and favor coextraction of organic interference compounds (Fatoki *et al*., 2000). Therefore, the current recommended procedures (Wade *et al.*, 1988; Abalos *et al.,* 1997) are based on the extraction

advantage of SPE include a higher pre-concentration factor and ease of application in the field and in on- line systems, while a drawback has been observed in that only filtered samples can be analyzed. More recently, the solid phase microextraction (SPME) technique has been applied to organotin speciation (Attar, 1996; Abalos *et al.*, 1997). SPME offers an attractive alternative method, which minimizes some problems associated with other

Using Liquid-liquid extraction methods, many non-polar and polar solvents have been used for extraction in water, sediment and biological samples. During early techniques applications for organotins determination in water samples, speciation analysis were based on acidification (hydrochloric acid-HCL, hydrobromic acid-HBr or acetic acid-HOAc), to release alkyl tin compounds from the sample matrix, then converting them to the halides or acetate forms (Forstner, 1983). Relatively high polarity solvents are now being used for extraction. These methods (Abalos *et al*., 1997) succeeded for TBT, TPT and tricyclohexyl

For sediment analysis with regard to organotins speciation, acid leaching (Kan-atireklap *et al.*, 1997; Chau *et al.*, 1995; Tanabe *et al.*, 1998) to release organotin compounds from sediment was the basic approach in early use. Hydrochloric acids (HCl), Hdrobromic acid (HBr) and acetic acids (HOAc) are used. This is done in an aqueous or methanolic medium by sonification, stirring, shaking or Soxhlet extraction with an organic solvent (Dirkx *et al.*, 1994). As mentioned above, organotin compounds are not involved in biogeochemical process. They rather bind onto the surface of the sediment, hence, the complete dissolution of the later prior to the analysis is, therefore, not considered necessary. Extraction yield is increased by the addition of complexing agents such as tropolone or DDTC (Fatoki *et al.,*  2000). While the tri- and di-substituted compounds can be extracted quantitatively, only about 60 % or less of the mono-substituted compounds are recovered. Two approaches have been evaluated to improve the extraction efficiency of mono- and di-organotin species: (i) the addition of complexing agents (e.g. diethylammonium-diethylthiocarbamate, DEA-DDC or (DDC), and (ii) alkylation in a reaction cell with Grignard reagent prior to the extraction (Fatoki *et al.,* 2000; Abalos *et al.*, 1997). On one hand, recoveries obtained by the first approach are satisfactory for di- and tri-organotin species but a clean-up step is usually needed. On the other hand, the second method yields satisfactory recoveries only for TBT and TPT. Further developments are needed to bring these methods to routine analysis. Apparently, no reliable and efficient method for extracting all organotins from sediment has

Hexane, benzene, toluene or dichloromethane (DCM) are non-polar solvents used for the extraction of organotins with complexing agent (Cai and Bayona, 1995; Jiang *et al.*, 2000; Abalos *et al.*, 1997). The efficiency with which butyltins are extracted from spiked sediment with non-polar solvents in the presence of complexing agents is satisfactory. In contrast, very poor recovery is obtained with monobutyl tin and dibutyltin with DCM without a complexing agent (Abalos *et al.*, 1997). With volatile solvents such as hexane, hexaneacetone, dichloromethane (Abalos *et al.*, 1997; Tanabe *et al.*, 1998; Dirkx *et al.,* 1992), Soxhlet extraction is applied without complexing agent (Willis, 1965), since the more polar solvents are incompatible with the Grignard reagents used later for derivatization and favor coextraction of organic interference compounds (Fatoki *et al*., 2000). Therefore, the current recommended procedures (Wade *et al.*, 1988; Abalos *et al.,* 1997) are based on the extraction

(TCyT) and failed for other species due to their high polarity.

methods (Fatoki *et al*., 2000).

yet been developed (Abalos *et al.*, 1997).

of low polar organotin complexes with tropolone or diethyldithiocarbanate (DDTC). Tropolone is preferred to DDTC (Dirkx *et al.*, 1994), as under acidic condition, this undergoes decomposition, giving rise to extractable interference (http://h2osparc.wq.ncsu.edu/info/hmetals.html). Sample preparation procedures before analyses, such as liquid-liquid extraction of organotin chelates with fresh tropolone or diethyl dithiocarbamate (DDTC) (Fatoki et al, 2000) are also convenient.

Non-polar solvent plus acids are used in complex matrices' extraction (Tanabe *et al.*, 1998). For example, sediment sample is treated with hydrochloric acid with shaking or sonification, followed by sequential solvent extraction (Wade *et al.*, 1988). Hydrobromic acid or acetic acid or a mixture is also used (Tolosa *et al.*, 1992; Martin *et al.*, 1994). Sonification has become the most widely used stirring method for sediment matrix, whereas, energymixing methods are used for biotic materials. Mixtures of solvents have been used to increase the polarity of the medium, hexane-ethyl acetate, hexane -diethyl ether, and chloroform-ethyl acetate (Fatoki, 2000). The salting out effect or ion-pairing effect is used to increase the efficiency of extraction of organotins (OTs) from aqueous phase to the organic phase, when HCl is used (Dirk *et al.,* 1992; Tao *et al.*, 1999). Polar solvents have also been used to achieve extraction. The polar solvents which have been used are aqueous (i) HCl (Ceulemans and Adams, 1995); (ii) HCl or HOAc in polar organic solvents (MeOH, acetone) (Kuballa *et al.*, 1995; Cai *et al.*, 1993); (iii) acetic acid (Quevauviller, 1996); (iv) net polar organic solvents (MeOH, DCM-MeOH, butanol, MeOH-EtOAc (Han and Weber, 1988; Apte and Gardner, 1998); (v) polar organic solvents in basic conditions (Pawliszyn, 1997). In this case, sonification is used in most procedures. Very recently, a focused microwave field has been introduced to reduce extraction time from hours to several munites (Donard *et al.*, 1995). In some cases, after the acid or polar solvent extraction, a liquid-liquid extraction (LLE) with a non-miscible solvent (benzene, CH3Cl-DCM, ETOAc-MeOH, DCM, hexane, cyclohexane, toluene, hexane-ETOAc) is used to recover OTs from the extract. Several authors (Abalos *et al.*, 1997) have used tropolone and salting out effect to increase the solubility of OTs in the organic solvent. Quite recently, as already mentioned, a more environmentally-friendly extraction technique developed is the supercritical fluid extraction (SFE). Advantages of these methods are shortened extraction time and limited amount of toxic solvents and acids used.

For biological samples analysis, tetramethyl ammonium (TMAH) hydrolysis is currently applied (Fatoki *et al.*, 2000), above room temperature (60oC) for several hours, for example 1- 2 h. (Leal *et al.*, 1995). The TMAH hydrolysis can be reduced from hours to minutes when the digestion is carried out under focused microwave irradiation. OTs are isolated from the hydrolyzed tissue by hexane (LLE) in the presence of tropolone. Alternatively, after a pH adjustment, simultaneous extraction derivatization with NaBEt4 is used to reduce the numbers of LLE, compared to Grignard reagents. Also, ethanolic- KOH at 60oC for 90 min or NaOH at 40oC for 20 min followed by pH adjustment and LLE has also been applied to the determination of OTs from biotic matrices (Nagase *et al.*, 1998). The digestion time in basic extraction conditions is critical due to the lack of stability of mono- and di-organotin compounds. Basic and enzymatic hydrolysis methods, which are restricted to biotic samples, lead to tissue solubilization. This makes the embedded organotin more available to extracting agent (Tanabe *et al.*, 1998).

In the case of heavy metals speciation, many extraction techniques have been reported with differences in methods approach depending on the aqueous, solid or gaseous nature of the

Speciation Methods for the Determination of Organotins (OTs)

cleanup or harsh instrumentation conditions (Lespes *et al.*, 1998).

The last step, which allows for the compound under investigation to be speciated, is detection. Many techniques have been developed although most of these methods are not commonly used due to their poor sensitivity or cost. From the detection point of view, GC is highly flexible (Fatoki *et al.,* 2000; Ndibewu *et al*., 2002]. In this respect, the following detector have been used for OTs speciation, GC (Fatoki *et al.,* 2000) coupled to flame ionization detection (FID), flame photometric (FPD) detection (Brickman, 1978); liquid chromatography (LC) (Fatoki *et al*., 2000), or supercritical fluid chromatography (SFC)

methylbutyltins.

and Heavy Metals (HMs) in the Freshwater and Marine Environments 37

*al.*, 1996). However, the method is time consuming, and requires strict anhydrous conditions and non-protic solvents, which necessitate solvent exchange when polar solvents are used as extracting agents. Furthermore, the LLE step becomes necessary to isolate the derivatized OTs. Cai *et al.,* (1994; 1995) found the formation of dialkyl mono- and disulfide when the derivatization is performed in-situ on a sediment sample before the SFE, which necessitates large excess of Grignard reagents. Similar side reactions occur when the derivatization is performed on the extracts. A wide range of reaction times is reported (Ashby and Craig, 1989) but too long exposure of phenyl to Grignard reagent can lead to deproportionation reactions. Some workers have reported substantial losses of the most volatile tin species when the derivatization is performed with methyl and ethyl Grignard reagents. It is, thus, advisable to avoid evaporation to dryness of derivatized OTs. Another limitation of the methyl derivatives is that they do not allow for the determination of the naturally occurring

The next step, usually after the derivatization step is the clean-up phase. Most of the analytical procedures based on GC determination require a clean-up step or process, usually after the derivatization step as mentioned above. Silica is the adsorbent mostly used. Other adsorbent candidates in use are: florisil (Harino *et al*., 1992), alumina (Dirkx and Adams, 1992), alumina-silica (Willis, 1965), amino and C18 cartridges, florisil-alumina, and florisilsilica (Harino *et al*., 1992; Dirkx and Adams, 1992). In most of the methods applied to sediments that use GC-MS or GC-flame photometric detection (FPD), a desulfurization with activated copper following a clean-up is performed. However, alkylsulfides generated during the Grignard derivatization from elemental sulfur occurring in the sediment are not removed by this procedure. Alternatively, other desulfurization reagents such as tetrabutyl ammonium hydogensulfate and sodium sulfide have been successfully applied (Okamura *et al*., 2000). Florisil is a preferred adsorbent for biotic matrix with a high lipid content. Hexane or hexane-Et2O mixtures are the most widely used eluents during the clean-up step because they allow GC determination without evaporation to dryness. More volatile solvent such as pentane is used to minimize the evaporation losses of the most volatile species. Other analytical procedures perform the clean-up before derivatization. Since underivatized OTs have a strong interaction in these adsorbents; polar eluents are needed to achieve quantitative recovery, which leads to poor clean-up efficiency. Tropolone in hexane has been used as an eluent in this case. Today, the most preferred approach, gradually gaining popularity, is the extraction of the analyte earlier derivatized in situ, preferably using sodium tetraethylborate (NaBEt4) (Thompson *et al*., 1998) and sodium tetrahydroborate (NaBH4) as a derivatization reagent (Balls, 1987). Hydride generation is more prone to interference and in the case of mono substituted organotin; it leads to very volatile derivatives, which can hardly be further pre-concentrated by evaporation of the extracting solvent. In addition, organotins are relatively reactive and decompose when subject to

species. Most of these techniques are in use today depending on individual situations and analytical goals. For the analysis of many metals in seawater or other matrices with strongly interfering elements, several different extraction techniques have been developed using coprecipitation with Co-APDC (ammonium pyrrolidine dithiocarbamate) or FeOH, or reductive precipitation using APDC, NaBH4, Fe, and Pd. These techniques allow quantitative extraction of metals from the interfering matrix. In addition, the extraction serves to pre-concentrate the metals, thus, improving detection limits.

After the above discussion, it is understood that, generally, testing and analysis of environmental pollutants demands the highest quality reagents for calibration and validation. Solvents used in the preparation of standard solutions must be validated free of interfering substances (Chemika-BioChemika Analytika–1995/96, 1905-1923). Quality parameters that need checking are: the physical characteristics and purity of the analytes, gravimetric data pertaining to solution preparation, actual concentration of analyte, chromatographic analysis of finished standard, and the expiration date or scheduled reassay date. The second step involving derivatization is reviewed. For organotins speciation, GC methods require a derivatization reaction to produce volatile OT compounds for separation (Attar, 1996). The methods of conversion of ionic alkyl tins into gas chromatographable species include: (i) in situ hybridization using NaBH4; (ii) ethylation with NaBEt4; (iii) derivatization using Grignard reagents: methyl-, ethyl-, propyl-, pentyl-, hexyl-magnesium chlorides/bromides.

In in-situ process, Hydride generation with NaBH4 has seldom been used in off-line methods, owing to the lack of hydride stability. However, this derivatization technique combined with CT-QFAAS allows for the determination of butyltins and highly volatile OTs (i.e methyltin), which cannot be determined by most off-line methods (Bayona, 1994). Furthermore; phenyltin cannot be analyzed by this method. The on-line HG-CT-QFAAS methodology allows for the reduction of the sample handling steps to a minimum, which makes this approach tone of the most rapid alternatives for the analysis of OTs (Quevauviller *et al.*, 1989). The amount of derivatization reagent needed to be optimized according to the matrix characteristics, since the matrix can inhibit the hybridization reaction. In this regard, the uncomplexed tropolone suppresses the hydride generation reaction. SPME technique has recently been used for speciation analysis of the hydride derivatives (Bayona and Cai, 1994). This method is also used in generating the hydride volatiles in the analysis of mercury using the cold vapor technique (CVAAS or CVAFS). Boron tetra-ethyl reagents have been developed (Schrader *et al.*, 1983; Leal *et al.*, 1995; Nagase *et al.*, 1995; Mueler, 1984) to minimize analyzing time. This allows carrying out the reaction in aqueous media under buffer conditions. In spiked river sediments, the derivatization yield of MBT using NaBEt4 is lower than that given by hybridization methods, but matrix effects are reduced (Fatoki *et al.*, 2000; Cai *et al.*, 1993). The method is particularly successful for aqueous samples, but lower derivatization yields than those given by the Grignard reactions are observed in complexed matrix containing large amounts of co-extracted compounds. The NaBEt4 procedure allows a simultaneous extractionderivatization in the buffer medium. The ethylated derivatives are recovered with non-polar solvents (Tao *et al.*, 199). SPME technique has recently been used for speciation analysis of ethyl organotin derivatives (Millán and Pawliszyn, 2000). Alkylation with a variety of Grignard reagents (e.g. methylation, ethylation, propylation, pentylation and hexylation) is the most widely used derivatization technique for water, sediment and the biota (Tolosa et

species. Most of these techniques are in use today depending on individual situations and analytical goals. For the analysis of many metals in seawater or other matrices with strongly interfering elements, several different extraction techniques have been developed using coprecipitation with Co-APDC (ammonium pyrrolidine dithiocarbamate) or FeOH, or reductive precipitation using APDC, NaBH4, Fe, and Pd. These techniques allow quantitative extraction of metals from the interfering matrix. In addition, the extraction

After the above discussion, it is understood that, generally, testing and analysis of environmental pollutants demands the highest quality reagents for calibration and validation. Solvents used in the preparation of standard solutions must be validated free of interfering substances (Chemika-BioChemika Analytika–1995/96, 1905-1923). Quality parameters that need checking are: the physical characteristics and purity of the analytes, gravimetric data pertaining to solution preparation, actual concentration of analyte, chromatographic analysis of finished standard, and the expiration date or scheduled reassay date. The second step involving derivatization is reviewed. For organotins speciation, GC methods require a derivatization reaction to produce volatile OT compounds for separation (Attar, 1996). The methods of conversion of ionic alkyl tins into gas chromatographable species include: (i) in situ hybridization using NaBH4; (ii) ethylation with NaBEt4; (iii) derivatization using Grignard reagents: methyl-, ethyl-, propyl-, pentyl-,

In in-situ process, Hydride generation with NaBH4 has seldom been used in off-line methods, owing to the lack of hydride stability. However, this derivatization technique combined with CT-QFAAS allows for the determination of butyltins and highly volatile OTs (i.e methyltin), which cannot be determined by most off-line methods (Bayona, 1994). Furthermore; phenyltin cannot be analyzed by this method. The on-line HG-CT-QFAAS methodology allows for the reduction of the sample handling steps to a minimum, which makes this approach tone of the most rapid alternatives for the analysis of OTs (Quevauviller *et al.*, 1989). The amount of derivatization reagent needed to be optimized according to the matrix characteristics, since the matrix can inhibit the hybridization reaction. In this regard, the uncomplexed tropolone suppresses the hydride generation reaction. SPME technique has recently been used for speciation analysis of the hydride derivatives (Bayona and Cai, 1994). This method is also used in generating the hydride volatiles in the analysis of mercury using the cold vapor technique (CVAAS or CVAFS). Boron tetra-ethyl reagents have been developed (Schrader *et al.*, 1983; Leal *et al.*, 1995; Nagase *et al.*, 1995; Mueler, 1984) to minimize analyzing time. This allows carrying out the reaction in aqueous media under buffer conditions. In spiked river sediments, the derivatization yield of MBT using NaBEt4 is lower than that given by hybridization methods, but matrix effects are reduced (Fatoki *et al.*, 2000; Cai *et al.*, 1993). The method is particularly successful for aqueous samples, but lower derivatization yields than those given by the Grignard reactions are observed in complexed matrix containing large amounts of co-extracted compounds. The NaBEt4 procedure allows a simultaneous extractionderivatization in the buffer medium. The ethylated derivatives are recovered with non-polar solvents (Tao *et al.*, 199). SPME technique has recently been used for speciation analysis of ethyl organotin derivatives (Millán and Pawliszyn, 2000). Alkylation with a variety of Grignard reagents (e.g. methylation, ethylation, propylation, pentylation and hexylation) is the most widely used derivatization technique for water, sediment and the biota (Tolosa et

serves to pre-concentrate the metals, thus, improving detection limits.

hexyl-magnesium chlorides/bromides.

*al.*, 1996). However, the method is time consuming, and requires strict anhydrous conditions and non-protic solvents, which necessitate solvent exchange when polar solvents are used as extracting agents. Furthermore, the LLE step becomes necessary to isolate the derivatized OTs. Cai *et al.,* (1994; 1995) found the formation of dialkyl mono- and disulfide when the derivatization is performed in-situ on a sediment sample before the SFE, which necessitates large excess of Grignard reagents. Similar side reactions occur when the derivatization is performed on the extracts. A wide range of reaction times is reported (Ashby and Craig, 1989) but too long exposure of phenyl to Grignard reagent can lead to deproportionation reactions. Some workers have reported substantial losses of the most volatile tin species when the derivatization is performed with methyl and ethyl Grignard reagents. It is, thus, advisable to avoid evaporation to dryness of derivatized OTs. Another limitation of the methyl derivatives is that they do not allow for the determination of the naturally occurring methylbutyltins.

The next step, usually after the derivatization step is the clean-up phase. Most of the analytical procedures based on GC determination require a clean-up step or process, usually after the derivatization step as mentioned above. Silica is the adsorbent mostly used. Other adsorbent candidates in use are: florisil (Harino *et al*., 1992), alumina (Dirkx and Adams, 1992), alumina-silica (Willis, 1965), amino and C18 cartridges, florisil-alumina, and florisilsilica (Harino *et al*., 1992; Dirkx and Adams, 1992). In most of the methods applied to sediments that use GC-MS or GC-flame photometric detection (FPD), a desulfurization with activated copper following a clean-up is performed. However, alkylsulfides generated during the Grignard derivatization from elemental sulfur occurring in the sediment are not removed by this procedure. Alternatively, other desulfurization reagents such as tetrabutyl ammonium hydogensulfate and sodium sulfide have been successfully applied (Okamura *et al*., 2000). Florisil is a preferred adsorbent for biotic matrix with a high lipid content. Hexane or hexane-Et2O mixtures are the most widely used eluents during the clean-up step because they allow GC determination without evaporation to dryness. More volatile solvent such as pentane is used to minimize the evaporation losses of the most volatile species. Other analytical procedures perform the clean-up before derivatization. Since underivatized OTs have a strong interaction in these adsorbents; polar eluents are needed to achieve quantitative recovery, which leads to poor clean-up efficiency. Tropolone in hexane has been used as an eluent in this case. Today, the most preferred approach, gradually gaining popularity, is the extraction of the analyte earlier derivatized in situ, preferably using sodium tetraethylborate (NaBEt4) (Thompson *et al*., 1998) and sodium tetrahydroborate (NaBH4) as a derivatization reagent (Balls, 1987). Hydride generation is more prone to interference and in the case of mono substituted organotin; it leads to very volatile derivatives, which can hardly be further pre-concentrated by evaporation of the extracting solvent. In addition, organotins are relatively reactive and decompose when subject to cleanup or harsh instrumentation conditions (Lespes *et al.*, 1998).

The last step, which allows for the compound under investigation to be speciated, is detection. Many techniques have been developed although most of these methods are not commonly used due to their poor sensitivity or cost. From the detection point of view, GC is highly flexible (Fatoki *et al.,* 2000; Ndibewu *et al*., 2002]. In this respect, the following detector have been used for OTs speciation, GC (Fatoki *et al.,* 2000) coupled to flame ionization detection (FID), flame photometric (FPD) detection (Brickman, 1978); liquid chromatography (LC) (Fatoki *et al*., 2000), or supercritical fluid chromatography (SFC)

Speciation Methods for the Determination of Organotins (OTs)

sulfonate is used as an ion-pair (Kumar *et al*., 1993).

reaction is performed after chromatographic separation (McKie, 1987).

(Kuballa *et al*., 1995).

and Heavy Metals (HMs) in the Freshwater and Marine Environments 39

siloxanes (DB-1710) allow the resolution between specific OTs (phenyl and cyclohexyl)

The use of liquid chromatographic separation is not very popular in speciation procedures. Most of the published works with LC have been done on standards (Fatoki *et al.*, 2000), with few on environmental samples (Yang *et al*., 1995; Suyani *et al*., 1989). In spite of the advantage of avoiding derivatization step, LC has some limitation arising from the insufficient sensitivity of the most common detector for the levels found in environmental samples. Butyltin are the most species considered but in some cases phenyltin is considered. Ion exchange chromatography is performed in the silica-based cation-exchange column and it has been the most applied (Rivaro *et al*., 1995; Leal *et al*., 1995). Mobile phases consist of mixtures of methanol or acetonitrile and water containing ammonium acetate or citrate. The separation of TBTs and DBTs amongst the other OTs is achieved at the same pH. In the separation of di- and triorganotin compounds based on normal phase mode, cyanopropyl have been used. The mobile phase consisted of high percentage of hexane together with polar solvent such as ethyl acetate, tetrahydrofuran (THF) and HOAc. A mobile phase consisting of tropolone in toluene has been used (Fatoki *et al.,* 2000). On one hand, reversedphase with octadecyl silane stationary phase (C18) has been used (Fatoki *et al.,* 2000) in the separation of butyltin compounds in sediments using a polar mobile phase containing complexing agent such as tropolone. On the other hand, reversed-phase ion pair approach has been used in the separation of tri-organotin compounds (Beyer *et al*., 1997). Polymeric based column (PRP-1) or octylsilane column was used, where pentane sulfonate or hexane

Several detectors or hyphenated techniques have been used in LC: AAS, ICP-MS (Beyer *et al*., 1997), fluorimetry, MS, laser-enhanced ionization (LEI) and ICP-AES. Among different AAS modes, flame AAS with pulse nebulization and off -line GFAAS were the earliest (Fatoki *et al.,* 2000). When ICP-MS is coupled to LC, pneumatic nebulizers and spray chambers are the common systems for sample introduction. ICP methods suffer incompatibility of most of the mobile phases. When fluorimetric detection is used, derivatization with fluoregenic reagent such as flavone derivatives is mandatory. The

In any analytical method, a few important parameters are important to assure quality and reliability of the method involved. Some of these parameters are: detection limits, calibration, accuracy and precision. Bearing this in mind during methods development, any analytical methods developed for speciation should, therefore, provide sufficient sensitivity allowing for the determination of individual organotin compounds and elemental heavy metals below set limits. Selected absolute detection limits according to analytical techniques and analytes are reviewed below. Among the non-chromatographic techniques, ion spray mass spectrometry (ISMS-MS) is ca. 4-order of magnitude more sensitive than GFAAS (Fatoki, 2000). In the group of the GC detection techniques, AED, MS in the electron impact (selected ion monitoring) and FPD have the detection limit in the sub-to-low picogram range (Fatoki *et al., 2000;* Ndibewu *et al.,* 2002*)*. The FPD configuration can lead to a remarkable difference in sensitivity. Filterless operation and quartz surface-induced luminescence are the most sensitive detection mode in FPD (Attar, 1996). Unfortunately, the dramatic deterioration of the selectivity due to sulfur emission at 390 nm was found in these operational modes. Also, oxidant flames can lead to poor selectivity since the luminescence

(Martin and Donard, 1994) with spectrometric (AAS) detection (DWAF, 1992; Prudente *et al*., 1999), atomic emission (AES) spectrometry (Fatoki *et al*., 2000; Ombaba and Barry, 1992), flame photometric (FPD) detection (Kumar *et al*., 1993; Fatoki and Ngassoum, 2000; Jiang *et al*., 2000), electron capture detection (ECD), mass (MS) spectrometry (Fatoki *et al.,* 2000) or induced coupled-plasma (ICP-MS) spectrometry (Fatoki *et al*., 2000). Most of these techniques are based on an extraction (Ndibewu *et al*., 2002) step followed by derivatization using Grignard reagents (Fatoki *et al.,* 2000, Abalos *et al*., 1997), sodium tetrahydroborate (Abalos *et al*., 1997) or sodium ethylborate (Abalos *et al*., 1997). However, some analytical techniques allow TBT determination by GFAAS after hybridization and selective extraction in water (Balls, 1987), sediments (Lespes *et al*., 1998) and biological samples (Prudente *et al*., 1992, Lespes *et al*., 1998). ECD and FID were used in the earlier speciation studies but seldom used during the last decade. The lack of selectivity and/or sensitivity of those detection systems for organotins led to their replacement by more sensitive low cost detector such as MS in the electron impact mode, FPD equipped with an interference filter at 610 nm or AAS. The low molecular masses of diagnostic ions in the electron impact or chemical ionization modes impair moderate selectivity in case of complex matrices (Morabito *et al*., 1995). Similarly, FPD suffers some interference associated with co-extracted sulfur species (Cai and Bayona, 1995). AED is one of the most sensitive and selective detection systems coupled to GC used in OT speciation. However the high cost and maintenance operation of the GC-microwave induced plasma (MIP)-AED system makes it unsuitable to monitoring studies involving a large number of samples.

Despite the more complex sample preparation procedure often required in GC because of insufficient volatility of the ionic organic compounds, GC is preferred (Sasaki *et al.*, 1988; Arakawa *et al*., 1983) to the liquid chromatography-based technique (Fatoki *et al.,* 2000) which suffers from poor resolution and usually a lack of sensitivity (Yang *et al*., 1995). Another advantage of GC over LC is the possibility of using several internal standards (IS) and surrogates, which allows the steps of analytical procedure to be traced. The main disadvantage of GC methods is that they usually require production of volatile OT derivatives to perform their separation. Packed columns are used exclusively in cold temperature (CT) when hydride derivatization is carried out. The hydrides are purged with a helium stream and trapped in a U-shaped packed column cooled by liquid N2. The column is then heated rapidly until the purging step is complete. This method is only successful for the determination of methyl and butyl tins. Capillary column methods gained acceptance during the 1990's (Fatoki *et al.,* 2000) and nowadays, they are commonly used rather than packed or megabore columns. Sample is usually introduced into the column by splitless injection because non-volatile co-injected compound is retained in the liner. Its limitation is the low sample capacity (up to 2 µL) and the discrimination of low volatile OTs against the high volatile tin species. Cold on-column and temperature programmable injectors avoid some of the limitations of the splitless mode and then allow up to 5 µL to be injected. In order to prevent column contamination, GC Tenax packing in the injection port or uncoated deactivated tubing has been used.

The high efficiency achieved by capillary GC (cGC) allows satisfactory resolution of OTs according to carbon number even with non-polar, non-selective stationary phases, such as dimethylpolysiloxane or 5 % diphenyldimethylpolysiloxane (DB-1, HP-1, SE-30). OTs with equal number of carbon co-elute (Mueller, 1987). The mid-polarity stationary-phases such as 50 % diphenyldimethylsiloxane (OV-17) or 14 % cyanopropylphenyl 86 % dimethyl

(Martin and Donard, 1994) with spectrometric (AAS) detection (DWAF, 1992; Prudente *et al*., 1999), atomic emission (AES) spectrometry (Fatoki *et al*., 2000; Ombaba and Barry, 1992), flame photometric (FPD) detection (Kumar *et al*., 1993; Fatoki and Ngassoum, 2000; Jiang *et al*., 2000), electron capture detection (ECD), mass (MS) spectrometry (Fatoki *et al.,* 2000) or induced coupled-plasma (ICP-MS) spectrometry (Fatoki *et al*., 2000). Most of these techniques are based on an extraction (Ndibewu *et al*., 2002) step followed by derivatization using Grignard reagents (Fatoki *et al.,* 2000, Abalos *et al*., 1997), sodium tetrahydroborate (Abalos *et al*., 1997) or sodium ethylborate (Abalos *et al*., 1997). However, some analytical techniques allow TBT determination by GFAAS after hybridization and selective extraction in water (Balls, 1987), sediments (Lespes *et al*., 1998) and biological samples (Prudente *et al*., 1992, Lespes *et al*., 1998). ECD and FID were used in the earlier speciation studies but seldom used during the last decade. The lack of selectivity and/or sensitivity of those detection systems for organotins led to their replacement by more sensitive low cost detector such as MS in the electron impact mode, FPD equipped with an interference filter at 610 nm or AAS. The low molecular masses of diagnostic ions in the electron impact or chemical ionization modes impair moderate selectivity in case of complex matrices (Morabito *et al*., 1995). Similarly, FPD suffers some interference associated with co-extracted sulfur species (Cai and Bayona, 1995). AED is one of the most sensitive and selective detection systems coupled to GC used in OT speciation. However the high cost and maintenance operation of the GC-microwave induced plasma (MIP)-AED system makes it unsuitable to monitoring

Despite the more complex sample preparation procedure often required in GC because of insufficient volatility of the ionic organic compounds, GC is preferred (Sasaki *et al.*, 1988; Arakawa *et al*., 1983) to the liquid chromatography-based technique (Fatoki *et al.,* 2000) which suffers from poor resolution and usually a lack of sensitivity (Yang *et al*., 1995). Another advantage of GC over LC is the possibility of using several internal standards (IS) and surrogates, which allows the steps of analytical procedure to be traced. The main disadvantage of GC methods is that they usually require production of volatile OT derivatives to perform their separation. Packed columns are used exclusively in cold temperature (CT) when hydride derivatization is carried out. The hydrides are purged with a helium stream and trapped in a U-shaped packed column cooled by liquid N2. The column is then heated rapidly until the purging step is complete. This method is only successful for the determination of methyl and butyl tins. Capillary column methods gained acceptance during the 1990's (Fatoki *et al.,* 2000) and nowadays, they are commonly used rather than packed or megabore columns. Sample is usually introduced into the column by splitless injection because non-volatile co-injected compound is retained in the liner. Its limitation is the low sample capacity (up to 2 µL) and the discrimination of low volatile OTs against the high volatile tin species. Cold on-column and temperature programmable injectors avoid some of the limitations of the splitless mode and then allow up to 5 µL to be injected. In order to prevent column contamination, GC Tenax packing in the injection port or uncoated

The high efficiency achieved by capillary GC (cGC) allows satisfactory resolution of OTs according to carbon number even with non-polar, non-selective stationary phases, such as dimethylpolysiloxane or 5 % diphenyldimethylpolysiloxane (DB-1, HP-1, SE-30). OTs with equal number of carbon co-elute (Mueller, 1987). The mid-polarity stationary-phases such as 50 % diphenyldimethylsiloxane (OV-17) or 14 % cyanopropylphenyl 86 % dimethyl

studies involving a large number of samples.

deactivated tubing has been used.

siloxanes (DB-1710) allow the resolution between specific OTs (phenyl and cyclohexyl) (Kuballa *et al*., 1995).

The use of liquid chromatographic separation is not very popular in speciation procedures. Most of the published works with LC have been done on standards (Fatoki *et al.*, 2000), with few on environmental samples (Yang *et al*., 1995; Suyani *et al*., 1989). In spite of the advantage of avoiding derivatization step, LC has some limitation arising from the insufficient sensitivity of the most common detector for the levels found in environmental samples. Butyltin are the most species considered but in some cases phenyltin is considered.

Ion exchange chromatography is performed in the silica-based cation-exchange column and it has been the most applied (Rivaro *et al*., 1995; Leal *et al*., 1995). Mobile phases consist of mixtures of methanol or acetonitrile and water containing ammonium acetate or citrate. The separation of TBTs and DBTs amongst the other OTs is achieved at the same pH. In the separation of di- and triorganotin compounds based on normal phase mode, cyanopropyl have been used. The mobile phase consisted of high percentage of hexane together with polar solvent such as ethyl acetate, tetrahydrofuran (THF) and HOAc. A mobile phase consisting of tropolone in toluene has been used (Fatoki *et al.,* 2000). On one hand, reversedphase with octadecyl silane stationary phase (C18) has been used (Fatoki *et al.,* 2000) in the separation of butyltin compounds in sediments using a polar mobile phase containing complexing agent such as tropolone. On the other hand, reversed-phase ion pair approach has been used in the separation of tri-organotin compounds (Beyer *et al*., 1997). Polymeric based column (PRP-1) or octylsilane column was used, where pentane sulfonate or hexane sulfonate is used as an ion-pair (Kumar *et al*., 1993).

Several detectors or hyphenated techniques have been used in LC: AAS, ICP-MS (Beyer *et al*., 1997), fluorimetry, MS, laser-enhanced ionization (LEI) and ICP-AES. Among different AAS modes, flame AAS with pulse nebulization and off -line GFAAS were the earliest (Fatoki *et al.,* 2000). When ICP-MS is coupled to LC, pneumatic nebulizers and spray chambers are the common systems for sample introduction. ICP methods suffer incompatibility of most of the mobile phases. When fluorimetric detection is used, derivatization with fluoregenic reagent such as flavone derivatives is mandatory. The reaction is performed after chromatographic separation (McKie, 1987).

In any analytical method, a few important parameters are important to assure quality and reliability of the method involved. Some of these parameters are: detection limits, calibration, accuracy and precision. Bearing this in mind during methods development, any analytical methods developed for speciation should, therefore, provide sufficient sensitivity allowing for the determination of individual organotin compounds and elemental heavy metals below set limits. Selected absolute detection limits according to analytical techniques and analytes are reviewed below. Among the non-chromatographic techniques, ion spray mass spectrometry (ISMS-MS) is ca. 4-order of magnitude more sensitive than GFAAS (Fatoki, 2000). In the group of the GC detection techniques, AED, MS in the electron impact (selected ion monitoring) and FPD have the detection limit in the sub-to-low picogram range (Fatoki *et al., 2000;* Ndibewu *et al.,* 2002*)*. The FPD configuration can lead to a remarkable difference in sensitivity. Filterless operation and quartz surface-induced luminescence are the most sensitive detection mode in FPD (Attar, 1996). Unfortunately, the dramatic deterioration of the selectivity due to sulfur emission at 390 nm was found in these operational modes. Also, oxidant flames can lead to poor selectivity since the luminescence

Speciation Methods for the Determination of Organotins (OTs)

*al.*, 2000; Garcia-Romero *et al*., 1993).

such as phenyl tin species is necessary.

certification of MBT in CRM-462 (Martin *et al.*, 1994).

achieved with natural samples (Abalos *et al.*, 1997).

and Heavy Metals (HMs) in the Freshwater and Marine Environments 41

(TePeT), and triphenyltin (TPT). Generally, only IS and or surrogate is used but some alternative approaches have been proposed: (i) the used of different IS's such as monophenyltriethyltin (MPTT), diphenyldiethyltin (DPDT), triphenylethyltin (TPEtT) and tributylmethyltin (TBMEtT), depending on the nature of OTs being determined. This has been shown to be the more accurate way for correcting variations of the alkylation step (Stab *et al*., 1994), (ii) the use of several surrogates with different degrees of alkylation (Tripropyltin (TPrT), monophenyltin (MPT), diphenyltin (DPT) and triphenyltin (TPT)), in order to mach the behaviour of the different OTs in moieties in the extraction step (Fatoki *et* 

After considering both the detection limits and calibration, the accuracy of the analytical procedures is mostly evaluated through the analysis of either certified reference materials (CRMs) or spiked samples. In the field of OTs in sediments, nowadays, there are two CMRs available (Fatoki *et al*., 2000): the harbor sediments, PACS-1 with certified value of MBT (280 ± 170, ng g-1 as Sn), DBT (1160 ± 180 ng g-1 as Sn) and TBT (1270 ± 220 ng g-1 as Sn), and the coastal sediment CRM-462 with certified values for DBT (63 ± 8 ng g-1 as Sn) and TBT (24 ± 6 ng g-1 as Sn) (Fatoki et al., 2000; Abalos *et al*., 1997). There is also the reference material RM-424, with a reference value for TBT (8 5 ng g-1 as Sn) and indicative value for DBT (27 ± 10 ng g-1 as Sn) and MBT (174 ± 36 ng g-1 as Sn) (Fatoki *et al.*, 2000). The situation now is that, the CRM only allows for the assessment of the accuracy of butyltin compounds, and thus, the need for more CRMs with certified values for other OTs of environmental relevance,

Although the analysis of CRMs is preferable to that of spiked samples, only a few papers have reported the use of sediments certified reference materials, and in most cases PACS-1 is the CRM analyzed (Abalos *et al.*, 1997). This is probably due to the fact that PACS-1 was the first CRM available for organotins, or because concentration levels of OTs are higher in PACS-1 than in CRM-462. In relation to the analysis of MBT in PACS-1, some problems have been reported though. None of the ten methods evaluated by Zhang *et al*. (1996) could recover MBT from PACS-1 satisfactorily. A higher scatter of results has also prevented the

In the field of biological samples, only one CRM for OTs has been available since 1991. This is a fish tissue (sea bass) from the National Institute for Environmental Studies (NIES) in Japan, with a certified value for TBT (475 ± 36 ng g-1 as Sn), and indicative value of TPHT (1942 ng g-1 as Sn (Abalos *et al.*, 1997). Another method for the assessment of the accuracy of the analytical methods is based on the analyses of spiked samples, and the determination of the recoveries obtained for each analyte (Fatoki *et al.*, 2000). The analyses of spiked samples are carried out in most of the papers reviewed for quality assurance. In this case the main problem lies in how the spiking has been performed, probably because; this is one of the most critical points. In any case, experiments should be performed with several kinds of matrices, and at several concentration levels, always in the range of concentrations usually found in environmental samples (Cai *et al.*, 1994). Moreover, it should be taken into account that the availability of spiked analytes in the extraction step can be higher than that of the same substances incorporated into the matrices in the environment. So, using spiked samples can lead to an overestimation of the extraction efficiency, and, therefore, quantitative recoveries from spiked materials do not ensure that the same result will be

at 610 nm is attributed to tin hydride (Gomez *et al*., 1994; Martin *et al*., 1987). The sensitivity of the AES is strongly dependent on the plasma source. In this regard, alternating current plasma (ACP-AES) (Ombaba and Barry, 1992) has detection limit at least two orders of magnitude higher than MIP-AES.

The GC-QFAAS techniques have LODs ca. two orders of magnitude higher than the former detection systems (ECD, FPD, MS, AED) coupled to GC techniques. Nevertheless, the suitable design of the interface and GCs columns can improve the sensitivity of AAS by at least one order of magnitude (Kuballa *et al*., 1995). Among the LC methods, ICP-MS detection (Yang *et al*., 1995; Suyani *et al*., 1989), either with ultrasonic or pneumatic nebulization is the most sensitive for all the OTs, and is comparable to the most sensitive GC methods (Kumar *et al*., 1993). Concerning LC-MS interface, only thermospray has been applied to environmental studies; it has moderate sensitivity with a detection limit about 2 or 3-orders of magnitude higher than ICP-MS. The sensitivity attained with fluorimetric detection depend both on the species and the fluoregenic reagents used, and in some cases very low detection limits are achieved, only improved by LC-ICP-MS by one order of magnitude.

Calibration is another essential operation in the analytical method procedures. In some papers, especially in those devoted to environmental monitoring, little information, if any is provided on this aspect (Abalos *et al*., 1997) is very limited. In methods based on GC, calibration is generally carried out with an internal standard; however, external standards are almost extensively used. In contrast, the standard addition method is seldom used (Abalos *et al*., 1997). In those methods that involved cold trapping of volatile species, (hydride or ethyl derivatives) quantitation is usually performed by the method of standard addition or with matrix matched standards (Abalos *et al*., 1997, Fatoki *et al., 2000)*. When the technique applied is LC, calibration is usually performed with external standards, although the standard addition method is sometimes used (Abalos *et al*., 1997). When the external standards are used, the standard solution must be subjected to an entire extraction procedure (Tam and Wong, 1995; Sasaki *et al*., 1988). In other cases, in order to account for the matrix effects, matrix- matched standards are proposed (Han and Weber, 1988). However, suitable analyte-free matrices to match sample matrices may not be available. When using internal standard methods, several approaches are proposed. In the most common approach, the substance used as internal standard (IS) is added to the extracts before the derivatization step, usually as trialkyltin, or just before the injection to the chromatograph as the tetraalkyltin (Fatoki *et al., 2000*, Abalos *et al.*, 1997). In the first case, the IS affords a compensation for the incompleteness of the derivatizatioin reaction, for the possible losses occurring in the operations subsequent to derivatization (extractions, evaporations, clean up) and for the instrumental variability. In the second case, it only compensates for the uncontrolled variations in the chromatographic measurements (Abalos *et al*., 1997). A second approach consists of the addition of IS (in this case also called surrogate) at the beginning of the extraction process, providing the compensation for the losses taking place in the whole process, including the variability in the determination step (Arakawa *et al*., 1983; Pereira *et al*., 1999). Some authors used both the surrogate and IS. This allows for the calculation of the recovery of the substance added as surrogate and, on this basis, correction of the amount of the analytes recovered (Willis, 1965). The substances most commonly used as IS or surrogates are tripropyltin (TPT), tetrabutyltin (TBT), tetraphenyltin

at 610 nm is attributed to tin hydride (Gomez *et al*., 1994; Martin *et al*., 1987). The sensitivity of the AES is strongly dependent on the plasma source. In this regard, alternating current plasma (ACP-AES) (Ombaba and Barry, 1992) has detection limit at least two orders of

The GC-QFAAS techniques have LODs ca. two orders of magnitude higher than the former detection systems (ECD, FPD, MS, AED) coupled to GC techniques. Nevertheless, the suitable design of the interface and GCs columns can improve the sensitivity of AAS by at least one order of magnitude (Kuballa *et al*., 1995). Among the LC methods, ICP-MS detection (Yang *et al*., 1995; Suyani *et al*., 1989), either with ultrasonic or pneumatic nebulization is the most sensitive for all the OTs, and is comparable to the most sensitive GC methods (Kumar *et al*., 1993). Concerning LC-MS interface, only thermospray has been applied to environmental studies; it has moderate sensitivity with a detection limit about 2 or 3-orders of magnitude higher than ICP-MS. The sensitivity attained with fluorimetric detection depend both on the species and the fluoregenic reagents used, and in some cases very low detection limits are achieved, only improved by LC-ICP-MS by one order of

Calibration is another essential operation in the analytical method procedures. In some papers, especially in those devoted to environmental monitoring, little information, if any is provided on this aspect (Abalos *et al*., 1997) is very limited. In methods based on GC, calibration is generally carried out with an internal standard; however, external standards are almost extensively used. In contrast, the standard addition method is seldom used (Abalos *et al*., 1997). In those methods that involved cold trapping of volatile species, (hydride or ethyl derivatives) quantitation is usually performed by the method of standard addition or with matrix matched standards (Abalos *et al*., 1997, Fatoki *et al., 2000)*. When the technique applied is LC, calibration is usually performed with external standards, although the standard addition method is sometimes used (Abalos *et al*., 1997). When the external standards are used, the standard solution must be subjected to an entire extraction procedure (Tam and Wong, 1995; Sasaki *et al*., 1988). In other cases, in order to account for the matrix effects, matrix- matched standards are proposed (Han and Weber, 1988). However, suitable analyte-free matrices to match sample matrices may not be available. When using internal standard methods, several approaches are proposed. In the most common approach, the substance used as internal standard (IS) is added to the extracts before the derivatization step, usually as trialkyltin, or just before the injection to the chromatograph as the tetraalkyltin (Fatoki *et al., 2000*, Abalos *et al.*, 1997). In the first case, the IS affords a compensation for the incompleteness of the derivatizatioin reaction, for the possible losses occurring in the operations subsequent to derivatization (extractions, evaporations, clean up) and for the instrumental variability. In the second case, it only compensates for the uncontrolled variations in the chromatographic measurements (Abalos *et al*., 1997). A second approach consists of the addition of IS (in this case also called surrogate) at the beginning of the extraction process, providing the compensation for the losses taking place in the whole process, including the variability in the determination step (Arakawa *et al*., 1983; Pereira *et al*., 1999). Some authors used both the surrogate and IS. This allows for the calculation of the recovery of the substance added as surrogate and, on this basis, correction of the amount of the analytes recovered (Willis, 1965). The substances most commonly used as IS or surrogates are tripropyltin (TPT), tetrabutyltin (TBT), tetraphenyltin

magnitude higher than MIP-AES.

magnitude.

(TePeT), and triphenyltin (TPT). Generally, only IS and or surrogate is used but some alternative approaches have been proposed: (i) the used of different IS's such as monophenyltriethyltin (MPTT), diphenyldiethyltin (DPDT), triphenylethyltin (TPEtT) and tributylmethyltin (TBMEtT), depending on the nature of OTs being determined. This has been shown to be the more accurate way for correcting variations of the alkylation step (Stab *et al*., 1994), (ii) the use of several surrogates with different degrees of alkylation (Tripropyltin (TPrT), monophenyltin (MPT), diphenyltin (DPT) and triphenyltin (TPT)), in order to mach the behaviour of the different OTs in moieties in the extraction step (Fatoki *et al.*, 2000; Garcia-Romero *et al*., 1993).

After considering both the detection limits and calibration, the accuracy of the analytical procedures is mostly evaluated through the analysis of either certified reference materials (CRMs) or spiked samples. In the field of OTs in sediments, nowadays, there are two CMRs available (Fatoki *et al*., 2000): the harbor sediments, PACS-1 with certified value of MBT (280 ± 170, ng g-1 as Sn), DBT (1160 ± 180 ng g-1 as Sn) and TBT (1270 ± 220 ng g-1 as Sn), and the coastal sediment CRM-462 with certified values for DBT (63 ± 8 ng g-1 as Sn) and TBT (24 ± 6 ng g-1 as Sn) (Fatoki et al., 2000; Abalos *et al*., 1997). There is also the reference material RM-424, with a reference value for TBT (8 5 ng g-1 as Sn) and indicative value for DBT (27 ± 10 ng g-1 as Sn) and MBT (174 ± 36 ng g-1 as Sn) (Fatoki *et al.*, 2000). The situation now is that, the CRM only allows for the assessment of the accuracy of butyltin compounds, and thus, the need for more CRMs with certified values for other OTs of environmental relevance, such as phenyl tin species is necessary.

Although the analysis of CRMs is preferable to that of spiked samples, only a few papers have reported the use of sediments certified reference materials, and in most cases PACS-1 is the CRM analyzed (Abalos *et al.*, 1997). This is probably due to the fact that PACS-1 was the first CRM available for organotins, or because concentration levels of OTs are higher in PACS-1 than in CRM-462. In relation to the analysis of MBT in PACS-1, some problems have been reported though. None of the ten methods evaluated by Zhang *et al*. (1996) could recover MBT from PACS-1 satisfactorily. A higher scatter of results has also prevented the certification of MBT in CRM-462 (Martin *et al.*, 1994).

In the field of biological samples, only one CRM for OTs has been available since 1991. This is a fish tissue (sea bass) from the National Institute for Environmental Studies (NIES) in Japan, with a certified value for TBT (475 ± 36 ng g-1 as Sn), and indicative value of TPHT (1942 ng g-1 as Sn (Abalos *et al.*, 1997). Another method for the assessment of the accuracy of the analytical methods is based on the analyses of spiked samples, and the determination of the recoveries obtained for each analyte (Fatoki *et al.*, 2000). The analyses of spiked samples are carried out in most of the papers reviewed for quality assurance. In this case the main problem lies in how the spiking has been performed, probably because; this is one of the most critical points. In any case, experiments should be performed with several kinds of matrices, and at several concentration levels, always in the range of concentrations usually found in environmental samples (Cai *et al.*, 1994). Moreover, it should be taken into account that the availability of spiked analytes in the extraction step can be higher than that of the same substances incorporated into the matrices in the environment. So, using spiked samples can lead to an overestimation of the extraction efficiency, and, therefore, quantitative recoveries from spiked materials do not ensure that the same result will be achieved with natural samples (Abalos *et al.*, 1997).

Speciation Methods for the Determination of Organotins (OTs)

Biological material (biota) (Also suitable for water and sediment)

and Heavy Metals (HMs) in the Freshwater and Marine Environments 43

TECHNIQUES

SPECIES

MBT.

GC/FPD TBT, DBT,

SAMPLE SAMPLE TREATMENT STEPS DETERMINATION

Table 1. Recommended analytical procedures for the speciation analysis of organotins

Precision is the hardcore of all analytical operations. The precision of analytical methods reviewed in this chapter corresponds to the whole analytical procedure, that is to say: extraction, the derivatization and the determination technique (Cai *et al.*, 1994). An attempt is made to point out some trends about the precision of the reviewed methods. Two groups including the method commonly applied have been considered: those based on GC-FPD as determination technique and those based on CT-QFAAS. Independent of the analytical method used, the analysis of OTs in biological material gives more precise results than in sediments (Fatoki, 2000; Abalos *et al.*, 1997). For instance, in the case of TBTs, relative standard deviations (R.S.Ds) calculated as the mean of the different methods and concentration, are 12% and 8.5% for sediments and biological materials, respectively. The precision of GC-FPD seems to be, somewhat, better than those using CT-QFAAS: 10.5 versus 13 for sediment and 7 versus10 for biological materials (Fatoki, 2000). This trend is also noticeable in the results of the certification campaign of coastal sediments (CRM-462) of the European Commission (Quevauviller, 1996). One of the reasons for the higher precision for GC-FPD may be the fact that ISs are used in the calibration step of this technique, whereas, in the case of CT-QFAAS, the standard addition method is usually applied. The MIP-AED was introduced in the '90s (nineties) in the field of GC. It has not been widely applied to the analysis of OTs. However, some trend in the precision of this technique can be pointed out. An inter-laboratory study carried out in the USA among ten laboratories (Sharron *et al.*, 1995) that analyzed fourteen OTs compounds in three pentylated extracts of soils and sediments gave inter-laboratory R.S.Ds between 2 and 4 % for most compounds. Some researchers have applied GC-MS to OTs analysis, the R.S.D given are higher than the

Various techniques (Unicam, 1992; Prudente *et al.*, 1999; Nriagu, 1996; Fatoki *et al.*, 2002),) have been developed and used for speciation analysis of heavy metals or analysis of total metals. The modification and development of these methods sometimes not only require acute analytical skills but do involve several steps. In order to reduce the number of steps and for optimization purposes], instrumentation design and set-ups nowadays use the combination of a number of the techniques. The validity of an analytical method is based, in part, on the procedures used for sample collection and analysis, and data interpretation. In many instances these procedures use approaches that have been refined over many years and are accepted by professionals as good practice. However, the multitude of variables within a specific workplace requires the professional to exercise judgment in the design of a

1) 500-1000l mL water or 1-5 g sediment or 1g fish tissue in 10 ml TMAH + 10 mL hexane or 10-500 l air trapping on carbosive, eluted with hexane

2) Acetate buffer 4 + 10 mL hexane + 4 ml NaBEt4 (2% sol.), reaction 30 min, separation, evaporation to 1 mL 3) Clean up, 0.5 g silica gel 80-100 or alumina elution 5mL hexane

techniques previously commented upon by Fatoki *et al.* (2000).

particular assessment.


TECHNIQUES

GC/MS (if possible)

> GC/FPD GC/MS

SPECIES

MBT.

MBT

MBT.

(Sn)

TBT, DBT, MBT.

MBT,

TBT

GC/FPD TBT, DBT,

GC/FPD TBT, DBT,

GC/FPD TBT, DBT,

GFAAS Total tin

GC/FPD TBT, DBT,

SAMPLE SAMPLE TREATMENT STEPS DETERMINATION

Water 1) 50 ml water + 50 acetate buffer pH 3.3 2) 1 mL NaBH4 (3 %) in water)

Water 1) 1000 mL water 70 oC + 0.6 g NaBH4

Water 1) 250 mL water + adjust pH = 6 tris +

Water 1) 500 mL H2O + 5-20 mL HBr (48 %) +

2) 3 mL MeMgBr (2.5 M sol. In diethylether), for 30 min., 3) Cooling + 25 mL 1 N H2SO4, shaking, separation, evaporation

Water 1) 25 ml water + standard 2) + Pd

shaking for 15 min.,

2) air purged and trapping Porapak

3) CH2Cl2 elution, evaporation to 0.1 mL

AcOH + 1 mL isooctane 0.1 ml NaBEt4 (2 % sol.) stirred 30 min, separation

standards + 25 mL benzene tropolone,

1) 200 mL H2O + 20 mL buffer pH 5 or 3 g wet sediment + 15 g Na2SO4, mixing, Soxhlet extraction 9 hrs (110 mL hexane-

2) acetate buffer 4 + 10 mL hexane + 4 mL NaBTEt4 (2% sol.), reaction 30 min, separation, evaporation to 1 mL 3) clean up, 0.5 g silica gel 80-100 or alumina elution 5 mL hexane

1) 1000 mL H2O + 5 mL conc. HCl 2) + 3\* 10 mL pentane extraction by shaking, separation and evaporation to 5

4) + 5 mL H2O + 0.5 conc. HCl,

3) + 2 mL MeMgBr sol. reaction 10 min

separation + CaCl2 evaporation to 1 ml 5) clean up, 0.5 g silica gel 60 elution

1') 20 g wet sediment + conc. HCl to pH 2 2') + 4 - 10 ml diethyl ether extraction by shaking separation and evaporation to 5 ml 3' – 6') following the same steps

3) SPME 15 min

cartridge,

solution

to 25 mL

acetone 9:1,

mL

pentane

as for water sample

Sediment

Sediment

(Also suitable for water)

(Also suitable for water)


Table 1. Recommended analytical procedures for the speciation analysis of organotins

Precision is the hardcore of all analytical operations. The precision of analytical methods reviewed in this chapter corresponds to the whole analytical procedure, that is to say: extraction, the derivatization and the determination technique (Cai *et al.*, 1994). An attempt is made to point out some trends about the precision of the reviewed methods. Two groups including the method commonly applied have been considered: those based on GC-FPD as determination technique and those based on CT-QFAAS. Independent of the analytical method used, the analysis of OTs in biological material gives more precise results than in sediments (Fatoki, 2000; Abalos *et al.*, 1997). For instance, in the case of TBTs, relative standard deviations (R.S.Ds) calculated as the mean of the different methods and concentration, are 12% and 8.5% for sediments and biological materials, respectively. The precision of GC-FPD seems to be, somewhat, better than those using CT-QFAAS: 10.5 versus 13 for sediment and 7 versus10 for biological materials (Fatoki, 2000). This trend is also noticeable in the results of the certification campaign of coastal sediments (CRM-462) of the European Commission (Quevauviller, 1996). One of the reasons for the higher precision for GC-FPD may be the fact that ISs are used in the calibration step of this technique, whereas, in the case of CT-QFAAS, the standard addition method is usually applied. The MIP-AED was introduced in the '90s (nineties) in the field of GC. It has not been widely applied to the analysis of OTs. However, some trend in the precision of this technique can be pointed out. An inter-laboratory study carried out in the USA among ten laboratories (Sharron *et al.*, 1995) that analyzed fourteen OTs compounds in three pentylated extracts of soils and sediments gave inter-laboratory R.S.Ds between 2 and 4 % for most compounds. Some researchers have applied GC-MS to OTs analysis, the R.S.D given are higher than the techniques previously commented upon by Fatoki *et al.* (2000).

Various techniques (Unicam, 1992; Prudente *et al.*, 1999; Nriagu, 1996; Fatoki *et al.*, 2002),) have been developed and used for speciation analysis of heavy metals or analysis of total metals. The modification and development of these methods sometimes not only require acute analytical skills but do involve several steps. In order to reduce the number of steps and for optimization purposes], instrumentation design and set-ups nowadays use the combination of a number of the techniques. The validity of an analytical method is based, in part, on the procedures used for sample collection and analysis, and data interpretation. In many instances these procedures use approaches that have been refined over many years and are accepted by professionals as good practice. However, the multitude of variables within a specific workplace requires the professional to exercise judgment in the design of a particular assessment.

Speciation Methods for the Determination of Organotins (OTs)

acid (Unicam, 1992).

analysis environment.

chromatographic separation (Wade *et al.,* 1988).

and Heavy Metals (HMs) in the Freshwater and Marine Environments 45

is extracted with solvent, the inorganic tin is determined by AAS after digestion with nitric

For reliability, efficiency and sensitivity, most trace metals analyses currently involve inductively coupled plasma–mass spectrometry (ICP-MS). The standard ICP-MS technique works quite well for many matrices. But for some element/matrix combinations, it gives poor detection limits or accuracy because of elemental or molecular interferences. In choosing the most appropriate methods to analyze any element/matrix combination, ICP-MS has been coupled to dynamic reaction cell (IC-DRC-MS) (Han and Weber, 1988). This combined technique can detect trace levels in complex matrices, where the standard ICP-MS would be prone to interferences; the standard has also been coupled to a micro-mass platform ICP-MS with collision cell technology (IC-ICP-MS) (Han and Weber, 1988). Extremely low detection limits (especially for the higher mass elements) even in matrices traditionally considered difficult are achievable in an ultra-clean sample preparation and

An analytical technique utilizing hydride generation and atomic fluorescence spectrometry (HG-AFS) (Unicam, 1992) or atomic absorption (HG-AAS) has been developed for the analysis of arsenic, antimony, and selenium at either ultra-trace levels or in complex matrices. With HG-AFS, we can accurately measure total arsenic, antimony, and selenium in nearly all matrices at single-digit parts-per-trillion levels [http://www.frontiergeosciences.com/ebru]. Speciation information could be determined using modifications of this technique, including cryogenic trapping/GC and ion

For mercury speciation, relatively poor sensitivity is provided by transitional flame absorption. Alternate atomization techniques for the AA determination of this element have been developed (Cai and Bayona, 1995; Shrader *et al.,* 1983; Maguire *et al.,* 1982; Wade *et al.*, 1988). Amongst them, the cold vapor atomic absorption technique has received the greatest attention (Schrader *et al.*, 1983, Wade *et al.,* 1988). Other techniques employ cold vapor atomic fluorescence spectrometers (CVAFS) (http://www.frontiergeosciences.com/ebru/), which give unparalleled sensitivity for the determination of low-level total mercury. Using this detector in combination with gold amalgamation or aqueous phase ethylation plus gas chromatographic separation allows determination of Hg speciation at the parts-per-quadrillion level (Wade *et al.*, 1988). Furnace methods for mercury are not recommended due to the extreme volatility of mercury, which has a significant vapor pressure even at room temperature (Stab *et al.,* 1992; Maenpa *et al.,* 2002). Although the first cold vapor principle was proposed by Poluekov and co-workers in 1963, the most popular method credited to Hatch and Ott was published in 1968 (Shrader *et al.,* 1983). In this method, an acidified solution containing mercury is reacted with stannous chloride in a vessel external to the AA instrument. Ground state mercury atoms are produced which subsequently are transported by an air or inert gas flow to an absorption cell installed in the AA instrument. This method provides sensitivities approximately four orders of magnitude better than flame AA (Schrader *et al.*, 1983; Wade *et al*., 1988). It is critical to note that, unlike in the case of organotins speciation where a global approach is favorable, heavy metals speciation will very much require methods choice to consider the thermodynamics of the elemental compound (solid/liquid or gaseous state at room temperature). Therefore, methods for each metal speciation will be specific although the principle underlining the different steps are similar.


Table 2. Comparative extraction and detection techniques for speciation of organotins

For the analysis of total metals (including all metals, organically and inorganically bound, both dissolved and particulate) most samples will require digestion before analysis to reduce organic matter interference and to convert metal to a form that can be analyzed by atomic absorption spectroscopy or inductively coupled plasma spectrometry (APHA, 1992),

For speciation analysis, a few common direct determination methods have earlier been employed (Unicam, 1992). These are the dissolved metals by air/acetylene and direct determination with nitrous oxide/acetylene. In water samples analysis, elemental Cd, Zn and Sn have been determined directly by AAS after dissolving in air/acetylene (dissolved metals by air/acetylene). In another technique, Cd and Zn, if present in low levels is chelated, and aspired into the flame prior to detection by AAS (Mueler, 1984). This method consists of chelation with ammonium pyrrolidine dithiocarbamate (APDC) and extraction with methyl isobutyl ketone (MIBK), followed by aspiration into the flame (Liquid-Liquid Extraction Prior to Flame AAS) (Lucinda *et al.*, 1983). Results are achievable by adjusting the pH of the sample and the water blank, to the sample pH as the standard. While organic tin

TECHNIQUES (Comparative methods and procedures) Samples and species detected

> GC – AAS or GC – FPD

> GC – AAS or GC –

GC – AAS or GC – FPD

GC – MS (compared with GC-AAS/GC-FPD)

technique

Slight modifications where necessary

FPD As above

Tri-, Di- and Moclusters of the butyl compounds (TPh, DPhT, MPhT) can be monitored in water, sediment

As above Water and sediment

monitor both the butyltins and phenyltins (TPh, DPhT, MPh & TPh, DPhT, MPhT) ion species Water, sediment and biota

Monitoring of butyl and phenyl tins species in water, sediment and biota

Routine reason Sediment and biota

Sodium tetrahydroborate (NaBH4)

Sodium tetraethylborate (NaBEt4)

Best reagent out of three: NaBH4, NaBEt4 and EtMgBr

EtMgBr (Grignard reagent)

(best solvent) Best reagent Best detection

Above protocol followed (slight modifications where necessary)

Table 2. Comparative extraction and detection techniques for speciation of organotins

For the analysis of total metals (including all metals, organically and inorganically bound, both dissolved and particulate) most samples will require digestion before analysis to reduce organic matter interference and to convert metal to a form that can be analyzed by atomic absorption spectroscopy or inductively coupled plasma spectrometry (APHA, 1992), For speciation analysis, a few common direct determination methods have earlier been employed (Unicam, 1992). These are the dissolved metals by air/acetylene and direct determination with nitrous oxide/acetylene. In water samples analysis, elemental Cd, Zn and Sn have been determined directly by AAS after dissolving in air/acetylene (dissolved metals by air/acetylene). In another technique, Cd and Zn, if present in low levels is chelated, and aspired into the flame prior to detection by AAS (Mueler, 1984). This method consists of chelation with ammonium pyrrolidine dithiocarbamate (APDC) and extraction with methyl isobutyl ketone (MIBK), followed by aspiration into the flame (Liquid-Liquid Extraction Prior to Flame AAS) (Lucinda *et al.*, 1983). Results are achievable by adjusting the pH of the sample and the water blank, to the sample pH as the standard. While organic tin

Liquid-liquid (LLE)

Liquid-liquid (LLE)

Liquid-liquid (best solvent from above)

Liquid-liquid (best solvent from above)

Liquid-liquid

Solid-phase Extraction (SPME) is extracted with solvent, the inorganic tin is determined by AAS after digestion with nitric acid (Unicam, 1992).

For reliability, efficiency and sensitivity, most trace metals analyses currently involve inductively coupled plasma–mass spectrometry (ICP-MS). The standard ICP-MS technique works quite well for many matrices. But for some element/matrix combinations, it gives poor detection limits or accuracy because of elemental or molecular interferences. In choosing the most appropriate methods to analyze any element/matrix combination, ICP-MS has been coupled to dynamic reaction cell (IC-DRC-MS) (Han and Weber, 1988). This combined technique can detect trace levels in complex matrices, where the standard ICP-MS would be prone to interferences; the standard has also been coupled to a micro-mass platform ICP-MS with collision cell technology (IC-ICP-MS) (Han and Weber, 1988). Extremely low detection limits (especially for the higher mass elements) even in matrices traditionally considered difficult are achievable in an ultra-clean sample preparation and analysis environment.

An analytical technique utilizing hydride generation and atomic fluorescence spectrometry (HG-AFS) (Unicam, 1992) or atomic absorption (HG-AAS) has been developed for the analysis of arsenic, antimony, and selenium at either ultra-trace levels or in complex matrices. With HG-AFS, we can accurately measure total arsenic, antimony, and selenium in nearly all matrices at single-digit parts-per-trillion levels [http://www.frontiergeosciences.com/ebru]. Speciation information could be determined using modifications of this technique, including cryogenic trapping/GC and ion chromatographic separation (Wade *et al.,* 1988).

For mercury speciation, relatively poor sensitivity is provided by transitional flame absorption. Alternate atomization techniques for the AA determination of this element have been developed (Cai and Bayona, 1995; Shrader *et al.,* 1983; Maguire *et al.,* 1982; Wade *et al.*, 1988). Amongst them, the cold vapor atomic absorption technique has received the greatest attention (Schrader *et al.*, 1983, Wade *et al.,* 1988). Other techniques employ cold vapor atomic fluorescence spectrometers (CVAFS) (http://www.frontiergeosciences.com/ebru/), which give unparalleled sensitivity for the determination of low-level total mercury. Using this detector in combination with gold amalgamation or aqueous phase ethylation plus gas chromatographic separation allows determination of Hg speciation at the parts-per-quadrillion level (Wade *et al.*, 1988). Furnace methods for mercury are not recommended due to the extreme volatility of mercury, which has a significant vapor pressure even at room temperature (Stab *et al.,* 1992; Maenpa *et al.,* 2002). Although the first cold vapor principle was proposed by Poluekov and co-workers in 1963, the most popular method credited to Hatch and Ott was published in 1968 (Shrader *et al.,* 1983). In this method, an acidified solution containing mercury is reacted with stannous chloride in a vessel external to the AA instrument. Ground state mercury atoms are produced which subsequently are transported by an air or inert gas flow to an absorption cell installed in the AA instrument. This method provides sensitivities approximately four orders of magnitude better than flame AA (Schrader *et al.*, 1983; Wade *et al*., 1988). It is critical to note that, unlike in the case of organotins speciation where a global approach is favorable, heavy metals speciation will very much require methods choice to consider the thermodynamics of the elemental compound (solid/liquid or gaseous state at room temperature). Therefore, methods for each metal speciation will be specific although the principle underlining the different steps are similar.

Speciation Methods for the Determination of Organotins (OTs)

methylmercury concentrations to allow the smaller aliquot size.

formation that was present in the commonly used distillation procedure.

standards. pH range for maximum extraction is 1-6 for Cd and 2-4 for Zn.

the optical path of the AA instrument for recording.

and Heavy Metals (HMs) in the Freshwater and Marine Environments 47

overcome matrix interference. However, this procedure is preferred as it allows for complete digestion. This allows more accurate results than are possible for other methodologies such as distillations, and most tissues samples (especially from higher organisms) have sufficient

Sediment for methyl mercury analysis is prepared by extraction into methylene chloride from sulfuric acid/potassium bromide/copper sulfate slurry. After extraction, an aliquot of the methylene chloride layer is placed in reagent water and the solvent is purged completely from the solution with an N2 gas stream, leaving the methylmercury in the relatively clean water matrix. A sub-aliquot of the water matrix is treated with sodium tetraethyl borate and is analyzed in the same manner as methylmercury in tissue (described above). This procedure is particularly useful as it isolates the methylmercury from the interfering matrix, and allows a large sample aliquot to be analyzed, which yields low detection limits. The extraction method was developed (Harino *et al.*, 1992) in order to overcome artifact

For water sample analysis, samples can be treated with 25 mL of 4% KMnO4 to break up the organo-mercury compounds. Adding excess hydroxylamine sulphate and passing clean nitrogen through the sample removes free chlorine gas formed during the oxidation step. Reduction of mercury can be carried out in a similar manner to the other hydride forming elements. Samples can be placed in a reaction vessel (normally 20 mL). 1-2 mL of 20 % by weight NaBH4 in concentrated HNO3 can be placed in another vessel outside the cold vapor kit. The solutions can be conveyed into an enclosed system by a circulating peristaltic pump. The mercury vapor formed can then be flushed out of the system into a T piece aligned in

For the analysis of cadmium (Cd) and lead (Pb) using the flame atomic absorption method, the sample is aspirated into a flame and atomized. The amount of light emitted is measured. Detection range may be extended (1) downward by scale expansion or by integrating the absorption signal over a long time and (2) upward by dilution of sample, using a lesssensitive wavelength, rotating the burner head, or by linearizing the calibration curve at high concentrations. Chemical interference occurs by a lack of absorption by atoms that are bound in molecular combination by the flame. By using electrothermal atomic absorption spectrometry, the high heat of a graphite furnace atomizes the element being determined and use of a larger sample volume or reduced flow rate of the purge gas increases sensitivity (detection limit: 0.1µg L-1). Interferences by broadband molecular absorption and chemical (formation of refractory carbides) and matrix effects are common. The use of inductively coupled plasma (ICP) method with ionization of an argon gas stream by an oscillating radio frequency and high temperature dissociates molecules, creating ion emission spectra yielding detection limit > 4.0 1µg L-1. Spectral interference from light emissions originating elsewhere (other than the source) and other physical interference from changes in sample viscosity and surface tension can affect sensitivity. In the determination of Cd, and Zn by liquid-liquid extraction prior to flame AAS, ammonium pyrrolidine dithiocarbamate (APDC) is used to chelate the compounds. Aspiration into the flame follows after extraction with methyl isobuthyl ketone (MIBK). To achieve results in normal conditions, the pH of the sample and the water blank are adjusted to the same pH as the

#### **9.1 Methods and techniques for the determination of heavy metals in water, sediment and biota samples**

Generally, for direct atomic absorption spectroscopy or inductively coupled plasma spectrometry, the sample must be colourless, transparent, odourless, single phase, and have a turbidity of < 1 Nephelometric Turbidity Unit. Otherwise, the sample must first be digested. The following digestion methods are generally used:


For individual metals analysis, requirements vary with the metal and the concentration range to be determined (APHA, 1992) as follows:

Dithizone Method: Mercury ions react with dithizone solution to form an orange solution that is measured in the spectrophotometer. This method is most accurate for samples with [Hg] > 2µ L-1. Known interferences are: Copper, gold, palladium, divalent platinum, and silver react with dithizone in acid solution.

Mercury: Cold vapor atomic adsorption method (CVAAS): detection Limits: Choice of method for all samples with [Hg] < 2µ L-1. Here, there are no known interferences.

For the analysis of solid samples such as sediments and tissues, direct determination is not possible due to the very large matrix effects that are encountered. In order to provide accurate determination in these complex matrices, specialized digestion procedures is required that not only brings the analyte of interest into solution, but also diminishes the interfering compounds present in the matrix as much as possible. The relative extreme low detection limits that are achievable with cold vapor atomic absorption spectrometry (CVAAS) offer the option of dilution or smaller aliquot sizes to overcome sample matrix issues.

Due to the volatile nature of mercury compounds, wet digestion methods are preferred over other trace metal preparation techniques such as dry ashing. Tissue samples can be prepared for total mercury analysis using a heated mixture of nitric and sulfuric acids. After the tissues have been fully solubilized, the digestate can be further oxidized by the addition of BrCl, to bring all the mercury to the Hg2+ oxidation state. At the time of analysis, a subaliquot of the digested sample can be reacted with SnCl2 to reduce the mercury to its elemental form, which can then be concentrated on a trap filled with gold-plated sand and introduced into the CVAFS instrument by thermal desorption with argon as a carrier gas. Tissues can be digested for methylmercury using a heated mixture of potassium hydroxide and methanol. A small aliquot of this digestate can be reacted with sodium tetraethyl borate to produce the volatile methyl-ethyl mercury species. All forms of mercury can be collected on activated carbon traps, and can be introduced into an AAS with an argon carrier gas. The detection limit is somewhat elevated due to the small analytical aliquot required to

**9.1 Methods and techniques for the determination of heavy metals in water, sediment** 

Generally, for direct atomic absorption spectroscopy or inductively coupled plasma spectrometry, the sample must be colourless, transparent, odourless, single phase, and have a turbidity of < 1 Nephelometric Turbidity Unit. Otherwise, the sample must first be



For individual metals analysis, requirements vary with the metal and the concentration

Dithizone Method: Mercury ions react with dithizone solution to form an orange solution that is measured in the spectrophotometer. This method is most accurate for samples with [Hg] > 2µ L-1. Known interferences are: Copper, gold, palladium, divalent platinum, and

Mercury: Cold vapor atomic adsorption method (CVAAS): detection Limits: Choice of

For the analysis of solid samples such as sediments and tissues, direct determination is not possible due to the very large matrix effects that are encountered. In order to provide accurate determination in these complex matrices, specialized digestion procedures is required that not only brings the analyte of interest into solution, but also diminishes the interfering compounds present in the matrix as much as possible. The relative extreme low detection limits that are achievable with cold vapor atomic absorption spectrometry (CVAAS) offer the option of dilution or smaller aliquot sizes to overcome sample matrix

Due to the volatile nature of mercury compounds, wet digestion methods are preferred over other trace metal preparation techniques such as dry ashing. Tissue samples can be prepared for total mercury analysis using a heated mixture of nitric and sulfuric acids. After the tissues have been fully solubilized, the digestate can be further oxidized by the addition of BrCl, to bring all the mercury to the Hg2+ oxidation state. At the time of analysis, a subaliquot of the digested sample can be reacted with SnCl2 to reduce the mercury to its elemental form, which can then be concentrated on a trap filled with gold-plated sand and introduced into the CVAFS instrument by thermal desorption with argon as a carrier gas. Tissues can be digested for methylmercury using a heated mixture of potassium hydroxide and methanol. A small aliquot of this digestate can be reacted with sodium tetraethyl borate to produce the volatile methyl-ethyl mercury species. All forms of mercury can be collected on activated carbon traps, and can be introduced into an AAS with an argon carrier gas. The detection limit is somewhat elevated due to the small analytical aliquot required to

method for all samples with [Hg] < 2µ L-1. Here, there are no known interferences.

digested. The following digestion methods are generally used:

solution is clear and white HClO4 fumes appear.

range to be determined (APHA, 1992) as follows:

silver react with dithizone in acid solution.

issues.

**and biota samples** 


white HClO4 fumes appear.

overcome matrix interference. However, this procedure is preferred as it allows for complete digestion. This allows more accurate results than are possible for other methodologies such as distillations, and most tissues samples (especially from higher organisms) have sufficient methylmercury concentrations to allow the smaller aliquot size.

Sediment for methyl mercury analysis is prepared by extraction into methylene chloride from sulfuric acid/potassium bromide/copper sulfate slurry. After extraction, an aliquot of the methylene chloride layer is placed in reagent water and the solvent is purged completely from the solution with an N2 gas stream, leaving the methylmercury in the relatively clean water matrix. A sub-aliquot of the water matrix is treated with sodium tetraethyl borate and is analyzed in the same manner as methylmercury in tissue (described above). This procedure is particularly useful as it isolates the methylmercury from the interfering matrix, and allows a large sample aliquot to be analyzed, which yields low detection limits. The extraction method was developed (Harino *et al.*, 1992) in order to overcome artifact formation that was present in the commonly used distillation procedure.

For water sample analysis, samples can be treated with 25 mL of 4% KMnO4 to break up the organo-mercury compounds. Adding excess hydroxylamine sulphate and passing clean nitrogen through the sample removes free chlorine gas formed during the oxidation step. Reduction of mercury can be carried out in a similar manner to the other hydride forming elements. Samples can be placed in a reaction vessel (normally 20 mL). 1-2 mL of 20 % by weight NaBH4 in concentrated HNO3 can be placed in another vessel outside the cold vapor kit. The solutions can be conveyed into an enclosed system by a circulating peristaltic pump. The mercury vapor formed can then be flushed out of the system into a T piece aligned in the optical path of the AA instrument for recording.

For the analysis of cadmium (Cd) and lead (Pb) using the flame atomic absorption method, the sample is aspirated into a flame and atomized. The amount of light emitted is measured. Detection range may be extended (1) downward by scale expansion or by integrating the absorption signal over a long time and (2) upward by dilution of sample, using a lesssensitive wavelength, rotating the burner head, or by linearizing the calibration curve at high concentrations. Chemical interference occurs by a lack of absorption by atoms that are bound in molecular combination by the flame. By using electrothermal atomic absorption spectrometry, the high heat of a graphite furnace atomizes the element being determined and use of a larger sample volume or reduced flow rate of the purge gas increases sensitivity (detection limit: 0.1µg L-1). Interferences by broadband molecular absorption and chemical (formation of refractory carbides) and matrix effects are common. The use of inductively coupled plasma (ICP) method with ionization of an argon gas stream by an oscillating radio frequency and high temperature dissociates molecules, creating ion emission spectra yielding detection limit > 4.0 1µg L-1. Spectral interference from light emissions originating elsewhere (other than the source) and other physical interference from changes in sample viscosity and surface tension can affect sensitivity. In the determination of Cd, and Zn by liquid-liquid extraction prior to flame AAS, ammonium pyrrolidine dithiocarbamate (APDC) is used to chelate the compounds. Aspiration into the flame follows after extraction with methyl isobuthyl ketone (MIBK). To achieve results in normal conditions, the pH of the sample and the water blank are adjusted to the same pH as the standards. pH range for maximum extraction is 1-6 for Cd and 2-4 for Zn.

Speciation Methods for the Determination of Organotins (OTs)

the following plan is necessary and should be strictly implemented.


of these techniques is of topical importance to analytical scientists.

critical review. J. Chromatogr. A. 788, 1- 49.

EPA, Region VI; Murmansk Marine Biological Institute.


Both organotins (OTs) and Heavy metals (HMs) have been implicated in endocrine disrupting activities (Mueller, 1987; Fatoki *et al*., 2000; Ndibewu *et al*., 2002). Despite their potential danger to man and the ecosystem, the manufacture and uses of these compounds are not currently controlled in many developing countries. TBT-based antifouling paints are still currently being manufactured in some developing countries and there appears to be no legislation regulating use in the environment. Thus, there is the potential for significant contamination of marine water environments by TBT and heavy metals; hence, they need to be regularly monitored to prevent potential danger to man and the ecosystem due to their endocrine disrupting activities. Also, it is observed that there is a shortage of research capacity in this field, particularly in Africa, explaining why data are very scanty on the occurrence and levels of these toxic compounds. The toxicity of OTs and heavy metals and the ulta-trace levels at which they exist in the aquatic environment make it extremely important to have sensitive and reliable analytical methods available for their determination. Such techniques are not yet commonly available. The need to develop some

Information presented in this chapter was partly researched through the endocrine disrupting contaminants (EDCs) global initiative funded by the National Research Foundation (NRF) of South Africa. The remainder of the chapter has been written based on

[1] Abalos, M., Bayona, J.M., Compano, R., Leal, C. and M. D. Prat. (1997). Analytical

[2] Advanced Technology Research Project Corporation (2000, 2001, 2002). ATRP Corp.; U.S.

procedures for the determination of organotin compounds in sediment and biota:

**9.3 Quality assurance planning** 


**10. Conclusion and recommendations** 


**11. Acknowledgements** 

**12. References** 

the authors' own scientific endeavour.




and Heavy Metals (HMs) in the Freshwater and Marine Environments 49

The accuracy of the method should be demonstrated by analyzing the samples and by performing spiking experiments with water samples and reference sediment materials as outlined in the methods above. In order to carry out a successful quality assurance program,

Samples are placed in a 200 or 250 mL separatory funnel fitted with a teflon stopcock and 4 ml of acetate buffer of pH 6.2 added. The mixture is well agitated. 5 mL of 1% w/v mixed solution of ammonium pyrrolidinedithiocarbamate and diethylammonium diethyldithiocarbamate in water (chelating agent) is added. The total mixture is briefly agitated and 10 – 20 mL of methyl isobutyl ketone (MIBK) is be added. The mixture is vigorously agitated for 60 seconds. The layers are allowed to separate. The lower aqueous layer is removed while the MIBK layer is retained in the tightly capped glass bottles until sample is ready for analysis. A standard of Cd and Zn stock solution are prepared so that 200 mL of water that is extracted would contain 1-20 µg Cd or Zn L-1. In this way, a direct concentration relationship would exist with samples. A reagent blank is run and the sample is analyzed (AAS) under instruments recommended conditions (Van Loon, 1985).

For the determination of arsenic (As) by continuous flow hydride generation (CF-HG-AAS), suitable for volatile metals that produce a metal hydride, the sample is treated with sodium borohydride in the presence of HCL, and then detected by AAS. If recovery is poor, interring organics could be removed by passing the acidified sample through a resin.


Table 3. Summary of techniques for speciation and determination of selected heavy metals

#### **9.2 Sampling and sample location**

In the investigation of the freshwater and marine waters environment, water, sediment and the biota, samples should be taken within the study program-site time schedule from selected locations that reflect different sea regions (for example, Atlantic and Indian oceans) and related shipping activities of that particular location. Sampling for heavy metals analysis, apart from the marine sites, should include other sites such as from rivers (sampling sites can be fixed). Locations such as upstream, midstream and downstreams the rivers should be targeted. Lakes or municipal stream water environments can also be considered. For biological materials, biota, more logically, sourcing should be matched as much as possible to the various sites chosen for freshwater and marine sampling.

About 2.5 L subsurface water samples should be collected at each sampling site. Before sampling, sample bottles should be cleaned by washing with detergent and then soaked in 50 % HCl for 24 h. Finally, bottles should be washed with water and then rinsed with doubly distilled or deionized water. Core sediment samples should be collected at the same site used for water samples by divers. Both sample types should be kept at about 4 oC until analyzed. The biota should be fresh and bought from catchmen direct from source.

#### **9.3 Quality assurance planning**

48 Environmental Health – Emerging Issues and Practice

Samples are placed in a 200 or 250 mL separatory funnel fitted with a teflon stopcock and 4 ml of acetate buffer of pH 6.2 added. The mixture is well agitated. 5 mL of 1% w/v mixed solution of ammonium pyrrolidinedithiocarbamate and diethylammonium diethyldithiocarbamate in water (chelating agent) is added. The total mixture is briefly agitated and 10 – 20 mL of methyl isobutyl ketone (MIBK) is be added. The mixture is vigorously agitated for 60 seconds. The layers are allowed to separate. The lower aqueous layer is removed while the MIBK layer is retained in the tightly capped glass bottles until sample is ready for analysis. A standard of Cd and Zn stock solution are prepared so that 200 mL of water that is extracted would contain 1-20 µg Cd or Zn L-1. In this way, a direct concentration relationship would exist with samples. A reagent blank is run and the sample

is analyzed (AAS) under instruments recommended conditions (Van Loon, 1985).

interring organics could be removed by passing the acidified sample through a resin.

TECHNIQUES (Methods and Procedures)

Microsolid phase extraction (MSE) Flame photometry

Liquid-Liquid Extraction (LLE): Open beaker digestion and extractive Conc. methods

**9.2 Sampling and sample location** 

For the determination of arsenic (As) by continuous flow hydride generation (CF-HG-AAS), suitable for volatile metals that produce a metal hydride, the sample is treated with sodium borohydride in the presence of HCL, and then detected by AAS. If recovery is poor,

Extraction/Separation Detection Observations

Flame photometry AAS

Hydride Generation CF-HG-AAS As in water, sediment

Cold Vapor Technique CV – AAS Hg in water, sediment

Table 3. Summary of techniques for speciation and determination of selected heavy metals

In the investigation of the freshwater and marine waters environment, water, sediment and the biota, samples should be taken within the study program-site time schedule from selected locations that reflect different sea regions (for example, Atlantic and Indian oceans) and related shipping activities of that particular location. Sampling for heavy metals analysis, apart from the marine sites, should include other sites such as from rivers (sampling sites can be fixed). Locations such as upstream, midstream and downstreams the rivers should be targeted. Lakes or municipal stream water environments can also be considered. For biological materials, biota, more logically, sourcing should be matched as

About 2.5 L subsurface water samples should be collected at each sampling site. Before sampling, sample bottles should be cleaned by washing with detergent and then soaked in 50 % HCl for 24 h. Finally, bottles should be washed with water and then rinsed with doubly distilled or deionized water. Core sediment samples should be collected at the same site used for water samples by divers. Both sample types should be kept at about 4 oC until

much as possible to the various sites chosen for freshwater and marine sampling.

analyzed. The biota should be fresh and bought from catchmen direct from source.

AAS Water, sediment, biota

Cd and Zn in water and sediment samples, Detection of Hg

and Biota

and biota

The accuracy of the method should be demonstrated by analyzing the samples and by performing spiking experiments with water samples and reference sediment materials as outlined in the methods above. In order to carry out a successful quality assurance program, the following plan is necessary and should be strictly implemented.


#### **10. Conclusion and recommendations**

Both organotins (OTs) and Heavy metals (HMs) have been implicated in endocrine disrupting activities (Mueller, 1987; Fatoki *et al*., 2000; Ndibewu *et al*., 2002). Despite their potential danger to man and the ecosystem, the manufacture and uses of these compounds are not currently controlled in many developing countries. TBT-based antifouling paints are still currently being manufactured in some developing countries and there appears to be no legislation regulating use in the environment. Thus, there is the potential for significant contamination of marine water environments by TBT and heavy metals; hence, they need to be regularly monitored to prevent potential danger to man and the ecosystem due to their endocrine disrupting activities. Also, it is observed that there is a shortage of research capacity in this field, particularly in Africa, explaining why data are very scanty on the occurrence and levels of these toxic compounds. The toxicity of OTs and heavy metals and the ulta-trace levels at which they exist in the aquatic environment make it extremely important to have sensitive and reliable analytical methods available for their determination. Such techniques are not yet commonly available. The need to develop some of these techniques is of topical importance to analytical scientists.

#### **11. Acknowledgements**

Information presented in this chapter was partly researched through the endocrine disrupting contaminants (EDCs) global initiative funded by the National Research Foundation (NRF) of South Africa. The remainder of the chapter has been written based on the authors' own scientific endeavour.

#### **12. References**


Speciation Methods for the Determination of Organotins (OTs)

AAS). Mikrochim. Acta. 109, 79.

1996, 12, 461R.

Chem. 67, 4250.

Forestry, Pretoria.

Chichester, 1989, p. 188.

Wat.Res. 31(8), 1996-2004.

and Safety Information Sheet on Cd.

chromatography. Anal. Chim. Acta, 264, 91.

Pollution. 81, 103 – 111.

Marine Chem., 25, 279.

51.

and Heavy Metals (HMs) in the Freshwater and Marine Environments 51

[21] Crighton, J. S., Carroll, J., Fairman, B., Haines, J. and M. Hinds. (1996). Industrial

[22] Determination of tin species in environmental samples. (Technical Report).

[23] Dirkx, W. M. R. and Adams, F. C. (1992). Speciation of organotin compounds in water

[24] Dirkx, W. M. R.; Lobinski, R.; and F. C. Adams. (1994). Speciation analysis of organotin

[26] Donard, O. F. X. In: R. M. Harrison and S. Rapsomanikis (Eds.), Environmental Analysis

[27] DWAF (1992) Analytical Methods Manual, TR 151. Department of Water Affairs &

[28] Espadaler, I., Caixach, J., Om, J., Ventura, F., Cortina, M., Paune, F. and J. Rivera. (1997).

[29] Evans, S. M., Kerrigan, E. and N. Palmer. (2000). Causes of Imposex in the Dogwhelk

[30] Fatoki, O. S. and M. B. Ngassoum. (2000). Tributyltin Analysis in Water and Sediments from Port Elizabeth and East London Harbours by GC FPD.(Tech. note). [31] Fatoki, O. S., Ngassoum, M. B and A. O. Ogunfowokan. (2000). Speciation Analysis of

[34] Garcia-Romero, B., Wade, T. L., Salata, G. G. and J. M. Brooks. (1993).Butyltin

[35] Gomez Ariza, J. L., Morales, E., Beltran, R., Giraldez, I. and M. R. Benitez. (1994).

[36] Han, J. S. And J. H. Weber. (1988). Speciation of methyl- and butyl-tin compounds and

[37] Harino, H, Fukushima, M. and M. Tanaka. (1992). Simultaneous determination of

[32] Fent, K. (1996).Ecotoxicology of organotin compounds.Crit.Rev.Toxicol.,26(1),1-117. [33] Forstner, Ulrich and Gottfried T. W. Wittmann "Metal Pollution in the Aquatic

Contamination. Marine Pollution Bulletin, 40, 212-219.

Organotins in Water, Sediment and Biota (Literature Review).

detection after extraction separation. Anal. Chim. Acta, 286, 309-318. [25] Donard, O. F. X., Lalere, B., Martin, F. and R. Lobinski. (1995). Microwave-assisted

International Union of Pure and Applied Chemistry Oct 1998. .

Analysis: Metals, Chemicals and Advanced Materials. J. Atom. Absorpt. Spetr.,

and sediments by gas chromatography - atomic-absorption spectrometry (GC -

in water and sediments by gas chromatography with optical spectrometric

leaching of organotin compounds from sediments for speciation analysis. Anal.

Using Chromatography Interfaced with Atomic Spectroscopy, Horwood,

Identification of organic pollutants in Ter river and its system of reservoirs supplying water to Barcelona (Catalona, Spain): A study by GC/MS and FAB/MS.

Nucella lapillus (L.) and its Use as a Biological Indicator of Tributyltin

Environment" Springre-Verlag: Berlin, 1983. National Toxicology Program's Health

concentration in oysters from the Gulf of Mexico from 1989 to 1991. Environmental

Sampling and storage of sediment samples for organotin.Appl.Organomet.Chem.9,

inorganic tin in oysters by hydride-generation atomic-absorption spectrometry.

butyltin and phenyltin compounds in the aquatic environment by gas


[3] Apte, S. C. and M. J. Gardner. (1998). Determination of organotins in natural waters by toluene extraction and graphite-furnace AAS. Anal. Chem., 1989, 61, 2320. [4] Arakawa, Y., Wada, O., and M. Manabe. (1983). Extraction and fluorimetric determination of organotin compounds with morin. J. Chromatogr. 207, 237. [5] Ashby, J. R. and P. J. Craig. (1989). New method for the production of volatile

[6] ATSDR-ToxFAQs - Cadmium http://www.atsdr.cdc.gov/tfacts5.html[Accessed 2011] [7] ATSDR-ToxFAQs - Mercury http://www.atsdr.cdc.gov/tfacts46.html[Accessed 2002] [8] Attar, K. M. (1996). Analytical Methods for Speciation of Organotins in the Environment.

[9] Balls, P. W. (1987).Tributyltin (TBT) in the waters of a Scottish sea loch arising from the use of antifoulant treated netting by salmon farms. Aquaculture, 65, 227 – 237. [10] Barjaktarovic, L., Elliott, J. E. and A. M. Scheuhammer. (2002). Metal and

[13] Bryan, G. W. And P. E. Gibbs. (1991). Impact of low concentrations of tributyltin (TBT)

[15] Cai, Y., and J. Bayona. (1995). Determination of methlmercury in fish and river water

[16] Cai, Y., Rapsomanikis, S. and O. A. Meinrat. (1993). Determination of butylin

[17] Ceulemans, M. and F. C. Adams. (1995). Evaluation of sample preparation methods for

[18] Chau, Y. K, Yang, F. and R. J. Maguire. (1996). Improvement of extraction recovery for the mono butlyltin species from sediment. Anal. Chim. Acta 320, 165 – 169. [19] Coello, W. F. and M. A. Q. Khan. (1996). Protection Against Heavy Metal Toxicity by Mucus and Scales in Fish. Arch. Environ. Contam. Toxicol. 30, 319-326. [20] Corr, J. J. and E. H. Larsen. (1996). Arsenic Speciation by Liquid Chromatography

derivatization methods. Analytica Chimica Acta, 274(3), 243-251.

ed. M. C. Newman & A. W. McIntosh. Lewis Punlishers, pp. 323–361. [14] Cai, Y., Alzaga, R. and J. M. Bayona. (1994). In situ derivatization and supercritical-fluid

UK sediments. Sci. Total Environ., 78, 219.

Applied Organometallic Chemistry. 10(5). 317-337.

Feathers of Wading Birds from Florida. Ambio. 26, 2.

compounds in sediment. Anal. Chem. 66, 1161 – 1167.

A. 696, 113-122.

1225.

Anal. Chim. Acta. 317, 161 - 170.

organometallic species for analysis from the environment: some butyltin levels in

Metallothionein Concentrations in Scoter (*Melanitta* spp.) from the Pacific Northwest of Canada, 1989-1994. Arch. Environ. Contam. Toxicol. 43, 486-491. [11] Bayona, J. M. and Y. Cai. (1994). The role of supercritical fluid extraction and chromatography in organotin speciation. Trends in Anal. Chem. 13(8), 327 – 332. [12] Beyer, W. N., Spalding, M. and M. Morrison. (1997). Mercury Concentrations in

on marine organisms: A review. In Metal Ecotoxicology Concepts & Applications,

extraction for the simultaneous determination of butyltin and phenyltin

samples using in situ sodium tetraethlborate derivatization following by solidphase microextraction and gas chromatography-mass spectrometry. J. Chromatogr.

compounds in sediment using gas chromatography-atomic absorption spectrometry: comparison of sodium tetrahydroborate and sodium tetraethylborate

organotin speciation analysis in sediments – focus on monobutyltin extraction.

Coupled with Ion spray Tandem Mass Spectrometry. J. Atom. Absorpt. Spetr., 12,


Speciation Methods for the Determination of Organotins (OTs)

GC/MS. Anal. Chem. 72, 5136 - 5141.

Absorpt. Spetr., 12, 203.

Absorption. No. AA-35(9).

Sc. Technol. 16, 698 – 702.

Chem., 11, 17.

Ambio. 26, 6.

197, 303-308.

203-231.

Appl. Organomet. Chem., 5, 399.

Analytica Chimica Acta, 286(2), 343-355.

and Heavy Metals (HMs) in the Freshwater and Marine Environments 53

[57] Lespes, G., Pinasseau,C. C., Gautier, M. P. and M. Astruc. (1998). Direct Determination

[58] Looser, P. W., Berg, M., Fent, K., Muhlemann, J., and R. P. Schwarzenbach. (2000).

[59] Lu, J. Y., Chakrabarti, C. L., Back, M. H. Sekaly, A. L. R., Gregoire, D. C. and W. H.

[60] Lucinda, M. V. (1983). Determination of the priority pollutant metals – Regulations and Methodology.Varian Instruments at Work, Varian Atomic Absorption. AA-34,(9). [61] Lucinda, M. V. (1983). Dealing with matrix interferences in the determination of priority

[62] Lucinda, M. V., Covick, L. A and D. E. Shrader. (1983). The Determination of the

[63] Maenpa, K. A., Kukkonen, J. V. K. and M. J. Lydy. (2002). Remediation of Heavy Metal-

[64] Maguire, R. J., Chau, Y. K., Bengert, G. A., Hale, E. J.,Wong, P. T. S. and O. Kramar.

[65] Martin-Landa, I., De Pablos, F., and I. L. Marr. (1989). Determination of organotins in

[66] Martin, F. M., Tseng, C. M., Belin, C., Quevauviller, P. and F. X. Donard. (1994).

*[67]* Martin, F. M. and O. F. X. Donard. (1994). Interference mechanisms and reduction

[68] Martin, P. J. and P. S. Rainbow. (1998). The kinetics of zinc and cadmium in the

[69] Matschullat, J. (1997).Trace Element Fluxes to the Baltic Sea: Problems of Input Budgets.

[70] MCkie, J. C. (1987). Determination of total Tin and Tributyltin in marine biological

[71] Meech, J. A., Veiga, M. M. and D. Tromans. (1998). Reactivity of Mercury from Gold

Mining Activities in Darkwater Ecosystems. Ambio. 26, 2.

Instruments at Work, Varian Atomic Absorption. No. AA-33(9).

Earthworm Bioassay. Arch. Environ. Contam. Toxicol. 43, 389-398.

Flame Photometric Detection. Analyst. 123 (5), 1091 – 1094.

of Butyl- and Phenyltin Compounds as Chlorides Using Gas Chromatography and

Phenyl- and Butyltin Analysis in Small Samples by Cold Methanolic Digestion and

Schroeder. (1996). Speciation of Some Metals in River Surface Water, Rain and Snow, and the Interactions of These Metals With Selected Soil Matrices. J. Atom.

pollutant metals by furnace AA. Varian Instruments at Work, Varian Atomic

priority pollutant metals using the CRA-90 Carbon Rod Atomizer. Varian

Contaminated Soils Using Phosphorus: Evaluation of Bioavailability Using an

(1982). Occurrence of organotin compounds in Ontario lakes and rivers. Environ.

fish and sediments by gas chromatography with flame photometric detection.

Interferences generated by organic and inorganic compounds during organotin speciation using hydride generation coupled with cryogenic trapping, gas chromatographic separation and detection by atomic absorption spectrometry.

during the speciation of organotin compounds by hydride generation, cryoseparation and detection by atomic absorption spectrometry. Trends in Anal.

haemolymph of the shore crab *Carcinus maenas* (L.), Aquatic Toxicology. 40, 2-3,

materials by electrothermal Atomic Absorption Spectrometry. Anal. Chim. Acta


[44] Ikeda, M., Zhang, Z.W., Moon, C. S., Imai,Y., Watanabe, T., Shimbo, S.M, W.C. Lee, C.C.

Lead in Tainan City, Taiwan. Arch. Environ. Contam. Toxicol., 30, 121 – 126. [45] Inma, F. E. and J. M. Bayona. (1997). Supercritical fluid extraction of priority organotin contaminants from biological matrices. Anal. Chim. Acta 355, 269 – 276. [46] Jiang, G. B., Ceulemans, M. and F. C. Adams. (1996). Optimization study for the

surface-induced luminescence. Journal of Chromatography A, 727, 119-129. [47] Jiang, G. B., Liu, J. Y. and K. W. Yang. (2000). Speciation analysis of butyltin compounds

[48] Jiang, G., Liu, J. and K. Yang. (2000). Speciation analysis of butyltin compounds in

[49] Kan-atireklap, S., Tanabe, S. and J. Sanguansin. (1997). Contamination by Butyltin

[50] Kan-atireklap, S., Tanabe, S. and Sanguansin, J., Tabucanon, M. S. and Hungspreugs

[51] Kan-atireklap, S., Yen, N. T. H., Tanabe, S. and A. N. Subramanian. (1998). Butyltins,

[53] Kuballa, J., Wilken, R. D., Jantzen, E., Kwan, K. K. and Y. K. Chau. (1995). Speciation

[54] Kumar, T. U., Dorsey, G. J., Caruso, J. A. and E. H. Evans. (1993). Speciation of inorganic

[55] Leal, C., Grandos, M., Prat, M. D. and R. Compano. (1995). Labelling of organotin

[56] Lespes, G., Desauziers, V, Montigny, C. and M. P. Gautier. (1998). Optimization of solid-

and genotoxicity of butyltin compounds. Analyst 120 (3), 667 – 673.

compounds for fluorometric detection. Anal. Chim. Acta, 314, 175.

experimental designs. J. Chromatographr. A. 826(1), 67 – 76.

from India. Toxicological and Environmental Chemistry. 67, 409 – 424. [52] Krull, I. S. And K. W. Panaro. (1985). Trace analysis and speciation for methylated

and Y.L. Guo. (1996). Background Exposure of General Population to Cadmium &

speciation analysis of organotin and organogermanium compounds by on-column capillary gas chromatography with flame photometric detection using quartz

in Chinese seawater by capillary gas chromatography with flame photometric detection using insitu hydride derivatization followed by headspace sold-phase

Chinese seawater by capillary gas chromatography with flame photmetric detection using in-situ hdride derivatization followed by headspace solid-phase

Compounds in Sediments from Thailand. Marine Pollution Bulletin. 34(11),894 -

(1997). Contamination by Butyltin Compounds and Orgamochlorine Residues in Green Mussel (*Perna viridis, L.)* from Thailand Coastal Waters. Environmental

organochlorines and Organochlorine residues in green mussels (*Perna viridis, L.)* 

organotins by HPLC - hydride generation – direct-current plasma emission

and organotin compounds in biological samples by liquid chromatography with

phase microextraction for the speciation of butyl- and phenyltins using

[38] http://h2osparc.wq.ncsu.edu/info/hmetals.html[Accessed 2002]. [39] http://www.frontiergeosciences.com/ebru/[Accessed 2011].

[43] http://www.osha.gov/SLTC/cadmium/index.html[Accessed 2003].

microextraction. Anal. Chim. Acta 421, 67 – 74.

microextraction. Anal. Chim. Acta 421, 67 – 74.

spectroscopy. Appl. Organomet. Chem., 3,295.

ICP-MS detection. J. Chromatogr. A, 654, 261-268.

899.

Pollution. 97(1-2), 79 - 89.

[40] http://www.igc.org/[Accessed[2011].

[41] http://www.mercurypolicy.org/[Accessed 2010]. [42] http://www.msceast.org/hms/[Accessed 2001].


Speciation Methods for the Determination of Organotins (OTs)

149 – 166.

and Heavy Metals (HMs) in the Freshwater and Marine Environments 55

[88] Prudente, M., Ichihashi, H., Kan-atireklap, S., Watanabe, I. and S. Tanabe. (1999).

[89] Quevauviller, P., Lavigne, R., Pinel, R. and Astruc, M. (1989). Organotins in sediments

[90] Quevauviller, Ph. (1996). Atomic Spectrometry Hyphenated to Chromatography for

[91] Regoli, F.,Nigro,M. and E.Orlando.(1998).Lysosomal & antioxidant responses to metals in Antarcticscallop *Adamussium colbecki*, AquaticToxicology.40(4),375-392.

[93] Richardson and S. Gangolli. In: The Dictionary of Substances and Their Effects Royal

[94] Rivaro, P., Zaratin, L., Frache, R. and A. Mazzucotelli. (1995). Determination of

[95] Roos, J. T. H. and W. J. Price. (1971). Mechanisms of interference and releasing action in

[96] Samson, J. C. and J. Shenker. (2000). The teratogenic effects of methylmercury on early development of the zebrafish, *Danio rerio.* Aquatic Toxicology.40, 2,3,203- 231. [97] Sasaki, K., Suzuki, T. and Y. Saito. (1988). Determination of tributyltin and dibutyltin

[98] Sasaki, K., Ishizaka, T., Suzuki, T. and Saito, Y. (1988). Determination of tributyltin and

[99] Sharron, D., Swami, K. And R. L. Jansing. *(*1995). Rapid ultratrace analysis of tributyltin

[100] Shrader, D. E, and W. B. Hobbins. (1983). The Determination of Mercury by Cold

[101] Shrader, D. E, Lucinda, M. V. and L. A. Covick. (1983). The Determination of toxic

[102] Stab, J. A., Van Hattum, B., De Voogt, P. and U. A. T. Brinkman. (1992). Preparation of

[103] Stab, J. A. Cofino, W. P., Van Hattum, B. and U. A. T. Brinkman. (1994). Assessment of

compounds in yellowtails. Bull. Environ. Contam. Toxicol. 41, 888.

photometric detection. J. AOAC Int., 78(5), 1317 – 1321.

Work, Varian Atomic Absorption. No. AA-31, June.

organotin compounds in marine mussel samples by using high-performance liquid chromatography - hydride-generation inductively coupled plasma atomic-emission

atomic absorption spectroscopy-III, Interference of iron on chromium absorption.

dibutyltin compounds in fish by gas chromatography with flame photometric

in aqueous matrixes by purge-and-trap gas chromatography with flame-

Vapor Atomic Absorption. Varian Instruments at Work, Varian Atomic

metals in waters and wastes by furnace Atomic Absorption. Varian Instruments at

pentylated organotin standards for use in trace analysis with gas chromatography.

transport routes of triphenyltin used in potato culture in the Netherlands. Appl.

[92] Reisch, M. S. Paints and coatings. Chemical Engineering News 1996, 44049.

Society of Chemistry, Cambridge (1994), p. 288 and 344.

waters of the Phillipines. Fisheries Science 65(3), 44 –447.

European Union. J. Atom. Absorpt. Spetr., 12.

`pectrometry. Analyst, 120, 1937.

Spectrochimica Acta. 26B, 441 – 444.

detection. Assoc. of Anal. Chem. 126.

Absorption. No. AA-32, September.

J. Chromtogr. A, 609, 1955.

Organomet. Chem. 8,577.

Butyltins, organochlorines and metals in green mussels, *perna viridis L.* from coastal

and mussels from the Sado Estuarine system (Portugal). Environ. Pollut., 1989, 57,

Elemental Speciation: Performance Assessment Within the Standards, Measurements and Testing Programme (Community Bureau of Reference) of the


[72] Millán, E. and J. Pawliszyn. (2000). Determination of butyltin species in water and

[73] Morabito, R., Chiavarini, S. and C. Cremisini, in Ph. Quevauviller (ed.), Quality

[74] Mueler, M. D. (1984). Tributyltin Detection at Trace levels in water and sediments using GC with Flame-Photometric Detection and GC-MS. Anal. Chem. 317, 32 – 36. [75] Mueller, M. D. (1987). Comprehensive trace level determination of organotin

[76] Muyssen, B. T. A. and C. R., Janssen. (2002). Accumulation and Regulation of Zinc in

[77] Nagase, M., Kondo, H. and K. Hasebe. (1995). Determination of tributyltin and

[79] National Toxicology Program's Health and Safety Information Sheet on Cd. http://ntp-

[80] Ndibewu, P. P. And O. S. Fatoki. (2002). Development of speciation methods for the

[83] Ombaba, J. M. and E. F. Barry. (1992). Determination of organotin species by capillary

[84] Organometals and Organometalloids, ocurrence and fate in environment, ACS

[85] Pascoe, G. A., Blanchet, R. J. and G. Linder. (1996). Food Chain Analysis of Exposures

[86] Pawliszyn, J. (1997). Solid Phase Microextraction. John Wiley & sons Ltd, New York. [87] Pereira, W. E., Wade, T. L., Hostettler, F. D. and F. Parchaso. (1999). Accumulation of

[81] Nriagu, J. O. (1996). History of Global Metal Pollution. Science,Vol. 272 (4),223-224. [82] Okamura, H., Aoyama, I., Takami, T., Maruyama, Y., Suzuki, M., Matsumoto, I.,

b.niehs.nih.gov/NTP\_Reports/NTP\_Chem\_H&S/NTP\_Chem7/

chromatography with flame photometric detection. Analyst. 120, 1923. [78] Nagase, M., Toba, M., Kondo, H. and K. Hasebe. (1998). Determination of dibutyltin

detection. Journal of Chromatography A, 873, 63 -71.

with flame photometric detection. Anal. Chem. 59, 617.

photometric detection. Analyst. 1091 – 1094.

p.435.

Toxicol. 43, 492-496.

Radian7440.

763.

environments. Unpublished.

Chromatography A, 598(5), 97-103.

Toxicol. 30, 306-318.

Bulletin, 38 1005-1010.

PR, E11 HV, Wilkinson RR,-Washington, p 388.

sediment by solid-phase microextraction–gas chromatography–flame ionization

Assurance for Environmental Analysis, Elsevier,Amsterdam, 1995, Chapter 19,

compounds in environmental samples using high-resolution gas chromatography

*Daphnia magna*: Links with Homeostasis and Toxicity. Arch. Environ. Contam.

triphenyltin compounds in hair and fish using a hydrolysis technique and gas

compounds in soft polyurethane foam by gas chromatography with flame

determination of organotins and heavy metals in the freshwater and marine

Katsuyama, J., Hamada, T., Beppu, O., Tanaka, R. J., Maguire, D., Liu, Y. L., Lau and G. J. Pacepavicius. (2000). Phytotoxicity of the New Antifouling Compound Irgarol 1051 and a Major Degradation Product. Marine Pollution Bulletin.40, 754-

gas chromatography with alternating current plasma emission detection. Journal of

Symposium Series 82 (Brickman Bella J. M. (eds) (1978) Zuckermann JJ, Reisdorf

and Risks to Wildlife at a Metals-Contaminated Wetland. Arch. Environ. Contam.

Butyltins in Sediments and Lipid Tissues of the Asian Clam, Potamocorbula amurensis, Near Mare Island Naval Shipyard, San Francisco Bay. Marine Pollution


**3** 

**Use of** *Enterococcus,* **BST and** 

Vesna Furtula1, Charlene R. Jackson2,

*North Vancouver, British Columbia,* 

*USDA-ARS, Athens, Georgia,* 

*1Canada 2USA 3Malaysia* 

Rozita Osman3 and Patricia A. Chambers1

**Sterols for Poultry Pollution Source** 

**Tracking in Surface and Groundwater** 

*1Aquatic Ecosystem Impacts Research Division, Environment Canada,* 

*2Bacterial Epidemiology and Antimicrobial Resistance Research Unit,* 

*3Faculty of Applied Sciences, Universiti Teknologi MARA, Shah Alam, Selangor,* 

Maintaining and preserving the quality of surface and ground waters involves many challenges, one of the most serious being bacteriological contamination caused by discharge of human and animal waste. Water resources may become contaminated with pathogens from human or animal feces as a result of malfunctioning wastewater operations (treatment plants or septic systems), stormwater or combined sewer overflows, poor management practices for storing or land-applying livestock manure, and defecation by livestock and wildlife in or near surface waters. Pollution source identification is crucial in order to improve best management practices and eliminate consequent health risks to the general public and aquatic ecosystems. Distinguishing between human and animal sources of fecal pollution in water has been a subject of many studies (Tyagi et al., 2009a). Microbial source tracking methods have employed a wide range of micro-organisms (e.g., fecal coliforms, total coliforms, bifidobacteria, *E. coli*, enterococci) for identifying sources of water pollution, but each has certain limitations (Tyagi et al., 2009a). Moreover, many microbes are not hostspecific, making them ineffective for source identification. Chemical methods for fecal source tracking include analysis of sterols, bile acids, caffeine, whitening agents etc., with sterols being the most widely used indicator compound (Bull et al., 2002; Saim et al., 2009; Tyagi et al., 2009b). Both classes of methods have been somewhat successful in identifying pollution sources but not fully evaluated and accepted as established methods in

Enterococci are the second most studied group of bacteria in the field of microbial source tracking (following *E. coli*) due to their connection to humans and animals as well as their

**1. Introduction** 

environmental studies.


### **Use of** *Enterococcus,* **BST and Sterols for Poultry Pollution Source Tracking in Surface and Groundwater**

Vesna Furtula1, Charlene R. Jackson2, Rozita Osman3 and Patricia A. Chambers1 *1Aquatic Ecosystem Impacts Research Division, Environment Canada, North Vancouver, British Columbia, 2Bacterial Epidemiology and Antimicrobial Resistance Research Unit, USDA-ARS, Athens, Georgia, 3Faculty of Applied Sciences, Universiti Teknologi MARA, Shah Alam, Selangor, 1Canada 2USA 3Malaysia* 

#### **1. Introduction**

56 Environmental Health – Emerging Issues and Practice

[104] Stewart, F. M., Furness, R. W., Monteiro, L. R. and L. R. Monteiro. (1996). Relationships

Gulls, *Larus fuscus,* & Cory's. Arch. Environ. Contam. Toxicol. 30, 299-305. [105] Strategic plan for the Health Related Water Issues (HRWI) research field. Version 1.2B

[106] Suyani, H., Creed, J., Davidson, T. and J. Caruso. (1989). Inductively coupled plasma

[107] Tam, N. F. Y., and Y. S. Wong. (1995). Spatial and temporal variations of heavy metal

[108] Tanabe, S., Prudente,M., Mizuno., Hasegawa, T., Iwata, H. and N. Miyazaki. (1998).

[109] Tao, H., Rajendran, R. B., Quetel, C. R., Nakazato, T., Tominaga, M. and A. Miyazaki.

[110] Thomaidis, N. S., Adams, F. C., and Lekkas, T. D. (2001). A simple method for the

[111] Thompson, J. A, J., Douglas, Y. K. S., Chau, Y. K. and R. J. Maguire. (1998). Recent

[112] Tolosa, I., Merlini, L., de Bertrand, N., Bayona, J. and J. Albageis. (1992). Occurrence

[115] Wade, T. L., Garcia-Romero, B., Brooks, J. M. (1988). Tributyltin contamination in bivalves from United States coastal estuaries. Environ. Sci. Technol. 22,1488. [116] Willis, J. B. `The analysis of biological materials by Atomic-absorption Spectroscopy`,

Organization, Melbourne, Australia, Vol. 11, No. 2, Supp.,1965. pp251-258. [117] Woller, A., Garraud, H., Martin, F., Donard, O. F. X. and Péter. (1996). Determination

enclosures. Environmental and Toxicological Chemistry. 11, 145-155. [113] Toth, S., Becker-van Slooten, K., Spack, L., de Alencastro, L. F. and J. Tarradellas.

Lake Geneva. Bull. Environ. Contam. Toxicol. 57, 426-433.

waters. Environmental Science and Technology. 32, 193 –198.

normal plasma conditions. Anal. Chem. 71, 4208 – 4215.

Organometallic Chemistry. 129(8-9), 643-650.

[114] Unicam AAS Methods Manual (1992). 1.1-27. 207.

detection. Anal. Chim. Acta 312, 141 – 148.

(7/02/2001). p9.

Bulletin 31, 54–261.

QFAAS.

compounds. Appl. Spectr.6, 962 – 967.

Between Heavy Metal and Metallothionein Concentrations in Lesser Black-Backed

mass spectrometry and atomic-emission spectrometry coupled to highperformance liquid chromatography for speciation and detection of organotin

concentration in sediments of a mangrove swamp in Hong Kong. Marine Pollution

Bututyltin contamination in marine mammals from north Pacific and Asian coastal

(1999). Tin speciation in the femtogram range in open ocean seawater by gas chromatography/inductively coupled plasma spectrometry using a sheild torch at

speciation of organotin compounds in water samples using ethylation & GC-

studies of residual tributyltin in coastal British Columbia sediments. Applied

and fate of tributyl-andtriphenyltin compounds in western Mediterranean coastal

(1996). Irgarol 1051, an antifouling compound in freshwater, sediment, and biota of

from the Division of Chemical Physics, Commonwealth Scientific and Industrial

of Total Mercury in Sediments by Microwave-assisted Digestion-Flow Injection-Inductively Coupled Plasma Mass Spectrometry. J. Atom. Absorpt. Spetr., 2, 461R. [118] Yang, H. J., Jiang, S., Yang, Y. and C. Hwang. (1995). Speciation of tin by reversed

liquid gas chromatography with inductivedly coupled plasma mass spectrometric

Maintaining and preserving the quality of surface and ground waters involves many challenges, one of the most serious being bacteriological contamination caused by discharge of human and animal waste. Water resources may become contaminated with pathogens from human or animal feces as a result of malfunctioning wastewater operations (treatment plants or septic systems), stormwater or combined sewer overflows, poor management practices for storing or land-applying livestock manure, and defecation by livestock and wildlife in or near surface waters. Pollution source identification is crucial in order to improve best management practices and eliminate consequent health risks to the general public and aquatic ecosystems. Distinguishing between human and animal sources of fecal pollution in water has been a subject of many studies (Tyagi et al., 2009a). Microbial source tracking methods have employed a wide range of micro-organisms (e.g., fecal coliforms, total coliforms, bifidobacteria, *E. coli*, enterococci) for identifying sources of water pollution, but each has certain limitations (Tyagi et al., 2009a). Moreover, many microbes are not hostspecific, making them ineffective for source identification. Chemical methods for fecal source tracking include analysis of sterols, bile acids, caffeine, whitening agents etc., with sterols being the most widely used indicator compound (Bull et al., 2002; Saim et al., 2009; Tyagi et al., 2009b). Both classes of methods have been somewhat successful in identifying pollution sources but not fully evaluated and accepted as established methods in environmental studies.

Enterococci are the second most studied group of bacteria in the field of microbial source tracking (following *E. coli*) due to their connection to humans and animals as well as their

Use of *Enterococcus,* BST and Sterols for Poultry

**2. Materials and methods** 

**2.1 Sample collection** 

collection.

Pollution Source Tracking in Surface and Groundwater 59

Surface water was collected from 12 sites and ground water was sampled at 28 sites in the Abbotsford area of British Columbia, Canada, near poultry farms and berry farms that use poultry litter as fertilizer (Figure 1). Surface water was sampled for bacteriological water quality and for sterol analysis in December 2009; all samples were grab water collections. Ground water was sampled in April, August and December 2009 for microbial water quality and sterols. At each site, three full well volumes were pumped out of the piezometer (purged) using a submersive Hydrofit pump prior to sampling with low density polyethalene (LDPE) tubing (dedicated for each well) located close to the well screen. A minimum of three line volumes were purged from the sample tubing prior to sample

Fig. 1. Aerial map of the sampling area located in the Lower Fraser Valley region of British Columbia, Canada. "S" and "GS" indicate surface water and groundwater sampling sites,

Water samples were collected in: (1) 250-mL sterile polypropylene bottles for bacteriological analyses (total and fecal coliform, *E. coli* and *Enterococcus*); (2) certified clean one liter amber glass bottles for sterol analysis and (3) one liter sterile polypropylene bottles for BST analysis. Samples were placed on ice packs in coolers (~ 4°C) and shipped to the laboratory where they were kept in a cold-room (≤4°C) until analyzed. Samples for BST analysis were

Litter samples were collected from two different poultry farms: a broiler farm and a layer farm, in at least three different locations in the poultry barn. Samples from broiler barns (total of 4) were collected at the beginning (day 3) and end of the production cycle (day 35). Samples were collected by an analyst wearing gloves and using a sterile scoop, and placed into sterile falcon tubes. The samples were kept on ice until analysis, which was performed within 24 h of collection (samples from broiler farm). Samples from layer farms were

respectively. The map was generated using Google Maps

collected once, frozen after collection and analyzed at a later date.

filtered within 24 h of collection.

recent significance as a clinical pathogen (Figueras et al., 2000; Aarestrup et al., 2002; Scott et al., 2005) and ability to persist in the environment (Harwood et al., 2000). Different strains of enterococci populate the digestive tracts of humans and animals, making them a good indicator of water contamination. A metabolic fingerprint database developed by Ahmed et al., (2005) for enterococci was able to distinguish between human and animal sources despite the fact that a number of biochemical phenotypes were found in multiple host groups. Bacterial Source Tracking (BST) uses unique genetic markers in *Bacteroides* (naturally occurring bacteria in the intestinal flora) to identify organisms responsible for fecal pollution in aquatic environments, and has been used for detection of bacteriological contamination from different sources such as humans, ruminant animals, dogs, pigs, horses, and elk (Bernhard et al., 2000a,b; Dick et al., 2005a,b). Unfortunately, primers for poultry and other birds are not available.

Sterols are organic molecules, a family of compounds that occur naturally in animals, plants and fungi. They have a steroid ring structure and varying functional groups that confer specific characteristics (such as polarity, bioactivity and lipophilicity) to the molecule. Coprostanol, the major compound in human faeces and a product of the microbial reduction of cholesterol in the higher animal guts, has been considered an indicator of faecal pollution. Other cholesterol congeners (campesterol, sitosterol and stigmasterol) are also degraded by bacteria in the intestinal tract of higher mammals to stanols. Higher concentrations of coprostanol have been found in human sewage than in animal wastewater and concentrations of stigmastanol and epicoprostanol were usually higher in animal (cows, pigs, and poultry) than in human wastewater (Blanch et al., 2004). Sterol analysis, a widely used chemical method for identifying fecal pollution sources is based on the fact that different sterol compounds are associated with human or animal waste and their presence/absence and relative concentrations and ratios can be used as an indication of the origin of water contamination (Chou & Liu, 2004; Devane et al., 2006; Gilpin et al., 2003; Bull et al., 2002; Jardé et al., 2007a, b; Saim et al., 2009).

Interpretation of findings from these and other markers can be improved by the application of chemometric techniques which are gaining ground in evaluation of environmental data (Brodnjak-Voncina et al., 2002; Mendiguchía et al., 2004; Singh et al., 2005; Terrado et al., 2011). The most common chemometric methods are cluster analysis (CA) and principal component analysis (PCA) with factor analysis (FA). The goal of CA is to identify groups of objects (such as sampling sites) that give similar, homogenous results with respect to extent or type of fecal pollution, whereas PCA enables a reduction in data and description of a given multidimensional system by a smaller number of new variables (Loska & Wiechula, 2003). Pollution sources and dischargers can also be identified using PCA (Einax et al., 1998; Loska & Wiechula, 2003).

The Fraser River valley of British Columbia is considered the poultry capital of Canada. The poultry waste generated from the industry is used as fertilizer and spread onto the fields, thus creating a non-point source run-off type of surface and ground water pollution. The objective of this study was to determine the extent and sources of fecal contamination in surface and ground water in this poultry dominated agricultural area. In particular, we tested *Enterococcus* isolates as source tracking indicators for poultry in combination with chemical indicators sterols, BST and chemometric analysis.

### **2. Materials and methods**

#### **2.1 Sample collection**

58 Environmental Health – Emerging Issues and Practice

recent significance as a clinical pathogen (Figueras et al., 2000; Aarestrup et al., 2002; Scott et al., 2005) and ability to persist in the environment (Harwood et al., 2000). Different strains of enterococci populate the digestive tracts of humans and animals, making them a good indicator of water contamination. A metabolic fingerprint database developed by Ahmed et al., (2005) for enterococci was able to distinguish between human and animal sources despite the fact that a number of biochemical phenotypes were found in multiple host groups. Bacterial Source Tracking (BST) uses unique genetic markers in *Bacteroides* (naturally occurring bacteria in the intestinal flora) to identify organisms responsible for fecal pollution in aquatic environments, and has been used for detection of bacteriological contamination from different sources such as humans, ruminant animals, dogs, pigs, horses, and elk (Bernhard et al., 2000a,b; Dick et al., 2005a,b). Unfortunately, primers for poultry

Sterols are organic molecules, a family of compounds that occur naturally in animals, plants and fungi. They have a steroid ring structure and varying functional groups that confer specific characteristics (such as polarity, bioactivity and lipophilicity) to the molecule. Coprostanol, the major compound in human faeces and a product of the microbial reduction of cholesterol in the higher animal guts, has been considered an indicator of faecal pollution. Other cholesterol congeners (campesterol, sitosterol and stigmasterol) are also degraded by bacteria in the intestinal tract of higher mammals to stanols. Higher concentrations of coprostanol have been found in human sewage than in animal wastewater and concentrations of stigmastanol and epicoprostanol were usually higher in animal (cows, pigs, and poultry) than in human wastewater (Blanch et al., 2004). Sterol analysis, a widely used chemical method for identifying fecal pollution sources is based on the fact that different sterol compounds are associated with human or animal waste and their presence/absence and relative concentrations and ratios can be used as an indication of the origin of water contamination (Chou & Liu, 2004; Devane et al., 2006; Gilpin et al., 2003; Bull

Interpretation of findings from these and other markers can be improved by the application of chemometric techniques which are gaining ground in evaluation of environmental data (Brodnjak-Voncina et al., 2002; Mendiguchía et al., 2004; Singh et al., 2005; Terrado et al., 2011). The most common chemometric methods are cluster analysis (CA) and principal component analysis (PCA) with factor analysis (FA). The goal of CA is to identify groups of objects (such as sampling sites) that give similar, homogenous results with respect to extent or type of fecal pollution, whereas PCA enables a reduction in data and description of a given multidimensional system by a smaller number of new variables (Loska & Wiechula, 2003). Pollution sources and dischargers can also be identified using PCA (Einax et al., 1998;

The Fraser River valley of British Columbia is considered the poultry capital of Canada. The poultry waste generated from the industry is used as fertilizer and spread onto the fields, thus creating a non-point source run-off type of surface and ground water pollution. The objective of this study was to determine the extent and sources of fecal contamination in surface and ground water in this poultry dominated agricultural area. In particular, we tested *Enterococcus* isolates as source tracking indicators for poultry in combination with

and other birds are not available.

et al., 2002; Jardé et al., 2007a, b; Saim et al., 2009).

chemical indicators sterols, BST and chemometric analysis.

Loska & Wiechula, 2003).

Surface water was collected from 12 sites and ground water was sampled at 28 sites in the Abbotsford area of British Columbia, Canada, near poultry farms and berry farms that use poultry litter as fertilizer (Figure 1). Surface water was sampled for bacteriological water quality and for sterol analysis in December 2009; all samples were grab water collections. Ground water was sampled in April, August and December 2009 for microbial water quality and sterols. At each site, three full well volumes were pumped out of the piezometer (purged) using a submersive Hydrofit pump prior to sampling with low density polyethalene (LDPE) tubing (dedicated for each well) located close to the well screen. A minimum of three line volumes were purged from the sample tubing prior to sample collection.

Fig. 1. Aerial map of the sampling area located in the Lower Fraser Valley region of British Columbia, Canada. "S" and "GS" indicate surface water and groundwater sampling sites, respectively. The map was generated using Google Maps

Water samples were collected in: (1) 250-mL sterile polypropylene bottles for bacteriological analyses (total and fecal coliform, *E. coli* and *Enterococcus*); (2) certified clean one liter amber glass bottles for sterol analysis and (3) one liter sterile polypropylene bottles for BST analysis. Samples were placed on ice packs in coolers (~ 4°C) and shipped to the laboratory where they were kept in a cold-room (≤4°C) until analyzed. Samples for BST analysis were filtered within 24 h of collection.

Litter samples were collected from two different poultry farms: a broiler farm and a layer farm, in at least three different locations in the poultry barn. Samples from broiler barns (total of 4) were collected at the beginning (day 3) and end of the production cycle (day 35). Samples were collected by an analyst wearing gloves and using a sterile scoop, and placed into sterile falcon tubes. The samples were kept on ice until analysis, which was performed within 24 h of collection (samples from broiler farm). Samples from layer farms were collected once, frozen after collection and analyzed at a later date.

Use of *Enterococcus,* BST and Sterols for Poultry

**2.4 Bacterial Source Tracking (BST)** 

Table 1. Sterols and limits of quantification (LOQ)

Pollution Source Tracking in Surface and Groundwater 61

mm x 0.25 µm film thickness) by the following temperature gradient: initial temperature 70°C hold for 1 min, 30°C/min to 180°C, 5°C/min to 310°C and hold on 310°C for 4 min. Eluting compounds were analyzed by mass spectrometer and ChemStation software (revision A.01.01, Palo Alto, CA) and sterols quantitated using internal standard method. List of sterols and their limits of quantification are presented in Table 1. Quality control blanks and spikes were run with each batch of samples. Various sterol ratios were calculated

BST analysis was conducted according to the BST method used at the Pacific Environmental Science Centre, North Vancouver BC, Canada (Environment Canada, 2006). One liter water samples were filtered through AP15 prefilters (Millipore Corporation, Billerica, MA) to remove large pieces of material. Prefiltrate was split into two aliquots (500 ml each), which were then filtered through 0.22 μm filters (Supor-200, PALL Corporation, Ann Arbour, MI). The filters were stored individually in 15 mL tubes containing 0.5 mL of GITC lysis buffer

*α* -Methyl-5-cholesten-3*β* -ol Campesterol C28H48O 0.005 Cholest-5-en-3 *β* -ol Cholesterol C27H46O 0.009 *β* -Cholestan-3*β* -ol Coprostanol C27H48O 0.005 *β* -cholesta-5,24-dien-3-ol Desmosterol C27H44O 0.008

Cholest-5-en-3 *α* -ol Epicoprostanol C27H48O 0.005 1,3,5,7-Estratetraen-3-ol-17-one Equilin C18H20O2 0.07 3,4-Dihydro-3-(4-hydroxyphenyl)-2H-1-benzopyran-7-ol Equol C15H15O3 0.1

3- *β* -5- *β* -cholestan-3-ol, C27H48O 0.007

1,3,5(10)-Estratriene-3,16α,17β-triol Estriol C18H24O3 0.01 3-Hydroxyestra-1,3,5(10)-trien-17-one Estrone C18H22O2 0.02 19-Norpregna-1,3,5(10)-trien-20-yne-3,17-diol 17-Ethinylestradiol C20-H24-O2 0.1 17α-Ethynyl-1,3,5(10)-estratriene-3,17β-diol 3-methyl ether Mestranol C21H26O2 0.01 13β-Ethyl-17α-ethynyl-17β-hydroxygon-4-en-3-one Norgestrel C21H28O2 0.07 19-nor-17alpha-ethynyl-17beta-hydroxy-4-androsten-3-one Norethindrone C20H26O2 0.08 5-Stigmasten-3β-ol -Sitosterol C29H50O 0.007 24-Ethylcolesta-5,22E-dien-3*β* -ol Stigmasterol C29H48O 0.007

For each sample, DNA was extracted from one AP15 pre-filter and one 0.22 µm filter. DNA extraction was performed with the Qiagen DNeasy kit (Mississauga, ON), and the manufacturer's instructions were followed with the following exception: for the first steps,

Sterol Common names Formula LOQ

Dihydrocholesterol (cholestanol)

> 17-Estradiol C18H2402 0.01 17-Estradiol C18H2402 0.01

(µg/L)

to determine the presence of fecal contamination and its likely source (Table 2).

(5M guanidine isothiocyanate, 100 mM EDTA and 0.5% sarkosyl) at -20°C.

#### **2.2 Bacteriological analysis**

Analysis of enterococci in water samples was performed using a membrane filtration technique whereby samples retained on filter paper were incubated on mE agar for 48 h at 41°C followed by incubation on Esculin Iron Agar (EIA) for 20 minutes at 41°C (USEPA, 2000). Colonies that appeared pink to red with dark precipitation on EIA were verified using Biolog Microbial ID system in combination with Biolog Gram Positive Aerobic Bacteria Database (Release 6.01, Biolog, Hayward, CA) Results are reported as colony-forming units (cfu) per unit volume.

For enterococci in poultry litter samples, 5-6 g of litter was weighed into 10-ml of 0.85% sterile saline in a sterile 50-mL falcon tube. The tube was vortexed on high for one minute and serial dilutions were plated on KF streptococcal agar (Difco, Detroit, MI). Red or pink colonies on the KF agar were verified using Biolog Microbial ID system in combination with Biolog Gram Positive Aerobic Bacteria Database (Release 6.01, Biolog). Isolated colonies of confirmed *Enterococcus* were inoculated into 5 ml of tryptic soy broth containing 6.5% NaCl and incubated for 5 – 12 hours at 35°C; one milliliter of this culture was then combined with 325 µL 80% glycerol (20% glycerol final concentration) and stored at -40°C until further analysis. Confirmed *Enterococcus* isolates were identified to species level using multiplex PCR (Jackson et al., 2004).

Total and fecal coliform and *E coli* analyses of water samples were performed using procedures based on "British Columbia Environmental Laboratory Manual for the Analysis of Water, Wastewater, Sediment and Biological Materials" (2005 Edition) (Horvath, 2009).

#### **2.3 Sterol analysis**

Analytical grade standards were purchased from Sigma–Aldrich (Oakville, ON) for 17 compounds (mestranol, norethindrone, equol, estrone, equilin, norgestrel, 17 αethinylestradiol, 17 α-estradiol, 17 β-estradiol, estriol, coprostanol, epicoprostanol (cholestanol), cholesterol, desmosterol, campesterol, stigmasterol and β-sitosterol); equol was purchased from Fluka (Oakville, ON). Primary standards were made in acetone at a concentration of 1 mg/ml and stored at -20°C. Acetylated mixture calibration standards of 0.02 to 0.5 µg/L were made every two months and stored at -20°C. Surrogate 17 β-estradiold3 and internal standard p-terphenyl-d14 were added to every sample. Solvents, sodium chloride and potassium carbonate were purchased from VWR (Edmonton, AL) and all chemical reagents were of analytical grade.

Sterol extraction and detection were conducted according to the sterol method used at the Pacific Environmental Science Centre, North Vancouver BC, Canada (Environment Canada, 2005). Briefly, 800 mL of unfiltered sample was acidified with sulfuric acid to pH ~ 3 and surrogate β-estradiol-d3 was added. After stirring samples with 100 ml of dichloromethane for two hours, they were transferred into separatory funnels and the organic layers separated. Samples were then concentrated and derivatized with pyridine/acetic acid and re-extracted with petroleum ether in the presence of 10% potassium carbonate solution. The organic layers were concentrated to near dryness and reconstituted in 200 µl of internal standard (p-terphenyl-d14).

Extracted samples for sterol analysis were injected into Agilent 5973 MS system (injector 280°C), carried by helium flow of 1.2 mL/min, separated on Rtx-5ms column (30 m x 0.25

Analysis of enterococci in water samples was performed using a membrane filtration technique whereby samples retained on filter paper were incubated on mE agar for 48 h at 41°C followed by incubation on Esculin Iron Agar (EIA) for 20 minutes at 41°C (USEPA, 2000). Colonies that appeared pink to red with dark precipitation on EIA were verified using Biolog Microbial ID system in combination with Biolog Gram Positive Aerobic Bacteria Database (Release 6.01, Biolog, Hayward, CA) Results are reported as colony-forming units

For enterococci in poultry litter samples, 5-6 g of litter was weighed into 10-ml of 0.85% sterile saline in a sterile 50-mL falcon tube. The tube was vortexed on high for one minute and serial dilutions were plated on KF streptococcal agar (Difco, Detroit, MI). Red or pink colonies on the KF agar were verified using Biolog Microbial ID system in combination with Biolog Gram Positive Aerobic Bacteria Database (Release 6.01, Biolog). Isolated colonies of confirmed *Enterococcus* were inoculated into 5 ml of tryptic soy broth containing 6.5% NaCl and incubated for 5 – 12 hours at 35°C; one milliliter of this culture was then combined with 325 µL 80% glycerol (20% glycerol final concentration) and stored at -40°C until further analysis. Confirmed *Enterococcus* isolates were identified to species level using multiplex

Total and fecal coliform and *E coli* analyses of water samples were performed using procedures based on "British Columbia Environmental Laboratory Manual for the Analysis of Water, Wastewater, Sediment and Biological Materials" (2005 Edition) (Horvath, 2009).

Analytical grade standards were purchased from Sigma–Aldrich (Oakville, ON) for 17 compounds (mestranol, norethindrone, equol, estrone, equilin, norgestrel, 17 αethinylestradiol, 17 α-estradiol, 17 β-estradiol, estriol, coprostanol, epicoprostanol (cholestanol), cholesterol, desmosterol, campesterol, stigmasterol and β-sitosterol); equol was purchased from Fluka (Oakville, ON). Primary standards were made in acetone at a concentration of 1 mg/ml and stored at -20°C. Acetylated mixture calibration standards of 0.02 to 0.5 µg/L were made every two months and stored at -20°C. Surrogate 17 β-estradiold3 and internal standard p-terphenyl-d14 were added to every sample. Solvents, sodium chloride and potassium carbonate were purchased from VWR (Edmonton, AL) and all

Sterol extraction and detection were conducted according to the sterol method used at the Pacific Environmental Science Centre, North Vancouver BC, Canada (Environment Canada, 2005). Briefly, 800 mL of unfiltered sample was acidified with sulfuric acid to pH ~ 3 and surrogate β-estradiol-d3 was added. After stirring samples with 100 ml of dichloromethane for two hours, they were transferred into separatory funnels and the organic layers separated. Samples were then concentrated and derivatized with pyridine/acetic acid and re-extracted with petroleum ether in the presence of 10% potassium carbonate solution. The organic layers were concentrated to near dryness and reconstituted in 200 µl of internal

Extracted samples for sterol analysis were injected into Agilent 5973 MS system (injector 280°C), carried by helium flow of 1.2 mL/min, separated on Rtx-5ms column (30 m x 0.25

**2.2 Bacteriological analysis** 

(cfu) per unit volume.

PCR (Jackson et al., 2004).

**2.3 Sterol analysis** 

chemical reagents were of analytical grade.

standard (p-terphenyl-d14).

mm x 0.25 µm film thickness) by the following temperature gradient: initial temperature 70°C hold for 1 min, 30°C/min to 180°C, 5°C/min to 310°C and hold on 310°C for 4 min. Eluting compounds were analyzed by mass spectrometer and ChemStation software (revision A.01.01, Palo Alto, CA) and sterols quantitated using internal standard method. List of sterols and their limits of quantification are presented in Table 1. Quality control blanks and spikes were run with each batch of samples. Various sterol ratios were calculated to determine the presence of fecal contamination and its likely source (Table 2).

#### **2.4 Bacterial Source Tracking (BST)**

BST analysis was conducted according to the BST method used at the Pacific Environmental Science Centre, North Vancouver BC, Canada (Environment Canada, 2006). One liter water samples were filtered through AP15 prefilters (Millipore Corporation, Billerica, MA) to remove large pieces of material. Prefiltrate was split into two aliquots (500 ml each), which were then filtered through 0.22 μm filters (Supor-200, PALL Corporation, Ann Arbour, MI). The filters were stored individually in 15 mL tubes containing 0.5 mL of GITC lysis buffer (5M guanidine isothiocyanate, 100 mM EDTA and 0.5% sarkosyl) at -20°C.


Table 1. Sterols and limits of quantification (LOQ)

For each sample, DNA was extracted from one AP15 pre-filter and one 0.22 µm filter. DNA extraction was performed with the Qiagen DNeasy kit (Mississauga, ON), and the manufacturer's instructions were followed with the following exception: for the first steps,

Use of *Enterococcus,* BST and Sterols for Poultry

Table 3. Bacterial Source Tracking (BST) primers

**2.5 Chemometric approach** 

XLStat2009 statistical program.

**3.1 Bacterial contamination** 

**3. Results** 

Pollution Source Tracking in Surface and Groundwater 63

program GeneSnap was used to capture the image from a CCD camera. Positive matches were made by correlating the bands with the DNA ladder and the known size of the positive bands as published (Bernhard et al., 2000a,b, Dick et al., 2005a,b). All negative controls

(included at every stage) were blank and all positive controls worked appropriately.

Organism Primer Set reference

Human HF134F / HF654R Bernhard et al., 2000b

Ruminant Animal CF128F / Bac708R Bernhard et al., 2000b

Pig PF134F / Bac708R Donation from K. Field

Horse HoF597F / Bac708R Dick et al., 2005b Dog DF475F / Bac708R Dick at al., 2005a Elk EF447F / EF990R Dick at al., 2005a General *Bacteroides* Bac32F / Bac708R Bernhard et al., 2000a

The goal of the chemometrics approach is to display the most significant patterns in the complex data sets. The most popular statistical methods are principal component analysis (PCA) which provides information on the most meaningful parameters to describe a large data set and cluster analysis (CA) which identifies natural groupings within a data set. Sterol data were used in both PCA and CA; the statistical analyses were performed by

Principal component analysis (PCA) generated principal components (PCs). Varimax rotation was applied on the PCs with eigenvalues greater than 1 (Kim & Mueller, 1987) in order to obtain new groups of variables called varimax factors (VFs) that better interpret the data set (Juahir et al., 2009). Cluster analysis applied on surface water samples data identified similarities in the sterol composition and grouped sampling sites accordingly.

Total coliform, fecal coliform and *E. coli* in surface water ranged from 100-17,000 cfu/100 mL, <1-700 cfu/100 ml and <1-690 cfu/100 mL, respectively (Table 4). Several groundwater locations also tested positive for total coliform ranging from 25-17000 cfu/100 ml, although the majority of groundwater samples showed no evidence of total coliform contamination

*Enterococcus* was detected in all surface water samples and at 3 groundwater sites. *Enterococcus* counts ranged from 2 to 2100 cfu/100 mL for surface water samples (Table 4) and 1 to 5 cfu/100 ml for groundwater samples (Table 5). Seven enterococci were isolated

(Table 5). Only one location, BC-008, consistently tested positive for total coliform.

HF183F / Bac708R Bernhard et al., 2000b

CF193F / Bac708R Bernhard et al., 2000b

PF163F / Bac708R Dick et al., 2005b

Buffer AL (provided with the kit) and 100% ethanol (Commercial Alcohols, Langley, BC) were added to the 15 ml tubes (containing filters and GITC buffer) in 1:1:1 ratios with 1 minute of vortexing after each addition of a liquid.



Table 2. Sterol ratios for identifying (a) human fecal contamination and (b) differentiating sources of fecal contamination

Three aliquots (600 µl each) were loaded onto the DNeasy columns and washed according to the manufacturer's protocol. The pure genomic DNA samples were stored at -20°C in sterile 1.5 ml tubes (Fisher Scientific, Ottawa, ON). Extracted DNA was amplified by Polymerase Chain Reaction (PCR) carried out with a DNA engine Tetrad 2 (Bio-Rad Laboratories Canada, Toronto, ON). The samples were tested with all *Bacteroides* primers available (Table 3), which identify feces from humans, ruminant animals, pigs, horses, dogs, elk, and general *Bacteroides*. After agarose gel electrophoresis of PCR samples, gels were visualized and scored in a bio-imaging system (Gene Genius, Fisher Scientific, Ottawa, ON) and the

Buffer AL (provided with the kit) and 100% ethanol (Commercial Alcohols, Langley, BC) were added to the 15 ml tubes (containing filters and GITC buffer) in 1:1:1 ratios with 1

**Human Fecal Contamination (a)**

Yes Unsure No

3 Epicoprostanol / Coprostanol < 0.2 0.2 - 0.8 > 0.8 Froehner 2009; de Castro Martin 2007

4 Coprostanol / Cholesterol > 0.5 - < 0.5 Gilpin 2003; Fattore 1996; Carreira 2004;

7 Coprostanol / Epicoprostanol > 1.5 - < 1.5 Fattore 1996; Marvin 2001; Patton 1999; Reeves

<0.7 chicken and/or cow

>0.1 cattle/horse/deer

Three aliquots (600 µl each) were loaded onto the DNeasy columns and washed according to the manufacturer's protocol. The pure genomic DNA samples were stored at -20°C in sterile 1.5 ml tubes (Fisher Scientific, Ottawa, ON). Extracted DNA was amplified by Polymerase Chain Reaction (PCR) carried out with a DNA engine Tetrad 2 (Bio-Rad Laboratories Canada, Toronto, ON). The samples were tested with all *Bacteroides* primers available (Table 3), which identify feces from humans, ruminant animals, pigs, horses, dogs, elk, and general *Bacteroides*. After agarose gel electrophoresis of PCR samples, gels were visualized and scored in a bio-imaging system (Gene Genius, Fisher Scientific, Ottawa, ON) and the

<1 human

Table 2. Sterol ratios for identifying (a) human fecal contamination and (b) differentiating

>0.4 Shah 2007

**Ratios for Differentiating Sources of Fecal Contamination (b)**

5 Coprostanol / Cholestanol > 0.5 0.3 - 0.5 < 0.3 Devane 2006; Roser 2006

6 Coprostanol / (Cholestanol + Cholesterol) > 0.2 0.15 - 0.2 < 0.15 Chan 1998

8 >3.7 pig Jardé 2007

9 >1.5 pig/chicken/cow Jardé 2007

10 <0.01 human Standley 2005
